Preface

Sleep-related breathing disorders: new developments

Obstructive sleep apnea (OSA) is defined as

recurrent episodes of airflow cessation during sleep

despite persistence of respiratory effort. It is common

in the general population—an estimated 15 million

people in the United States are believed to suffer from

the disorder. Furthermore, it is widely suspected that

sleep-disordered breathing is underdiagnosed in both

adults and children.

There are a variety of ways in which sleep-related

breathing disorders are classified. In one simple

schema, OSA may be considered the extreme end of

a spectrum of repetitive sleep-related upper airway

(UA) obstructions that includes, in order of severity,

intermittent snoring, continuous snoring, UA resist-

ance syndrome, asymptomatic hypopnea, and symp-

tomatic apnea-hypopnea. An American Academy of

Sleep Medicine Task Force Report published in 1999

defined four separate syndromes associated with

abnormal respiratory events during sleep among

adults, namely obstructive sleep apnea-hypopnea syn-

drome, central sleep apnea-hypopnea syndrome,

Cheyne-Stokes breathing syndrome, and sleep hypo-

ventilation syndrome. According to this classification,

the UA resistance syndrome was not regarded as a

distinct disease; rather, respiratory event related

arousals (RERAs) were considered part of OSA.

Sleep state dependency is one of the most impor-

tant and central features of OSA. During wakefulness,

ventilation and oxygenation are generally normal, only

to be disrupted during sleep by repetitive UA narrow-

ing or obstruction. The diminished tone of the muscles

maintaining UA patency is part of the generalized

muscle hypotonia that occurs during sleep. Sleep

apnea is terminated by an arousal accompanied by

restoration of UA patency and airflow. Sleep state–

dependent changes in UA biomechanics and neuro-

physiology may lead to alterations in the balance

between inward forces that favor collapse of the air-

ways and outward forces that counter the former. Not

only do persons with OSA tend to have anatomically

narrower and physiologically more collapsible UAs,

theymay also have decreased activity of the UAdilator

muscles with which to compensate for the collapse.

Persons with OSA commonly have alternating

episodes of loud snoring and periods of silence

during sleep due to marked diminution or total

absence of airflow. Blood oxygen saturation may

drop during the apneic phase. Respiratory events

typically recur throughout the evening, at times

reaching numbers substantial enough to produce

sleep fragmentation and subsequent daytime sleepi-

ness. There is increasing recognition of the potential

consequences of this disorder: neuropsychological

impairment, adverse effects on quality of life, and

seizure disorders, in addition to specific cardiovas-

cular diseases such as hypertension, atherosclerosis,

stroke, pulmonary hypertension, cardiac arrhythmia,

and congestive heart failure.

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00028-5

Teofilo L. Lee-Chiong, Jr, MD Vahid Mohsenin, MD

Guest Editors

Clin Chest Med 24 (2003) xi–xiii

Technological innovations will likely transform

the recognition and diagnosis of sleep-disordered

breathing events. The overnight, attended, laboratory

polysomnography is the generally accepted reference

standard for diagnosis. Its limited availability and

high costs have prompted the search for alternative

sleep study protocols, such as portable sleep monitor-

ing. Accurate monitoring of respiration during sleep,

including measurements of airflow, respiratory effort,

oxygenation, and ventilation, is indispensable in the

identification of sleep-related respiratory events. In

particular, measurement of respiratory effort using

either esophageal pressure monitoring or surface

diaphragmatic electromyography is vital in distin-

guishing central from obstructive apneas. Today, the

sleep clinician has a wide variety of devices available

to monitor oro-nasal airflow, including pneumota-

chometers, nasal pressure monitors, thermal or ex-

pired carbon sensors, strain gauges, and respiratory

inductance plethysmography.

As we explore the indications for treatment and

various options for managing persons with OSA,

including behavioral modifications, pharmacological

interventions, positive airway pressure devices, oral

appliances, and surgery, the challenge is to provide a

framework within which we can integrate basic

research and clinical data with future therapies for

this disorder.

Obesity is strongly correlated with the prevalence

of sleep apnea, and weight reduction can be a highly

effective short-term treatment. However, recurrence

of sleep apnea appears to be common during long-

term follow-up, either because of a failure to maintain

weight loss or, for reasons that are unclear, despite

successful maintenance of weight loss.

Positional modification, using a variety of devices

such as posture alarms and wedge pillows to avoid

the supine sleep position, appear to be most effective

in persons with milder disease. Again, long-term data

are sparse and dishearteningly variable.

The search for effective pharmacological targets

continues. Currently, none of the agents that has

been evaluated to reduce sleep-disordered breathing

events is consistently effective to be considered as

standard therapy. Identification of excitatory neuro-

transmitters of the UA dilator motoneurons is ac-

tively being pursued. Another area of research is

pharmacological intervention using stimulant medi-

cations to attenuate residual daytime sleepiness that

may persist despite regular use of positive airway

pressure (PAP) therapy.

Since its first description in 1981, continuous

positive airway pressure (CPAP) therapy has become

the main therapy for OSA. It is highly effective, safe,

and reliable. PAP therapy most likely acts primarily as

a pneumatic splint; it may also decrease pharyngeal

collapsibility by augmenting lung volume as well as

increase UA length and tension. CPAP is typically

titrated during a formal sleep study, determining the

pressure at which it will effectively abolish all sleep-

disordered breathing in the supine position and in

REM sleep. Nevertheless, significant intra- and inter-

night variability exists in the severity of sleep-disor-

dered breathing and the corresponding corrective PAP

settings. A new generation of PAP devices, referred to

as automated PAPs, are capable of detecting signals

serving as surrogates of UA obstruction (eg, snores,

apneas, hypopneas, or airflow limitation) and, using

model-specific diagnostic and therapeutic algorithms,

responding to changes in airway resistance by either

increasing or decreasing the pressures generated.

Whether or not they are appropriate, automated-PAPs

are being increasingly used to diagnose and treat OSA

or to titrate pressures for conventional CPAP devices.

Oral devices, including tongue repositioning

devices and mandibular repositioning appliances,

are established therapies for primary snoring and

milder forms of OSA. Some persons with more

severe sleep-disordered breathing may also respond

favorably to these devices. Oral appliances are

becoming increasingly popular because of their ease

of use, portability, and reversibility. Increased under-

standing of their mechanisms of actions (including

effects of UA patency and muscle function), indica-

tions of therapy, predictors of treatment outcome,

and complications will help clarify their roles in the

management of patients.

Surgery remains an option for many patients,

especially those who are either unwilling to try, or

are intolerant of, positive pressure therapy. Advances

in surgical techniques have significantly improved

outcomes. Selection among the various surgical pro-

cedures is individualized, tailored primarily to the

anatomical region of narrowing or obstruction. Thus,

uvulopalatopharyngoplsty is commonly performed

for oropharyngeal obstruction, whereas surgical alter-

ations of the tongue, hyoid, and maxillomandibular

complex are attempted for hypopharyngeal airway

obstruction. The role of radiofrequency UA soft

tissue ablation is still being debated.

Dionysius of Heracleia (born 360 BC) was

described by Athenaeus as ‘‘ . . . an unusually fat

man . . . sleepy, difficult to arouse and had problems

breathing . . .so [his] physicians prescribed . . . fineneedles, long enough that they thrust through his ribs

and belly when he happened to fall into a very deep

sleep . . . ’’ Could this be how OSAwas treated then?

If so, we would like to believe that over the past

T.L. Lee-Chiong, Jr, V. Mohsenin / Clin Chest Med 24 (2003) xi–xiiixii

2400 years there has been some progress in our

understanding and management of this disorder.

The purpose of this issue of the Clinics in Chest

Medicine is to provide a comprehensive discussion of

the various aspects of OSA, focusing on new devel-

opments and controversies and emphasizing trends

that may potentially offer a glimpse of the future of

the science and practice of sleep medicine. We hope

that readers find this issue to be clinically useful, and

we welcome all feedback.

We wish to acknowledge our sincere gratitude to

the outstanding authors who have generously pro-

vided us with an array of excellent texts. We are

especially indebted to Sarah Barth and the editorial

staff at W.B. Saunders for their expert counsel and

unwavering support. Finally, we would like to thank

our families: Grace and Zoe, Shahla, Amir, and

Neda—it is to them that we dedicate this issue.

Teofilo L. Lee-Chiong, Jr, MD

Sleep Medicine Center

Division of Pulmonary and Critical Care Medicine

University of Arkansas for Medical Sciences

Central Arkansas Veterans Healthcare System

4301 West Markham Street, Slot 555

Little Rock, AR 72205, USA

E-mail address: leechiongteofilol@uams.edu

Vahid Mohsenin, MD

Director

Yale Center for Sleep Medicine

Associate Professor of Medicine

Yale University

40 Temple Street, Suite 3C

New Haven, CT 06511, USA

T.L. Lee-Chiong, Jr, V. Mohsenin / Clin Chest Med 24 (2003) xi–xiii xiii

Molecular and physiologic basis of obstructive sleep apnea

Sigrid Carlen Veasey, MD*

Division of Sleep Medicine, University of Pennsylvania School of Medicine, 3600 Spruce Street, Philadelphia, PA 19104, USA

This is an exciting time to be involved in the

study of the obstructive sleep apnea-hypopnea syn-

drome (OSAHS) because characterization of the

diverse manifestations of this disorder continues to

evolve. One may be certain that the characterization

of this highly prevalent and disabling disorder is not

complete. There are many reasons why the defini-

tions and descriptions of the OSAHS will continue to

evolve. The syndrome-in-progress status may be

attributed, in part, to the relative newness of the

initial characterization of the OSAHS three decades

ago [1,2]. A more important reason, however, is that

this disease process, with repeated systemic oxy-

hemoglobin desaturations followed by reoxygenation

events and sleep disruption, has the potential to place

a substantial oxidative burden on many, if not all,

physiologic systems. Recently, researchers have be-

gun to recognize that included in the physiologic

systems impacted on by the repeated airway occlu-

sions and hypoxia/reoxygenation events are the

upper airway soft tissues and muscles and neural

control mechanisms. The disease process itself may

alter the molecular and physiologic mechanisms in-

volved in OSAHS.

This article summarizes the pathophysiologic

mechanisms of OSAHS and complements the phys-

iologic information with data concerning molecular

mechanisms involved in OSAHS and newer informa-

tion regarding the mechanisms through which the

disease process may alter obstructive sleep-disor-

dered breathing. An understanding of the pathophysi-

ology [3,4] has brought therapies such as continuous

positive airways pressure [5,6], surgical therapies for

the upper airway [7 –10], and oral mandibular

advancement devices [11,12]. An understanding of

the molecular mechanisms may provide unique

approaches to therapies for this prevalent disorder,

including pharmacotherapies, and at the same time, a

comprehension of the molecular mechanisms may

afford insight into the differential vulnerability in

the severity and diverse manifestations of OSAHS,

so that we may better understand who is at risk for

this disease and its many morbidities.

An overview of the pathophysiology of obstructive

sleep apnea-hypopnea syndrome

One of the most remarkable features of the

OSAHS is the state dependency of this disorder.

Specifically, in persons with isolated OSAHS, ven-

tilatory patterns and arterial oxygen values during

wakefulness are completely normal. In contrast, dur-

ing sleep, the upper airway of persons with OSAHS

narrows or collapses or both [4], which results in

upper airway occlusion with large intrathoracic and

upper airway intraluminal pressure swings [13,14],

oxyhemoglobin desaturations [4], hypercapnia [15],

increases in sympathetic drive [16–18], and ulti-

mately, arousal with larger sympathetic surges [4]

and massive increases in upper airway dilator muscle

activity, which restores airway patency [4].

This state dependency in upper airway patency

and respiratory function suggests that state-dependent

changes in neural drive to the upper airway dilator

and pump muscles prompt obstructive upper airway

events. It is important to recognize that state-depen-

dent changes in neural drive to respiratory muscles

are not unique to sleep apnea. State-dependent reduc-

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00019-4

This work was supported in part by NIH HL 60287.

* Center for Sleep and Respiratory Neurobiology,

987 Maloney Building, 3600 Spruce Street, Philadelphia,

PA 19104.

E-mail address: veasey@mail.med.upenn.edu

Clin Chest Med 24 (2003) 179–193

tions in respiratory muscle activity are a normal

phenomenon of sleep [19,20]. The unique features

in individuals with OSAHS are a reliance on upper

airway dilator muscle activity and lung volume and a

greater magnitude of sleep state-dependent reductions

in muscle activity [20,21]. For these reasons, the

impact on airway caliber is larger [22–25].

State dependency of disease is a unique and

clinically important characteristic because it implies

that this disorder should be readily amenable to

pharmacologic therapies that effectively target state-

dependent neural changes. Understanding the mech-

anisms underlying OSAHS is essential for elucidating

safe, effective therapies for this syndrome. The key

components of this pathophysiology are (1) state-

dependent changes in respiratory drive to the upper

airway and pump muscles, (2) upper airway anatomy,

(3) upper airway mechanics, and (4) upper airway

muscle function.

Neural mechanisms underlying state-dependent

changes in upper airway and pump muscle activity

State-dependent upper airway obstruction in

OSAHS occurs most commonly within the pharynx

in the retropalatal or retroglossal regions or both

[26,27]. During inspiration, negative intraluminal

pressures in these regions exert centripetal forces that

must be countered by centrifugal forces of the upper

airway dilator muscles, particularly in persons with

upper airway narrowing or increased collapsibility.

The neurochemical control of upper airway motoneu-

rons is complex, and in this section the author works

backward from the upper airway dilator motoneurons

to reflexes and upper respiratory neural drive to

describe what is known of the neural and neurochem-

ical mechanisms that may contribute to state depen-

dency of the upper airway for each neural mechanism.

Many muscles contribute to centrifugal forces in

the upper airway (Fig. 1), whereas other muscles

that are important in phonation, deglutition, and

respiratory breaking act as constrictors within the

upper airway. When discussing neural mechanisms,

it is important to recognize that most upper airway

motor nuclei (cranial nuclei, V, VII, X, XI, and XII)

house motoneurons for upper airway dilators and

constrictors. Many researchers, including the author,

have chosen to focus first on XII, the hypoglossal

nucleus, because this collection of motoneurons

innervates the largest upper airway dilator muscles

in humans with OSAHS: the genioglossus and

geniohyoid [4,20]. The hypoglossal motoneurons

also innervate many more dilators than constrictors

[28]. The information gained concerning the state-

dependent control of hypoglossal motoneurons ulti-

mately must be addressed for other populations of

motoneurons, however. Recently, Kuna showed that

Fig. 1. Schematic of potential upper airway dilators in humans. Muscles surrounding the upper airway have the potential to dilate

or stent the upper airway in many different directions. Represented in this drawing are the force vectors for activation of specific

muscle groups. As a collapsible tube (gray), oropharyngeal patency is most effectively achieved by simultaneous activation of

muscles with vectors in different directions. As discussed in the text, elongation of the airway along with widening of the lateral

walls may be most effective in rendering the airway less collapsible.

S.C. Veasey / Clin Chest Med 24 (2003) 179–193180

electrical stimulation of the glossopharyngeal nerve

causes marked dilation of the pharynx (Fig. 2) [29].

Many excitatory and inhibitory neurotransmitters

and neuromodulators contribute to the activity of

hypoglossal and other upper airway dilator motoneu-

rons. Serotonin and its co-localized neuropeptides,

substance P, thyrotropin-releasing hormone, and nor-

adrenaline, orexin, acetylcholine (nicotinic receptors),

and glutamate may contribute to upper airway moto-

neuronal excitation, whereas acetylcholine (through

different receptor subtypes), glycine, GABA, and

perhaps enkephalin may contribute to upper airway

motoneuronal suppression [30–44]. Sleep state-

dependent reductions in upper airway motoneuronal

activity may reflect changes in inhibitory, excitatory,

or both inputs. One model used to explore the neuro-

chemical changes in motoneurons during sleep has

been the pontine carbachol model of rapid eye

movement (REM)-associated atonia. This model pro-

duces many of the phenomena of REM sleep, includ-

ing suppression in respiratory muscles in a manner

similar to natural, or spontaneous, REM sleep (upper

airway muscle activity is more suppressed than dia-

phragmatic activity) [32–35].

In models of carbachol REM atonia, serotonin and

noradrenaline delivery are reduced to hypoglossal

motor neurons coincident with upper airway motor

neuron suppression [32,35]. Kubin et al have shown

that carbachol suppression of hypoglossal nerve

activity may be largely prevented by pretreating the

hypoglossal nucleus with serotonin [31]. Further

evidence that sleep-dependent serotonin withdrawal

contributes to suppression of upper airway dilator

activity is shown in research on adult rats, in which

serotonin delivered by way of a chronic microdialysis

probe into the hypoglossal nucleus largely prevents

genioglossus suppression in spontaneous NREM

sleep and reduces the suppression in REM sleep,

albeit to a lesser extent [45]. Serotonin is important

for the maintenance of patent airways in an animal

model of obstructive sleep-disordered breathing, the

English bulldog [46], and a combination of seroto-

nergic drugs that increase serotonin production and

release within the brain and target multiple serotonin

receptor subtypes reduces obstructive sleep-disor-

dered events in the bulldogs [47]. Serotonin may

have excitatory and inhibitory effects at motoneurons

and on respiration [48,49], and there are at least 15

Fig. 2. The effects of glossopharyngeal nerve stimulation on pharyngeal patency in a decerebrate cat. Caudal view from

pharyngoscopy in a tracheostomized cat. The left panel shows velopharyngeal and oropharyngeal patency in the absence of nerve

stimulation. The right panel illustrates the large effect of glossopharyngeal nerve stimulation on the patency of the velopharynx and

oropharynx. Stimulation of the glossopharyngeal nerve extends most pharyngeal dimensions. Although this nerve innervates

primarily the stylophyryngeus, it also contributes to the innervation of the anterior digastric, levator palatine, and stylohyoid, a

collection of muscle that would extend all airway dimensions. (Courtesy of Sam Kuna, MD, University of Pennsylvania,

Philadelphia, PA)

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 181

unique serotonin receptor subtypes within the central

nervous system in mammals [50]. Researchers re-

cently identified which excitatory serotonin receptor

subtypes are involved in postsynaptic serotonergic

excitation of hypoglossal motor neurons [51,52].

5-HT2A and 2C are the excitatory 5-HT receptors

transcribed in hypoglossal motoneurons and the only

functional excitatory receptors [51,52]. Kubin et al

identified a1B as the main postsynaptic noradrenergic

receptor subtype that mediates excitation [53]. Sub-

stance P excites hypoglossal motoneurons through

activation of the natural killer-1 receptor [42].

Glutamatergic excitation of hypoglossal moto-

neurons involves multiple receptor subtypes in the

hypoglossal nucleus [54,160–163], and although

reports have identified the presence of many different

receptor subtypes [55–59], it remains unclear what

the relative role of each subtype is. Recent studies

suggested that N-methyl-D-aspartate (NMDA) recep-

tor subtypes are particularly vulnerable to nitrosative

and oxidative stress and that the excitability of this

receptor is reduced in oxidative stress through nitro-

sative and oxidative changes in the sodium channel, a

mechanism believed to be protective in preventing

glutamatergic excitotoxicity [60]. Because glutamate

is involved in the respiratory drive to hypoglossal and

other respiratory neurons, it is essential to understand

glutamatergic control of upper airway motoneurons

and how OSAHS impacts on glutamate receptor

function [159].

Although glycine plays a major role in REM sleep

postural muscle suppression, it does not seem to

contribute significantly to either the pontine carba-

chol REM suppression of hypoglossal activity [61] or

spontaneous REM suppression of brain stem motor

reflex activity [62]. There are large hyperpolariza-

tions of brain stem motor neurons during phasic REM

sleep [63], which may occur when glycine contrib-

utes to upper airway dilator muscle suppression [64].

In summary, recent studies have identified the

subtypes for monoaminergic excitatory inputs to

hypoglossal motoneurons in an effort to identify drug

targets. Unfortunately, the predominant and non-

rapidly desensitizing serotonin receptor subtype

involved in excitation of hypoglossal motoneurons

in normal mammals, 5-HT2A, is not an ideal target for

pharmacotherapies, because activation of this receptor

subtype has been implicated in vasoconstriction of the

systemic and pulmonary beds, bronchospasm, and

thromboembolic disease [65]. The adrenergic receptor

subtype involved in hypoglossal excitation, alpha1B,

is also implicated in vasoconstriction [66]. A thyrot-

ropin-releasing hormone analog has been tried in the

English bulldog model of sleep-disordered breathing

and found to increase wakefulness without improving

sleep-disordered breathing (S.C. Veasey, unpublished

observations). To date, there are no ideal receptor

targets for pharmacotherapeutics for OSAHS.

With the certainty that the clinical description of the

manifestations of OSAHS is not complete, one also

may be sure that the list of neurochemicals involved

directly in the control of upper airway motoneurons is

not complete. Many ‘‘orphan’’ G protein-coupled

receptors and other potential targets for drug therapies

for OSAHS exist. Researchers currently are probing

upper airway motoneuronal tissue for novel receptors

with activity at upper airway dilator motoneurons

because they may provide additional avenues for

pharmacotherapies for this disorder.

It is crucial to determine how OSAHS alters

neuronal function. There are recent reports of long-

term intermittent hypoxia inducing neuronal injury

and reducing excitatory responsiveness in hippocam-

pal neurons [67,68]. There is at least one report of

patients with OSAHS showing delayed phrenic nerve

conduction, which is associated with severe oxy-

hemoglobin desaturations [69], suggesting that per-

haps oxidative injury occurs to the respiratory motor

neurons with severe OSAHS. Motor neurons are

sensitive to oxidative stress, and one likely mech-

anism of disease progression in persons with OSAHS

is oxidative injury to respiratory neurons and upper

airway dilator motor neurons. Advancing knowledge

concerning the neurochemical control of upper air-

way dilator motor neurons in sleep requires an un-

derstanding of the major inputs to motor neurons.

Respiratory neural inputs to the upper airway

motoneurons are numerous and include reflexes,

respiratory drive, and other central inputs. Responses

to reflexes may be excitatory or inhibitory, fast or

slow adapting responses. There is evidence that sleep

may modulate upper airway activity through many of

these mechanisms [70–88]. Readers are referred to

excellent review chapters [89]. One example in which

a rapid reflex response may play an important role in

upper airway patency in sleep is the immediate

response (first 200–300 milliseconds) to increased

respiratory loads. This augmentation of upper airway

muscle activity is not evident in non-REM sleep [70].

Loss of an initial powerful excitatory drive to the

upper airway muscles could reduce substantially the

effectiveness of pump muscle activity. In the English

bulldog model of obstructive sleep-disordered breath-

ing, the lead-time for upper airway muscles before

diaphragmatic activation that occurs upon waking is

lost in non-REM and REM sleep [21]. The relative

role that this reflex plays in waking respiratory drive

to upper airway muscles in persons with OSAHS is

S.C. Veasey / Clin Chest Med 24 (2003) 179–193182

largely unknown. There is evidence for a significant

contribution of a slow adapting reflex response,

mechanoreceptor reflex activation, to waking dilator

muscle activity in persons with OSAHS. When

topical anesthesia is applied to the pharyngeal

mucosa, electromyographic activity of upper airway

dilators and airway caliber declines in normal persons

and persons with OSAHS [71,72]. In both groups, the

apnea-hypopnea index increases [71,72].

Sleep also affects the pharyngeal muscle reflex

response to negative pressure [73–81]. Evidence that

this reflex contributes to waking genioglossus activity

is apparent because the application of positive pres-

sure abruptly (within a reflex latency) drops genio-

glossus activity in persons with OSAHS [75,81].

Effects of sleep on suppression of the negative

pressure reflex are more pronounced in REM sleep

than in non-REM sleep [77,78]. It is unclear, how-

ever, whether the sleep effect is a primary effect on

reflex inactivation or whether this is secondary to

sleep-induced reductions in upper airway motor neu-

ron excitability [79].

There is some evidence that reflex responses may

be impaired in persons with OSAHS. One recent

report suggested that long-term severe OSAHS is

associated with swallowing dysfunction [79]. The

swallowing reflex impairment was associated with

more frequent severe oxyhemoglobin desaturations

and is improved in patients after successful continu-

ous positive airway pressure (CPAP) therapy [79].

The negative pressure reflex response is also impaired

in OSAHS and improves with CPAP therapy [80]. It

is likely that in addition to impairments in respiratory

motor neurons, OSAHS may result in impairments in

important upper airway reflex responses. This is an

area in need of further exploration.

Another group of neurons affected by sleep and

likely by OSAHS is the upper respiratory neurons.

Collectively, the work from many studies suggests

that sleep may have larger suppressive effects on

pontine respiratory neurons [84,87], some of which

rely on serotonergic inputs [86]. There are little to no

suppressive effects on medullary neurons; in cats,

medullary respiratory neurons may increase firing

during REM sleep [83,85]. The large changes in

upper airway motor activity in sleep are most con-

sistent with tonic reductions in monoaminergic inputs

from nonrespiratory groups and perhaps phasic

increases in glycinergic drive through activation of

glycinergic interneurons. The reduced chemosensitiv-

ity in sleep is also poorly understood. It is not

because of sleep-related effects on nucleus tractus

solitarius response to hypercarbia [88]. Medullary

serotonergic neurons are chemosensitive, and because

firing of these neurons occurs less during sleep, this

could contribute to reduced chemosensitivity in sleep.

OSAHS may injure upper respiratory neurons and

alter drive to dilator and pump muscles. In neonatal

rats exposed to intermittent hypoxia, nucleus tractus

solitarius neurons show substantial injury, including

apoptosis [68]. Functional magnetic imaging in adults

with OSAHS reveals loss of grey matter in brain

regions involved in respiratory drive [90]. Whether

this is a consequence of OSAHS, or whether the

lesions render persons more vulnerable to OSAHS, is

presently unknown. The above referenced study in

young rats suggested a narrow window of increased

vulnerability, and whether clinically significant injury

may occur at later stages is presently unknown.

Overall, upper airway and other protective respi-

ratory reflexes are lost in sleep, and reduced or absent

reflex responses and respiratory neuronal injury may

contribute to the pathogenesis of OSAHS. How much

of a role these reflexes play remains unknown. It is

important to determine how much waking upper

airway dilator muscle activity is present because of

specific reflex activation in humans with OSAHS.

This is important to determine in persons with

OSAHS because the neurochemical control of reflex

activity may differ significantly from the neurochem-

ical control for central mechanisms. If reflexes con-

tribute substantially to upper airway activity in persons

with OSAHS, then the neurochemical basis for sig-

nificant reflexes may be determined in animals and

targeted to provide therapeutic targets. Differences

among patients in relative roles of reflex and central

inputs may explain partly the differential responses to

pharmacotherapies. At the same time, it is important to

understand which neurons are injured by OSAHS and

how this injury occurs.

The neurochemical control of upper airway re-

flexes is not well delineated, but it seems that nor-

adrenaline and serotonin may contribute to inhibitory

[91,92] and excitatory upper airway motor responses

for trigeminal nerve reflexes [93]. Serotonin does not

seem to contribute to the superior laryngeal nerve

stimulatory response of hypoglossal motor neurons

[94]. Glutamate contributes to excitatory responses

[95,96]; however, few other upper airway motoneuro-

nal excitatory receptor targets have been excluded

from reflex contribution, and this is an area in need

of further study.

Upper airway anatomy

One of the challenges for studying upper airway

anatomy in persons with OSAHS has been the state

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 183

dependency of the upper airway anatomy. Specif-

ically, the upper airway is sufficiently patent in

wakefulness to allow normal ventilatory function,

and it is only during sleep, or anesthesia, that airway

collapse manifests. The following studies describe the

anatomy of the upper airway in awake normal sub-

jects and distinguish the unique features of the

waking upper airway in persons with OSAHS before

characterizing the features of the sleeping upper air-

way anatomy in persons with OSAHS.

The upper airway extends from the nares to the

vocal cords. Upper airway collapse, however, occurs

most frequently within the oropharynx, which

extends from the posterior edge of the hard palate

to the level of the cervical esophagus and glottic

inlet [97,164]. The anatomy described in this section

is the anatomy of the oropharynx with an emphasis

on the two more collapsible segments, the retropal-

atal and retroglossal airway, both of which are

surrounded by abundant soft tissues. The hypophar-

ynx has been identified as a site of collapse. Typ-

ically, however, the hypopharynx is not a primary

site of collapse. The posterior wall of the oropharynx

is comprised of mucosal tissue encompassed by

various posterior pharyngeal constrictors (muscles

that narrow the airway somewhat but also stiffen

the wall). The lateral walls of the oropharynx include

mucosal folds, a continuation of the constrictor

muscles, tonsils, tonsillar pillars, other lymphoid

tissue, and the parapharyngeal fat pads. The anterior

wall of the oropharynx consists of mucosa, the soft

palate, and the tongue. Many of the soft tissues that

surround the upper airway are surrounded, in turn,

by fixed skeletal structures, including the skull base,

maxilla, mandible, and cervical vertebral column.

There are many potential causes of upper airway

compromise, and many anatomic variations have

been associated with OSAHS, including retrogna-

thia, maxillary retropositioning, intranasal obstruc-

tion, caudal displacement of the hyoid bone,

macroglossia, a low-lying or enlarged soft palate,

enlarged lymphoid tissue in the upper oropharynx,

and brachycephalic posture [97–99].

Evidence supports the hypothesis that genetic

variations in skeletal head and neck structures con-

tribute to the likelihood of OSAHS. Several genetic

disorders with craniofacial anomalies are associated

with an increased risk of OSAHS, including cranio-

facial microsomia, Down syndrome, Pierre Robin

syndrome, Nager syndrome, Treacher Collins syn-

drome, and cri du chat syndrome [100–102]. There

are racial differences in the skeletal anomalies asso-

ciated with OSAHS. Hispanics, relative to white

adults, have on average smaller anteroposterior and

lateral dimensions for the maxilla and mandible [98].

Support that the smaller facial bones may contribute

to a predisposition to OSAHS stems from the in-

creased prevalence for OSAHS in Hispanics [98,103].

In many patients with OSAHS, however, obvious

craniofacial anomalies are not evident [98]. For

example, African Americans have on average larger

mandibular and maxillary inner dimensions relative

to whites, but the median respiratory disturbance index

is higher in African-American adult men compared

with white adult men [104]. Collectively, these data

suggest that the skeletal predispositions to OSAHS are

multifactorial; there are genetic influences on facial

skeletal structure that might increase the likelihood of

developing OSAHS, but skeletal structural variances

cannot explain all cases of OSAHS.

In addition to skeletal anatomic variations, there

are soft tissue differences in persons with OSAHS

(Fig. 3), and significant evidence supports the hy-

pothesis that changes in the upper airway soft tissue

anatomy also may predispose an individual to the

pathogenesis of OSAHS [98,99]. As with skeletal

changes, the sources of soft tissue abnormalities in

persons with OSAHS are numerous. It is difficult,

however, to determine which of the soft tissue

changes contribute to the disease process and which

are secondary to repeated upper airway obstruction.

For example, one tissue change in OSAHS is edema,

not only in the mucosa and submucosa but also in the

upper airway muscles, as evidenced by MRI of the

pharynx and neck muscles with T2 relaxation mea-

surements [105]. Edema could be caused by upper

airway negative pressure trauma but also could wor-

sen OSAHS by reducing airway caliber. Fatty infil-

tration of upper airway soft tissues is likely to play a

causal role in upper airway compromise. Obesity is a

significant risk factor for OSAHS [106], and signifi-

cant weight loss in obese persons with OSAHS

reduces the severity of sleep-disordered breathing

[107]. Of obesity parameters, neck size is the stron-

gest predictor of OSAHS [108,109], and neck cir-

cumference correlates with increased dimensions of

the parapharyngeal fat pads [110]. Increased weight

gain not only augments fat in mucosal tissue but also

increases adipose tissue within upper airway muscles

[111]. Weight gain may jeopardize the upper airway

caliber by increasing soft tissue confined by skeletal

structures surrounding the airway and causing poten-

tially deleterious effects on muscle function. A larger

upper airway soft tissue volume in men may contrib-

ute to the increased prevalence of OSAHS in men

compared to women [112].

One of the most striking differences in persons

with OSAHS in wakefulness is a marked narrowing

S.C. Veasey / Clin Chest Med 24 (2003) 179–193184

of the lateral airway walls (Fig. 3) [113]. An increase

in the size of the parapharyngeal fat pads may

contribute to airway narrowing, but because the

increase in fat pad size cannot explain fully the

marked narrowing, there also must be an increase in

soft tissue edema or mucosa [158]. It is conceivable

that persons with mild upper airway narrowing mani-

fest a progression of OSAHS from soft tissue stress-

induced mucosal growth. Several growth factors in

mucosa elsewhere in the body respond to tissue

distortion with increased growth factor transcription

[114]. This concept has not been explored in human

upper airway soft tissues, however. Increased surface

area of mucosa would increase tissue collapsibility.

CPAP clearly affects soft tissue structures, and at

pressures effective to treat OSAHS, CPAP increases

the lateral wall soft tissue cross-sectional area more

so than anterior or posterior soft tissue, which sug-

gests that this region is more distensible in humans

with OSAHS [115]. An increase in upper airway

mucosal surface area may contribute to lateral wall

increased collapsibility in persons with OSAHS.

State-dependent imaging of the upper airway has

provided more clues concerning the pathogenesis of

OSAHS. By imaging persons during sleep, it is

possible to discern which structures surrounding the

upper airway might contribute to airway collapse or

narrowing. In normal persons, consistent with the

reduced upper airway muscle activity during sleep,

the upper airway dimensions decline in sleep [116].

The decline may be attributed to posterior positioning

of the tongue and soft palate and narrowing or folding

in of the lateral walls [116]. The posterior and lateral

changes are less likely to be explained by activity

reduction in one muscle. Presumably the narrowing

results from simultaneous reductions in several of the

following muscles: genioglossus, geniohyoid, tensor

veli palatini, and levator palatini. Similar dimensional

changes have been observed in persons with OSAHS

[117–119]. The reductions in upper airway caliber,

however, are more pronounced in persons with

OSAHS [117]. The larger changes in persons with

OSAHS may occur because of larger reductions in

upper airway muscle activity but also may occur as a

consequence of smaller lung volumes, which may

shorten the upper airway and allow the lateral walls to

collapse inward [120].

Imaging studies of the upper airway in persons

with and without OSAHS, particularly imaging stud-

ies performed during sleep, have provided a char-

acterization of many abnormalities of skeletal and

soft tissue origin that may contribute to OSAHS. The

abnormalities in waking are not sufficient to allow

diagnosis or consistently reliable predictions concern-

Fig. 3. Axial MRIs of the pharynx and all surrounding skeletal, soft tissue structures in a normal individual (left) and a person

with severe OSAHS (right). Notice the increased fat pads (white) in the person with OSAHS and compromise of the anterior

posterior and lateral pharyngeal walls. (Courtesy of Richard Schwab, MD, University of Pennsylvania, Philadelphia, PA)

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 185

ing which patients may benefit from various surgical

and nonsurgical therapies. Future imaging studies in

sleeping persons with OSAHS will be tremendously

insightful when measurement of specific muscle

activity and lung volume may be acquired simulta-

neously with dynamic breath-to-breath imaging

across states. The insight gained into neural control

of the upper airway and upper airway anatomy in

persons with OSAHS must be complemented with

data on mechanics to begin to approach unanswered

questions concerning state-dependent changes in

upper airway mechanics, because muscle activity

over several breaths before upper airway collapse

may not change in parallel with progressive reduc-

tions in upper airway caliber.

Upper airway mechanics

This article highlights the sleep state–dependent

reductions in upper airway dilator activity as normal

neurologic phenomena and phenomena that are more

pronounced in persons with OSAHS and result in

repetitive upper airway occlusions only in persons

with OSAHS. The author has discussed several

anatomic changes, including several genetically

determined bone and soft tissue features that may

predispose an individual to require increased upper

airway dilator activity to maintain a patent upper

airway. However, anatomy and muscle activity alone

are insufficient to explain fully the complicated

pathogenesis of OSAHS [121]. The mechanics of

the upper airway, particularly forces that alter com-

pliance and upper airway collapsibility, are equally

important in determining which patients snore and

which patients have occlusive apneas [122–124]. It is

difficult to predict reliably OSAHS severity with

either imaging or electromyographic studies. In con-

trast, several studies of upper airway biomechanics

help to distinguish snorers from persons with hypo-

pnea and persons with apnea [124–126].

The retropalatal and retroglossal regions of the

upper airway act much as a Starling resistor, a col-

lapsible passageway [127]. The clinical significance of

Starling properties is that variations in intraluminal

pressures, resistance, and airway collapsibility influ-

ence upper airway flow so that despite a high pulling

pressure (from inspiratory muscle activity), flow may

become limited [127,128]. Several factors influence

maximal flow in the upper airway through the col-

lapsible area [128]. First, a greater upstream (nasal)

driving pressure increases flow, because flow is some-

what proportional to the pressure gradient (nasal

pressure minus the critical closing pressure) [129].

Through this mechanism, positive airway pressure

therapies (CPAP, BiPAP, mask ventilation) work.

The increased driving pressure increases inspiratory

flow [130,131]. Nasal pressure does not differ in

normal persons and persons with OSAHS, however;

at end-expiration, this is simply atmospheric pressure.

One factor that varies among persons with and

without OSAHS is nasal or upstream resistance, and

as a Starling resistor, maximal flow is limited by

upstream resistance. If this resistance is too great,

flow ceases. In this manner, nasal obstruction may

contribute to OSAHS [132,133], although correction

of nasal resistance only rarely results in substantial

reductions in apnea/hypopnea frequencies [134]. The

third—and perhaps most influential—factor in per-

sons with OSAHS is the specific collapsing pressure

of the Starling segment [127,128]. This pressure is

termed the critical pressure, Pcrit, and is defined as the

upper airway pressure (nasal pressure) at which air

flow ceases in the collapsible segment. The upper

airway muscles come into play, and Pcrit is affected

by sleep state [121]. The dilator muscles act with

centrifugal force to produce a more negative closing

pressure, a less collapsible segment. Even in normal

persons, the effects of sleep are pronounced on upper

airway collapsibility and may change the Pcrit from

� 40 cm H2O when awake to � 15 cm H2O during

sleep [121]. In sleep the Pcrit can be used to distin-

guish types of obstructive sleep-disordered breathing.

Snorers have a Pcrit closer to � 6 cm H2O, whereas in

persons with hypopnea, the Pcrit is more positive,

closer to � 2 cm H2O. In persons with predominantly

apneas, the Pcrit actually may be above atmospheric

pressure during sleep [121]. The frequency of

obstructive sleep-disordered breathing events corre-

lates somewhat with the Pcrit [123].

Collapsibility of a Starling resistor also may vary

with lengthening or shortening of the tube (pharyngeal

mucosa/submucosa). The collapsible portion of the

upper airway may be thought of as a tube that, under

some circumstances, is too long for the space within it

is housed, and under these circumstances the walls of

the tube are redundant with many folds of tissue. The

upper airway space may be shortened by reductions in

lung volume [135–142]. Sleep may impose reduced

lung volume through two mechanisms: reducing end-

expiratory lung volume and reducing tidal volume

[120,143]. Functional residual volume or end-expira-

tory volume may be reduced in sleep because of

supine posturing and less activity to tonic respiratory

muscles, including the external intercostals [143].

Phillipson et al examined the upper airway in awake

subjects with OSAHS and in controls at several lung

volumes using acoustic reflection, and they found

S.C. Veasey / Clin Chest Med 24 (2003) 179–193186

reductions in pharyngeal cross-sectional area in nor-

mal persons and in persons with OSAHS from total

lung capacity to residual volume [120]. The reduction

was greater in persons with OSAHS [120].

Begle et al extended these findings to show that

increasing lung volume (0.5 L) reduces the pharyn-

geal resistance in non-REM sleep despite reductions

in genioglossus electromyographic phasic and tonic

activity [137]. Increasing the functional residual

capacity reduces obstructive sleep-disordered breath-

ing event frequency [135]. A major effect of CPAP

therapy is pneumatic splinting [140]. The second

mechanism through which sleep reduces lung volume

is reduction in tidal volume [143]. Tidal volume is

reduced in non-REM sleep and reduced even further

in REM sleep in persons with OSAHS [144]. Sleep-

related reductions in lung volume impose additional

challenges on an already highly vulnerable airway in

persons with OSAHS. Through reduction in lung

volume it is possible to reduce the upper airway

caliber profoundly.

The effect of supine positioning on the pharyngeal

cross-sectional area is independent of the lung vol-

ume and is likely additive [145,146]. It is surprising

that little is known about the effects of upright

posturing on OSAHS (many patients prefer this

sleeping position). In one small study, resolution of

OSAHS was shown in half of the subjects, whereas

the rest of the subjects had significant reductions in

sleep-disordered breathing [147]. It is more likely that

upright posture for sleep might represent a supple-

mental therapy for patients in whom high positive

airway pressures are required or in whom other

therapies are only partially effective.

An additional factor for upper airway mechanics

is upper airway hysteresis. This is a minimally

explored area, with the exception of several topical

oropharyngeal lubricant therapy studies for sleep-

disordered breathing. In the upper airway, particularly

in the oropharynx, there are redundant folds. With

airway collapse and even with end-expiration when

the upper airway is smallest, the number of folds or

contact areas increases. Each of these folds represents

a potential contact area for the development of

hysteresis. Part of the airway compromise relates to

sleep state –dependent changes in upper airway

dilator activity [148]. Progressive hysteresis within

the upper airway would partially explain the dissoci-

ation between upper airway dilator activity and upper

airway caliber in the last few breaths preceding an

apneic event [121,149,150]. Lubricants that may

reduce surface tension on pharyngeal mucosa have

been shown to reduce apneic and hypopneic events

and snoring [151,152].

Upper airway muscle function

Many muscle disorders predispose to sleep apnea,

including OSAHS [153]. Evidence also exists that the

disease process itself may result in injury to the upper

airway dilator muscles. In individuals with OSAHS,

upper airway dilator muscle activity is required for

airway patency. In quiet wakefulness, the drive to

upper airway muscles is relatively constant compared

to sleep. During sleep, the drive to upper airway

muscles fluctuates with each obstructive event, some-

times reaching tremendously high levels of activity at

the termination of an event. Intense activation of

upper airway muscle activity at a time when intra-

luminal pressure is low may cause muscle injury.

That is, the centrifugal force of the dilator muscles is

opposed by the centripetal force of negative intra-

luminal pressure. Mechanical lengthening of a muscle

during contraction (eccentric contraction) may injure

the muscle [154].

Petrof hypothesized that eccentric contraction may

occur in persons and in English bulldogs with

OSAHS and that evidence of eccentric contraction

injury should be seen on biopsy specimens of upper

airway dilator muscles. Petrof also observed an

increased proportion of fast twitch fibers, increased

inflammation throughout the upper airway dilator

muscles, increased connective tissue, and a signifi-

cant reduction in muscle fibers in bulldog compared

to control dog airway muscles [155]. These findings

are consistent with an overuse injury [154] to upper

airway muscles in the bulldog. The increase in

myosin type II fibers in the sternohyoid muscle is

consistent with resistive load training of this dilator

muscle [156]. There were no differences in myosin

type in a non–upper airway striated muscle, the

anterior tibialis. Petrof concluded that eccentric con-

traction of upper airway muscles over a long time,

seen particularly in older dogs, may result in muscle

injury, which could help explain progression of

disease. Injury specific to upper airway muscles

rather than diffusely has been shown by Dr. Schot-

land and colleagues [165]. Intermittent hypoxia also

may increase fatigability of upper airway dilator

muscle, as shown recently in adult rats exposed to

5 weeks of intermittent hypoxia [157].

Summary

Obstructive sleep apnea-hypopnea syndrome

occurs because of various combinations of anatomic,

mechanical, and neurologic anomalies that jeopardize

ventilation only when normal state-dependent reduc-

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 187

tions in drive to upper airway respiratory muscles and

pump muscles occur. Awell thought out and carefully

described infrastructure of the normal and abnormal

physiology in persons with OSAHS has been

developed over the past few decades, which enables

the development of innovative and largely effective

therapies. The most recent data complement the infra-

structure with the neurochemical changes underlying

the state-dependent respiratory disorder and observa-

tions that the disease process itself can impair muscles,

neural inputs, and soft tissue in a manner that has the

potential to worsen disease. Oxidative and nitrosative

stress from the repeated oxyhemoglobin desaturations

and re-oxygenations is implicated in the injury to these

tissues. An improved understanding of the mecha-

nisms through which OSAHS progresses may lead to

alternative therapies and aid in the identification of

persons at risk for disease progression.

References

[1] Gastaut H, Tassinari CA, Duron B. Etude polygraphi-

que des manifestations episodiques (hypniques et res-

piratoires) diurnes et nocturnes, du syndrome de

Pickwick. Rev Neurol 1965;112:568–79.

[2] Guilleminault C, Eldridge FL, Dement WC. Insomnia

with sleep apnea: a new syndrome. Science 1973;

181:856–8.

[3] Guilleminault C, Eldridge FL, Tilkian A, Simmons

FB, Dement WC. Sleep apnea syndrome due to upper

airway obstruction: a review of 25 cases. Arch Intern

Med 1977;137:296–300.

[4] Remmers JE, deGroot WJ, Sauerland EK, Anch AM.

Pathogenesis of upper airway occlusion during sleep.

J Appl Physiol 1978;44:931–8.

[5] Sanders MH, Moore SE, Eveslage J. CPAP via nasal

mask: a treatment for occlusive sleep apnea. Chest

1983;83:144–5.

[6] Issa FG, Sullivan CE. Upper airway closing pressures

in obstructive sleep apnea. J Appl Physiol 1984;57:

520–7.

[7] Kuo PC, West RA, Bloomquist DS, McNeil RW. The

effect of mandibular osteotomy in three patients with

hypersomnia sleep apnea. Oral Surg Oral Med Oral

Pathol 1979;48:385–92.

[8] Guilleminault C, Simmons FB, Motta J, Cummiskey

J, Rosekind M, Schroeder JS, et al. Obstructive sleep

apnea syndrome and tracheostomy: long-term follow-

up experience. Arch Intern Med 1981;141:985–8.

[9] Guilleminault C, Hayes B, Smith L, Simmons FB.

Palatopharyngoplasty and obstructive sleep apnea

syndrome. Bull Eur Physiopathol Respir 1983;19:

595–9.

[10] Rubin AH, Eliaschar I, Joachim Z, Alroy G, Lavie P.

Effects of nasal surgery and tonsillectomy on sleep

apnea. Bull Eur Physiopathol Respir 1983;19:612–5.

[11] Cartright R, Samelson C. The effects of nonsurgical

treatment for obstructive sleep apnea: the tongue re-

taining device. JAMA 1983;248:705–9.

[12] Meier-Ewert K, Schafer H, Kloe W. Treatment of

sleep apnea by a mandibular protracting device. Be-

richtsband 7th European Congress on Sleep Apnea

Research. Munich; 1984. p. 217.

[13] Shepard Jr JW. Gas exchange and hemodynamics

during sleep. Med Clin North Am 1985;69:1243–64.

[14] Chaudhary BA, Nadimi M, Chaudhary TK, Speir

WA. Pulmonary edema due to obstructive sleep ap-

nea. South Med J 1984;77:499–501.

[15] Tilkian AG, Guilleminault C, Schroeder JS, Lehrman

KL, Simmons FB, Dement WC. Hemodynamics in

sleep-induced apnea: studies during wakefulness

and sleep. Ann Intern Med 1976;85:714–9.

[16] Jennum P, Wildschiodtz G, Christensen NJ, Schwartz

T. Blood pressure, catecholamines, and pancreatic

polypeptide in obstructive sleep apnea with and

without nasal continuous positive airway pressure

(nCPAP) treatment. Am J Hypertens 1989;2(11 Pt 1):

847–52.

[17] Somers VK, Mark AL, Abboud FM. Sympathetic

activation by hypoxia and hypercapnia: implications

for sleep apnea. Clin Exp Hypertens 1988;10(Suppl):

413–22.

[18] Fletcher EC, Miller J, Schaaf JW, Fletcher JG. Uri-

nary catecholamines before and after tracheostomy in

patients with obstructive sleep apnea and hyperten-

sion. Sleep 1987;10:35–44.

[19] Wiegand DA, Latz B, Zwillich CW, Wiegand L.

Geniohyoid muscle activity in normal men during

wakefulness and sleep. J Appl Physiol 1990;69:

1262–9.

[20] Mezzanotte WS, Tangel DJ, White DP. Influence of

sleep onset on upper-airway muscle activity in apnea

patients versus normal controls. Am J Respir Crit

Care Med 1996;153(6 Pt 1):1880–7.

[21] Hendricks JC, Petrof BJ, Panckeri K, Pack AI. Upper

airway dilating muscle hyperactivity during non-rapid

eye movement sleep in English bulldogs. Am Rev

Respir Dis 1993;148:185–94.

[22] Morrell MJ, Arabi Y, Zahn B, Badr MS. Progressive

retropalatal narrowing preceding obstructive apnea.

Am J Respir Crit Care Med 1998;158:1974–81.

[23] Morrell MJ, Badr MS. Effects of NREM sleep on

dynamic within-breath changes in upper airway pa-

tency in humans. J Appl Physiol 1998;84:190–9.

[24] Schwab RJ, Goldberg AN. Upper airway assessment:

radiographic and other imaging techniques. Otolar-

yngol Clin North Am 1998;31:931–68.

[25] Veasey SC, Panckeri KA, Hoffman EA, Pack AI,

Hendricks JC. The effects of serotonin antagonists

in an animal model of sleep-disordered breathing.

Am J Respir Crit Care Med 1996;153:776–86.

[26] Hudgel DW. Variable site of airway narrowing among

obstructive sleep apnea patients. J Appl Physiol 1986;

61:1403–9.

[27] Shepard JW, Thawley SE. Localization of upper air-

S.C. Veasey / Clin Chest Med 24 (2003) 179–193188

way collapse during airway collapse during sleep in

patients with obstructive sleep apnea. Am Rev Respir

Dis 1990;141:1350–5.

[28] Dobbins EG, Feldman JL. Differential innervation of

protruder and retractor muscles of the tongue in rat.

J Comp Neurol 1995;357:376–94.

[29] Kuna ST. Effects of pharyngeal muscle activation on

airway size and configuration. Am J Respir Crit Care

Med 2001;164:1236–41.

[30] Hwang BH, Chang KJ, Severs WB. Increased delta,

but not mu, opiate receptor binding in the medulla

oblongata of Long-Evans rats following 5-day water

deprivation. Brain Res 1986;371:345–9.

[31] Kubin L, Davies RO, Pack AI. Control of upper air-

way motor neurons during REM sleep. News Physiol

Sci 1998;13:91–7.

[32] Lai YY, Kodama T, Siegel JM. Changes in mono-

amine release in the ventral horn and hypoglossal

nucleus linked to pontine inhibition of muscle tone:

an in vivo microdialysis study. J Neurosci 2001;21:

7384–91.

[33] Donato R, Nistri A. Relative contribution by GABA

or glycine to Cl(-)-mediated synaptic transmission on

rat hypoglossal motoneurons in vitro. J Neurophysiol

2000;84:2715–24.

[34] Yamuy J, Fung SJ, Xi M, Morales FR, Chase MH.

Hypoglossal motoneurons are postsynaptically in-

hibited during carbachol-induced rapid eye movement

sleep. Neuroscience 1999;94:11–5.

[35] Kubin L, Reignier C, Tojima H, Taguchi O, Pack AI,

Davies RO. Changes in serotonin level in the hypo-

glossal nucleus region during carbachol-induced ato-

nia. Brain Res 1994;645:291–302.

[36] Berger AJ, Bayliss DA, Viana F. Modulation of neo-

natal rat hypoglossal motoneuron excitability by se-

rotonin. Neurosci Lett 1992;143:164–8.

[37] Kubin L, Tojima H, Davies RO, Pack AI. Seroto-

nergic excitatory drive to hypoglossal motor neurons

in the decerebrate cat. Neurosci Lett 1992;139:

243–8.

[38] Douse MA, White DP. Serotonergic effects on hypo-

glossal neural activity and reflex responses. Brain Res

1996;726:213–22.

[39] Bayliss DA, Viana F, Talley EM, Berger AJ. Neuro-

modulation of hypoglossal motor neurons: cellular and

developmental mechanisms. Respir Physiol 1996;

110:139–50.

[40] Fung SJ, Yamuy J, Sampogna S, Morales FR, Chase

MH. Hypocretin (orexin) input to trigeminal and

hypoglossal motoneurons in the cat: a double-label-

ing immunohistochemical study. Brain Res 2001;

903(1-2):257–62.

[41] Zaninetti M, Tribollet E, Bertrand D, Raggenbass M.

Presence of functional neuronal nicotinic acetylcho-

line receptors in brainstem motoneurons of the rat.

Eur J Neurosci 1999;11:2737–48.

[42] Yasuda K, Robinson DM, Selvaratnam SR, Walsh

CW, McMorland AJ, Funk GD. Modulation of hypo-

glossal motoneuron excitability by NK1 receptor acti-

vation in neonatal mice in vitro. J Physiol 2001;

534(Pt. 2):447–64.

[43] Parkis MA, Bayliss DA, Berger AJ. Actions of nor-

epinephrine on rat hypoglossal motoneurons. J Neu-

rophysiol 1995;74:1911–9.

[44] Volgin DV, Saghir M, Kubin L. Developmental

changes in the orexin 2 receptor mRNA in hypoglos-

sal motoneurons. Neuroreport 2002;13:433–6.

[45] Jelev A, Sood S, Liu H, Nolan P, Horner RL. Micro-

dialysis perfusion of 5-HT into hypoglossal motor

nucleus differentially modulates genioglossus activity

across natural sleep-wake states in rats. J Physiol

2001;532:467–81.

[46] Veasey SC, Panckeri KA, Hoffman EA, Pack AI,

Hendricks JC. The effects of serotonin antagonists

in an animal model of sleep-disordered breathing.

Am J Respir Crit Care Med 1996;153:776–86.

[47] Veasey SC, Fenik P, Packeri K, Pack AI, Hendricks

JC. The effects of Trazodone with L-tryptophan on

sleep-disordered breathing in the English bulldog.

Am J Respir Crit Care Med 1999;160:1659–67.

[48] Singer JH, Berger AJ. Presynaptic inhibition by

serotonin: a possible mechanism for switching mo-

tor output of the hypoglossal nucleus. Sleep 1996;

19(10 Suppl):S146–9.

[49] Bayliss DA, Viana F, Talley EM, Berger AJ. Neuro-

modulation of hypoglossal motoneurons: cellular and

developmental mechanisms. Respir Physiol 1997;

10(2-3):139–50.

[50] Kroeze WK, Kristiansen K, Roth BL. Molecular bi-

ology of serotonin receptors structure and function at

the molecular level. Current Topics in Medical Chem-

istry 2002;2:507–28.

[51] Zhan G-X, Shaheen F, Mackiewicz M, Fenik P, Vea-

sey SC. Serotonergic receptor subtypes mRNA in

hypoglossal motoneurons. Neuroscience 2002;113:

145–54.

[52] Fenik P, Veasey SC. Pharmacological characterization

of serotonergic receptor activity in the hypoglossal

nucleus. Am J Respir Crit Care Med 2003;167:

563–9.

[53] Volgin DV, Mackiewicz M, Kubin L. Alpha(1B)

receptors are the main postsynaptic mediators of

adrenergic excitation in brainstem motoneurons:

a single-cell RT-PCR study. J Chem Neuroanat

2001;22:157–66.

[54] Rekling JC, Funk GD, Bayliss DA, Dong XW, Feld-

man JL. Synaptic control of motoneuronal excitabil-

ity. Physiol Rev 2000;80:767–852.

[55] Essin K, Nistri A, Magazanik L. Evaluation of GluR2

subunit involvement in AMPA receptor function of

neonatal rat hypoglossal motoneurons. Eur J Neurosci

2002;15:1899–906.

[56] Garcia Del Cano G, Gerrikagoitia I, Sarasa M, Matute

C, Martinez-Millan L. Ionotropic glutamate receptor

subunits are differentially regulated in the motoneuro-

nal pools of the rat hypoglossal nucleus in response to

axotomy. J Neurocytol 2000;29:509–23.

[57] Paarmann I, Frermann D, Keller BU, Hollmann M.

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 189

Expression of 15 glutamate receptor subunits and var-

ious splice variants in tissue slices and single neurons

of brainstem nuclei and potential functional implica-

tions. J Neurochem 2000;74:1335–45.

[58] Del Cano GG, Millan LM, Gerrikagoitia I, Sarasa M,

Matute C. Ionotropic glutamate receptor subunit dis-

tribution on hypoglossal motoneuronal pools in the

rat. J Neurocytol 1999;28:455–68.

[59] O’Brien JA, Isaacson JS, Berger AJ. NMDA and non-

NMDA receptors are co-localized at excitatory syn-

apses of rat hypoglossal motoneurons. Neurosci Lett

1997;227:5–8.

[60] Choi YB, Lipton SA. Redox modulation of the

NMDA receptor. Cell Mol Life Sci 2000;57:

1535–41.

[61] Kubin L, Kimura H, Tojima H, Davies RO, Pack AI.

Suppression of hypoglossal motoneurons during the

carbachol-induced atonia pf REM sleep is not

caused by fast synaptic inhibition. Brain Res 1993;

611:300–12.

[62] Soja PJ, Finch DM, Chase MH. Effect of inhibitory

amino acid antagonists on masseteric reflex sup-

pression during active sleep. Exp Neurol 1987;96:

178–93.

[63] Pedroarena C, Castillo P, Chase MH, Morales FR.

The control of jaw-opener motoneurons during active

sleep. Brain Res 1994;653(1-2):31–8.

[64] Chirwa SS, Stafford-Segert I, Soja PJ, Chase MH.

Strychnine antagonizes jaw-closer notoneuron IPSPs

induced by reticular stimulation during active sleep.

Brain Res 1991;547:323–6.

[65] Kihara H, Hirose K, Koganei H, Sasaki N, Yama-

moto H, Kimura A, et al. AT-1015, a novel serotonin

(5-HT)2 receptor antagonist, blocks vascular and

platelet 5-HT2A receptors and prevents the laurate-

induced peripheral vascular lesion in rats. J Cardio-

vasc Pharmacol 2000;35:523–30.

[66] Civantos Calzada B, Aleixandre de Artinano A. Al-

pha-adrenoceptor subtypes. Pharmacol Res 2001;44:

195–208.

[67] Gu XQ, Haddad GG. Decreased neuronal excitability

in hippocampal neurons of mice exposed to cyclic

hypoxia. J Appl Physiol 2001;91:1245–50.

[68] Gozal E, Row BW, Schurr A, Gozal D. Developmen-

tal differences in cortical and hippocampal vulnerabil-

ity to intermittent hypoxia in the rat. Neurosci Lett

2001;305:197–201.

[69] Lu Z, Tang X, Huang X. Phrenic nerve conduction

and diaphragmatic motor evoked potentials: evalua-

tion of respiratory dysfunction. Chin Med J (Engl)

1998;111:496–9.

[70] Henke KG, Badr MS, Skatrud JB, Dempsey JA. Load

compensation and respiratory muscle function during

sleep. J Appl Physiol 1992;72:1221–34.

[71] Berry RB, McNellis MI, Kouchi K, Light RW. Upper

airway anesthesia reduces phasic genioglossal activity

during sleep apnea. Am J Respir Crit Care Med

1997;156:127–32.

[72] McNicholas WT, Coffey M, McDonnell T, O’Regan

R, Fitzgerald MX. Upper airway obstruction during

sleep in normal subjects after selective topical oro-

pharyngeal anesthesia. Am Rev Respir Dis 1987;

135:1316–9.

[73] Malhotra A, Fogel RB, Edwards JK, Shea SA, White

DP. Local mechanisms drive genioglossus activation

in obstructive sleep apnea. Am J Respir Crit Care

Med 2000;161:1746–9.

[74] Horner RL, Innes JA, Morrell MJ, Shea SA, Guz A.

The effect of sleep on reflex genioglossus muscle

activation by stimuli of negative airway pressure in

humans. J Physiol 1994;476:141–51.

[75] Mathew OP, Abu-Osba YK, Thach BT. Influence of

upper airway pressure changes on genioglossusmuscle

respiratory activity. J Appl Physiol 1982;52:438–44.

[76] Henke KG, Sullivan CE. Effects of high-frequency

pressure waves applied to upper airway on respiration

in central apnea. J Appl Physiol 1992;73:1141–5.

[77] Shea SA, Edwards JK, White DP. Effect of wake-

sleep transitions and rapid eye movement sleep on

pharyngeal muscle response to negative pressure in

humans. J Physiol 1999;520(Pt 3):897–908.

[78] Eastwood PR, Curran AK, Smith CA, Dempsey JA.

Effect of upper airway negative pressure on inspira-

tory drive during sleep. J Appl Physiol 1998;84:

1063–75.

[79] Teramoto S, Ishii T, Matsuse T. Relationship between

swallowing function and gas exchange during day

and night in patients with obstructive sleep apnea

syndrome. Dysphagia 2001;16:249–53.

[80] Mortimore IL, Douglas NJ. Palatal muscle EMG re-

sponse to negative pressure in awake sleep apneic and

control subjects. Am Med J Respir Crit Care Med

1997;156(3 Pt 1):867–73.

[81] Janczewski WA. Muscle relaxation attenuates the re-

flex response to laryngeal negative pressure. Respir

Physiol 1997;107:210–30.

[82] Deegan PC, Nolan P, Carey M, McNicholas WT. Ef-

fects of positive airway pressure on upper airway

dilator muscle activity and ventilatory timing. J Appl

Physiol 1996;81:470–9.

[83] Orem J. Medullary respiratory neuron activity: rela-

tionship to tonic and phasic REM sleep. J Appl Phys-

iol 1980;48:54–65.

[84] Lee LH, Friedman DB, Lydic R. Respiratory nuclei

share synaptic connectivity with pontine reticular re-

gions regulating REM sleep. Am J Physiol 1995;268

(2 Pt 1):L251–62.

[85] Kubin L, Kimura H, Tojima H, Pack AI, Davies RO.

Behavior of VRG neurons during the atonia of REM

sleep induced by pontine carbachol in decerebrate

cats. Brain Res 1992;592(1-2):91–100.

[86] Al-Zubaidy ZA, Erickson RL, Greer JJ. Serotonergic

and noradrenergic effects on respiratory neural dis-

charge in the medullary slice preparation of neonatal

rats. Pflugers Arch 1996;431:942–9.

[87] Ni HF, Zhang JX, Harper RM. Respiratory-related

discharge of periaqueductal gray neurons during

sleep-waking states. Brain Res 1990;511:319–25.

S.C. Veasey / Clin Chest Med 24 (2003) 179–193190

[88] Nattie EE, Li A. CO2 dialysis in nucleus tractus sol-

itarius region of rat increases ventilation in sleep and

wakefulness. J Appl Physiol 2002;92:2119–30.

[89] Kubin L, Davies RO. Mechanisms of airway hypoto-

nia. In: Pack AI, editor. Sleep apnea: pathogenesis,

diagnosis and treatment. New York: Marcel Dekker;

2000. p. 99–107.

[90] Macey PM, Henderson LA, Macey KE, Alger JR,

Frysinger RC, Woo MA, et al. Brain morphology

associated with obstructive sleep apnea. Am J Respir

Crit Care Med 2002;166:1382–7.

[91] Kessler JP, Jean A. Effect of catecholamines on the

swallowing reflex after pressure microinjections into

the lateral solitary complex of the medulla oblongata.

Brain Res 1986;386(1-2):69–77.

[92] Kessler JP, Jean A. Inhibition of the swallowing re-

flex by local application of serotonergic agents into

the nucleus of the solitary tract. Eur J Pharmacol

1985;118(1-2):77–85.

[93] Ribeiro-do-Valle LE, Metzler CW, Jacobs BL. Facil-

itation of masseter EMG and masseteric (jaw-closure)

reflex by serotonin in behaving cats. Brain Res

1991;550:197–204.

[94] Jiang C, Mitchell GS, Lipski J. Prolonged augmenta-

tion of respiratory discharge in hypoglossal motoneu-

rons following superior laryngeal nerve stimulation.

Brain Res 1991;538:215–25.

[95] Ambalavanar R, Purcell L, Miranda M, Evans F, Lud-

low CL. Selective suppression of late laryngeal

adductor responses by N-methyl-D-aspartate receptor

blockade in the cat. J Neurophysiol 2002;87:1252–62.

[96] Karius DR, Speck DF. Short-latency excitation of

phrenic motor output mediated by non-NMDA recep-

tors. Brain Res 1995;682(1-2):235–8.

[97] Schwab RJ, Goldberg AN. Upper airway assessment:

radiographic and other imaging techniques. Otola-

ryngol Clin North Am 1998;31:931–68.

[98] Will MJ, Ester MS, Ramirez SG, Tiner BD, McAnear

JT, Epstein L. Comparison of cephalometric analysis

with ethnicity in obstructive sleep apnea syndrome.

Sleep 1995;18:873–5.

[99] Schwab RJ. Imaging for the snoring and sleep apnea

patient. Dent Clin North Am 2001;45:759–96.

[100] Guilleminault C, Khramtsov A. Upper airway resist-

ance syndrome in children: a clinical review. Semin

Pediatr Neurol 2001;8:207–15.

[101] Suzuki K, Yamamoto S, Ito Y, Baba S. Sleep apnea

associated with congenital diseases and moderate hy-

pertrophy of tonsils. Acta Otolaryngol Suppl 1996;

523:225–7.

[102] Argenta LC, Iacobucci JJ. Treacher Collins syn-

drome: present concepts of the disorder and their

surgical correction. World J Surg 1989;13:401–9.

[103] Kripke DF, Ancoli-Israel S, Klauber MR, Wingard

DL, Mason WJ, Mullaney DJ. Prevalence of sleep-

disordered breathing in ages 40–64 years: a popu-

lation-based survey. Sleep 1997;20:65–76.

[104] Buxbaum SG, Elston RC, Tishler PV, Redline S. Ge-

netics of the apnea hypopnea index in caucasians and

African Americans: I. Segregation analysis. Genet

Epidemiol Suppl 2002;22:243–53.

[105] Schotland HM, Insko EK, Schwab RJ. Quantitative

magnetic resonance imaging demonstrates alterations

of the lingual musculature in obstructive sleep apnea.

Sleep 1999;22:605–13.

[106] Young T, Peppard PE, Gottlieb DJ. Epidemiology of

obstructive sleep apnea: a population health per-

spective. Am J Respir Crit Care Med 2002;165:

1217–39.

[107] Livingston EH. Obesity and its surgical management.

Am J Surg 2002;184:103–13.

[108] Schellenberg JB, Maislin G, Schwab RJ. Physical

findings and the risk for obstructive sleep apnea:

the importance of oropharyngeal structures. Am J

Respir Crit Care Med 2000;162(2 Pt 1):740–8.

[109] Young T, Palta M, Dempsey J, Skatrud J, Weber S,

Badr S. The occurrence of sleep-disordered breathing

among middle-aged adults. N Engl J Med 1993;328:

1230–5.

[110] Shelton KE, Woodson H, Gay S, Suratt PM. Pharyn-

geal fat in obstructive sleep apnea. Am Rev Respir

Dis 1993;148:462–6.

[111] Janssen I, Heymsfield SB, Wang ZM, Ross R. Skeletal

muscle mass and distribution in 468 men and women

aged 18–88 yr. J Appl Physiol 2000;89:81–8.

[112] Rowley JA, Sanders CS, Zahn BR, Badr MS. Gender

differences in upper airway compliance during

NREM sleep: role of neck circumference. J Appl

Physiol 2002;92:2535–41.

[113] Schwab RJ. Properties of tissues surrounding the

upper airway. Sleep 1996;19(10 Suppl):S170–4.

[114] Konturek PC, Brzozowski T, Konturek SJ, Taut A,

Sliwowski Z, Stachura J, et al. Activation of genes

for growth factors and cyclooxygenases in rat gastric

mucosa during recovery from stress damage. Eur J

Pharmacol 1998;342:55–65.

[115] Schwab RJ, Pack AI, Gupta KB, Metzger LJ, Oh E,

Getsy JE, et al. Upper airway and soft tissue structural

changes induced by CPAP in normal subjects. Am J

Respir Crit Care Med 1996;154(4 Pt 1):1106–16.

[116] Trudo FJ, Gefter WB, Welch KC, Gupta KB, Maislin

G, Schwab RJ. State-related changes in upper airway

caliber and surrounding soft-tissue structures in nor-

mal subjects. Am J Respir Crit Care Med 1998;158:

1259–70.

[117] Horner RL, Shea SA, McIvor J, Guz A. Pharyngeal

size and shape during wakefulness and sleep in pa-

tients with obstructive sleep apnoea. Q J Med 1989;

72:719–35.

[118] Suto Y, Matsuo T, Kato T, Hori I, Inoue Y, Ogawa S,

et al. Evaluation of the pharyngeal airway in patients

with sleep apnea: value of ultrafast MR imaging. AJR

Am J Roentgenol 1993;160:311–4.

[119] Stein MG, Gamsu G, de Geer G, Golden JA, Crumley

RL, Webb WR. Cine CT in obstructive sleep apnea.

AJR Am J Roentgenol 1987;148:1069–74.

[120] Hoffstein V, Zamel N, Phillipson EA. Lung volume

dependence of pharyngeal cross-sectional area in pa-

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 191

tients with obstructive sleep apnea. Am Rev Respir

Dis 1984;130:175–8.

[121] Schwartz AR, O’Donnell CP, Baron J, Schubert N,

Alam D, Samadi SD, et al. The hypotonic upper air-

way in obstructive sleep apnea: role of structures and

neuromuscular activity. Am J Respir Crit Care Med

1998;157(4 Pt 1):1051–7.

[122] Fouke JM, Wolin AD, Strohl KP, Galbraith GM. Elas-

tic characteristics of the airway wall. J Appl Physiol

1989;66:962–7.

[123] Sforza E, Bacon W, Weiss T, Thibault A, Petiau C,

Krieger J. Upper airway collapsibility and cephalo-

metric variables in patients with obstructive sleep ap-

nea. Am J Respir Crit Care Med 2000;161(2 Pt 1):

347–52.

[124] Gleadhill IC, Schwartz AR, Schubert N, Wise RA,

Permutt S, Smith PL. Upper airway collapsibility in

snorers and in patients with obstructive hypopnea and

apnea. Am Rev Respir Dis 1991;143:1300–3.

[125] Condos R, Norman RG, Krishnasamy I, Peduzzi N,

Goldring RM, Rapoport DM. Flow limitation as a

noninvasive assessment of residual upper-airway re-

sistance during continuous positive airway pressure

therapy of obstructive sleep apnea. Am J Respir Crit

Care Med 1994;150:475–80.

[126] Liistro G, Stanescu DC, Veriter C, Rodenstein DO,

Aubert-Tulkens G. Pattern of snoring in obstructive

sleep apnea patients and in heavy snorers. Sleep

1991;14:517–25.

[127] Gleeson K, Zwillich CW, Bendrick TW, White DP.

Effect of inspiratory nasal loading on pharyngeal re-

sistance. J Appl Physiol 1986;60:1882–6.

[128] Schwartz AR, Smith PL, Gold AR, Wise RA, Permutt

S. Induction of upper airway occlusion in sleeping

individuals with subatmospheric nasal pressure.

J Appl Physiol 1988;64:535–42.

[129] Schwartz AR, Smith PL, Wise RA, Permutt S. Effect

of nasal pressure on upper airway pressure flow rela-

tionships. J Appl Physiol 1989;66:1626–34.

[130] Smith PL, Wise RA, Gold AR, Schwartz AR, Permutt

S. Upper airway pressure-flow relationships in

obstructive sleep apnea. J Appl Physiol 1988;64:

789–95.

[131] Suratt PM, Wilhoit SC, Cooper K. Induction of air-

way collapse with subatmospheric pressure in awake

patients with sleep apnea. J Appl Physiol 1984;57:

140–6.

[132] Metes A, Ohki M, Cole P, Haight JS, Hoffstein V.

Snoring, apnea and nasal resistance in men and wom-

en. J Otolaryngol 1991;20:57–61.

[133] De Vito A, Berrettini S, Carabelli A, Sellari-France-

schini S, Bonanni E, Gori S, et al. The importance of

nasal resistance in obstructive sleep apnea syndrome:

a study with positional rhinomanometry. Sleep Breath

2001;5:3–11.

[134] Verse T, Maurer JT, Pirsig W. Effect of nasal surgery

on sleep-related breathing disorders. Laryngology

2002;112:64–8.

[135] Series F, Cormier Y, Lampron N, La Forge J. Increas-

ing the functional residual capacity may reverse ob-

structive sleep apnea. Sleep 1988;11:349–53.

[136] Series F, Cormier Y, Couture J, Desmeules M.

Changes in upper airway resistance with lung infla-

tion and positive airway pressure. J Appl Physiol

1990;68:1075–9.

[137] Begle RL, Badr S, Skatrud JB, Dempsey JA. Effect

of lung inflation on pulmonary resistance during

NREM sleep. Am Rev Respir Dis 1990;141(4 Pt 1):

854–60.

[138] Aronson RM, Carley DW, Onal E, Wilborn J, Lopata

M. Upper airway muscle activity and the thoracic

volume dependence of upper airway resistance.

J Appl Physiol 1991;70:430–8.

[139] Burger CD, Stanson AW, Daniels BK, Sheedy II PF,

Shepard Jr JW. Fast-CT evaluation of the effect of

lung volume on upper airway size and function in

normal men. Am Rev Respir Dis 1992;146:335–9.

[140] Series F, Marc I. Influence of lung volume depen-

dence of upper airway resistance during continuous

negative airway pressure. J Appl Physiol 1994;77:

840–4.

[141] Eastwood PR, Szollosi I, Platt PR, Hillman DR. Col-

lapsibility of the upper airway during anesthesia with

isoflurane. Anesthesiology 2002;97:786–93.

[142] Kamal I. Lung volume dependence of pharyngeal

cross-sectional area by acoustic pharyngometry. Oto-

laryngol Head Neck Surg 2002;126:164–71.

[143] Ballard RD, Irvin CG, Martin RJ, Pak J, Pandey R,

White DP. Influence of sleep on lung volume in asth-

matic patients and normal subjects. J Appl Physiol

1990;68:2034–41.

[144] Becker HF, Piper AJ, Flynn WE, McNamara SG,

Grunstein RR, Peter JH, et al. Breathing during sleep

in patients with nocturnal desaturation. Am J Respir

Crit Care Med 1999;159:112–8.

[145] Hoffstein V, Zamel N, Phillipson EA. Lung volume

dependence of pharyngeal cross-sectional area in pa-

tients with obstructive sleep apnea. Am Rev Respir

Dis 1984;130:175–8.

[146] Fouke JM, Strohl KP. Effect of position and lung

volume on upper airway geometry. J Appl Physiol

1987;63:375–80.

[147] McEvoy RD, Sharp DJ, Thornton AT. The effects of

posture on obstructive sleep apnea. Am Rev Respir

Dis 1986;133:662–6.

[148] Mezzanotte WS, Tangel DJ, White DP. Influence of

sleep onset on upper-airway muscle activity in apnea

patients versus normal controls. Am J Respir Crit

Care Med 1996;153(6 Pt 1):1880–7.

[149] Katz I, Zamel N, Slutsky AS, Rebuck AS, Hoffstein

V. Tracheal hysteresis in sleep apnea. J Appl Physiol

1989;67:1349–53.

[150] Praud JP, D’Allest AM, Delaperche MF, Bobin S,

Gaultier C. Diaphragmatic and genioglossus electro-

myographic activity at the onset and at the end of

obstructive apnea in children with obstructive sleep

apnea syndrome. Pediatr Res 1988;23:1–4.

[151] Jokic R, Klimaszewski A, Mink J, Fitzpatrick MF.

S.C. Veasey / Clin Chest Med 24 (2003) 179–193192

Surface tension forces in sleep apnea: the role of a

soft tissue lubricant. A randomized double-blind, pla-

cebo-controlled trial. Am J Respir Crit Care Med

1998;157(5 Pt 1):1522–5.

[152] Miki H, Hida W, Kikuchi Y, Chonan T, Satoh M,

Iwase N, et al. Effects of pharyngeal lubrication on

the opening of obstructed upper airway. J Appl Phys-

iol 1992;72:2311–6.

[153] Guilleminault C, Stoohs R, Quera-Salva MA. Sleep-

related obstructive and nonobstructive apneas and

neurologic disorders. Neurology 1992;42(7 Suppl 6):

53–60.

[154] Dennett X, Fry HJH. Overuse syndrome: a muscle

biopsy study. Lancet 1988;2:905–8.

[155] Petrof BJ, Pack AI, Kelly AM, Hendricks JC. Pha-

ryngeal myopathy of loaded upper airway in dogs

with sleep apnea. J Appl Physiol 1994;76:1746–52.

[156] Prince FP, Hikida RS, Hagerman FC. Human muscle

fiber types in power lifters, distance runners and un-

trained subjects. Pflugers Arch 1977;370:227–32.

[157] McGuire M, MacDermott M, Bradford A. Effects of

chronic episodic hypoxia on rat upper airway muscle

contractile properties and fiber-type distribution.

Chest 2002;122:1012–7.

[158] Schwab RJ, Gefter WB, Hoffman EA, Gupta KB,

Pack AI. Dynamic upper airway imaging during

awake respiration in normal subjects and patients

with sleep disordered breathing. Am Rev Respir Dis

1993;148:1385–400.

[159] Yang CC, Chan JY, Chan SH. Excitatory innervation

of caudal hypoglossal nucleus from nucleus reticula-

ris gigantocellularis in the rat. Neuroscience 1995;

65:365–74.

[160] Essin K, Nistri A, Magazanik L. Evaluation of GluR2

subunit involvement in AMPA receptor function of

neonatal rat hypoglossal motoneurons. Eur J Neurosci

2002;15:1899–906.

[161] Talley EM, Lei Q, Sirois JE, Bayliss DA. TASK-1,

a two-pore domain K+ channel, is modulated by mul-

tiple neurotransmitters in motoneurons. Neuron 2000;

25:399–410.

[162] Hulsmann S, Oku Y, Zhang W, Richter DW. Metab-

otropic glutamate receptors and blockade of glial

Krebs cycle depress glycinergic synaptic currents of

mouse hypoglossal motoneurons. Eur J Neurosci

2000;12:239–46.

[163] Del Cano GG, Millan LM, Gerrikagoitia I, Sarasa M,

Matute C. Ionotropic glutamate receptor subunit dis-

tribution on hypoglossal motoneuronal pools in the

rat. J Neurocytol 1999;28:455–68.

[164] Goldberg AN, Schwab RJ. Identifying the patient

with sleep apnea: upper airway assessment and phys-

ical examination. Otolaryngol Clin North Am 1998;

31:919–30.

[165] Schotland HM, Insko EK, Panckeri KA, Leigh JS,

Pack AI, Hendricks JC. Quantitative magnetic reso-

nance imaging of upper airway musculature in an

animal model of sleep apnea. J Appl Physiol 1996;

81:1339–46.

S.C. Veasey / Clin Chest Med 24 (2003) 179–193 193

Cardiovascular consequences of obstructive sleep apnea

Robert Wolk, MD, PhD, Virend K. Somers, MD, PhD*

Mayo Clinic, Department of Medicine, Division of Cardiovascular Diseases and Division of Hypertension,

200 First Street Southwest, Rochester, MN 55905, USA

Sleep disorders are common, with an estimated

prevalence of approximately 40 million cases in the

United States alone. Fifteen million persons in the

United States are believed to have sleep apnea, which

is defined as recurrent episodes of cessation of respi-

ratory airflow during sleep, with a consequent decrease

in oxygen saturation. Sleep apnea can be considered as

central or obstructive. Central sleep apnea (CSA) is

characterized by periodic apneas and hypopneas sec-

ondary to diminution or cessation of respiratory

efforts. In contrast, obstructive sleep apnea (OSA) is

secondary to upper airway collapse during inspiration

and is accompanied by strenuous breathing efforts.

CSA and OSA often may coexist. There is an increas-

ing recognition of the widespread prevalence of OSA

and its potential cardiovascular consequences. CSA

also has been implicated in cardiovascular disease,

primarily in patients with heart failure. This article

addresses the association between OSA and specific

cardiovascular disease conditions and examines the

evidence that implicates OSA in the pathophysiology

and progression of these disorders.

Hypertension

Much work has focused on the link between sleep

apnea and hypertension, and the evidence that sug-

gests a causal association between these two condi-

tions is compelling. The prevalence of hypertension

is greater in patients with OSA, and hypertensive pa-

tients (especially the nondippers) have a higher inci-

dence of OSA [1,2], which suggests that OSA may be

etiologically linked to chronic daytime hypertension.

The evidence for a causal relationship between OSA

and daytime hypertension has been strengthened by

recent epidemiologic studies. The Wisconsin Sleep

Cohort Study demonstrated a dose-response associa-

tion between sleep-disordered breathing at baseline

(diagnosed by in-hospital polysomnography) and the

development of new hypertension 4 years later, inde-

pendent of other known risk factors [3]. Specifically,

the odds ratios for the presence of hypertension at fol-

low-up were 1.42, 2.03, and 2.89 with an apnea-hy-

popnea index of less than 5, 5 to 15, and more than

15 events/hour at baseline, respectively (Fig. 1). A

similar relationship between OSA and the risk of

hypertension was seen in other studies [4,5]. Further

support for some causal interaction between OSA and

hypertension is provided by evidence that successful

treatment of OSA with continuous positive airway

pressure (CPAP) reduces blood pressure, especially in

patients with hypertension [6–11]. Taken together,

these data suggest that OSA is likely to contribute to

hypertension in some patients and that the manage-

ment of hypertension in these patients may be aug-

mented by treating the underlying sleep apnea.

Neurogenic mechanisms may contribute impor-

tantly to the acute and chronic hypertensive effects of

OSA. Acute nocturnal surges in blood pressure occur

in response to chemoreflex-mediated hypoxic stimu-

lation of sympathetic activity [12–14]. These re-

sponses are potentiated in hypertensive subjects

[15]. Activation of the chemoreflex leads to an

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00020-0

Work for this article was funded by theMayo Foundation,

HL-61560, HL-65176, HL-70302, MO1-RR00585.

* Corresponding author. Mayo Foundation, St. Mary’s

Hospital, DO-4-350, 1216 Second Street SW, Rochester,

MN 55902.

E-mail address: somers.virend@mayo.edu

(V.K. Somers).

Clin Chest Med 24 (2003) 195–205

increase in vascular sympathetic nerve activity and

circulating catecholamines, which increase peripheral

vascular resistance. Upon termination of apnea, car-

diac output increases (caused by changes in intra-

thoracic pressures) in the presence of a constricted

peripheral vascular bed, which leads to dramatic

surges in blood pressure (sometimes to levels as high

as 240/120 mm Hg) [14]. There seems to be a

‘‘carry-over’’ effect, such that sympathetic activity

remains elevated even in normoxic conditions, serv-

ing as one of several possible mechanisms that

maintain elevated blood pressure even during day-

time wakefulness.

Daytime hypertension in OSA may be mediated

by enhanced sympathetic activity, as evidenced by

elevated circulating catecholamine levels, increased

sympathetic nerve activity [14,16–21], and other

mechanisms. Normotensive OSA patients, who are

free of any overt cardiovascular disease, have de-

creased heart rate variability and increased blood

pressure variability [22]—characteristics that may

predispose to the development of hypertension [23]

and end-organ damage [24]. These abnormalities in

daytime neural circulatory control may be related to

chemoreceptor resetting and tonic chemoreceptor

activation (even in normoxia) [21,25]. By attenuating

apneas, acute CPAP therapy prevents blood pressure

surges and nocturnal sympathetic activation. Long-

term CPAP therapy results in lower daytime sym-

pathetic traffic in OSA patients [26].

Other mechanisms are also important in contrib-

uting to hypertension in OSA. One such potential

mechanism is endothelial dysfunction, with a decrease

in endothelium-dependent vasodilatation [27–29]

(Fig. 2). OSA also may enhance production of vaso-

constrictor and trophic agents, such as endothelin

[30,31], and attenuate production of nitric oxide

[32,33], further favoring vasoconstriction. Metabolic

factors, such as those related to obesity, insulin

resistance, or hyperleptinemia [34–41], are also likely

to play a role. Finally, an intriguing but unprov-

en possibility is that OSA-induced neuroendo-

crine activation, together with the mechanical effects

of blood pressure surges, may lead to vascular re-

modeling, increased wall-to-lumen ratio, and sus-

tained hypertension.

From the clinical standpoint, OSA always should

be considered in the differential diagnosis of causes

of refractory hypertension, particularly in obese hy-

pertensive patients and in patients in whom there is a

blunted nocturnal blood pressure decline (nondip-

pers). Appropriate therapy is effective in decreasing

blood pressure acutely at night [14] and even during

the daytime [11].

Atherosclerosis

In patients with established coronary artery dis-

ease, severe OSA may trigger acute nocturnal cardiac

ischemia with ST-segment depression (predominantly

Fig. 1. Odds ratios for the presence of incident hypertension

at 4-year follow-up according to the apnea-hypopnea index

(AHI) at baseline. The odds ratios are adjusted for baseline

hypertension status, age, gender, body habitus (body mass

index, waist and neck circumference), alcohol consumption,

and cigarette use. Data are shown as odds ratio (lower and

upper 95% confidence interval). P for trend = 0.002. (Data

from Peppard PE, Young T, Palta M, Skatrud J. Prospective

study of the association between sleep-disordered breathing

and hypertension. N Engl J Med 2000;342:1378–84.)

Fig. 2. Percent change in forearm blood flow (FBF) during

infusion of acetylcholine (ACh) and verapamil (VER)

in patients with OSA (circles) and matched normal control

subjects (squares). Data are mean F SEM. (Modified from

Kato M, Roberts-Thomson P, Phillips BG, Haynes WG,

Winnicki M, Accurso V, et al. Impairment of endotheli-

um-dependent vasodilation of resistance vessels in pa-

tients with obstructive sleep apnea. Circulation 2000;102:

2607–10; with permission.)

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205196

in rapid eye movement sleep) that is often resistant to

traditional therapy [42–44]. ST-segment depression

in association with OSA was also noted in patients

without clinically significant coronary artery disease

and was reduced by CPAP treatment [45]. Nocturnal

ischemia in these patients is probably a result of

simultaneous oxygen desaturation, increased sym-

pathetic activity, tachycardia and increased systemic

vascular resistance (all increasing cardiac oxygen

demand), a prothrombotic state (see later discussion),

and any underlying subclinical coronary artery dis-

ease and impaired coronary reserve. Cardiac ischemia

may be exacerbated further by left ventricular hyper-

trophy, especially in patients with OSA who have

long-standing hypertension. Conceivably, the hemo-

dynamic stress induced by apneas and arousals may

increase the risk of coronary plaque rupture.

Whether nocturnal ischemia is directly related to

cardiovascular endpoints or mortality in patients with

OSA has not been established. The observation that

untreated OSA may be associated with an increased

risk of cardiovascular mortality in patients with

coronary artery disease [46,47] argues for the recog-

nition and treatment of any sleep apnea in these

patients, however.

Clinical and epidemiologic evidence suggests a

possible direct role for OSA in the pathophysiology

of atherosclerosis and ischemic heart disease. First,

several studies have reported a high prevalence of

OSA in patients with coronary artery disease [48–

51]. Second, several case-control studies of patients

with myocardial infarction or angina pectoris sug-

gested that the presence of sleep apnea is an indepen-

dent predictor of coronary artery disease [50–54].

Third, patients with OSA have a greater prevalence

of increased carotid wall thickness (a marker of gen-

eralized atherosclerosis) and calcified carotid artery

atheromas [55,56]. Finally, in a large cross-sectional

study of 6424 free-living individuals, sleep apnea

(diagnosed by unattended polysomnography at home)

was associated with increased multivariable-adjusted

relative odds of self-reported coronary heart disease

[57]. This observation has been supported by another

prospective study [5]. These findings suggest that

sleep apnea perhaps may be associated with, or even

predispose to, coronary artery disease. Any such

predisposition may be indirect (eg, through hyperten-

sion, dyslipidemia) or may be directly related to pro-

moting the process of atherogenesis independent of

other comorbidities.

Experimental studies lend further support to the

notion that there might be a cause-and-effect relation-

ship between OSA and atherosclerosis. In OSA, repet-

itive surges in blood pressure, sympathetic activity,

and increased oxidative stress [58,59] may lead to

vascular injury. Increased plasma endothelin levels

[30,31], decreased nitric oxide production [32,33],

and endothelial dysfunction [27–29] also may con-

tribute to the initiation and progression of atherogenic

lesions and vascular damage.

Atherogenic processes can be initiated and po-

tentiated by endothelial damage and the ensuing and

coexisting inflammatory response [60]. Specifically,

leukocyte accumulation and adhesion to the endo-

thelium (with consequent leukocyte-endothelial cell

interactions) may impair endothelial function and

promote atherogenic processes. It is possible that

OSA may influence atherogenesis by inducing such

inflammatory reactions. C-reactive protein level (an

index of the presence of systemic inflammation and

probably a direct mediator of vascular dysfunction,

damage, and atherogenesis) is elevated in persons

with OSA (Fig. 3) [61]. Elevated plasma levels of

various adhesion molecules, increased expression

of adhesion molecules on leukocytes, and their

enhanced adherence to endothelial cells also have

been reported in patients with OSA [59,62–64].

The correlation between these changes and OSA

severity [63] and their reversal after CPAP therapy

[59,63] point to a possible causal relationship

between OSA and the systemic activation of inflam-

matory processes.

Stroke

Several studies have investigated the association

between sleep-related breathing disorders and the

incidence of stroke. A history of snoring seems to

increase the risk of stroke, independent of other

cardiovascular risk factors. A recent large prospective

study in women also supported this conclusion [65].

Similarly, many studies that used polysomnography

noted that the prevalence of OSA is greatly elevated

in patients with stroke [66–71].

A high incidence of OSA in patients with stroke

raises the possibility that perhaps stroke may cause

OSA (rather than being a result of it), especially when

the evidence is based on case-control studies of

patients with and without a history of stroke. This

possibility cannot be excluded. However, it seems

that breathing disorders consequent on a cerebrovas-

cular accident are more likely to cause changes in

respiratory pattern leading to primarily central sleep

apnea [70,72]. These breathing disorders are most

likely to manifest in the first hours after stroke, but

may aggravate preexisting OSA or even cause ob-

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205 197

structive apnea secondary to changes in tone of the

upper airway muscles and upper airway resistance.

The concept that OSA actually precedes and

predisposes to stroke is based on several lines of

evidence. First, in some studies the prevalence of

OSA has been shown to be equally high in patients

with transient ischemic attacks, which suggests the

possibility that OSA precedes stroke events [69,70].

Second, patients with stroke and OSA demonstrate

persistence of OSAwhen repeated polysomnographic

studies are performed several months after the acute

event (although the incidence of central apnea may

actually decrease) [67,70]. Third, the obstructive

events are independent of the type of stroke and its

location [70]. Finally, a possible causal relationship

between OSA and stroke is supported by several

pathophysiologic studies that investigated the actual

mechanisms whereby OSA may predispose to stroke.

For example, Doppler measurements of cerebral blood

flow suggest that obstructive apneas are associated

with blood flow reduction in association with in-

dividual apneic episodes [73–75] and, probably,

impairment of cerebrovascular autoregulation and

diminished cerebral vasodilator reserve. The decreases

in cerebral blood flow are most likely related to the

presence of negative intrathoracic pressures and

increased intracranial pressure. Ischemic effects of

decreased cerebral blood flow would be further poten-

tiated by hypoxemia secondary to apnea. Indeed,

cerebral tissue hypoxia has been recorded during

episodes of OSA [76]. OSA also is a prothrombotic

state that is characterized by higher levels of platelet

aggregation and activation [77–80], elevated fibrino-

gen levels (correlating with the severity of OSA) [81],

decreased fibrinolytic activity [82], and increased

whole blood viscosity—all of which may contribute

to thrombosis and ischemic stroke. It is relevant that

the cerebral hemodynamic changes may be reversed

[83], platelet aggregability can be decreased [79,80],

and the increase in morning fibrinogen levels can be

blunted [84] by CPAP treatment.

Atherosclerosis also may be an important factor

predisposing to stroke. Increased carotid wall thick-

ness (a marker of generalized atherosclerosis and a

risk factor for stroke) and calcified carotid artery

atheromas are significantly more prevalent in indi-

viduals with OSA [55,56]. Finally, hypertension, the

prevalence of which is high in OSA, is a known risk

factor for stroke and may contribute substantially to

any association between OSA and stroke.

Although OSA is an attractive potential contrib-

utor to stroke, the evidence that links OSA to stroke

is primarily observational, and any causality is

inferred from these data and the experimental data

that suggest that OSA contributes to abnormalities in

cerebral blood flow and a prothrombotic state. There

is a clear need for more definitive longitudinal

studies of stroke risk in patients with OSA, inde-

pendent of other risk factors, particularly hyperten-

sion and hyperlipidemia. Importantly, there is some

indication that OSA in stroke survivors may be

associated with increased mortality and a worse

long-term functional outcome [67,68,85]. Hence, it

may be prudent to use CPAP therapy in compliant

Fig. 3. Plasma CRP levels in OSA patients and controls. Middle horizontal line inside box indicates median. Bottom and top of the

box are 25th and 75th percentiles, respectively. (From Shamsuzzaman AS, Winnicki M, Lanfranchi P, Wolk R, Kara T, Accurso V,

et al. Elevated C-reactive protein in patients with obstructive sleep apnea. Circulation 2002;105:2462–4; with permission.)

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205198

patients after stroke with documented evidence of

sleep-disordered breathing.

Heart failure

Patients with systolic heart failure have a signifi-

cant prevalence of sleep apnea (primarily CSA) [86–

90]. OSA may be especially common in patients with

left ventricular diastolic dysfunction [91,92], al-

though not all studies are consistent [93]. The relative

contribution of CSA and OSA to sleep-disordered

breathing varies in different congestive heart failure

(CHF) populations studied, with a general predom-

inance of CSA. Recent observations suggest that

there may be an important pathophysiologic link

between OSA and CSA. Namely, it has been

observed that in heart failure patients the proportion

of OSA decreases and the proportion of CSA

increases from the first to the last quarter of the night,

with an accompanying decrease in transcutaneous

carbon dioxide levels and a significant lengthening

of circulation time [94]. This overnight shift from

OSA to CSA may be related to a deterioration of

cardiac function (caused by the assumption of a

recumbent position and by the detrimental hemody-

namic effects of OSA), with a subsequent increase in

left ventricular filling pressures.

The significance of OSA in CHF is twofold. First,

OSA might predispose a person to CHF. Some

preliminary epidemiologic data suggest that the pres-

ence of OSA is associated with a relative odds for

self-reported CHF of 2.38 (independent of other risk

factors) [57]. Such a causal relationship between

OSA and CHF may be explained by the association

of OSA with other direct or indirect risk factors for

CHF (eg, hypertension, ischemic heart disease, ven-

tricular hypertrophy, oxidative tissue damage, sys-

temic inflammation, neuroendocrine activation, or

autonomic dysfunction).

Second, CHF might contribute to new-onset OSA,

especially in susceptible individuals. In this case,

OSA may be caused by the collapse of the upper

airway because of soft tissue edema and changes in

upper airway muscle tone. OSA superimposed on

CHF may lead to further deterioration of cardiac

function (caused by hypoxemia, sympathetic activa-

tion, vasoconstriction, endothelial dysfunction) and

set up a vicious cycle of progressing, refractory CHF.

An independent association between the severity of

sleep apnea and depression of left ventricular ejection

fraction has been reported [95]. In small study sam-

ples, treatment of OSA with CPAP has been shown

to substantially improve left ventricular ejection frac-

tion and functional class in patients with CHF [96]

(Fig. 4).

Pulmonary hypertension

Apnea and hypoxemia also may elicit acute el-

evations of pulmonary artery pressure during sleep.

Conceivably, these nocturnal events of hypoxia and

pulmonary hypertension might contribute to endothe-

lial damage and vascular remodeling, which may

further lead to sustained pulmonary hypertension.

Several studies have reported the presence of daytime

pulmonary hypertension in patients with OSA. In

many studies, however, other comorbidities were also

present (most notably lung disease, heart failure, or

systemic hypertension), so that any independent con-

tribution of OSA to chronic pulmonary hypertension

remains unclear.

Several studies have investigated the occurrence

of daytime pulmonary hypertension in patients with

OSA in the absence of lung and heart disease. These

Fig. 4. Effects of nasal continuous positive airway pressure

(nCPAP) therapy on improving left ventricular ejection

fraction (LVEF) and functional class (NYHA) in patients

with congestive heart failure. (Modified from Malone S, Liu

PP, Holloway R, Rutherford R, Xie A, Bradley TD. Ob-

structive sleep apnea in patientswith idiopathic dilated cardio-

myopathy: effects of continuous positive airway pressure.

Lancet 1991;338:1480–4; with permission.)

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205 199

studies generally support the concept that OSA is

associated with daytime pulmonary hypertension

[97–100]. The frequency of pulmonary hypertension

varies among various populations studied. It should

be noted that in several studies there was no differ-

ence between pulmonary hypertensive and normo-

tensive OSA subjects with respect to nocturnal

oxygenation and OSA severity [97,98,100], which

suggests that individual variation in pulmonary vas-

cular sensitivity to hypoxic stimuli may be important

or, alternatively, that factors other than OSA per se

may be responsible for the apparent increased pul-

monary artery pressures in patients with OSA.

Patients with OSA with daytime pulmonary hyper-

tension have been reported to have greater elevations

of pulmonary artery vascular tone during rapid eye

movement sleep, independent of the degree of hy-

poxia [101]. In some [100,102,103], although not all

[97,104] studies, patients with OSA and pulmonary

hypertension have been suggested to differ from their

nonhypertensive counterparts in that they tend to

have a greater body mass index and lower daytime

arterial oxygen saturation. It is possible, at least in

some patients with OSA, that mild daytime hypoxe-

mia caused by the obesity-hypoventilation syndrome

might play a role in increasing daytime pulmonary

artery pressures. Interestingly, CPAP therapy seems

to reduce pulmonary pressures in OSA patients with

either pulmonary hypertension or with normal pul-

monary pressures [100,105], which suggests the

possibility that in many cases even ‘‘normal’’ pul-

monary pressures may be elevated compared with

individual baseline values.

A recent report on subjects drawn from the general

population suggested that sleep-disordered breathing

is associated with increased right ventricular wall

thickness [106]. Right ventricular hypertrophy has

been found in selected subjects with OSA [107,108].

Depressed right ventricular ejection fraction and clin-

ical evidence of right ventricular failure also have been

reported in patients with OSA [109–111]. Echocardio-

graphic studies of right ventricular morphology and

function and Doppler estimates of right ventricular

systolic pressure (and hence pulmonary artery systolic

pressure) in OSA patients are limited by several

factors, however, including (1) the difficulty in obtain-

ing high-quality images in a population that is often

obese, (2) the potential influence of comorbidities and

medication on these measurements, (3) the difficulties

in selecting appropriate control subjects for compar-

ison, and (4) the natural scatter of measurements

within a population, together with margins of error

inherent in these measurements. As with other cardi-

ovascular conditions, there is a clear need for more

stringent longitudinal studies before any definitive

assessment of the risk of chronic pulmonary hyperten-

sion in patients with OSA can be made. Further studies

also are needed to investigate the relationship between

OSA, pulmonary hypertension, right ventricular

hypertrophy, and right ventricular failure and to estab-

lish whether these pathologic changes have any impact

on prognosis and require specific treatment.

Cardiac arrhythmias

Most studies that investigate the association

between OSA and cardiac arrhythmias have meth-

odologic limitations related to small sample sizes and

lack of control groups. The exact prevalence of

arrhythmias in patients with sleep apnea is also

difficult to assess because of comorbidities, medica-

tion, and differences among the populations studied.

There is nevertheless a general perception that sleep

apnea is associated with an increased incidence of

bradyarrhythmias and tachyarrhythmias (both supra-

ventricular and ventricular).

The most frequent arrhythmias described in asso-

ciation with sleep apnea are severe sinus bradycardia

and atrioventricular block (including sinus arrest and

complete heart block). These arrhythmias are purely

functional because they have been reported in the

absence of any primary disease of the cardiac con-

duction system and they readily respond to atropine.

The most important pathophysiologic mechanism of

bradyarrhythmias in OSA is a reflex (chemoreceptor

mediated) increase in vagal tone, which is elicited by a

combination of apnea and hypoxemia [112–115] that

activates the diving reflex (increased sympathetic

traffic to peripheral blood vessels and increased vagal

drive to the heart). The occurrence of OSA-related

bradycardia seems to be linked to apnea severity

[116–118]. Bradyarrhythmias also may be more

likely to occur in patients with impaired baroreflex

function (eg, persons with hypertension or heart fail-

ure) [115]. The number of bradyarrhythmias seems to

be greater in rapid eye movement sleep [118], which

may be related in part to greater OSA severity in this

sleep stage.

The causal relationship between these bradyar-

rhythmias and OSA is supported by the observation

that bradycardia occurs only during the night (in

association with nocturnal apnea episodes) in other-

wise asymptomatic subjects [119,120] and is readily

prevented by tracheostomy or CPAP therapy [116,

117,119–122]. CPAP therapy has been shown to

be curative in a sample of patients primarily referred

for pacemaker therapy with asymptomatic brady-

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205200

arrhythmias during sleep, most of whom were sub-

sequently diagnosed with OSA [123]. Although the

prognostic significance of severe nocturnal bradyar-

rhythmias in OSA is not known, it is prudent to

evaluate all patients with asymptomatic bradyarrhyth-

mias for the presence of sleep apnea, which should be

treated appropriately.

Cardiac tachyarrhythmias also have been reported

in OSA, including ventricular tachycardia [116,120,

124] and supraventricular tachycardias. The preva-

lence and severity of these rhythm disturbances are

low in otherwise healthy patients with OSA, how-

ever, and the clinical significance of these arrhyth-

mias is unclear. In contrast, sleep apnea may be an

important trigger for clinically significant arrhythmias

in the presence of serious comorbidities, such as

ischemic heart disease or heart failure. For example,

sleep apnea (central and obstructive) has been asso-

ciated with a greater prevalence of atrial fibrillation in

patients with heart failure [125,126] or after coronary

artery bypass surgery [127]. Similarly, CSA and OSA

are related to the occurrence of ventricular arrhyth-

mias in the heart failure population [126,128], with a

decrease in arrhythmias after CPAP therapy [129].

Summary

Sleep apnea is associated with several cardiovas-

cular disease conditions. A causal relationship be-

tween sleep apnea and each of these diseases is likely,

but remains to be proven. The clearest evidence

implicating OSA in the development of new cardio-

vascular disease involves data that show an increased

prevalence of new hypertension in patients with OSA

followed over 4 years [3]. Circumstantial evidence

and data from small study samples suggest that OSA,

in the setting of existing cardiovascular disease, may

exacerbate symptoms and accelerate disease progres-

sion. The diagnosis of OSA always should be con-

sidered in patients with refractory heart failure,

resistant hypertension, nocturnal cardiac ischemia,

and nocturnal arrhythmias, especially in individuals

with risk factors for sleep apnea (eg, central obesity,

age, and male gender). Treating sleep apnea may help

to achieve better clinical control in these diseases and

may improve long-term cardiovascular prognosis.

References

[1] Kales A, Bixler EO, Cadieux RJ, Schneck DW, Shaw

III LC, Locke TW, et al. Sleep apnoea in a hyper-

tensive population. Lancet 1984;2:1005–8.

[2] Portaluppi F, Provini F, Cortelli P, Plazzi G, Bertozzi

N, Manfredini R, et al. Undiagnosed sleep-disordered

breathing among male nondippers with essential hy-

pertension. J Hypertens 1997;15:1227–33.

[3] Peppard PE, Young T, Palta M, Skatrud J. Prospective

study of the association between sleep-disordered

breathing and hypertension. N Engl J Med 2000;342:

1378–84.

[4] Nieto FJ, Young TB, Lind BK, Shahar E, Samet JM,

Redline S, et al. Association of sleep-disordered

breathing, sleep apnea, and hypertension in a large

community-based study: Sleep Heart Health Study.

JAMA 2000;283:1829–36.

[5] Peker Y, Hedner J, Norum J, Kraiczi H, Carlson J.

Increased incidence of cardiovascular disease in mid-

dle-aged men with obstructive sleep apnea: a 7-year

follow-up. Am J Respir Crit Care Med 2002;166:

159–65.

[6] Suzuki M, Otsuka K, Guilleminault C. Long-term na-

sal continuous positive airway pressure administration

can normalize hypertension in obstructive sleep apnea

patients. Sleep 1993;16:545–9.

[7] Wilcox I, Grunstein RR, Hedner JA, Doyle J, Collins

FL, Fletcher PJ, et al. Effect of nasal continuous pos-

itive airway pressure during sleep on 24-hour blood

pressure in obstructive sleep apnea. Sleep 1993;16:

539–44.

[8] Akashiba T, Kurashina K, Minemura H, Yamamoto H,

Horie T. Daytime hypertension and the effects of short-

term nasal continuous positive airway pressure treat-

ment in obstructive sleep apnea syndrome. Intern Med

1995;34:528–32.

[9] Dimsdale JE, Loredo JS, Profant J. Effect of continu-

ous positive airway pressure on blood pressure: a pla-

cebo trial. Hypertension 2000;35:144–7.

[10] Pankow W, Lies A, Lohmann FW. Sleep-disordered

breathing and hypertension. N Engl J Med 2000;

343:966.

[11] Pepperell JC, Ramdassingh-Dow S, Crosthwaite N,

Mullins R, Jenkinson C, Stradling JR, et al. Ambula-

tory blood pressure after therapeutic and subthera-

peutic nasal continuous positive airway pressure for

obstructive sleep apnoea: a randomised parallel trial.

Lancet 2002;359:204–10.

[12] Somers VK, Mark AL, Zavala DC, Abboud FM. Con-

trasting effects of hypoxia and hypercapnia on ven-

tilation and sympathetic activity in humans. J Appl

Physiol 1989;67:2101–6.

[13] Somers VK, Mark AL, Zavala DC, Abboud FM. In-

fluence of ventilation and hypocapnia on sympathetic

nerve responses to hypoxia in normal humans. J Appl

Physiol 1989;67:2095–100.

[14] Somers VK, DykenME, ClaryMP, Abboud FM. Sym-

pathetic neural mechanisms in obstructive sleep apnea.

J Clin Invest 1995;96:1897–904.

[15] Somers VK, Mark AL, Abboud FM. Potentiation of

sympathetic nerve responses to hypoxia in borderline

hypertensive subjects. Hypertension 1988;11:608–12.

[16] Fletcher EC, Miller J, Schaaf JW, Fletcher JG. Uri-

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205 201

nary catecholamines before and after tracheostomy in

patients with obstructive sleep apnea and hyperten-

sion. Sleep 1987;10:35–44.

[17] Marrone O, Riccobono L, Salvaggio A, Mirabella A,

Bonanno A, Bonsignore MR. Catecholamines and

blood pressure in obstructive sleep apnea syndrome.

Chest 1993;103:722–7.

[18] Dimsdale JE, Coy T, Ziegler MG, Ancoli-Israel S,

Clausen J. The effect of sleep apnea on plasma and

urinary catecholamines. Sleep 1995;18:377–81.

[19] Hedner J, Ejnell H, Sellgren J, Hedner T, Wallin G. Is

high and fluctuating muscle nerve sympathetic activ-

ity in the sleep apnoea syndrome of pathogenetic

importance for the development of hypertension?

J Hypertens 1988;6:S529–31.

[20] Carlson JT, Hedner J, Elam M, Ejnell H, Sellgren J,

Wallin BG. Augmented resting sympathetic activity in

awake patients with obstructive sleep apnea. Chest

1993;103:1763–8.

[21] Narkiewicz K, van de Borne PJ, Montano N, Dyken

ME, Phillips BG, Somers VK. Contribution of tonic

chemoreflex activation to sympathetic activity and

blood pressure in patients with obstructive sleep ap-

nea. Circulation 1998;97:943–5.

[22] Narkiewicz K, Montano N, Cogliati C, van de Borne

PJ, Dyken ME, Somers VK. Altered cardiovascular

variability in obstructive sleep apnea. Circulation

1998;98:1071–7.

[23] Singh JP, Larson MG, Tsuji H, Evans JC, O’Donnell

CJ, Levy D. Reduced heart rate variability and new-

onset hypertension: insights into pathogenesis of

hypertension: the Framingham Heart Study. Hyper-

tension 1998;32:293–7.

[24] Parati G, Di Rienzo M, Ulian L, Santucciu C, Girard

A, Elghozi JL, et al. Clinical relevance blood pressure

variability. J Hypertens 1998;16:S25–33.

[25] Narkiewicz K, van de Borne PJ, Pesek CA, Dyken

ME, Montano N, Somers VK. Selective potentiation

of peripheral chemoreflex sensitivity in obstructive

sleep apnea. Circulation 1999;99:1183–9.

[26] Narkiewicz K, Kato M, Phillips BG, Pesek CA, Davi-

son DE, Somers VK. Nocturnal continuous positive

airway pressure decreases daytime sympathetic traf-

fic in obstructive sleep apnea. Circulation 1999;100:

2332–5.

[27] Carlson JT, Rangemark C, Hedner JA. Attenu-

ated endothelium-dependent vascular relaxation in

patients with sleep apnoea. J Hypertens 1996;14:

577–84.

[28] Kato M, Roberts-Thomson P, Phillips BG, Haynes

WG, Winnicki M, Accurso V, et al. Impairment of

endothelium-dependent vasodilation of resistance

vessels in patients with obstructive sleep apnea.

Circulation 2000;102:2607–10.

[29] Kraiczi H, Caidahl K, Samuelsson A, Peker Y, Hed-

ner J. Impairment of vascular endothelial function

and left ventricular filling: association with the se-

verity of apnea-induced hypoxemia during sleep.

Chest 2001;119:1085–91.

[30] Saarelainen S, Seppala E, Laasonen K, Hasan J. Cir-

culating endothelin-1 in obstructive sleep apnea. En-

dothelium 1997;5:115–8.

[31] Phillips BG, Narkiewicz K, Pesek CA, Haynes WG,

Dyken ME, Somers VK. Effects of obstructive sleep

apnea on endothelin-1 and blood pressure. J Hypertens

1999;17:61–6.

[32] Ip MS, Lam B, Chan LY, Zheng L, Tsang KW, Fung

PC, et al. Circulating nitric oxide is suppressed in

obstructive sleep apnea and is reversed by nasal con-

tinuous positive airway pressure. Am J Respir Crit

Care Med 2000;162:2166–71.

[33] Schulz R, Schmidt D, Blum A, Lopes-Ribeiro X,

Lucke C, Mayer K, et al. Decreased plasma levels

of nitric oxide derivatives in obstructive sleep ap-

noea: response to CPAP therapy. Thorax 2000;55:

1046–51.

[34] Chin K, Shimizu K, Nakamura T, Narai N, Masuzaki

H, OgawaY, et al. Changes in intra-abdominal visceral

fat and serum leptin levels in patients with obstructive

sleep apnea syndrome following nasal continuous pos-

itive airway pressure therapy. Circulation 1999;100:

706–12.

[35] Phillips BG, Hisel TM, Kato M, Pesek CA, Dyken

ME, Narkiewicz K, et al. Recent weight gain in pa-

tients with newly diagnosed obstructive sleep apnea.

J Hypertens 1999;17:1297–300.

[36] Phillips BG, Kato M, Narkiewicz K, Choe I, Somers

VK. Increases in leptin levels, sympathetic drive, and

weight gain in obstructive sleep apnea. Am J Physiol

2000;279:H234–7.

[37] Ip MS, Lam KS, Ho C, Tsang KW, Lam W. Serum

leptin and vascular risk factors in obstructive sleep

apnea. Chest 2000;118:580–6.

[38] Elmasry A, Lindberg E, Berne C, Janson C, Gislason

T, Awad Tageldin M, et al. Sleep-disordered breathing

and glucose metabolism in hypertensive men: a pop-

ulation-based study. J Intern Med 2001;249:153–61.

[39] Ip MS, Lam B, Ng MM, Lam WK, Tsang KW, Lam

KS. Obstructive sleep apnea is independently associ-

ated with insulin resistance. Am J Respir Crit Care

Med 2002;165:670–6.

[40] Punjabi NM, Sorkin JD, Katzel LI, Goldberg AP,

Schwartz AR, Smith PL. Sleep-disordered breathing

and insulin resistance in middle-aged and over-

weight men. Am J Respir Crit Care Med 2002;

165:677–82.

[41] Shimizu K, Chin K, Nakamura T, Masuzaki H, Oga-

wa Y, Hosokawa R, et al. Plasma leptin levels and car-

diac sympathetic function in patients with obstructive

sleep apnoea-hypopnoea syndrome. Thorax 2002;

57:429–34.

[42] Franklin KA, Nilsson JB, Sahlin C, Naslund U. Sleep

apnoea and nocturnal angina. Lancet 1995;345:

1085–7.

[43] Schafer H, Koehler U, Ploch T, Peter JH. Sleep-

related myocardial ischemia and sleep structure in pa-

tients with obstructive sleep apnea and coronary heart

disease. Chest 1997;111:387–93.

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205202

[44] Mooe T, Franklin KA, Wiklund U, Rabben T, Holm-

strom K. Sleep-disordered breathing and myocardial

ischemia in patients with coronary artery disease.

Chest 2000;117:1597–602.

[45] Hanly P, Sasson Z, Zuberi N, Lunn K. ST-segment

depression during sleep in obstructive sleep apnea.

Am J Cardiol 1993;71:1341–5.

[46] Peker Y, Hedner J, Kraiczi H, Loth S. Respiratory

disturbance index: an independent predictor of mor-

tality in coronary artery disease. Am J Respir Crit

Care Med 2000;162:81–6.

[47] Mooe T, Franklin KA, Holmstrom K, Rabben T, Wi-

klund U. Sleep-disordered breathing and coronary ar-

tery disease: long-term prognosis. Am J Respir Crit

Care Med 2001;164:1910–3.

[48] De Olazabal JR, Miller MJ, Cook WR, Mithoefer JC.

Disordered breathing and hypoxia during sleep in

coronary artery disease. Chest 1982;82:548–52.

[49] Andreas S, Schulz R, Werner GS, Kreuzer H. Preva-

lence of obstructive sleep apnoea in patients with cor-

onary artery disease. Coron Artery Dis 1996;7:541–5.

[50] Mooe T, Rabben T, Wiklund U, Franklin KA, Eriks-

son P. Sleep-disordered breathing in women: occur-

rence and association with coronary artery disease.

Am J Med 1996;101:251–6.

[51] Mooe T, Rabben T, Wiklund U, Franklin KA, Eriks-

son P. Sleep-disordered breathing in men with coro-

nary artery disease. Chest 1996;109:659–63.

[52] Hung J, Whitford EG, Parsons RW, Hillman DR.

Association of sleep apnoea with myocardial infarc-

tion in men. Lancet 1990;336:261–4.

[53] Peker Y, Kraiczi H, Hedner J, Loth S, Johansson A,

Bende M. An independent association between ob-

structive sleep apnoea and coronary artery disease.

Eur Respir J 1999;14:179–84.

[54] Schafer H, Koehler U, Ewig S, Hasper E, Tasci S,

Luderitz B. Obstructive sleep apnea as a risk mar-

ker in coronary artery disease. Cardiology 1999;92:

79–84.

[55] Friedlander AH, Yueh R, Littner MR. The prevalence

of calcified carotid artery atheromas in patients with

obstructive sleep apnea syndrome. J Oral Maxillofac

Surg 1998;56:950–4.

[56] Silvestrini M, Rizzato B, Placidi F, Baruffaldi R,

Bianconi A, Diomedi M. Carotid artery wall thick-

ness in patients with obstructive sleep apnea syn-

drome. Stroke 2002;33:1782–5.

[57] Shahar E, Whithney CW, Redilne S, Lee TE, Newman

AB, Nieto JF, et al, for the Sleep Heart Heath Study

Research Group. Sleep-disordered breathing and car-

diovascular disease. Am J Respir Crit Care Med 2001;

163:19–25.

[58] Schulz R, Mahmoudi S, Hattar K, Sibelius U, Ol-

schewski H, Mayer K, et al. Enhanced release of

superoxide from polymorphonuclear neutrophils in

obstructive sleep apnea: impact of continuous positive

airway pressure therapy. Am J Respir Crit Care Med

2000;162:566–70.

[59] Dyugovskaya L, Lavie P, Lavie L. Increased adhesion

molecules expression and production of reactive oxy-

gen species in leukocytes of sleep apnea patients. Am

J Respir Crit Care Med 2002;165:934–9.

[60] Ross R. Atherosclerosis: an inflammatory disease.

N Engl J Med 1999;340:115–26.

[61] Shamsuzzaman AS, Winnicki M, Lanfranchi P, Wolk

R, Kara T, Accurso V, et al. Elevated C-reactive pro-

tein in patients with obstructive sleep apnea. Circu-

lation 2002;105:2462–4.

[62] Ohga E, Nagase T, Tomita T, Teramoto S, Matsuse T,

Katayama H, et al. Increased levels of circulating

ICAM-1, VCAM-1, and L-selectin in obstructive

sleep apnea syndrome. J Appl Physiol 1999;87:10–4.

[63] Chin K, Nakamura T, Shimizu K, Mishima M, Naka-

mura T, Miyasaka M, et al. Effects of nasal continu-

ous positive airway pressure on soluble cell adhesion

molecules in patients with obstructive sleep apnea

syndrome. Am J Med 2000;109:562–7.

[64] El-Solh AA, Mador MJ, Sikka P, Dhillon RS, Am-

sterdam D, Grant BJ. Adhesion molecules in patients

with coronary artery disease and moderate-to-severe

obstructive sleep apnea. Chest 2002;121:1541–7.

[65] Hu FB, Willett WC, Manson JE, Colditz GA, Rimm

EB, Speizer FE, et al. Snoring and risk of cardiovas-

cular disease in women. J Am Coll Cardiol 2000;35:

308–13.

[66] Mohsenin V, Valor R. Sleep apnea in patients with

hemispheric stroke. Arch Phys Med Rehabil 1995;

76:71–6.

[67] Dyken ME, Somers VK, Yamada T, Ren ZY, Zimmer-

manMB. Investigating the relationship between stroke

and obstructive sleep apnea. Stroke 1996;27:401–7.

[68] Good DC, Henkle JQ, Gelber D, Welsh J, Verhulst S.

Sleep-disordered breathing and poor functional out-

come after stroke. Stroke 1996;27:252–9.

[69] Bassetti C, Aldrich MS. Sleep apnea in acute cere-

brovascular diseases: final report on 128 patients.

Sleep 1999;22:217–23.

[70] Parra O, Arboix A, Bechich S, Garcia-Eroles L, Mon-

tserrat JM, Lopez JA, et al. Time course of sleep-

related breathing disorders in first-ever stroke or

transient ischemic attack. Am J Respir Crit Care

Med 2000;161:375–80.

[71] Wessendorf TE, Teschler H, Wang YM, Konietzko N,

Thilmann AF. Sleep-disordered breathing among pa-

tients with first-ever stroke. J Neurol 2000;247:41–7.

[72] Parra O. Sleep-disordered breathing and stroke: is

there a rationale for treatment? Eur Respir J 2001;18:

619–22.

[73] Fischer AQ, Chaudhary BA, Taormina MA, Akhtar B.

Intracranial hemodynamics in sleep apnea. Chest

1992;102:1402–6.

[74] Balfors EM, Franklin KA. Impairment of cerebral

perfusion during obstructive sleep apneas. Am J Res-

pir Crit Care Med 1994;150:1587–91.

[75] Netzer N, Werner P, Jochums I, Lehmann M, Strohl

KP. Blood flow of the middle cerebral artery with

sleep-disordered breathing: correlation with obstruc-

tive hypopneas. Stroke 1998;29:87–93.

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205 203

[76] Hayakawa T, Terashima M, Kayukawa Y, Ohta T,

Okada T. Changes in cerebral oxygenation and he-

modynamics during obstructive sleep apneas. Chest

1996;109:916–21.

[77] Bokinsky G, Miller M, Ault K, Husband P, Mitchell J.

Spontaneous platelet activation and aggregation dur-

ing obstructive sleep apnea and its response to therapy

with nasal continuous positive airway pressure: a pre-

liminary investigation. Chest 1995;108:625–30.

[78] Eisensehr I, Ehrenberg BL, Noachtar S, Korbett K,

Byrne A, McAuley A, et al. Platelet activation, epi-

nephrine, and blood pressure in obstructive sleep ap-

nea syndrome. Neurology 1998;51:188–95.

[79] Sanner BM, Konermann M, Tepel M, Groetz J, Mum-

menhoff C, Zidek W. Platelet function in patients with

obstructive sleep apnoea syndrome. Eur Respir J 2000;

16:648–52.

[80] Geiser T, Buck F, Meyer BJ, Bassetti C, Haeberli A,

Gugger M. In vivo platelet activation is increased

during sleep in patients with obstructive sleep apnea

syndrome. Respiration 2002;69:229–34.

[81] Wessendorf TE, Thilmann AF, Wang YM, Schreiber

A, Konietzko N, Teschler H. Fibrinogen levels and

obstructive sleep apnea in ischemic stroke. Am J Res-

pir Crit Care Med 2000;162:2039–42.

[82] Rangemark C, Hedner JA, Carlson JT, Gleerup G,

Winther K. Platelet function and fibrinolytic activity

in hypertensive and normotensive sleep apnea pa-

tients. Sleep 1995;18:188–94.

[83] Diomedi M, Placidi F, Cupini LM, Bernardi G, Sil-

vestrini M. Cerebral hemodynamic changes in sleep

apnea syndrome and effect of continuous positive air-

way pressure treatment. Neurology 1998;51:1051–6.

[84] Chin K, Ohi M, Kita H, Noguchi T, Otsuka N, Tsuboi

T, et al. Effects of NCPAP therapy on fibrinogen levels

in obstructive sleep apnea syndrome. Am J Respir Crit

Care Med 1996;153:1972–6.

[85] SpriggsDA, French JM,Murdy JM, Curless RH, Bates

D, James OF. Snoring increases the risk of stroke and

adversely affects prognosis. QJM 1992;83:555–62.

[86] Lofaso F, Verschueren P, Rande JHD, Harf A, Golden-

berg F. Prevalence of sleep-disordered breathing

in patients on a transplant waiting list. Chest 1994;

106:1689–94.

[87] Javaheri S, Parker TJ, Liming JD, Corbett WS, Nishi-

yama H, Wexler L, et al. Sleep apnea in 81 ambula-

tory male patients with stable heart failure: types and

their prevalences, consequences, and presentations.

Circulation 1998;97:2154–9.

[88] Lanfranchi PA, Braghiroli A, Bosimini E, Mazzuero

G, Colombo R, Donner CF, et al. Prognostic value of

nocturnal Cheyne-Stokes respiration in chronic heart

failure. Circulation 1999;99:1435–40.

[89] Tremel F, Pepint JL, Veale D, Wuyam B, Siche JP,

Mallion JM, et al. High prevalence and persistence of

sleep apnea in patients referred for acute left ventric-

ular failure and medically treated over 2 months. Eur

Heart J 1999;20:1201–9.

[90] Sin DD, Fitzgerald F, Parker JD, Newton G, Floras JS.

Risk factors for central and obstructive sleep apnea in

450 men and women with congestive heart failure. Am

J Respir Crit Care Med 1999;160:1101–6.

[91] Chan JJ, Sanderson J, Chan W, Lai C, Choy D, Ho A,

et al. Prevalence of sleep-disordered breathing in dia-

stolic heart failure. Chest 1997;111:1488–93.

[92] Fung JW, Li TS, Choy DK, Yip GW, Ko FW, Sander-

son JE, et al. Severe obstructive sleep apnea is associ-

ated with left ventricular diastolic dysfunction. Chest

2002; 121:422–9.

[93] Niroumand M, Kuperstein R, Sasson Z, Hanly PJ.

Impact of obstructive sleep apnea on left ventricular

mass and diastolic function. Am J Respir Crit Care

Med 2001;163:1632–6.

[94] Tkacova R, Niroumand M, Lorenzi-Filho G, Bradley

DT. Overnight shift from obstructive to central sleep

apneas in patients with heart failure: role of PCO2 and

circulatory delay. Circulation 2001;103:238–43.

[95] Sanner BM, Konermann M, Doberauer C, Weiss T,

Zidek W. Sleep-disordered breathing in patients

referred for angina evaluation: association with

left ventricular dysfunction. Clin Cardiol 2001;24:

146–50.

[96] Malone S, Liu PP, Holloway R, Rutherford R, Xie A,

Bradley TD. Obstructive sleep apnea in patients with

idiopathic dilated cardiomyopathy: effects of contin-

uous positive airway pressure. Lancet 1991;338:

1480–4.

[97] Sajkov D, Cowie RJ, Thornton AT, Espinoza HA,

McEvoy RD. Pulmonary hypertension and hypoxe-

mia in obstructive sleep apnea syndrome. Am J Res-

pir Crit Care Med 1994;149:416–22.

[98] Sajkov D, Wang T, Saunders NA, Bune AJ, Neill AM,

DouglasMcevoy R. Daytime pulmonary hemodynam-

ics in patients with obstructive sleep apnea without

lung disease. Am J Respir Crit Care Med 1999;159:

1518–26.

[99] Sanner BM, Doberauer C, Konermann M, Sturm A,

Zidek W. Pulmonary hypertension in patients with

obstructive sleep apnea syndrome. Arch Intern Med

1997;157:2483–7.

[100] Alchanatis M, Tourkohoriti G, Kakouros S, Kosmas E,

Podaras S, Jordanoglou JB. Daytime pulmonary hy-

pertension in patients with obstructive sleep apnea:

the effect of continuous positive airway pressure on

pulmonary hemodynamics. Respiration 2001;68:

566–72.

[101] Niijima M, Kimura H, Edo H, Shinozaki T, Kang J,

Masuyama S, et al. Manifestation of pulmonary hyper-

tension during REM sleep in obstructive sleep apnea

syndrome. Am J Respir Crit Care Med 1999;159:

1766–72.

[102] Laaban JP, Cassuto D, Orvoen-Frija E, Iliou MC,

Mundler O, Leger D, et al. Cardiorespiratory conse-

quences of sleep apnoea syndrome in patients with

massive obesity. Eur Respir J 1998;11:20–7.

[103] Bady E, Achkar A, Pascal S, Orvoen-Frija E, Laaban

JP. Pulmonary arterial hypertension in patients with

sleep apnoea syndrome. Thorax 2000;55:934–9.

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205204

[104] Laks L, Lehrhaft B, Grunstein RR, Sullivan CE. Pul-

monary hypertension in obstructive sleep apnoea. Eur

Respir J 1995;8:537–41.

[105] Sajkov D, Wang T, Saunders NA, Bune AJ, Mcevoy

RD. Continuous positive airway pressure treatment

improves pulmonary hemodynamics in patients with

obstructive sleep apnea. Am J Respir Crit Care Med

2002;165:152–8.

[106] Guidry UC,Mendes LA, Evans JC, Levy D, O’Connor

GT, Larson MG, et al. Echocardiographic features of

the right heart in sleep-disordered breathing: the Fra-

mingham Heart Study. Am J Respir Crit Care Med

2001;164:933–8.

[107] Berman EJ, DiBenedetto RJ, Causey DE, Mims T,

Conneff M, Goodman LS, et al. Right ventricular hy-

pertrophy detected by echocardiography in patients

with newly diagnosed obstructive sleep apnea. Chest

1991;100:347–50.

[108] NodaA, Okada T, Yasuma F, NakashimaN, YokotaM.

Cardiac hypertrophy in obstructive sleep apnea syn-

drome. Chest 1995;107:1538–44.

[109] Nahmias J, Lao R, Karetzky M. Right ventricular dys-

function in obstructive sleep apnoea: reversal with na-

sal continuous positive airway pressure. Eur Respir J

1996;9:945–51.

[110] Sanner BM, Konermann M, Sturm A, Muller HJ, Zi-

dek W. Right ventricular dysfunction in patients with

obstructive sleep apnoea syndrome. Eur Respir J 1997;

10:2079–83.

[111] Blankfield RP, Hudgel DW, Tapolyai AA, Zyzanski SJ.

Bilateral leg edema, obesity, pulmonary hypertension,

and obstructive sleep apnea. Arch Intern Med 2000;

160:2357–62.

[112] de Burgh DalyM, Scott MJ. An analysis of the primary

cardiovascular reflex effects of stimulation of the ca-

rotid body chemoreceptors in the dog. Am J Physiol

1962;162:555–73.

[113] de Burgh Daly M, Angell-James JE, Elsner R. Role

of carotid-body chemoreceptors and their reflex inter-

actions in bradycardia and cardiac arrest. Lancet

1979;1:764–7.

[114] Zwillich C, Devlin T, White D, Douglas N, Weil J,

Martin R. Bradycardia during sleep apnea: char-

acteristics and mechanism. J Clin Invest 1982;69:

1286–92.

[115] Somers VK, Dyken ME, Mark AL, Abboud FM.

Parasympathetic hyperresponsiveness and bradyar-

rhythmias during apnoea in hypertension. Clin Auton

Res 1992;2:171–6.

[116] Guilleminault C, Connolly SJ, Winkle RA. Cardiac

arrhythmia and conduction disturbances during sleep

in 400 patients with sleep apnea syndrome. Am J

Cardiol 1983;52:490–4.

[117] Becker H, Brandenburg U, Peter JH, Von Wichert P.

Reversal of sinus arrest and atrioventricular conduc-

tion block in patients with sleep apnea during nasal

continuous positive airway pressure. Am J Respir Crit

Care Med 1995;151:215–8.

[118] Koehler U, Becker HF, Grimm W, Heitmann J, Peter

JH, Schafer H. Relations among hypoxemia, sleep

stage, and bradyarrhythmia during obstructive sleep

apnea. Am Heart J 2000;139:142–8.

[119] Grimm W, Koehler U, Fus E, Hoffmann J, Menz V,

Funck R, et al. Outcome of patients with sleep apnea-

associated severe bradyarrhythmias after continuous

positive airway pressure therapy. Am J Cardiol 2000;

86:688–92.

[120] Harbison J, O’Reilly P, McNicholas WT. Cardiac

rhythm disturbances in the obstructive sleep apnea

syndrome: effects of nasal continuous positive airway

pressure therapy. Chest 2000;118:591–5.

[121] Koehler U, Fus E, Grimm W, Pankow W, Schafer H,

Stammnitz A, et al. Heart block in patients with ob-

structive sleep apnoea: pathogenetic factors and ef-

fects of treatment. Eur Respir J 1998;11:434–9.

[122] Grimm W, Hoffmann J, Menz V, Kohler U, Heit-

mann J, Peter JH, et al. Electrophysiologic evaluation

of sinus node function and atrioventricular conduc-

tion in patients with prolonged ventricular asystole

during obstructive sleep apnea. Am J Cardiol 1996;

77:1310–4.

[123] Stegman SS, Burroughs JM, Henthorn RW. Asymp-

tomatic bradyarrhythmias as a marker for sleep apnea:

appropriate recognition and treatment may reduce the

need for pacemaker therapy. Pacing Clin Electrophy-

siol 1996;19:899–904.

[124] Tilkian AG, Guilleminault C, Schroeder JS, Lehrman

KL, Simmons FB, Dement WC. Sleep-induced apnea

syndrome: prevalence of cardiac arrhythmias and

their reversal after tracheostomy. Am J Med 1977;

63:348–58.

[125] Blackshear JL, Kaplan J, Thompson RC, Safford RE,

Atkinson EJ. Nocturnal dyspnea and atrial fibrillation

predict Cheyne-Stokes respirations in patients with

congestive heart failure. Arch Intern Med 1995;155:

1297–302.

[126] Javaheri S, Corbett WS. Association of low PaCO2

with central sleep apnea and ventricular arrhythmias

in ambulatory patients with stable heart failure. Ann

Intern Med 1998;128:204–7.

[127] Mooe T, Gullsby S, Rabben T, Eriksson P. Sleep-dis-

ordered breathing: a novel predictor of atrial fibrilla-

tion after coronary artery bypass surgery. Coron

Artery Dis 1996;7:475–8.

[128] Fichter J, Bauer D, Arampatzis S, Fries R, Heisel A,

Sybrecht GW. Sleep-related breathing disorders are

associated with ventricular arrhythmias in patients

with an implantable cardioverter-defibrillator. Chest

2002;122:558–61.

[129] Javaheri S. Effects of continuous positive airway pres-

sure on sleep apnea and ventricular irritability in pa-

tients with heart failure. Circulation 2000;101:392–7.

R. Wolk, V.K. Somers / Clin Chest Med 24 (2003) 195–205 205

Heart failure and sleep apnea: emphasis on practical

therapeutic options

Shahrokh Javaheri, MDa,b,*

aSleep Disorders Laboratory, Department of Veterans Affairs Medical Center, 3200 Vine Street, Cincinnati, OH 45220, USAbDivision of Pulmonary and Critical Care Medicine, Department of Medicine, University of Cincinnati College of Medicine,

Cincinnati, OH 45220, USA

Heart failure is approaching epidemic proportions

and has become a major public health problem. It is

the only cardiovascular disorder with increasing

incidence and prevalence, causing excessive morbid-

ity and mortality. The economic burden of heart

failure is also huge. It is the largest single Medicare

expenditure because it is the leading cause of hospi-

talization for individuals old than age 65.

Heart failure is a major risk factor for sleep-related

breathing disorders, which could adversely affect

cardiovascular function and contribute to morbidity

and mortality of heart failure. Unfortunately, in the

clinical management of heart failure, sleep-related

breathing disorders remain underdiagnosed. The

underdiagnosis is mostly caused by lack of education

and unfamiliarity with sleep apnea by primary care

physicians and cardiologists involved in care of

subjects with heart failure.

In this article, the author briefly reviews the

epidemiology of heart failure and sleep-related

breathing disorders and discusses some practical

therapeutic options. It is hoped that treatment of

sleep-related breathing disorders will decrease mor-

bidity and mortality and improve quality of life for

persons with heart failure. Treatment also may

decrease the economic burden. Long-term studies

with such endpoints as primary outcomes are needed.

Epidemiology of heart failure

Heart failure results from any cardiac disorder that

impairs the ability of the ventricle to eject blood [1].

Left heart failure may result from disorders of great

vessels, valves, myocardium, and pericardium. In

most adults with left heart failure, however, the

symptoms are caused by impairment of left ventricu-

lar function (myocardial failure). Left ventricular

failure could be predominantly diastolic in nature or

manifested by systolic and diastolic dysfunction. The

principal hallmark of diastolic dysfunction is an

elevation in left ventricular end-diastolic pressure

and pulmonary capillary pressure. The underlying

pathology in diastolic heart failure is a stiff non-

compliant left ventricle, with systolic function of the

left ventricle being preserved. In contrast, the hall-

mark of left ventricular systolic dysfunction is a

depressed ejection fraction, which is commonly asso-

ciated with an increase in left ventricular end-dia-

stolic and systolic volumes. Left ventricular systolic

dysfunction is most commonly caused by coronary

artery disease. There are several nonischemic causes

of left ventricular systolic dysfunction, such as myo-

carditis and alcohol ingestion. In idiopathic dilated

cardiomyopathy, no cause can be identified.

Coronary artery disease and heart failure are

progressive disorders. The progression of heart fail-

ure is associated with geometric remodeling of the

ventricle, characterized by the development of dilata-

tion, hypertrophy, and becoming more spherical.

Pattern of ventricular remodeling is load dependent,

with pressure overload resulting in systolic and

volume overload resulting in diastolic wall stress

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00026-1

* Pulmonary Section, Department of Medicine, Veter-

ans Affairs Medical Center, 3200 Vine Street, Cincinnati,

OH 45220.

E-mail address: Shahrokh.Javaheri@med.va.gov

Clin Chest Med 24 (2003) 207–222

[1]. Activation of several endogenous factors, such as

neurohormones and cytokines, accelerates the process

of remodeling and contributes to progression of heart

failure. Elevated levels of neurohormones, particu-

larly the vasoconstrictors, norepinephrine, compo-

nents of renin-angiotensin-aldosterone system, and

endothelin adversely affect cardiovascular structure

and function, which ultimately results in myocyte

apoptosis and fibrosis.

For multiple reasons (Box 1), heart failure has

major impact on public health [2–8]. It is estimated

that 1.5% to 2% of the United States population has

heart failure. Heart failure is a disorder of elderly

persons, and its prevalence increases to approxi-

mately 6% to 10% in individuals older than 65 years.

Heart failure is the only major cardiovascular disease

with increasing incidence and prevalence. In this

regard, it is estimated that 20 million people may

have asymptomatic cardiac dysfunction, and with

time, these individuals are likely to become symp-

tomatic. Because of increased average life span and

improved therapy of ischemic coronary artery disease

and hypertension, which are risk factors for heart

failure, it is predicted that incidence and prevalence

of heart failure will continue to rise into the twenty-

first century.

Heart failure accounts for approximately 11million

physician office visits. It is the leading cause of hos-

pitalization in people older than 65 years and accounts

for 3.5 million hospitalizations annually. The latter

may be underestimated by standard ICD coding sys-

tem [8]. Annually, heart failure may contribute directly

or indirectly to 250,000 deaths. The death rate in-

creases progressively, with advanced symptomatology

approaching 30% to 40% annually in patients with

heart failure in New York Heart Class IV.

The economic impact of heart failure is also huge

and accounts for approximately US$40 billion annu-

ally for direct cost. This cost accounted for approx-

imately 5.4% of the health care budget in 1991 [3].

The cost of hospitalization is approximately $8 bil-

lion to $15 billion annually and accounts for most of

the total cost of heart failure; the remainder of the

cost covers the care delivered by health care profes-

sionals, including physicians, cost of medications,

home health care, and other medical durables.

Epidemiology of sleep apnea in heart failure

Prevalence of sleep-related breathing disorders

has been studied in patients with heart failure due

to various causes [9], although most systematically in

heart failure caused by left ventricular systolic dys-

function [10–22]. The results of some of these

studies are depicted in Table 1. At least 45% have

an apnea-hypopnea index (AHI) of � 10/hour, and

40% to 80% have an AHI of � 15/hour.

Differences in prevalence rates of sleep-related

breathing disorders in systolic heart failure (Table 1)

can be attributed to differences in various aspects of

the studies, the various thresholds to define the

disorder, and the varied definitions of hypopnea.

For example, in one study [22], a minimum of a

2% drop in saturation was the criterion to define

hypopnea. In the author’s studies [12,13], a minimum

of 4% decrease in saturation or an arousal was

required for criteria to define hypopnea, and some-

what arbitrarily considered, a threshold index of

� 15/hour was considered to be clinically significant.

Regarding the AHI, in population studies of subjects

without heart failure, an index of � 5/hour has been

used to define presence of a significant number of

disordered breathing events in obstructive sleep

apnea-hypopnea syndrome [23]. Results of recent

population studies [24–26] suggest that lower indi-

Box 1. Heart failure in the United States,present and future

� 1.5% to 2% of population(5 million)� 6% to 10% of population older than65 years

� 400,000 to 700,000 newcases annually

� 11 million physician officevisits annually

� 3.5 million hospitalizations annually� Leading cause of hospitalization inpeople older than age 65

� 250,000 deaths annually (directand indirect)

� $40 billion direct cost annually� $8 billion to $15 billion cost ofhospitalization annually

� Only cardiovascular disorder increas-ing in incidence and prevalence� 20 million with asymptomatic car-diac impairment, many ofwhom will develop heart failure

� Increased average life span� Improved therapy for ischemiccoronary disease, hypertension,and stroke

� Prevalence and incidence will in-crease into twenty-first century

S. Javaheri / Clin Chest Med 24 (2003) 207–222208

ces of disordered breathing events are also associ-

ated with cardiovascular pathology. Despite a wide

range in the reported AHIs in systolic heart failure

(see Box 1), these studies [10–22] collectively

showed a high prevalence of sleep-related breathing

disorders, which made systolic heart failure one of the

leading risk factors for sleep apnea-hypopnea.

The largest prospective study [13] (see Table 1)

involved 81 ambulatory male patients with stable,

treated heart failure. In the study, 92 consecutive

eligible patients who were followed in a cardiology

clinic were asked to participate (88% recruitment).

Using an AHI of � 15/hour as the threshold, 41 pa-

tients (51% of all patients) had moderate to severe

sleep apnea-hypopnea, with an average index of

44 F 19 (1 standard deviation) per hour. The results

of that study [13] compared well with the results of

the largest retrospective study [16] using similar crite-

ria to define hypopnea. In the study of 450 patients

[16], which also included women, 61% had an AHI

of � 15/hour. This prevalence, 61%, was expectedly

higher than the 51% prevalence rate in the author’s

study [13], because in Sin’s study [16], risk factors

for sleep apnea were among reasons for referral to the

sleep laboratory, whereas the author’s study sought

no information about symptoms or risk factors for

sleep apnea to recruit subjects.

The prevalence of obstructive and central sleep

apnea also varies widely among different studies. In

each study, obstructive sleep apnea-hypopnea/central

sleep apnea-hypopnea ratio depends on several fac-

tors, including the pattern of recruitment (consecutive

recruitment versus referral because of risk factor for

obstructive sleep apnea-hypopnea, such as snoring),

number of overweight and obese subjects, and the

cut-off point to define predominant obstructive ver-

sus central sleep apnea-hypopnea. Accurate differ-

entiation of hypopneas into central versus obstructive

is difficult, much more so than differentiation of

obstructive from central sleep apnea. Undoubtedly

some degree of contamination occurs. With these

limitations in mind, in the author’s study [13],

approximately 40% of the patients had central sleep

apnea and 11% has obstructive sleep apnea. When

polysomnograms were decoded and matched with

demographics of subjects, patients whose polysom-

nograms were categorized to have obstructive sleep

apnea-hypopnea had a significantly higher preva-

lence of habitual snoring, obesity, and hypertension

[13]. In Sin’s study [16], the prevalence of obstruc-

tive sleep apnea (OSA) was 32%, compared with

11% in the author’s study [13]. This rate also was

expected because snoring, a risk factor for OSA, was

a reason for referral in Sin’s study [16] but not the

author’s [13].

Sleep-related breathing disorders in isolated

diastolic heart failure

A small study [27] reported that approximately

50% of persons with isolated diastolic heart failure

have sleep apnea-hypopnea defined by an AHI of

� 10/hour (see Table 1). Large-scale epidemiologic

studies are needed to define the prevalence of sleep

apnea-hypopnea in isolated diastolic heart failure.

This is important for two reasons. First, a large

number of old patients with symptoms of congestive

heart failure suffer from isolated diastolic heart failure

[28], and sleep-related breathing disorders also may

be prevalent in this cohort. Second, sympathetic

activation, nocturnal hypertension, and hypoxemia,

which are the immediate consequences of sleep-

related breathing disorders, could impair left ven-

tricular diastolic functions or contribute to diastolic

dysfunction [29,30]. In other words, sleep-related

Table 1

Prevalence of sleep-related breathing disorders in systolic and isolated diastolic heart failure

Reference n LVEF % AHI�10/h (%) AHI�15/h (%)

Systolic heart failure

[13]a 81 25 F 9 57 51

[14]a 34 30 F 10 — 82

[22]a 66 23 F 6 76 —

[18]a 20 < 25 45 45

[15]a 75 < 40 59 43

[16]b 450 27 F 16 72 61

Diastolic heart failure

[27]a 20 > 50 55 —

Abbreviation: LVEF, left ventricular ejection fraction.a Prospective.b Retrospective.

S. Javaheri / Clin Chest Med 24 (2003) 207–222 209

breathing disorders could be a cause of diastolic

dysfunction or contribute to its progression.

Pathophysiologic consequences of sleep apnea

and hypopnea

There are three major adverse cardiovascular con-

sequences of sleep apnea and hypopnea: (1) intermit-

tent alterations in arterial blood gases, (2) arousals and

shift to light sleep stages, and (3) large negative inspir-

atory deflections in intrathoracic pressure [31–33].

Intermittent alterations in arterial blood gases

Periodic breathing is characterized by episodes of

apnea and hypopnea, which cause hypoxemia and

hypercapnia, and hyperpnea, which results in reoxy-

genation and hypocapnia. Hypoxemia may affect the

cardiovascular system adversely in multiple ways,

such as by decreasing myocardial oxygen delivery,

promoting endothelial cell dysfunction, and increas-

ing sympathetic nervous system activity.

Decreased oxygen delivery is most detrimental to

myocardium if there is established coronary athero-

sclerosis, which could limit myocardial blood supply.

In this regard, myocardium has the highest oxygen

extraction, as evidenced by a low coronary sinus

partial pressure oxygen (PO2). Hypocapnia, which

occurs because of hyperpnea after apnea or hypo-

pnea, may further impair myocardial oxygen delivery

and uptake by coronary artery vasoconstriction [34]

and shifting the oxygen-hemoglobin dissociation

curve to the left. Decreased myocardial oxygen

supply may impair systolic and diastolic function

and cause myocardial ischemia and arrhythmias.

Hypoxia also may promote coronary endothelial

dysfunction. Endothelial dysfunction has been dem-

onstrated in several cardiovascular disorders, includ-

ing hypertension, myocardial infarction, and stroke

[35–37], disorders that also have been associated with

OSAH [24–27]. Hypoxia causes an imbalance in

vasoregulatory agents and promotes coagulation and

inflammation. As an example, through activation of

hypoxia-inducible factor-1 [38,39], hypoxia increases

the expression of several genes, such as genes that

encode endothelin-1, a potent vasoconstrictor with

proinflammatory properties. In contrast, hypoxia sup-

presses transcriptional rate of endothelial nitric oxide

synthase [40,41] and results in decreased production

of nitric oxide (NO), which is vasodilatory and has

antimitogenic properties. By enhancing expression of

adhesion molecules and promoting leukocyte rolling

and endothelial adherence [42], hypoxia may mediate

coronary artery inflammation. Hypoxia is also in-

volved in induction of cardiac and endothelial cells

apoptosis [43,44].

Most studies that show adverse effects of hypoxia

have been performed with sustained and severe hy-

poxia. Because in sleep apnea-hypopnea hypoxemia is

intermittent, the results of studies with sustained hy-

poxia might not be necessarily applicable to sleep-

related breathing disorders. Recent studies [45–47]

have shown that intermittent hypoxia (ie, hypoxia-

reoxygenation) also results in gene activation. In this

context, intermittent hypoxia may be analogous to

ischemia-reperfusion syndrome, and it has been pro-

posed to be more deleterious than sustained hypoxia

[48–50]. Support for a causative role of intermittent

hypoxia in induction of these abnormalities stems from

studies on treatment of OSAH with nasal continuous

positive airway pressure (CPAP). Several adverse

effects of hypoxia (eg, platelet activation [51], hyper-

fibrinogenemia, increased factor VII activity [53],

abnormal endothelium-dependant vasodilation [54],

and leukocyte activation [55,56]) observed in OSAH

are reversed by treatment with nasal CPAP. Enhanced

sympathetic activity, another neurohormonal con-

sequence of sleep-related breathing disorders that

results in adverse structural and functional cardiac

alterations, decreases after treatment of OSA and

central sleep apnea in heart failure.

It is conceivable that endothelial dysfunction

caused by sleep-related breathing disorders contrib-

utes to worsening of atherosclerosis, atherothrombo-

sis, and left ventricular dysfunction (Fig. 1). In this

regard, one study [57] has shown that untreated OSA

is a risk factor for cardiovascular disease, and two

prospective studies [58,59] of persons with myocar-

dial infarction have shown increased mortality rates

in persons with sleep apnea when compared with

individuals without it.

Finally, hypoxemia by stimulation of the carotid

bodies [60] causes sympathetic activation. In contrast

to the inhibitory function of the baroreceptors,

increased carotid body activity augments central

nervous system sympathetic outflow. In heart failure

with left ventricular systolic dysfunction, sympathetic

activity may be increased partly because of blunting

of baroreceptor activity and partly because of

increased carotid body stimulation [29].

There are multiple adverse cardiac consequences

of increased sympathetic activity. At cellular level,

increased catecholamines may cause myocyte apop-

tosis and fibrosis [61–63], both of which are inhibited

by b-adrenergic blockade [61,62]. Hemodynamically,

sympathetic activation increases systemic vascular

resistance and left ventricular afterload, myocardial

S. Javaheri / Clin Chest Med 24 (2003) 207–222210

contractility, and heart rate, all of which increase

myocardial oxygen demand. As a consequence of

hypoxemia, myocardial oxygen delivery may de-

crease, whereas consumption may increase and result

in an imbalance in supply/demand ratio. Adverse

consequences include myocardial cell hypoxia, sys-

tolic and diastolic dysfunction, angina, myocardial

infarction, and arrhythmias.

Arousal and shift to light sleep stages

In addition to hypoxemia and hypercapnia causing

increased sympathetic activity, arousals also increase

sympathetic activity. Comparing wakefulness to

sleep, there is a reduction in sympathetic activity

and increased parasympathetic activity [64–66].

These changes in autonomic nervous system during

sleep are reflected in a decrease in heart rate, blood

pressure, and cardiac output [67]. Sleep is peaceful for

the heart; however, arousals and awakenings result in

reversal of autonomic nervous system activity [68].

In patients with heart failure and systolic dys-

function, sleep is disturbed. This was evidenced in

studies that included a first night stay in the sleep

laboratory for adaptation to minimize sleep frag-

mentation [12,13]. Considerable sleep fragmentation

was observed in the second night polysomnography.

Arousals, insomnia, and shift to light sleep stages

that were observed in patients with systolic heart

failure were further exaggerated by presence of

sleep-related breathing disorders. Approximately

half of the sleep-disordered breathing events caused

cortical arousals [13]. In patients with OSAH,

arousals commonly occurred immediately before

termination of the breathing disorder and resulted

in patency of the upper airway and resumption of

breathing. In central sleep apnea, however, arousals

occurred at the peak of hyperventilation and served

no purpose but to fragment sleep and increase

sympathetic activity.

In the presence of sleep apnea-hypopnea, for

various reasons such as arousals, shift to light sleep

stages, hypoxia, and hypercapnia, nocturnal sympa-

thetic activity is elevated, which makes sleep not so

peaceful for the cardiovascular system. Increased

sympathetic activity is a predictor of poor survival

in systolic heart failure. Sleep-related breathing dis-

orders, by augmenting sympathetic activity, may

contribute to mortality in systolic heart failure. Impor-

tantly, treatment of sleep-related breathing disor-

ders—OSA [69,70] and central sleep apnea [71]—

decreases sympathetic activity and conceivably

improves survival of subjects with systolic heart

failure, in a manner similar to b-blockers.

Fig. 1. Proposed mechanisms by which sleep-related breathing disorders may cause or contribute to progression of ather-

osclerosis. DO2, oxygen delivery; CBF, coronary blood flow; z, increase; #, decrease.

S. Javaheri / Clin Chest Med 24 (2003) 207–222 211

Large negative deflections in intrathoracic pressure

Sleep-related breathing disorders are associated

with exaggerated negative inspiratory intrathoracic

pressure deflections. High negative intrathoracic pres-

sures may be generated during episodes of obstruc-

tive apnea [33]. After central apnea, hyperpnea occurs

and relatively large negative pressure deflections

[33], particularly in the face of stiff lungs and chest

wall, also may be observed. Pleural pressure changes,

however, are generally less pronounced in central

sleep apnea than in OSA [33].

The exaggerated negative intrathoracic pressure

increases the transmural pressure (pressure inside

minus pressure outside) of the intrathoracic vascular

structures including aorta, ventricles, and pulmonary

vascular bed. The consequences of exaggerated nega-

tive intrathoracic pressure include increased venous

return, increased left ventricular afterload, and pul-

monary congestion and edema [72–74].

Treatment of sleep-related breathing disorders

Obstructive sleep apnea

Treatment of OSAH in heart failure is similar to

that in the absence of heart failure (Box 2). The two

main therapeutic approaches are weight loss and

nasal mechanical devices.

Obesity is the major known risk factor for OSAH

[23,75,76] in the general population and in persons

with heart failure [13,16]. Importance of weight loss

in heart failure is particularly evident from recent data

from the Framingham Heart Study [77], which

showed that excess weight and obesity are associated

with and presumably cause heart failure. Another

study [78] showed that obesity was associated with

increased mortality, primarily because of cardiovas-

cular causes. Undoubtedly, however, several obese

persons in these two studies suffered from OSAH.

Undiagnosed OSAH could have been an important

contributing factor, linking obesity to heart failure

and cardiovascular mortality reported in these two

studies [77,78]. OSAH was not mentioned in these

two reports [77,78]. This observation is consistent

with the author’s earlier remarks regarding unfamili-

arity of physicians with sleep-related breathing dis-

orders and importance of education. Overweight and

obese subjects with heart failure should get dietary

consultation and be encouraged to lose weight, which

has been shown to decrease OSAH index [76].

Noninvasive mechanical devices have been used

most successfully to treat OSAH in the general popu-

lation [79–81]. There are limited reports on the use

of nasal CPAP for treatment of OSA in heart failure

[82–84]. Application of nasal CPAP results in sig-

nificant improvement in obstructive disordered breath-

ing events and arterial oxyhemoglobin desaturation.

Left ventricular ejection fraction increases with long-

term use of CPAP [83]. This is an important finding

because left ventricular ejection fraction is a predictor

of survival in systolic heart failure. Application of

nasal CPAP to treat OSAH in the general population

reverses several neurohormonal abnormalities, such

as abnormal endothelium-dependent vasodilata-

tion, hypercoagulopathy, and leukocyte activation

[51–56]. If, as expected, treatment of OSA in heart

failure results in reversal of the aforementioned patho-

logical process, progression of coronary artery inflam-

mation, thrombosis, and atherosclerosis may subside.

Rarely in congestive heart failure may treatment with

CPAP convert OSA to central sleep apnea [82].

Central sleep apnea

The approach to treatment of central sleep apnea

in systolic heart failure is somewhat different from

that of OSA [31,33]. Of utmost importance is

Box 2. Potential treatment options forsleep apnea-hypopnea in heart failure

Obstructive sleep apnea-hypopnea

Optimization of medical therapy forheart failure

Weight lossMechanical devices (CPAP,

bi-level pressure forCPAP noncompliance)

Oxygen for subjects noncompliantwith mechanical devices

Central sleep apnea

Optimization of medical therapy forheart failure

Cardiac transplantationMechanical devices (CPAP, bi-level

pressure, and adaptive pressuresupport servoventilation)

Medications (eg, oxygen,theophylline)

S. Javaheri / Clin Chest Med 24 (2003) 207–222212

improving cardiorespiratory function before per-

forming polysomnography.

Optimization of cardiopulmonary function

Early studies by Harrison et al [85] and more

recent studies [86–88] showed that treatment of heart

failure may improve or even eliminate periodic

breathing. Given the limited manpower and the cost

of polysomnography, the author suggests that in

congestive heart failure polysomnography be per-

formed only after optimization of cardiorespiratory

functions. Optimal treatment of heart failure with

diuretics, angiotensin-converting enzyme inhibitors,

cardiotonic drugs, and b-blockers may improve or

eliminate periodic breathing by several mechanisms,

including normalization of partial pressure carbon

dioxide in the arterial blood (PaCO2), improved

arterial circulation time, and normalization of func-

tional residual capacity.

Although a low arterial PCO2 is not a prerequisite

for development of central sleep apnea [89,90],

several studies [89,91,92] have shown that a low

arterial PCO2 while awake highly predicts central

sleep apnea. In the author’s study [89], many patients

with central sleep apnea had normal PaCO2; however,

the predictive value of a low arterial PCO2 defined as

35 mm Hg or less was approximately 80%.

Multiple mechanisms may contribute to hyper-

ventilation in congestive heart failure. The most

commonly quoted factor is pulmonary congestion.

It is believed that stimulation of pulmonary juxta-

capillary receptors by pulmonary vascular congestion

and edema causes tachypnea. The rise in respiratory

rate may result in an increase in alveolar ventilation

and hypocapnia. Another cause of hyperventilation in

heart failure could be increased sympathetic activity.

Limited data in humans [93] show that intravenously

infused sympathomimetic agents increase ventilation

and lower PCO2; this action is blocked by pretreat-

ment with propranolol [93].

How does a low PaCO2 predispose subjects with

heart failure to central sleep apnea? The difference

between two PCO2 set points (the baseline prevailing

PCO2, PCO2 at the apneic threshold) is critical for the

genesis of central apnea [94,95]. The smaller the

difference, the greater the likelihood of the person

having central apnea. Normally, in transition from

wakefulness to sleep, the prevailing PCO2 increases

and the difference between the prevailing PCO2 and

the PCO2 at the apneic threshold increases. As long as

the prevailing PCO2 remains above the apneic thresh-

old, central apnea does not occur. Limited studies

[92,94,96] showed that patients with heart failure and

central sleep apnea, in contrast to patients without

central sleep apnea, are unable to increase their

prevailing PCO2 in transition from wakefulness to

sleep. Because sleep unmasks the apnea threshold,

such persons become prone to developing central

apnea during sleep. It is conceivable that because of

cardiorespiratory effects of advanced heart failure

(particularly in patients with hypocapnia), PaCO2 fails

to rise with sleep onset. If true, this occurrence may

relate specifically to severity of the left ventricular

diastolic dysfunction. In the supine position, as

venous return increases, left ventricular end-diastolic

and pulmonary capillary pressures rise if the left

ventricle is noncompliant. As a result of the rise in

pulmonary capillary pressure and consequent conges-

tion and edema, juxta-capillary receptors are stimu-

lated, which causes tachypnea and hyperventilation.

In this regard, a negative correlation between arterial

PCO2 and wedge pressure has been reported [15].

Another factor that increases the likelihood of

periodic breathing is increased arterial circulation

time, which delays the transfer of information regard-

ing pulmonary capillary PO2 and PCO2 to the con-

trollers (the chemoreceptors). Increased arterial

circulation time converts a negative feedback system

into a positive one. In heart failure, arterial circulation

time may be increased for various reasons, including

a low stroke volume and increased intrathoracic

blood volume (pulmonary congestion, increased left

atrial and left ventricular volumes).

A third factor that increases the likelihood of

developing periodic breathing in heart failure is a

low functional residual capacity, which results in

underdampening. In heart failure, functional residual

capacity could be low [13] for various reasons, such as

pleural effusion, pulmonary edema, and cardiomegaly.

Pharmacologic treatment of heart failure with

diuretics, angiotensin-converting enzyme inhibitors,

and b-blockers could normalize PCO2 by decreasing

pulmonary congestion and decreasing sympathetic

activity. Treatment also could decrease arterial cir-

culation time as stroke volume increases and cardio-

pulmonary blood volume decreases and increase

functional residual capacity as cardiac size, pleural

effusion, and intravascular and extravascular lung

water decrease, all of which should stabilize breath-

ing during sleep.

b-blockers have been added to pharmacologic

treatment of heart failure [3,4] and have been shown

to improve survival considerably. The additional

beneficial effect of b-blockers over angiotensin-con-

verting enzyme inhibitors relates to their counterbal-

ancing of increased sympathetic activity, which is

present in congestive heart failure and could be

augmented further by consequences of sleep-related

S. Javaheri / Clin Chest Med 24 (2003) 207–222 213

breathing disorders. The improvement in survival for

heart failure with the use of b-blockers may be partly

caused by counterbalancing the sympathetic activity

caused by sleep-related breathing disorders. If

increased sympathetic activity causes hyperventila-

tion, which promotes central sleep apnea, b-blockersmay normalize PCO2 by decreasing sympathetic

activity and decrease the likelihood of occurrence of

central sleep apnea. b-blockers could improve or

eliminate sleep-related breathing disorders by

improving cardiac function and normalizing PCO2.

Any residual breathing disorders result in failure of

maximal sympathetic deactivation by b-blockers.It is important to emphasize that aggressive treat-

ment of heart failure with various medications may

decrease or even eliminate central sleep apnea. This

result has been shown with use of salt restriction,

diuretics, ionotropes, and angiotensin-converting

enzyme inhibitors. Ironically, there are no studies

with b-blockers, although b-blockers may improve

cardiorespiratory functions more than angiotensin-

converting enzyme inhibitors and result in consid-

erable improvement in central sleep apnea. As left

heart structure and function deteriorate with time,

even in the presence of b-blockers, central sleep

apnea worsens or recurs. Patients who have heart

failure and whose central sleep apnea is initially

improved by b-blockers must be followed serially

for recurrence of central sleep apnea.

Regarding b-blockers, the author also emphasizes

one side effect that relates to their effect on melato-

nin. Secretion of melatonin, a sleep-promoting chem-

ical, is via cyclic AMP-mediated b-receptor signal

transduction system, and b-blockers have been shown

to decrease melatonin secretion [97,98]. Carvedilol is

an exception [97].

Oxygen

Systematic studies of subjects with systolic heart

failure [99–103] have shown that nocturnal admin-

istration of supplemental nasal oxygen improves

central sleep apnea, eliminates desaturation, and

may decrease arousals and light sleep. Pembrey

[104] should be credited with the observation approx-

imately 100 years ago, and Hanly et al [99] should be

credited for the first randomized, placebo-controlled

study of nocturnal oxygen versus compressed nasal

air. In a study [99] of nine subjects with systolic heart

failure (mean left ventricular ejection fraction was

12%F 5%) that compared one night of nasal oxygen

versus air, AHI (30F 5 versus 19F 2) and arousal

index (30 F 8 versus 14 F 2) decreased and sleep

architecture improved significantly. In the largest

study [103], with 36 subjects with systolic heart

failure (mean left ventricular ejection fraction approx-

imately 22%), central apnea index decreased signifi-

cantly from approximately 28/hour to 10/hour [100].

In a randomized, placebo-controlled, double-blind

study, Andreas et al [101] showed that short-term

(1 week) administration of supplemental nocturnal

oxygen improved maximum exercise capacity. This

is an important finding because VO2max is an inde-

pendent predictor of survival in heart failure [106] and

coronary artery disease [105]. Another randomized,

placebo-controlled study of 4 weeks’ duration showed

that nocturnal administration of oxygen decreased

overnight urinary norepinephrine excretion [102].

Overnight urinary norepinephrine excretion may be

a better indicator of the overall nocturnal sympathetic

activity than a single venous blood sample of norepi-

nephrine obtained in the morning. Recently, Andreas

et al showed that nasal oxygen decreased the aug-

mented muscle sympathetic activity caused by vol-

untary central apnea in persons with systolic heart

failure [107].

Potential adverse effects of oxygen. Long-term

nocturnal and diurnal nasal oxygen has been used

extensively in chronic obstructive pulmonary disease

and has been shown to increase survival [108,109]. It

is remarkably free from side effects [108,109]. In

heart failure, however, oxygen may have adverse

hemodynamic effects in heart failure. This was

studied in seven awake subjects with class III and

IV heart failure in a cardiac catheterization laboratory

[110]. After baseline (breathing room air) hemody-

namic variables were obtained, subjects breathed

graded amounts of nasal oxygen, 24%, 40%, and

100%, each for 5 minutes [110]. Oxygen breathing

resulted in a progressive dose-dependent increase in

systemic vascular resistance and pulmonary capillary

pressure and a decrease in stroke volume. Regarding

the results of this study, however, several important

issues must be emphasized. The study lacked appro-

priate control, and the hemodynamic effects of lying

supine for the same period of time in these patients

with class III and IV heart failure were not available.

The duration of each trial was 5 minutes, during

which time steady state may not be achieved. The

hemodynamic effects of oxygen were most pro-

nounced with administration of 40% to 100% oxy-

gen, and more therapeutic ranges of nasal oxygen (ie,

28%, 32%, and 36%, equivalent to 2–4 L/minute)

were not studied.

Careful studies (placebo controlled) are necessary

to determine any hemodynamic effects of therapeutic

amounts of nasal oxygen (24%–36%) in patients

with heart failure. Finally, the mechanisms for the

potential hemodynamic effects also must be studied

S. Javaheri / Clin Chest Med 24 (2003) 207–222214

because increased oxidative stress is present in con-

gestive heart failure, and administration of additional

oxygen provides substrate for production of oxygen

radicals. The results of this study [110] may not apply

to subjects with heart failure and periodic breathing

who may be treated with lesser amounts of oxygen

therapeutically because of desaturation during sleep.

Several studies have shown that nocturnal oxygen

results in improvement in central sleep apnea, sleep

characteristics, exercise tolerance, and a reduction in

sympathetic activity. In this regard, nocturnal and

diurnal use of oxygen has proven useful for heart

failure in patients with cor pulmonale secondary to

chronic obstructive pulmonary disease.

Mechanisms of therapeutic effect of nasal oxygen

on central sleep apnea are multiple and include a

small rise in PCO2 [100,111], which presumably

increases the difference between the prevailing

PCO2 and the PCO2 at the apneic threshold, a

reduction in ventilatory response to CO2 [112], and

increasing body tissue stores (eg, lung and blood

contents) of oxygen, which increase damping. Con-

sequently, breathing during sleep should stabilize.

Short-term studies show that nocturnal oxygen

improves or eliminates central sleep apnea and asso-

ciated arousals, eliminates arterial oxyhemoglobin

desaturation, improves sympathetic activity, and

increases exercise capacity. Prospective, placebo-con-

trolled, long-term studies are necessary to determine

if nocturnal oxygen therapy has the potential to

decrease morbidity and mortality of patients with

systolic heart failure [113].

Nasal positive airway pressure devices

Various positive airway pressure devices have

been used to treat central sleep apnea in congestive

heart failure [82,114–118]. Nasal CPAP has been

studied most extensively. Several laboratories have

reported on acute and chronic use of CPAP in patients

with central sleep apnea with differing results

[82,114,117–121].

The author’s experience with acute (one night)

effect of CPAP on central sleep apnea has been

reported elsewhere [82]. The author studied 21

patients with central sleep apnea, 9 of whom (43%)

responded to CPAP. In these patients, CPAP virtually

eliminated disordered breathing (AHI decreased from

36/hour F 15/hour to 4/hour F 3/hour) and arterial

oxyhemoglobin desaturation. An important finding in

the author’s study was the effect of CPAP on ven-

tricular irritability during sleep [82]. In patients whose

sleep apnea-hypopnea responded to CPAP, the number

of premature ventricular contractions, couplets, and

ventricular tachycardias decreased. In contrast, CPAP

had no significant effect on ventricular irritability in

patients whose disordered breathing did not improve.

Although the author’s study enrolled the largest num-

ber of patients in an acute CPAP trial [82], the number

of patients was small, and the electrocardiographic

findings must be confirmed in a large study.

Chronic effects of CPAP on central sleep apnea

have been studied by Naughton et al [114] and Sin et

al [120]. In a randomized, parallel design, controlled

trial [114], heart failure patients were assigned to

either nasal CPAP (n = 14) or served as controls

(n = 5). Patients were followed for 3 months, and 12

subjects (in each arm) completed the study. Com-

paring paired variables obtained initially and after

1 month use of CPAP at approximately 10 cm H2O,

the AHI (43/hour F 5/hour versus 15/hour F5/hour) and arousal index (36/hour F 6/hour versus

24/hour F 4/hour) decreased significantly with

CPAP. There was an increase in left ventricular

ejection fraction from 21% F 4% to 29% F 5%

noted at 3 months after use of CPAP. Ejection fraction

did not change significantly in the control group.

Naughton et al [71] reported that use of CPAP

decreases sympathetic activity as measured by

plasma norepinephrine level and urinary norepineph-

rine excretion. These are important observations

because left ventricular ejection fraction and plasma

norepinephrine are predictors of survival in systolic

heart failure.

Sudden death (presumably caused by ventricular

arrhythmias) and pump failure are the two major

causes of death in systolic heart failure. By decreas-

ing ventricular arrhythmias [82] and improving ejec-

tion fraction [120], nasal CPAP may improve survival

in patients with systolic heart failure. In this regard, in

a randomized, controlled trial [117] of 29 patients

with central sleep apnea (n = 15 in control and 14 in

CPAP group), treatment analysis (ie, excluding the 2

CPAP noncompliant patients) showed a significant

reduction in 3-year mortality-cardiac transplantation

(P = 0.017, n = 12 in CPAP group and 15 in control

group). With intention to treat analysis (which

includes all patients enrolled), a similar trend was

observed (P = 0.06) [120].

There are several unresolved issues about the use

of CPAP for central sleep apnea in heart failure, and

further large studies [122] are necessary to confirm the

effects of CPAP on central sleep apnea in heart failure.

Researchers [82] found that 57% of patients with

central sleep apnea did not respond to CPAP (one

night). These patients had the most severe central sleep

apnea and had a tendency to have a low PaCO2.

Negative studies from some other laboratories have

been reported [117–119]. Davies et al [117] random-

S. Javaheri / Clin Chest Med 24 (2003) 207–222 215

ized eight patients with mean left ventricular ejection

fraction of 18% to either 2 weeks of CPAP (7.5 cm

H2O) or placebo (sham CPAP). Two patients withdrew

from the CPAP trial because of worsening of heart

failure, and one patient died. There were no significant

changes in periodic breathing in the remaining sub-

jects. Buckle et al [118] reported that one night’s use of

CPAP (5–7.5 cm H2O) had no significant effect on

periodic breathing in eight patients with heart failure

and systolic dysfunction. Guilleminault et al [117]

studied nine patients with systolic heart failure and

central sleep apnea. Titration (5–12 cm H2O) with

CPAP failed to eliminate periodic breathing and

arousals. One study reported increased muscle sym-

pathetic activity with short-term use of CPAP in

subjects with chronic heart failure [123].

Another concern with long-term use of CPAP is

compliance. In OSAH syndrome, compliance varies

and is probably related to several factors, particularly

patient perception of improvement. Because imme-

diate improvement in symptoms is often not observed

in heart failure patients with central sleep apnea,

high-level compliance may not be achieved.

Because of an increase in intrathoracic pressure,

venous return may decrease with CPAP, which results

in decreased stroke volume and hypotension. Heart

failure patients with atrial fibrillation [124], intra-

vascular hypovolemia, and normal left ventricular

end-diastolic blood pressure may be more vulnerable

than others. For these reasons, successful use of

CPAP in heart failure is not easily achieved and

requires a skillful team. Acute (first night) titration

is not necessary. Gradual (during a few weeks)

titration, treatment of complications, particularly

nasal clogging, and repeated follow-ups with encour-

agement are key factors for success.

The mechanisms by which CPAP improves cen-

tral sleep apnea are complex and probably multi-

factorial. Upper airway closure has been shown to

occur in central sleep apnea [125,126], and in a full-

night polysomnographic study [11] and a nap study

[127], obstructed breaths were observed at the end of

some central apneas. CPAP could stabilize breathing

by increasing transmural pressure of upper airway, a

mechanism similar to that in obstructive apnea. It is

also possible that pressure stimulation of various

receptors in the upper airway could improve central

apnea because upper airway (laryngeal and pharyn-

geal) receptors are important in regulating the timing

and amplitude of breathing.

Another set of mechanisms by which CPAP may

improve central apnea may relate to improvement in

pathogenic factors that predispose subjects with sys-

tolic heart failure to central sleep apnea. Prolonged

arterial circulation, decreased functional residual

capacity, and a low arterial PCO2 are among predis-

posing factors that could be reversed toward normal

by application of CPAP. Nasal CPAP may shorten

arterial circulation time by decreasing afterload and

increasing stroke volume. By increasing intrathoracic

pressure, CPAP may decrease intrathoracic blood

volume (pulmonary intravascular and intracavitary

blood volume), which also should shorten circulation

time and stabilize breathing. Nasal CPAP increases

functional residual capacity, which should increase

damping and stabilize breathing. Although long-term

use of CPAP has been shown to increase PaCO2,

which should decrease the likelihood of developing

central sleep apnea, acute use of CPAP does not

significantly increase PaCO2 in patients with heart

failure and central sleep apnea [116]. By increasing

dead space and ventilation/perfusion ratio of some

areas of the lung, however, application of nasal CPAP

could increase PCO2 acutely. Further careful studies

with change in PaCO2 as the outcome parameter are

necessary to determine acute effects of CPAP on

PaCO2. The author noted that application of positive

end-expiratory pressure in adult respiratory distress

syndrome does not significantly change extravascular

lung water but affects its redistribution [128–132].

The same also could be true for effects of CPAP on

lung water in congestive heart failure, which may

further improve PO2 and increase damping.

Continuous positive airway pressure is uniformly

effective in treating OSAH in patients with heart

failure. It also could be effective in treating central

sleep apnea. The approach in the use of CPAP is

different in these two disorders, however. Acute

CPAP titration is necessary to determine the precise

upper airway pressure required to eliminate OSAH in

heart failure in a similar manner used to eliminate

OSAH in the absence of heart failure. For central

sleep apnea, the author suggests gradually increasing

the pressure from 5 to 10 to 12 cm H2O over many

days to weeks as tolerated by the patient. The

author’s protocol requires initiation of CPAP during

daytime in the laboratory. While the patient is in the

supine position, blood pressure and saturation are

monitored for 30 minutes to 1 hour as CPAP

increases from 5 to 7 cm H2O. Further increments

are made a few days apart, usually under similar

circumstances. Immediate overnight CPAP titration in

the sleep laboratory, particularly in persons with

severe central sleep apnea, is not necessary. Careful

follow-ups with frequent phone calls and visits for

aggressive treatment of complications (eg, nasal

clogging) and encouragement to use CPAP are neces-

sary to improve long-term compliance.

S. Javaheri / Clin Chest Med 24 (2003) 207–222216

Another noninvasive device used to treat central

sleep apnea is adaptive pressure support servoventi-

lation. This device provides varying amounts of

ventilatory support during different phases of periodic

breathing. The support is maximal during central

apnea and minimal during the hyperpneic phase of

periodic breathing [116]. The device provides a con-

stant positive expiratory pressure that should be

enough to eliminate obstructive apneas when present,

and pressure support ventilation is provided by super-

imposing additional, although variable, inspiratory

pressure. For example, if expiratory pressure is set at

6 cm H2O, when ventilation is stable, the inspiratory

pressure could be approximately 9 cm H2O; however,

inspiratory pressure quickly increases further if central

apnea develops. The inspiratory pressure returns to

9 cm H2O when breathing stabilizes.

This pattern of pressure support should be easier

to tolerate, particularly for patients with severe peri-

odic breathing and repetitive episodes of hyperpnea.

In an acute (one-night) study [116] of 14 subjects

with systolic heart failure and central sleep apnea,

the AHI decreased significantly from approximately

45/hour to 6/hour. The improvement by the adaptive

pressure support servoventilation was better than that

observed with either CPAP or oxygen. The author

believes that the adaptive pressure support servoven-

tilation device should be particularly helpful in heart

failure patients with severe central sleep apnea who

may be nonresponsive [82] or noncompliant to CPAP.

An adaptive pressure support servoventilation device

could be used initially and then replaced later by

CPAP as cardiopulmonary function and central sleep

apnea improve. Large-scale studies are needed.

Theophylline

Open studies [11,133] have shown the efficacy of

theophylline in the treatment of central sleep apnea

in heart failure. In a double-blind, randomized, pla-

cebo-controlled, cross-over study [134] of 15 patients

with treated, stable systolic heart failure, theophyl-

line, twice daily by mouth, at therapeutic plasma

concentrations (average, 11 mg/mL; range 7mg/mL–

15/mg/mL) decreased the AHI by approximately 50%

and improved arterial oxyhemoglobin saturation

[133]. Theophylline significantly decreased central

apnea but had no effect on OSA [134].

Mechanisms of action of theophylline in improv-

ing central apnea remain unclear [134]. At therapeutic

serum concentrations, theophylline increases ventila-

tion [134]. This action probably is caused by com-

petitive inhibition of adenosine, which is a respiratory

depressant [135,136]. An increase in ventilation by

theophylline could decrease the likelihood of occur-

rence of central apnea during sleep. Although theo-

phylline slightly decreases end-tidal and arterial

PCO2 [128,129], the assumption is that theophylline

also decreases PCO2 at the apneic threshold, and the

difference between the two PCO2 set point does not

decrease or may even increase. This could be similar

to the action of almitrine [95], another respiratory

stimulant. Theophylline does not increase ventilatory

response to CO2 [135].

Potential arrhythmogenic effects and phospho-

diesterase inhibition are common concerns with

long-term use of theophylline in patients with heart

failure. Further controlled studies are necessary to

ensure its safety. If theophylline is used to treat

central sleep apnea, frequent and careful follow-ups

are necessary.

Summary

Heart failure is a highly prevalent problem asso-

ciated with excess morbidity and mortality and eco-

nomic impact. Because of increased average life

span, improved therapy of ischemic coronary artery

disease and hypertension, the incidence and preva-

lence of heart failure will continue to rise into the

twenty-first century.

Multiple factors may contribute to the progres-

sively declining course of heart failure. One such

cause could be the occurrence of repetitive episodes

of apnea, hypopnea, and hyperpnea, which frequently

occur in patients with heart failure. Episodes of

apnea, hypopnea, and hyperpnea cause sleep disrup-

tion, arousals, intermittent hypoxemia, hypercapnia,

hypocapnia, and changes in intrathoracic pressure.

These pathophysiologic consequences of sleep-

related breathing disorders have deleterious effects

on cardiovascular system, and the effects may be

most pronounced in the setting of established heart

failure and coronary artery disease. Diagnosis and

treatment of sleep-related breathing disorders may

improve morbidity and mortality of patients with

heart failure [34]. Large-scale, carefully executed

therapeutic studies are needed to determine if treat-

ment of sleep-related breathing disorders changes the

natural history of left ventricular failure.

References

[1] Colucci WS, Braunwald E. Heart disease. In: Braun-

wald E, Zipes DP, Libby P, editors. Pathophysiology

of heart failure. Philadelphia: WB Saunders; 2001.

p. 503–33.

S. Javaheri / Clin Chest Med 24 (2003) 207–222 217

[2] American Heart Association. 2002 Heart and stroke

statistical update. Dallas: American Heart Associa-

tion; 2001.

[3] Hunt SA. ACC/AHA Guidelines for the evaluation

and management of chronic heart failure in the adult:

executive summary. J Am Coll Cardiol 2001;38:

2101–13.

[4] Consensus recommendations for the management

of chronic heart failure. On behalf of the member-

ship of the advisory council to improve outcomes

nationwide in heart failure. Am J Cardiol 1999;83:

2A–38A.

[5] McCullough PA, Philbin EF, Spertus JA, Kaatz S,

Sandberg KR, Weaver WD. Confirmation of a heart

failure epidemic: findings from the resource utiliza-

tion among congestive heart failure (REACH) study.

J Am Coll Cardiol 2002;39:60–9.

[6] Croft JB, Giles WH, Pollard RA, Keenan NL, Casper

ML, Anda RF. Heart failure survival among older

adults in the United States. Arch Intern Med 1999;

159:505–10.

[7] O’Connell JB. The economic burden of heart failure.

Clin Cardiol 2000;23(3 Suppl):III6–10.

[8] Goff DC, Pandey DK, Chan FA, Ortiz C, Nichaman

MZ. Congestive heart failure in the United States.

Arch Intern Med 2000;160:197–202.

[9] Javaheri S. Prevalence and prognostic significance of

sleep apnea in heart failure. In: Bradley TD, Floras

JS, editors. Sleep apnea: implications in cardiovascu-

lar and cerebrovascular disease. New York: Marcel

Dekker; 2000. p. 415–33.

[10] Hanly PJ, Millar TW, Steljes DG. Respiration and

abnormal sleep in patients with congestive heart fail-

ure. Chest 1989;96:480–8.

[11] Dowdell WT, Javaheri S, McGinnis W. Cheyne-

Stokes respiration presenting as sleep apnea syn-

drome: Clinical and polysomnographic features. Am

Rev Respir Dis 1990;141:871–9.

[12] Javaheri S, Parker TJ, Wexler L, Michaels SE, Stan-

berry E, Nishyama H, et al. Occult sleep-disordered

breathing in stable congestive heart failure. Ann In-

tern Med 1995;122:487–92 [Erratum appeared in

Ann Intern Med 1995;123:77].

[13] Javaheri S, Parker TJ, Liming JD, Corbett BS, Nish-

iyama H, Wexler L, et al. Sleep apnea in 81 ambula-

tory male patients with stable heart failure: types and

their prevalences, consequences, and presentations.

Circulation 1998;97:2154–9.

[14] Tremel F, Pepin J-L, Veale D, Wuyam B, Siche J-P,

Mallion JM, et al. High prevalence and persistence of

sleep apnoea in patients referred for acute left ven-

tricular failure and medically treated over 2 months.

Eur Heart J 1999;20:1201–9.

[15] Solin P, Bergin P, Richardson M, Kaye DM, Walters

EH, Naughton MD. Influence of pulmonary capillary

wedge pressure on central apnea in heart failure. Cir-

culation 1999;99:1574–9.

[16] Sin DD, Fitzgerald F, Parker JD, Newton G, Floras

JS, Bradley TD. Risk factors for central and obstruc-

tive sleep apnea in 450 men and women with con-

gestive heart failure. Am J Respir Crit Care Med

1999;160:1101–6.

[17] Yasuma F, Nomura H, Hayashi H, Okada T, Tsuzuki

M. Breathing abnormalities during sleep in pa-

tients with chronic heart failure. Jpn Circ J 1989;53:

1506–10.

[18] Lofaso F, Verschueren P, Rande JLD, Harf A, Gold-

enberg F. Prevalence of sleep-disordered breathing in

patients on a heart transplant waiting list. Chest 1994;

106:1689–94.

[19] Traversi E, Callegari G, Pozzoli M, Opasich C, Tavaz-

zi L. Sleep disorders and breathing alterations in pa-

tients with chronic heart failure. G Ital Cardiol 1997;

27:423–9.

[20] Fries R, Bauer D, Heisel A, Juhasz J, Fichter J,

Schieffer H, et al. Clinical significance of sleep-re-

lated breathing disorders in patients with implantable

cardioverter defibrillators. Pace 1999;22:223–7.

[21] Staniforth AD, Kinnear WJM, Starling R, Cowley AJ.

Nocturnal desaturation in patients with stable heart

failure. Heart 1998;79:394–9.

[22] Lanfranchi PA, Braghiroli A, Bosimini E, Mazzuero

G, Colombo R, Donner CF, et al. Prognostic value of

nocturnal Cheyne-Stokes respiration in chronic heart

failure. Circulation 1999;99:1435–40.

[23] Young T, Palta M, Dempsey J, Skatrud J, Wever S,

Safwan B. The occurrence of sleep-disordered breath-

ing among middle-aged adults. N Engl J Med 1993;

328:1230–5.

[24] Peppard PE, Young T, Palta M, Saktrud J. Prospective

study of the association between sleep-disordered

breathing and hypertension. N Engl J Med 2000;

342:1378–84.

[25] Shahar E, Whitney CW, Redline S, Lee ET, Newman

AB, Javier Nieto F, et al. Sleep-disordered breathing

and cardiovascular disease: cross-sectional results of

the Sleep Heart Health Study. Am J Respir Crit Care

Med 2001;163:19–25.

[26] Neito FJ, Young TB, Lind BK, Shahrar E, Samet JM,

Redline S, et al. Association of sleep-disordered

breathing, sleep apnea, and hypertension in a large

community-based study. JAMA 2000;283:1829–36.

[27] Chan J, Sanderson J, Chan W, Lai C, Choy D, Ho A,

et al. Prevalence of sleep-disordered breathing in dia-

stolic heart failure. Chest 1997;111:1488–93.

[28] Vasan RS, Benjamin EJ, Levy D. Prevalence, clinical

features and prognosis of DHF: an epidemiologic per-

spective. J Am Coll Cardiol 1995;26:1565–74.

[29] Floras JS. Clinical aspects of sympathetic activation

and parasympathetic withdrawal in heart failure.

J Am Coll Cardiol 1993;22:72A–84A.

[30] Cargill JI, Kiely DG, Lipworth BJ. Adverse effects of

hypoxemia on diastolic filling in humans. Clin Sci

1995;89:165–9.

[31] Javaheri S. Central sleep apnea-hypopnea syndrome

in heart failure: prevalence, impact and treatment.

Sleep 1996;19:S229–31.

[32] Leung RST, Bradley TD. Sleep apnea and cardiovas-

S. Javaheri / Clin Chest Med 24 (2003) 207–222218

cular disease. Am J Respir Crit Care Med 2001;164:

2147–65.

[33] Javaheri S. Sleep-related breathing disorders in heart

failure: diagnosis and treatment. Sleep Med Alert Na-

tional Sleep Foundation 2002;61:1–5.

[34] Nakao K, Ohgushi M, Yoshimura M, et al. Hyper-

ventilation as a specific test for diagnosis of coronary

artery spasm. Am J Cardiol 1997;80:545–9.

[35] Mombouli JV, Vanhoutte PM. Endothelial dysfunc-

tion: from physiology to therapy. J Mol Cell Cardiol

1999;31:61–74.

[36] Britten MB, Zeiher AM, Schachinger V. Clinical im-

portance of coronary endothelial vasodilator dysfunc-

tion and therapeutic options. J Intern Med 1999;245:

315–27.

[37] Biegelsen ES, Loscalzo J. Endothelial function and

atherosclerosis. Coron Artery Dis 1999;10:241–56.

[38] Yu AY, Shimoda LA, Iyer NV, Huso DL, Sun X,

McWilliams R, et al. Impaired physiological re-

sponses to chronic hypoxia in mice partially deficient

for hypoxia-inducible factor 1a. J Clin Invest 1999;

103:691–6.

[39] Koong AC, Chen EY, Giaccia AJ. Hypoxia causes the

activation of nuclear factor kB through the phospho-

rylation of IkBa on tyrosine residues. Cancer Res

1994;54:1425–30.

[40] Phelan MW, Faller DV. Hypoxia decreases constitu-

tive nitric oxide synthase transcript and protein in

cultured endothelial cells. J Cell Physiol 1996;167:

469–76.

[41] Faller DV. Endothelial cell responses to hypoxic

stress. Clin Exp Pharmacol Physiol 1999;26:74–84.

[42] Gonzalez NC, Wood JG. Leukocyte-endothelial inter-

actions in environmental hypoxia. Adv Exp Med Biol

2001;502:39–60.

[43] Tanaka M, Ito H, Adachi S, Akimoto H, Nishikawa T,

Kasajima T, et al. Hypoxia induces apoptosis with

enhanced expression of Fas antigen messenger RNA

in cultured neonatal rat cardiomyocytes. Circ Res

1994;75:426–33.

[44] AokiM, Nata T,Morishita R,Matsushita H, Nakagami

H, Yamamoto K, et al. Endothelial apoptosis induced

by oxidative stress through activation of NF-kB: anti-

apoptotic effect of antioxidant agents on endothelial

cells. Hypertension 2001;38:48–55.

[45] Muschel RJ, Bernhard EJ, Garza L, McKenna G,

Koch CJ. Induction of apoptosis at different oxygen

tension: evidence that oxygen radicals do not mediate

apoptotic signaling. Cancer Res 1995;55:995–8.

[46] Kanagy NL, Walker BR, Nelin LD. Role of endothe-

lin in intermittent hypoxia-induced hypertension. Hy-

pertension 2001;37:511–5.

[47] Adhikary G, Kline D, Yuan G, Kumar GK, Simonson

MS, Cherniack NS, et al. Gene regulation during in-

termittent hypoxia: evidence for the involvement of re-

active oxygen species. Adv Exp Med Biol 2001;499:

297–302.

[48] Ichikawa H, Flores S, Kvietys PR, Wolf RE, Yoshi-

kawa T, Granger DN, et al. Molecular mechanisms of

anoxia/reoxygenation-induced neutrophil adherence

to cultured endothelial cells. Circ Res 1997;81:

922–31.

[49] Samarasinghe DA, Tapner M, Farrell GC. Role of

oxidative stress in hypoxia-reoxygenation injury to

cultured rat hepatic sinusoidal endothelial cells. Hep-

atology 2000;31:1600–5.

[50] Prabhakar NR. Physiological and genomic conse-

quences of intermittent hypoxia. Invited review: Oxy-

gen sensing during intermittent hypoxia: cellular and

molecular mechanisms. J Appl Physiol 2001;90:

1986–94.

[51] Bokinsky G, Miller M, Ault K, Husband P, Mitchell

J. Spontaneous platelet activation and aggregation

during obstructive sleep apnea and its response to

therapy with nasal continuous positive airway pres-

sure. Chest 1995;108:625–30.

[52] Chin K, Ohi M, Kita H, Noguchi T, Otsuka N, Tsuboi

T, et al. Effects of NCPAP therapy on fibrinogen

levels in obstructive sleep apnea syndrome. Am J

Respir Crit Care Med 1996;153:1972–6.

[53] Chin K, Kita H, Noguchi T, Otsuka N, Tsuboi T,

Nakamura T, et al. Improvement of factor VII clotting

activity following long-term NCPAP treatment in ob-

structive sleep apnoea syndrome. Q J Med 1998;91:

627–33.

[54] Ip MSM, Lam B, Chan LY, Zheng L, Tasng KWT,

Fung PCW, et al. Circulating nitric oxide is sup-

pressed in obstructive sleep apnea and is reversed

by nasal continuous positive airway pressure. Am J

Respir Crit Care Med 2000;162:2166–71.

[55] Miyasaka M, Ohi M. Effects of nasal continuous

positive airway pressure on soluble cell adhesion

molecules in patients with obstructive sleep apnea

syndrome. Am J Med 2000;109:562–7.

[56] Dyugovskaya L, Lavie P, Lavie L. Increased adhesion

molecules expression and production of reactive oxy-

gen species in leukocytes of sleep apnea patients. Am

J Respir Crit Care Med 2002;165:934–9.

[57] Peker Y, Hedner J, Norum J, Kraiczi H, Carlson J.

Increased incidence of cardiovascular disease in mid-

dle-aged men with obstructive sleep apnea. Am J

Respir Crit Care Med 2002;166:159–65.

[58] Peker Y, Hedner J, Kraiczi H, Loth S. Respiratory

disturbance index: an independent predictor of mor-

tality in coronary artery disease. Am J Respir Crit

Care Med 2000;162:81–6.

[59] Mooe T, Franklin KA, Holmstrom K, Rabben T, Wi-

klundu U. Sleep-disordered breathing and coronary

artery disease. Am J Respir Crit Care Med 2001;164:

1910–3.

[60] Fletcher EC, Lesske J, Behm R, Miller CCL, Stauss

H, Unger T. Carotid chemoreceptors, systemic blood

pressure, and chronic episodic hypoxia mimicking

sleep apnea. J Appl Physiol 1992;72:1978–84.

[61] Mann DL, Kent RL, Parsons B, Cooper G. Adrener-

gic effects on the biology of the adult mammalian

cardiocyte. Circulation 1992;85:790–804.

[62] Communal C, Singh K, Pimental DR, Colucci WS.

S. Javaheri / Clin Chest Med 24 (2003) 207–222 219

Norepinephrine stimulates apoptosis in adult rat ven-

tricular myocytes by activation of the b-adrenergicpathway. Circulation 1998;98:1329–34.

[63] Kanai EI, Hasegawa K, Araki M, Kakita T, Morimoto

T, Sasayama S. a and b-Adrenergic pathways differ-

entially regulate cell type-specific apoptosis in rat

cardiac myocytes. Circulation 1999;100:305–11.

[64] Mancia G. Autonomic modulation of the cardiovas-

cular system during sleep. N Engl J Med 1993;238:

347–9.

[65] Baust W, Weidinger H, Kirchner F. Sympathetic ac-

tivity during natural sleep and arousal. Arch Ital Biol

1968;106:379–90.

[66] Somers VK, Dyken ME, Mark AL, Abboud FM.

Sympathetic-nerve activity during sleep in normal

subjects. N Engl J Med 1993;328:303–7.

[67] Khatri IM, Freis ED. Hemodynamic changes during

sleep. J Appl Physiol 1967;22:867–73.

[68] Horner RL, Brooks D, Kozar LF, Tse S, Phillipson

EA. Immediate effects of arousal from sleep on car-

diac autonomic outflow in the absence of breathing in

dogs. J Appl Physiol 1995;79:151–62.

[69] Fletcher EC, Miller J, Schaaf JW, Fletcher JG. Uri-

nary catecholamines before and after tracheostomy in

patients with obstructive sleep apnea and hyperten-

sion. Sleep 1987;10:35–44.

[70] Narkiewicz K, Kato M, Phillips BG, Pesesk CA, Da-

vison DE, Somers VK. Nocturnal continuous positive

airway pressure decreases daytime sympathetic traffic

in obstructive sleep apnea. Circulation 1999;100:

2332–5.

[71] Naughton MT, Benard DC, Liu PP, Rutherford R,

Rankin F, Bradley TD. Effects of nasal CPAP on

sympathetic activity in patients with heart failure

and central sleep apnea. Am J Respir Crit Care

Med 1995;152:473–9.

[72] Buda AJ, Pinsky MR, Ingels Jr NB, Daughters GT,

Stinson EB, Alderman EL. Effect of intrathoracic

pressure on left ventricular performance. N Engl J

Med 1979;301:453–9.

[73] Fessler HE, Brower RG, Wise RA, Permutt S. Mech-

anism of reduced LV afterload by systolic and dia-

stolic positive pleural pressure. J Appl Physiol 1988;

65:1244–50.

[74] Fletcher EC, Proctor M, Yu J, Zhang J, Guardiola JJ,

Hornung C, et al. Pulmonary edema develops after

recurrent obstructive apneas. Am J Respir Crit Care

Med 1999;160:1688–96.

[75] Redline S, Strohl KP. Recognition and consequences

of obstructive sleep apnea hypopnea syndrome. Clin

Chest Med 1998;19:1–20.

[76] Young T, Peppard PE, Gottlieb DJ. Epidemiology of

obstructive sleep apnea. Am J Respir Crit Care Med

2002;165:1217–39.

[77] Kenchaiah S, Evans JC, Levy D, Wilson PWF, Ben-

jamin EJ, Larson MG, et al. Obesity and the risk of

heart failure. N Engl J Med 2002;347:305–13.

[78] Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath

Jr CW. Body mass index and mortality in a prospec-

tive cohort of US adults. N Engl J Med 1999;241:

1097–105.

[79] Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Re-

versal of obstructive sleep apnea by continuous pos-

itive airway pressure applied through the nares.

Lancet 1981;1:862–5.

[80] Rappaport DM, Sorkin B, Garay SM, Goldring RM.

Reversal of ‘‘pickwickian syndrome’’ by long term

use of nocturnal airway pressure. N Engl J Med

1982;307:931–3.

[81] Sanders MH, Moore SE, Eveselage J. CPAP via nasal

mask: a treatment for occlusive sleep apnea. Chest

1983;83:144–5.

[82] Javaheri S. Effects of continuous positive airway

pressure on sleep apnea and ventricular irritability

in patients with heart failure. Circulation 2000;101:

392–7.

[83] Malone S, Liu PP, Holloway R, Rutherford R, Xie A,

Bradley TD. Obstructive sleep apnoea in patients with

dilated cardiomyopathy: effects of continuous posi-

tive airway pressure. Lancet 1991;338:1480–4.

[84] Naughton MT, Liu PP, Benard DC, Goldstein RS,

Bradley TD. Treatment of congestive heart failure

and Cheyne-Stokes respiration during sleep by con-

tinuous positive airway pressure. Am J Respir Crit

Care Med 1995;151:92–7.

[85] Harrison TR, King CE, Calhoun JA, Harrison WG.

Congestive heart failure: Cheyne-Stokes respiration

as the cause of paroxysmal dyspnea at the onset of

sleep. Arch Intern Med 1934;53:891–910.

[86] Walsh JT, Andrews R, Starling R, Cowley AJ, John-

ston IDA, Kinnear W. Effects of captopril and oxygen

on sleep apnea in patients with mild to moderate con-

gestive cardiac failure. Br Heart J 1995;73:237–41.

[87] Dark DS, Pingleton SK, Kerby CR, Crabb JE, Gollub

SB, Glatter TR. Breathing pattern abnormalities and

arterial oxygen desaturation during sleep in the con-

gestive heart failure syndrome. Chest 1987;91:833–6.

[88] Ribeiro JP, Knutzen A, Rocco MB, Hartley H, Co-

lucci WS. Periodic breathing during exercise in se-

vere heart failure: reversal with milrinone or cardiac

transplantation. Chest 1987;92:555–6.

[89] Javaheri S, Corbett WS. Association of low PaCO2

with central sleep apnea and ventricular arrhythmias

in ambulatory patients with stable heart failure. Ann

Intern Med 1998;128:204–7.

[90] Javaheri S. Central sleep apnea and heart failure. Cir-

culation 2000;342:293–4.

[91] Hanly P, Zuberi N, Gray R. Pathogenesis of Cheyne-

Stokes respiration in patients with congestive heart

failure: relationship to arterial PCO2. Chest 1993;

104:1079–84.

[92] Naughton H, Benard D, Tam A, Rutherford R, Brad-

ley TD. Role of hyperventilation in the pathogenesis

of central sleep apnea in patients with congestive

heart failure. Am Rev Respir Dis 1993;148:330–8.

[93] Heistad DD, Wheeler RC, Mark AL, Schmid PG,

Abboud FM. Effects of adrenergic stimulation on

ventilation in man. J Clin Invest 1972;51:1469–75.

S. Javaheri / Clin Chest Med 24 (2003) 207–222220

[94] Xie A, Skatrud JB, Puleo DS, Ranko PS, Dempsey

JA. Apnea-hypopnea threshold for CO2 in patients

with congestive heart failure. Am J Respir Crit Care

Med 2002;165:1245–50.

[95] Nakayama H, Smith CA, Rodman JR, Skatrud JB,

Dempsey JA. Effect of ventilatory drive on CO2 sen-

sitivity below apnea during sleep. Am J Crit Care

Med 2002;165:1251–9.

[96] Tkacova R, Hall ML, Luie PP, Fitzgerald FS, Bradley

TD. Left ventricular volume in patients with heart

failure and Cheyne-Stokes respiration during sleep.

Am J Respir Crit Care Med 1997;156:1549–55.

[97] Arendt J, Bojkowsxki C, Franey C, Wright J, Marks

V. Immunoassay of 6-hydroxymelatonin sulfate in

human plasma and urine: abolition of the urinary

24-hour rhythm with atenolol. J Clin Endocrinol

Metab 1985;60:1166–73.

[98] Stoschitzky K, Sakotnik A, Lercher P, Zweiker R,

Maier R, Liebmann P, et al. Influence of beta-blockers

on melatonin release. Eur J Clin Pharmacol 1999;

55:111–5.

[99] Hanly PF, Millar TW, Steljes DG, Baert R, Frais MA,

Kryger MH. The effect of oxygen on respiration and

sleep in patients with congestive heart failure. Ann

Intern Med 1989;111:777–82.

[100] Franklin KA, Eriksson P, Sahlin C, Lundgren R. Re-

versal of central sleep apnea with oxygen. Chest

1997;111:163–9.

[101] Andreas S, Clemens C, Sandholzer H, Figulla HR,

Kreuzer H. Improvement of exercise capacity with

treatment of Cheyne-Stokes respiration in patients

with congestive heart failure. J Am Coll Cardiol

1996;27:1486–90.

[102] Staniforth AD, Kinneart WJM, Starlight R, Hetman-

ski DJ, Cowley AJ. Effect of oxygen on sleep quality,

cognitive function and sympathetic activity in patients

with chronic heart failure and Cheyne-Stokes respira-

tion. Eur Heart J 1998;19:922–8.

[103] Javaheri S, Ahmed M, Parker TJ, with technical

assistance of Brown CR. Effects of nasal O2 on

sleep-related disordered breathing in ambulatory pa-

tients with stable heart failure. Sleep 1999;22:

1101–6.

[104] Pembrey MS. Observations on Cheyne-Stokes respi-

ration. J Pathol Bacteriol 1908;12:259–66.

[105] Myers J, Gullestad L, Vagelos R, Do D, Bellin D,

Ross H. clinical, hemodynamic, and cardiopulmonary

exercise test determinants of survival in patients re-

ferred for evaluation of heart failure. Ann Intern Med

1998;129:286–93.

[106] Myers J, Prakash M, Froelicher V, Do D, Partington

S, Atwood JE. Exercise capacity and mortality among

men referred for exercise testing. N Engl J Med 2002;

346:793–801.

[107] Andreas S, Bingeli C, Mohacsi P, Luscher TF, Noll G.

Nasal oxygen and muscle sympathetic nerve activity

in heart failure. Chest 2003;123:366–71.

[108] Long term domicillary oxygen therapy in chronic hy-

poxic cor pulmonale complicating chronic bronchitis

and emphysema. Report of the Medical Research

Council Working Party. Lancet 1981;1(8222):681–6.

[109] Neff TA, Petty TL. Long term continuous oxygen

therapy in chronic airway obstruction. Ann Intern

Med 1970;72:621–6.

[110] Haque WA, Boehmer J, Clemson BS, Leuenberger

UA, Silber DH, Sinoway LI. Hemodynamic effects

of supplemental oxygen administration in conges-

tive heart failure. J Am Coll Cardiol 1996;27:

353–7.

[111] Andreas S, Plock EH, Heindl S, Scholz KH. Nasal

oxygen effects on arterial carbon dioxide pressure and

heart rate in chronic heart failure. Am J Cardiol

1999;83:795–8.

[112] Andreas S, Von Zur Muhlen F, Stevens J, et al. Noc-

turnal oxygen and hypercapnic ventilatory response in

patients with congestive heart failure. Respir Med

1998;92:426–31.

[113] Javaheri S. Pembrey’s dream: time has come for a

long-term trial of nocturnal supplemental nasal oxy-

gen to treat CSA in CHF. Chest 2003;123:322–5.

[114] Naughton MT, Liu PP, Benard DC, et al. Treatment of

congestive heart failure and Cheyne-Stokes respira-

tion during sleep by continuous positive airway pres-

sure. Am J Respir Crit Care Med 1995;151:92–7.

[115] Wilson GN, Wilcox I, Piper AJ, Flynn WE, Grunstein

RR, Sullivan CE. Treatment of central sleep apnoea in

congestive heart failure with nasal ventilation. Thorax

1998;53:S41–6.

[116] Teschler H, Dohring J, Wang YM, Berthon-Jones M.

Adaptive pressure support servo-ventilation. Am J

Respir Crit Care Med 2001;164:614–9.

[117] Guilleminault C, Clerk A, Labanowski M, Simmons

J, Stoohs R. Cardiac failure and benzodiazepines.

Sleep 1993;16:524–8.

[118] Buckle P, Millar T, Kryger M. The effects of short-

term nasal CPAP on Cheyne-Stokes respiration in

congestive heart failure. Chest 1992;102:31–5.

[119] Davies RJO, Harrington KJ, Omerod JM, et al. Nasal

continuous positive airway pressure in chronic heart

failure with sleep-disordered breathing. Am Rev Res-

pir Dis 1993;147:630–4.

[120] Sin DD, Logan AG, Fitzgerald FS, Liu PP, Bradley

TD. Effects of continuous positive airway pressure on

cardiovascular outcomes in heart failure patients with

and without Cheyne-Stokes respiration. Circulation

2000;102:61–6.

[121] Krachman SL, D’Alonzo GE, Berger TJ, Eisen HJ.

Comparison of oxygen therapy with nasal continuous

positive airway pressure on Cheyne-Stokes respira-

tion during sleep in congestive heart failure. Chest

1999;116:1550–7.

[122] Bradley TD, Logan AG, Floras JS, CANPAP Inves-

tigators. Rationale and design of the Canadian con-

tinuous positive airway pressure trial for congestive

heart failure patients with central sleep apnea: CAN-

PAP. Can J Cardiol 2001;17:677–84.

[123] Heindl S, Dodt C, Krahwinkel M, Hasenfuss G, An-

dreas S. Short term effect of continuous positive air-

S. Javaheri / Clin Chest Med 24 (2003) 207–222 221

way pressure on muscle sympathetic nerve activity in

patients with chronic heart failure. Heart 2000;85:

185–90.

[124] Keily JL, Deegan P, Buckley A, Shiels P, Maurer B,

McNicholas WT. Efficacy of nasal continuous posi-

tive airway pressure therapy in chronic heart failure:

importance of underlying cardiac rhythm. Thorax

1998;53:956–62.

[125] Guilleminault C, Hill JH, Simmons SB, Powell N,

Riley R, Stoohs R. Passive constriction of the upper

airway during central apneas: fiberoptic and EMG

investigations. Respir Physiol 1997;108:11–22.

[126] Badr MS, Toiber F, Skatrud JB, Dempsey J. Pharyn-

geal narrowing/occlusion during central sleep apnea.

J Appl Physiol 1995;78:1806–15.

[127] Alex CG, Onal E, Lopata M. Upper airway occlusion

during sleep in patients with Cheyne-Stokes respira-

tion. Am Rev Respir Dis 1986;133:42–5.

[128] Slutsky A. Mechanical ventilation. In: Crystal RG,

West JB, Barnes PJ, Cherniack NS, Weibil ER, edi-

tors. The lung. New York: Raven Press; 1991.

p. 2163–74.

[129] Malo J, Ali J, Wood LDH. How does positive pres-

sure end-expiratory pressure reduce intrapulmonary

shunt in canine pulmonary edema? J Appl Physiol

1984;57:1002–10.

[130] Hopewell PC, Murray J. Effects of continuous pos-

itive pressure ventilation in experimental pulmonary

edema. J Appl Physiol 1976;40:568–74.

[131] Prewitt RM, McCarthy J, Wood LDH. Treatment of

acute low pressure pulmonary edema in dogs: relative

effects of hydrostatic and oncotic pressure, nitroprus-

side and positive end-expiratory pressure. J Clin In-

vest 1981;67:409–18.

[132] Pare PD, Warriner B, Baile EM, Hogg JS. Redistri-

bution of pulmonary extravascular water with positive

end expiratory pressure in canine pulmonary edema.

Am Rev Respir Dis 1983;127:590–3.

[133] Dowell AR, Heyman A, Sicker HO, et al. Effect of

aminophylline on respiratory-center sensitivity in

Cheyne-Stokes respiration and in pulmonary emphy-

sema. N Engl J Med 1965;273:1447–53.

[134] Javaheri S, Parker TJ, Wexler L, Liming JD, Lin-

dower P, Roselle GA. Effect of theophylline on

sleep-disordered breathing in heart failure. N Engl J

Med 1996;335:562–7.

[135] Javaheri S, Guerra L. Lung function, hypoxic and

hypercapnic ventilatory responses, and respiratory

muscle strength in normal subjects taking oral theo-

phylline. Thorax 1990;45:743–7.

[136] Javaheri S, Evers JAM, Teppama LJ. Increase in ven-

tilation caused by aminophylline in the absence of

changes in ventral medullary extracellular fluid pH

and carbon dioxide tension. Thorax 1989;44:121–5.

S. Javaheri / Clin Chest Med 24 (2003) 207–222222

Sleep-disordered breathing and stroke

Henry Yaggi, MD, MPHa, Vahid Mohsenin, MDa,b,*

aSection of Pulmonary and Critical Care Medicine, Yale University School of Medicine, 333 Cedar Street,

Post Office Box 208057, New Haven, CT 06520-8057, USAbYale Center for Sleep Medicine, 333 Cedar Street, LCI 105, Post Office Box 208057, New Haven, CT 06520-8057, USA

Care-charming Sleep, thou easer of all woes, Brother

to Death, sweetly thyself dispose. John Fletcher

(1579–1625), The Tragedy of Valentinian (V, ii).

It is fascinating to consider that something as

basic as the way we breath during sleep is associated

with conditions that account for several of the leading

causes of mortality in adults in this country: hyper-

tension, cardiovascular, and cerebrovascular disease.

When considered separately from other cardiovascu-

lar diseases, stroke ranks as the third leading cause

of death, and it is the leading cause of serious long-

term disability [1]. Stroke constitutes several different

types of cerebrovascular disease: ischemic stroke,

embolic stroke, transient ischemic attack (TIA), and

hemorrhagic stroke. There are currently few effective

therapies for stroke, so understanding underlying

pathophysiologies, promoting preventative behaviors,

and developing novel therapeutic approaches for the

treatment of stroke are of crucial importance.

Like stroke, sleep-disordered breathing is highly

prevalent [2] and constitutes a spectrum of diseases:

primary snoring, upper airway resistance syndrome,

obstructive sleep apnea (OSA), central sleep apnea,

and obesity-hypoventilation syndrome. The high prev-

alence of stroke and sleep apnea could cause an

overlap of these two diseases just by chance alone.

There are several reasons to suspect a direct relation-

ship between stroke and sleep-disordered breathing,

however. In the authors’ clinical experience, apneic

spells and snoring are frequently observed on the

stroke rehabilitation service. Patients who suffer from

cerebral infarction often complain of diffuse cerebral

symptoms and cognitive problems, such as impaired

memory, inability to concentrate, emotional instability,

and excessive daytime sleepiness [3]. In large part,

these symptoms have been attributed to structural

damage to brain tissue; however, many of these

symptoms are also pervasive in patients with sleep-

disordered breathing [4]. There are several overlap-

ping risk factors and consequences of both diseases,

such as gender, age, hypertension, obesity, smoking,

and alcohol use. Finally, some of the physiologic

consequences of OSA, such as cyclic oxygen desatu-

rations and labile blood pressure, are known to be

poorly tolerated in patients with stroke.

Identifying and treating underlying sleep-dis-

ordered breathing ultimately may represent a novel

management strategy for reducing the large morbidity

and mortality burden of stroke. Over the past decade,

the understanding of the strength of the association

between sleep-disordered breathing and stroke has

grown considerably, as has the understanding of the

physiologic, autonomic, humoral, and vascular con-

sequences of this breathing disorder. Several challeng-

ing questions persist with respect to any causal

inference between sleep-disordered breathing and

stroke, however: What is the temporal relationship

between sleep apnea and stroke? In other words, does

sleep-disordered breathing cause stroke, or does

stroke cause sleep disordered-breathing? Is sleep-

disordered breathing an independent risk factor for

the development of stroke in the setting of confound-

ing overlapping risk factors, or is the association with

stroke simply mediated by higher levels of cardio-

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00027-3

* Corresponding author. Section of Pulmonary and

Critical Care Medicine, Yale University School of Medi-

cine, 333 Cedar Street, Post Office Box 208057, New

Haven, CT 06520-8057.

E-mail address: vahid.mohsenin@yale.edu

(V. Mohsenin).

Clin Chest Med 24 (2003) 223–237

vascular risk factors in patients with sleep-disordered

breathing? Does the presence of sleep-disordered

breathing influence the morbidity or mortality from

stroke, and does treatment of sleep disordered breath-

ing influence this risk?

The primary objective of this article is to explore

these questions by critically reviewing the current

literature. First, epidemiologic studies about the rela-

tionship between sleep-disordered breathing and

stroke are analyzed with respect to issues regarding

the strength of the association, temporal relationship,

dose-response relationship, and consistency of the

association using different study designs and different

populations. Subsequently, the biologic plausibility of

the relationship is explored by reviewing studies that

examine the pathophysiology of sleep-disordered

breathing and stroke focusing on cerebral hemody-

namic and humoral mechanisms.

Epidemiologic studies

Several studies have sought to determine the

presence and extent of a causal interaction between

sleep-disordered breathing and stroke independent of

frequently coexisting and potentially confounding

variables common to both conditions. Established

modifiable risk factors for stroke include hyperten-

sion, hypercholesterolemia, smoking, and diabetes for

atherosclerotic cerebrovascular disease; atrial fibril-

lation and myocardial infarction for cardiogenic

embolism; and hypertension for intracerebral hemor-

rhage. Established risk factors for sleep-disordered

breathing include excess body weight, age, gender,

estrogen depletion, smoking, and alcohol. To com-

plicate matters further, the adjustment for potential

confounding factors is open to criticism, because

these factors may be on the causal pathway of the

relationship between OSA and stroke. This applies

especially to hypertension, because removal of its

effect might overadjust the apparent risk and negate a

true cause-effect relationship between sleep-dis-

ordered breathing and stroke.

An investigation from the Sleep Heart Health

Study, a cohort of 6440 men and women over age

40, explored the associations between sleep-dis-

ordered breathing and cardiovascular risk factors and

found that the respiratory disturbance index (RDI, the

number of apneas and hypopneas per hour sleep) was

cross sectionally associated with age, body-mass index

(BMI), waist-to-hip ratio, hypertension, diabetes, and

lipid levels [5]. This risk factor pattern of hyperten-

sion, diabetes, and hypertriglyceridemia is commonly

seen in people who are obese, and the multivariate

models in this study suggest that the degree of obesity,

age, and gender explain most of the elevation in these

cardiovascular risk factors, with the exception of

hypertension. The presence of an independent asso-

ciation of the RDI with hypertension suggests that it

may be in the causal pathway. As discussed elsewhere

in this issue, because of the acute and profound effects

of sleep-disordered breathing on vascular tone, hyper-

tension is believed to be a major mechanism by which

sleep-disordered breathing might influence future car-

diovascular and cerebrovascular disease risk [5].

Snoring and stroke

Early epidemiologic studies that examined the

relationship between sleep-disordered breathing and

cerebrovascular disease used self-reported snoring as

the primary exposure variable. Self-reported ‘‘habit-

ual snoring,’’ usually defined as subjects who snore

‘‘often’’ or ‘‘always,’’ is a sensitive measure of true

heavy snorers based on all night recordings [6]. The

specificity is low, however, with many patients being

misclassified as snorers. The consequence of such

misclassification is the reduction of a potential rela-

tionship and bias toward the null hypothesis.

Despite this failing, most of these studies clearly

show an association between snoring and stroke

(Table 1) and demonstrate that the strength of this

association is on the same order of magnitude as

traditional risk factors for stroke, such as hypertension,

smoking, cardiac arrhythmia, and hypercholesterol-

emia. Even when adjusted for confounding risk factors

such obesity, hypertension, age, and gender, an inde-

pendent association remained between snoring and

stroke. The designs of these initial studies were pre-

dominantly case control or cross-sectional [7–12] and

were subject to criticism with respect to recall bias and

establishing the temporal relationship between stroke

and sleep-disordered breathing, because snoring and

sleep apnea can be consequences of stroke [13].

More convincing evidence comes from several

large, prospective studies that seemed to corroborate

these case-control and cross-sectional studies. In an

early cohort study exclusively of men that used a

Finnish nationwide registry, there was a twofold

increase in the relative risk for the combined outcome

of stroke and ischemic heart disease in habitual

snorers compared with non-snorers [14]. A smaller

but still significant positive association (relative

risk = 1.33) between regular snoring and the com-

bined cardiovascular outcome of stroke and ischemic

heart disease was seen exclusively in women in the

Nurses Heath Study [15]. In this large cohort, the age-

adjusted relative risk for stroke alone in regular

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237224

Table 1

Selected studies of snoring and stroke

Study Study design No. of subjects Exposure assessment Disease assessment Confounding assessment Relative risk (95% CI)

Partinen [9] Case control 50 Personal interview,

habitual snoring

Stroke patients admitted

to hospital, neurologic

exam, CT/MRI

Age, BMI 10.3 (3.5–30.1)

Koskenvuo [14] Cohort, 3-year

follow-up

4388 Mailed questionnaire,

habitual snoring

Finish registry, death,

ischemic heart

disease, stroke

Age, BMI, hypertension,

smoking, alcohol

2.08 (1.5–3.77)

Spriggs [11,12] Case control 400 Personal interview,

regular snoring

Stroke patients admitted

to hospital, neurologic

exam, CT/MRI

Age, gender 3.2 (2.3–4.4)

Palomaki [8] Case control 177 Standardized

questionnaire,

habitual snoring

Stroke patients admitted

to hospital, neurologic

exam, CT/MRI

Age, gender, alcohol, hypertension,

ischemic heart disease

2.13 (1.29–3.52)

Smirne [10] Case control 330 Personal interview,

habitual snoring

Stroke patients admitted

to hospital, Neurologic

exam, CT/MRI

Age, gender, BMI, diabetes,

dyslipidemia, smoking,

alcohol, hypertension

1.86 (1.2–2.87)

Jennum [63] Cohort 6-year

follow-up

804 Personal interview,

habitual snoring

Cardiovascular outcome based

on Danish Health Registry

Hypertension, BMI, diabetes,

smoking, alcohol, hyperlipidemia,

catecholamines

1.26 (1.3–6.8)

Neau [7] Case control 133 Personal interview,

habitual snoring

Stroke patients admitted

to hospital, Neurologic

exam, CT/MRI

Gender, age, hypertension,

obesity, cardiac arrhythmia

2.9 (1.3–6.8)

Hu [15] Cohort 8-year

follow-up

(Nurses Health Study)

71,779 Mailed questionnaire,

regular snoring

Follow-up questionnaire to

determine cardiovascular

outcome confirmed by

medical record review

Smoking, age, BMI, alcohol,

physical activity, menopausal

status, family history of

myocardial infarction, diabetes,

high cholesterol

1.33 (1.06–1.67)

H.Yaggi,V.Mohsen

in/Clin

Chest

Med

24(2003)223–237

225

snorers was 1.88 (1.62–2.53), which became non-

significant when adjusted for BMI and other cardio-

vascular covariates.

These initial studies of the association between

snoring and stroke on balance supported a positive

association and served to raise some important meth-

odologic issues. First, if hypertension is on the

intermediate causal pathway between sleep-disor-

dered breathing and stroke, should it be adjusted for

and considered a confounder? Second, self-reported

habitual snoring may not be a reliable measurement

of true snoring. Although self-reported habitual snor-

ers seemed to be true heavy snorers when validated

against overnight recordings, a large percentage of

self-reported never-snorers were not aware of their

snoring, which resulted in exposure misclassification

and bias toward the null hypothesis [6]. The pre-

sumed mechanism for the association between snor-

ing and stroke is that snoring serves as a marker for

OSA. Although heavy snoring invariably accompa-

nies sleep apnea [16], most snorers do not have sleep

apnea. In some of the case-control studies discussed

previously [8,17], the authors attempted to identify

within their populations those snoring subjects more

likely to have OSA by identifying snorers who also

have apneas, excessive daytime sleepiness, and obe-

sity. The addition of these potential markers for OSA

increased the odds ratio in these studies.

A different approach to assessing exposure to

sleep-disordered breathing occurred in a study that

used data from the First National Health and Nutri-

tion Examination Survey (NHANES I) cohort [18].

Instead of self-reported snoring, other clues to pre-

existing OSA were used: self-reported sleep duration

and daytime somnolence. Sleep duration and symp-

toms of daytime somnolence were significantly asso-

ciated with the development of stroke and coronary

heart disease adjusted for potential confounding car-

diovascular risk factors. Although these symptoms

are assumed to serve as markers for sleep apnea, the

validity of this assumption is questionable, and it is

conceivable that these symptoms of increased sleep

duration and daytime somnolence serve as general

markers of disease and disability.

Sleep apnea and stroke (the temporal relationship)

Several studies have used overnight polysomno-

graphy to define OSA more precisely in an attempt to

sort out whether it is the minority of patients with

OSA who account for the apparent increased risk of

sleep-disordered breathing with stroke (Table 2).

These studies have focused on OSA as a risk factor

for the development of stroke and as an outcome and

consequence of stroke.

A study by one of the authors (V.M.) in 1995 of

ten patients who were recovering from hemispheric

stroke revealed a high prevalence (80%) of OSA

compared with age- and BMI-matched controls with

similar frequency of hypertension and smoking with-

out stroke [19]. The mean RDI for the control and

stroke group was 3 and 52 events per hour, respec-

tively. Predominantly obstructive events were found

in seven patients. Because none of the study subjects

had a previous history of significant snoring, apnea,

obesity, hypersomnolence, or neurologic impairment,

the conclusion was that OSA might be a sequela of

stroke. It is known that repeated upper airway

obstruction in patients with OSA occurs as a con-

sequence of reduction in pharyngeal muscle tone

during sleep. The pharyngeal muscles may be affect-

ed in stroke; neurologic dysphagia has been demon-

strated in 30% to 40% of patients admitted to the

hospital with unilateral hemispheric stroke [20,21].

A subsequent case-control study of consecutively

admitted inpatients with stroke [22] speculated that

the hypoxia and hemodynamic responses in OSA

may have predisposed to the development of stroke

rather than the other way around. This study com-

pared the polysomnograms of 27 healthy age- and

gender-matched controls recruited from the local pop-

ulation to 24 inpatients with recent stroke confirmed

by neurologic examination and imaging studies of the

brain. Overall, OSA was diagnosed in 19% of the

controls and 71% of the stroke patients. The mean

lowest oxygen saturation level was 91% in the con-

trol group and 85% in the stroke group, and the mean

RDI was 4 events per hour for controls and 26 events

per hour for stroke patients. Once again, predomi-

nantly obstructive apneas were found as opposed

to central or Cheyne-Stokes respirations. 4 stroke

patients were reevaluated at 5 months with polysom-

nography, and they demonstrated OSA on reevalua-

tion. The 4-year mortality rate for patients with stroke

was 21%, and all patients with stroke who died (of

various causes) had OSA. These findings led the

authors to propose that OSA predisposes patients

to stroke.

Although case-control studies generally are effi-

cient study designs for evaluating strength of asso-

ciation, they have a significant limitation in their

ability to establish the temporal course in a cause-

effect relationship. When comparing hospitalized

inpatients to healthy community-dwelling controls,

a selection bias known as Berkson’s Bias may distort

the actual association in that patients who are admit-

ted to the hospital or rehabilitation unit also may have

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237226

Table 2

Selected studies of sleep apnea and stroke using polysomnography

Study date Study design No. stroke/controls no. Mean RDI Study population Confounding assessment Prevalence sleep apnea in stroke (%)

Mohsenin [19] Case control 10/10 52 Predominantly hemispheric

stroke in a rehabilitation unit

Age, BMI,

hypertension, smoking

80% with RDI �20

Good [34] Descriptive 47 (19 underwent

polysomnography)

NA Rehabilitation patients

recently hospitalized for stroke

NA 32% had �10 desaturation events/h

based on computerized

overnight oximetry

Dyken [22] Case control 24/19 26 Recently hospitalized for stroke Age, gender 71% with RDI �10

Bassetti [24] Case control 128/25 (80 underwent

polysomnography)

28 Inpatients with stroke and TIA Age, BMI, diabetes,

severity of stroke

63% with RDI �10

Parra [28] Descriptive 161 21 Inpatients with stroke and TIA NA 71% with RDI �10 (acute phase)a

61% with RDI �10 (stable phase)

Shahar [23] Cross-sectional

(Sleep Heart

Health Study)

6424 NA

(see text)

Assembled from several ongoing

population based studies of

cardiovascular disease in the

United States

Age, race, gender,

smoking, diabetes,

hypertension, BMI,

cholesterol

NA (see text) relative risk of stroke

comparing lowest quartile to highest

quartile = 1.58 with 95% CI (1.02–2.46)

a ‘‘Acute phase’’ after admission and ‘‘stable phase’’ indicate > 3 months later.

H.Yaggi,V.Mohsen

in/Clin

Chest

Med

24(2003)223–237

227

pathology in addition to the stroke (ie, OSA), which

increases the chance of admission.

Perhaps the strongest epidemiologic evidence dem-

onstrating the association between sleep-disordered

breathing and cerebrovascular disease comes from

the initial results of the Sleep Heart Health Study

[23]. This study explored the cross-sectional associ-

ation between sleep-disordered breathing and preva-

lent self-reported cardiovascular disease (myocardial

infarction, angina, coronary revascularization proce-

dures, heart failure, or stroke) in a large cohort of 6424

individuals who underwent overnight polysomnogra-

phy at home. By comparing the upper apnea-hypopnea

index (AHI) severity quartile (>11) to the lower AHI

severity quartile (0–1.3), the most parsimonious

logistic regression model revealed an odds ratio of

1.58 (1.02–2.46) for the association of stroke with

sleep-disordered breathing adjusted for age, race, sex,

smoking status, self-reported diabetes, total choles-

terol, and HDL lipoprotein cholesterol. Unlike cor-

onary heart disease and congestive heart failure,

in which much of the risk associated with sleep-

disordered breathing came from mild sleep apnea

(RDI < 10), there seemed to be an incremental increase

in risk of stroke associated with increasing AHI

severity (Fig. 1). Support of this finding is limited,

however, by the small number of subjects at higher

AHI severity in this population-based study. Hypoxe-

mia seemed to explain 10% to 40% of the AHI effect,

and sleep fragmentation per se, as measured by the

arousal index, was not associated with cardiovascular

disease in these data. If the associations observed in

this study are causal, it seems that even a modestly

elevated risk of stroke coupled with the high preva-

lence of mild/moderate sleep-disordered breathingwill

have considerable public health implication.

Cross-sectional associations might reflect reverse

causal pathways, whereas sleep-disordered breathing

has been the consequence rather than the cause of

stroke. The direction of this arrow of causation

ultimately can be determined definitively only by

analysis of incident cerebrovascular disease events,

and it awaits the results of future prospective follow-

up studies. To the authors’ knowledge no study has

investigated prospectively the relationship between

polysomnographic indices of sleep-disordered breath-

ing and stroke, several investigations have taken

creative approaches to gaining insight into this tem-

poral relationship.

One study that provided some insight into the

causal pathway of stroke and OSA was a retrospec-

tive cohort study of patients who were diagnosed

with OSA by using polysomnography in the 1970s

before the availability of continuous positive airway

pressure (CPAP), when the only known aggressive

therapy for OSA consisted of tracheostomy [17].

7 years of follow-up was provided on 198 patients,

of whom 71 received tracheostomy (considered

‘‘effective treatment’’) and 127 received ‘‘conservative

treatment’’ that consisted of recommended weight loss

(the only alternative). Any new hypertension, myocar-

dial infarction, or stroke that occurred since the orig-

inal polysomnography was considered the main

vascular morbidity outcome. Despite the fact that at

study entry the tracheostomy group included more

patients with a history of hypertension, myocardial

infarction, or stroke, the conservatively treated group

presented with significantly more vascular morbidity.

Patients with TIA potentially represent another

unique opportunity to delineate the directionality of

the cause-effect relationship between OSA and cere-

brovascular disease. TIA represents an intermediate

stage of disease in the natural history of ischemic

stroke, and by definition, patients with TIA have no

residual neuromuscular side effects, which makes the

causal pathway of TIA leading to OSA less plausible.

Demonstrating an increased prevalence of OSA

among cases of TIA bolsters the theory that OSA

leads to the ultimate development of ischemic stroke

rather than the other way around. In follow-up studies

of patients with acute TIA or ischemic stroke [24,25]

researchers demonstrated a similar elevated frequency

and severity of OSA. In one of these studies [24],

adequate polysomnography was performed in 80

subjects (stroke = 48, TIA = 32) and the prevalence

Fig. 1. Predicted log odds (a measure of risk) of stroke as a

function of AHI. AHI indicates the number of apneas and

hypopneas per hour of sleep. The histogram is adapted from

regression of the log odds of stroke. (From Shahar E,Whitney

CW, Redline S, et al. Sleep-disordered breathing and

cardiovascular disease: cross-sectional results of the Sleep

Heart Health Study. Am J Respir Crit CareMed 2001;163:19;

with permission.)

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237228

and severity of OSA were compared in patients with

stroke, TIA and ‘‘normal’’ healthy controls from the

Michigan Alcohol Research Center. Stroke and TIA

patients differed significantly from normal controls in

measurements of AHI, maximal apnea duration, and

minimal oxygen saturation. Stroke and TIA patients

also were similar in all variables considered, including

habitual snoring, AHI, maximal apnea duration, mean

apnea duration, and minimal oxygen saturation.

Although the face validity of these studies of TIA

and OSA is high with respect to clarifying the

temporal relationship between OSA and cerebrovas-

cular disease, there are several limitations with

respect to internal and external validity. Most impor-

tantly, patients with TIA may represent a heteroge-

neous group of individuals. Symptoms of TIA, a

clinical diagnosis, are mimicked by multiple other

disease entities, which may result in disease misclas-

sification. Strictly defining and validating the defini-

tion of TIA for clinical research is of the utmost

importance. Because the traditional definition of TIA

requires the resolution of signs and symptoms within

a 24-hour period, generally it has been assumed that

TIAs leave no residual damage. Cerebral infarctions

have been demonstrated by neuroimaging techniques

in 5% to 10% of patients with clinically defined TIA,

however [26], and some estimates of unrecognized

infarctions by CT (32%) and MRI (77%) are even

higher [27].

Another approach used to gain some insight into

the temporal relationship between sleep-disordered

breathing and stroke prospectively followed 161

consecutive patients with first-ever stroke or TIA

admitted to a stroke unit [28]. TIA was strictly

defined according to the National Institute of Neuro-

logic Disease and Stroke classification [29]. In

this study, previously validated portable respiratory

recordings were performed within 48 to 72 hours

after admission (acute phase) and subsequently after

3 months (stable phase). The important findings of

this study included lack of significant differences

in OSA severity according to stroke subtype (TIA,

ischemic stroke, or hemorrhagic stroke) or brain

parenchymatous location. The study also found that

the frequency of obstructive apneas did not signifi-

cantly decline from the period immediately after stroke

to 3 months later. Because there were no significant

differences in obstructive apneas between baseline and

3 months later or between different stroke subtypes

and locations, the findings led the authors to conclude

that obstructive events seem to be a condition predat-

ing the development of cerebrovascular disease and

they act as a risk factor for rather than a consequence of

cerebrovascular disease.

Sleep-disordered breathing and hypertension

Further evidence in support of the causal pathway

of sleep-disordered breathing leading to stroke comes

from recent large cross-sectional and cohort studies in

support of OSA being an independent risk factor for

the development of hypertension. From the Sleep

Heart Health Study [30], sleep-disordered breathing

was associated with prevalent hypertension even after

controlling for potential confounders, such as age,

gender, BMI (and other measures of adiposity), alco-

hol, and smoking. The relative risk for the highest

category of AHI (>30/hour) compared with the lowest

category (< 1.5/hour) was 1.37 (95% CI, 1.03–1.83).

Overall, the odds of hypertension seemed to increase

with increases in AHI in a dose-response fashion.

More compelling data that lends support to the

evidence of a causal role of sleep-disordered breathing

in hypertension comes from the prospective findings

of the Wisconsin Sleep Cohort Study [31]. The

presence of sleep-disordered breathing at baseline

was accompanied by a substantially increased risk

for future hypertension at 4 years of follow-up. Even

after adjusting for baseline hypertension status, age,

gender, BMI, waist and neck circumference, and

weekly alcohol and cigarette use, the risk was ele-

vated, with an odds ratio of 2.89 (95% CI, 1.46–5.64)

for subjects with an AHI of more than 15/hour

compared with patients without any nocturnal apnea.

It should be noted that for many of the aforementioned

studies (see Tables 1, 2), the risk of stroke from sleep-

disordered breathing was independent of coexisting

hypertension. The presence of hypertension further

enhances the risk.

Functional outcome after stroke

Previous studies reported that up to 43% of stroke

patients will have a progression of their neurologic

deficit [32,33]. Regardless of whether OSA precedes

or follows stroke, it is associated with unfavorable

clinical outcomes after stroke, including early neu-

rologic worsening, delirium, depressed mood, poor

functional status, and impaired cognition [12,34–36].

In one study [34], the functional status, as assessed

by the Barthel Index (a multifaceted scale that

measures mobility and activities of daily living), in

patients with stroke and OSA was significantly lower

compared with patients with stroke but no evidence

of sleep-disordered breathing at discharge and 3 and

12 months (Fig. 2). Death at 1 year was negatively

correlated with percentage of time spent at less than

90% SaO2. Whether sleep-disordered breathing is an

independent predictor of poor functional outcome or

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237 229

simply a marker for more severe stroke is not clear

from this study.

Predictors of upper airway obstruction in stroke

Typical OSA-type risk factors, such as elevated

BMI and neck circumference, seem to be the best

predictors of the development of upper airway

obstruction in acute stroke. Limb weakness also

seemed to be an independent predictor of OSA in

acute stroke, but other stroke characteristics, such as

severity and subtype, do not seem to be associated

with the development of upper airway obstruction

[37]. Of clinical relevance in this study was that most

of the sleep-disordered breathing occurred while the

subjects were supine. Whether simple maneuvers

targeted at preventing upper airway obstruction, such

as position therapy, may improve outcomes in acute

stroke remains an important unanswered question.

Continuous positive airway pressure treatment trials

Two CPAP treatment trials of patients who exhibit

sleep-disordered breathing after stroke recently were

published and have provided insight into whether

sleep-disordered breathing is truly an independent

cause of worse outcome after stroke and the effective-

ness and acceptance of CPAP [12,34,35,38]. Although

the trials only reflect short-term use of CPAP, the

results are encouraging because they demonstrated

beneficial effects and comparable compliance rates to

OSA patients without stroke. In one trial [39],

although not randomized, there was a significant

reduction in nocturnal blood pressure (8 mm Hg) after

10 days of treatment in comparing CPAP-compliant

and CPAP-noncompliant patients. There was

improvement in subjective well-being (although this

later finding is based on less well-validated neuro-

psychiatric testing). In a logistic regression model,

aphasia and the severity of motor disability as quan-

tified by the Barthel index were significant negative

predictors of acceptance of CPAP.

The second CPAP study was a randomized treat-

ment trial [38], and although it was not double-

blinded, it demonstrated that depressive symptoms

are reduced in patients who are treated with nasal

CPAP at 7 and 28 days compared with controls who

are not treated. There was no significant improvement

in delirium, activities of daily living, or cognitive

functinoning. Compliance was lower in this study

(approximately 50%), perhaps partly related to the

fact that this was an older population.

Overall, the primary acceptance of CPAP (at least

in the first treatment study) seems comparable to

patients with OSA without stroke, and CPAP seems

to exert a beneficial influence in terms of well-being,

hypertension, and depression. Long-term compliance

is not certain, however, especially in a population of

patients with more functional and cognitive disability.

As suggested elsewhere [40], bearing in mind that

Fig. 2. The Barthel Index (BI) scores on admission and discharge and 3- and 12-month follow-up for patients with sleep-

disordered breathing (OSA group) compared with other stroke patients without sleep-disordered breathing. Lower BI scores

indicate worse cognitive impairment and activity of daily living impairment. (From Good DC, Henkle JQ, Gelber D, et al. Sleep-

disordered breathing and poor functional outcome after stroke. Stroke 1996;27:252–9; with permission.)

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237230

obstructive apneas result in recurrent hypoxemia and

cerebral blood flow fluctuations that could damage

the area of the ischemic penumbra, one possible

argument in favor of CPAP treatment is to prevent

stroke recurrence. Patients with TIA or minor non-

disabling stroke may represent an important target

group for CPAP treatment for secondary prophylaxis

because they may be a younger and more compliant

group with fewer deficits.

Circadian variation in ischemic events

The relation between the time of stroke symp-

toms and the time of day may relate to the under-

lying pathophysiology of stroke. Early studies of the

timing of acute stroke indicated that acute strokes

tend to occur either during the evening hours or

during sleep, and many afflicted patients reported

awakening with new neurologic deficits [41–43]. A

metaanalysis of 11,816 strokes revealed that similar

to myocardial infarction and sudden cardiac death, a

‘‘morning excess’’ of all types of stroke (including

TIA) is seen between 6:00 AM and 12:00 PM and is

significantly higher than would be expected by

chance (Fig. 3) [44].

It is of interest that the most prolonged rapid eye

movement (REM) sleep period occurs in close tem-

poral proximity to this circadian preference for ische-

mic stroke (the early morning hours). Specifically,

during REM sleep there are significant hemodynamic

changes [45] with increases in cerebral blood flow

[46] and blood pressures, which can reach near-

normal waking levels [47]. The early morning hours

are associated with decreased fibrinolytic activity

[48], increased platelet aggregability, and peak levels

of catecholamines [49].

As is described in the following sections, many of

these same autonomic, hemodynamic, and physiologic

mechanisms are heightened in patients with OSA.

Mechanism studies

During sleep in OSA, repetitive episodes of air-

way occlusion with resulting hypoxemia, hypercap-

nia, and significant changes in intrathoracic pressure

elicit a wide variety of autonomic, hemodynamic,

humoral, and vascular perturbations that serve as

plausible biologic mechanisms whereby OSA may

cause stroke (Table 3). Large variations in intratho-

racic pressure with nadirs during inspiratory effort

increase the filling of the right heart and cause a

leftward shift of the interventricular septum [50]. The

resulting reduction of stroke volume is one probable

cause of the decreased arterial pressure seen early

during apnea. Changes in autonomic activity influ-

ence blood pressure variability by vasoconstriction,

with increased levels of circulating catecholamines

and increased endothelin-1 production (a potent vaso-

constrictor) likely contributing to diurnal hyper-

tension [51]. Impaired endothelial function and

accelerated atherogenesis, which may theoretically

result from the repetitive hypoxia and pressure

surges, are also evident in patients with OSA. Finally,

altered cerebral blood flow, fluctuations in intracra-

nial pressure, impaired cerebrovascular autoregula-

tion combined with increased platelet aggregability,

increased fibrinogen, and increased plasma homocys-

teine levels are also likely contributory mechanisms.

Because autonomic mechanisms that contribute to

diurnal hypertension are discussed elsewhere in this

issue, the following discussion of physiologic mech-

Fig. 3. Circadian variation in ischemic stroke and cardiovascular events. (From Mohsenin V. Sleep-related breathing disorders

and risk of stroke. Stroke 2001;32:1271; with permission.)

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237 231

anisms focuses on cerebral hemodynamic and hu-

moral factors.

Cerebral hemodynamics and their changes during

normal sleep

Changes in blood flow to individual organs are

achieved by altering arteriolar resistance. The mech-

anisms that regulate blood flow are broadly catego-

rized as local (intrinsic) control and neural or

hormonal (extrinsic) control (ie, sympathetic innerva-

tion). The cerebral circulation is controlled almost

entirely by local control mechanisms. Many circulat-

ing vasoactive substances do not affect the cerebral

circulation because their large molecular size pre-

vents them from crossing the blood-brain barrier.

Mechanisms for the local control of blood flow

include autoregulation, active hyperemia, and reac-

tive hyperemia. Autoregulation is the maintenance of

constant blood flow to an organ in the face of

changing arterial pressure [52,53]. For example, if

arterial pressure in a cerebral artery suddenly

decreases, an attempt is made to maintain constant

blood flow through this artery by the immediate

compensatory dilation of cerebral arterioles decreas-

ing the resistance of the cerebral vasculature and

keeping flow constant in the face of decreased

pressure. Active hyperemia is the concept that blood

flow to an organ is proportional to its metabolic

activity. For example, if metabolic activity increases

as a result of strenuous activity, then blood flow

increases proportionately to meet the increased meta-

bolic demand. Finally, reactive hyperemia is an

increase in blood flow in response to or as a reaction

to a prior period of decreased blood flow. For

example, reactive hyperemia is the increase in blood

flow to an organ that follows a period of arterial

occlusion. During the occlusion, an oxygen ‘‘debt’’ is

accumulated. The longer the period of occlusion, the

greater the oxygen debt and the greater the sub-

sequent increase in blood flow above the preocclu-

sion levels. The increase in blood flow continues until

the oxygen debt is ‘‘repaid.’’ In the cerebral circula-

tion the major vasoactive metabolites are CO2 and

H + . In addition to these local control mechanisms,

mechanical effects, such as changes in intracranial

pressure, can cause changes in cerebral blood flow.

Sleep state has a profound effect on cerebral

hemodynamics. Multiple studies using various meth-

ods, including transcranial Doppler ultrasonography

[54], Xe inhalation, and single photon emission test-

ing, have shown a 5% to 28% reduction in cerebral

blood flow during non-REM sleep and a 4% to 41%

increase in REM sleep compared with wakefulness in

normal persons [46,54–61].

Intracranial hemodynamics in sleep apnea

Individual episodes of sleep apnea are accompan-

ied by marked episodic elevations of cerebrospinal

fluid pressure and decreases in SaO2 (Fig. 4) [62].

Cerebrospinal fluid pressure in patients with OSAwas

monitored via a pressure transducer and a plastic tube

inserted into the subarachnoid space at the lumbar

level. Another study that invasively monitored radial

artery pressure, central venous pressure, and intra-

cranial pressure (ICP) [63] confirmed the previous

findings and demonstrated that values of ICP were

also elevated in patients with OSA even while awake.

ICP increases further during sleep, and there was a

strong correlation between duration of apnea and ICP

elevations. These increases in ICP were attributed to

(1) increases in central venous pressure, which causes

an increase in cerebral vascular volume, (2) increased

systemic arterial pressure, which causes an increase in

cerebral perfusion pressure, and (3) hypoxic and

hypercapnic cerebral vasodilation, which causes an

increase of the intracranial blood volume. It was

suggested that these ICP elevations may be of impor-

tance in understanding the cerebral symptoms in

patients with sleep apnea, such as morning headache

and cognitive impairment. The mechanical effects of

increased ICP may impede cerebral blood flow and

predispose to cerebral ischemia.

Table 3

Mechanisms whereby sleep-disordered breathing may

cause stroke

Mechanism Consequence

Negative intrathoracic

pressure created

from inspiratory

effort against

closed airway

Decreased stroke

volume, increased

venous return/central

venous pressure

Autonomic mediated

increases in

circulating

catecholamines

and endothelin-1

Hypertension, increased

intracranial pressure

Impaired autoregulation,

active/reactive hyperemia

Alterations in cerebral

blood flow, increased

intracranial pressure

Increased platelet

aggregation, fibrinogen,

homocysteine, vascular

cell adhesion molecule-1,

intracellular adhesion

molecule-1, and L-selectin

Impaired endothelial

function, accelerated

atherogenesis, thrombosis

Right-to-left shunting through

a patent foramen ovale

Paradoxic embolism

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237232

Several recent studies have attempted to gain

insight into the regulation of cerebral flow during sleep

by measuring middle cerebral artery blood flow veloc-

ities noninvasively using transcranial Doppler. One

study [64] revealed that the overall cerebral blood flow

velocities in patients with sleep apnea were signifi-

cantly reduced during all phases of sleep compared

with control subjects with no polysomnographic evi-

dence of sleep apnea. They postulated that this may be

caused by impaired autoregulatory and active/reactive

hyperemic mechanisms in patients with OSA given

that PCO2 was noted to rise in these patients. Of

therapeutic interest is that impairment of cerebrovas-

cular reactivity to elevated CO2 in patients with OSA

may be reversed by treatment with nasal CPAP [55].

Another study [65] that examined more specif-

ically cerebral blood flow velocity in direct relation to

individual obstructive apneic events demonstrated a

biphasic pattern with a concomitant increase in mean

arterial pressure and cerebral blood flow velocity

during early apnea followed by a subsequent decrease

of almost 25% below baseline after apnea termina-

tion. The authors suggested that the period immedi-

ately after the apneas, after the resumption of

ventilation in combination with hypoxemia, poten-

tially would make individuals with OSAvulnerable to

nocturnal cerebral ischemia.

Obstructive apneas and hypopneas compared with

central apneas lead much more frequently to a reduc-

tion in cerebral blood flow, and the longer the

obstructive event the greater the likelihood for a

reduction in blood flow [66]. The relationship

between airway obstruction and decreased perfusion

of the middle cerebral artery was attributed to the

negative intrathoracic pressure generated by the

increased respiratory effort against an obstructed air-

way. Increased time of obstruction could lead to the

development of a high cardiac preload, lower cardiac

afterload, activation of carotid body receptors, and

vasodilation by increasing arterial carbon dioxide and

decreasing oxygenation, all of which can contribute

to a reduction in cerebral blood flow.

Effect of aging on cerebral blood flow

Several cross-sectional studies have demonstrated

an age-related reduction in regional cerebral blood

flow in the range of 20% to 24% in normal aging

individuals [67,68]. This reduction in regional blood

flow has been attributed to age-related brain atrophy

and increased cerebral vascular resistance secondary

to cerebral arteriosclerosis [68]. The mechanism

underlying this change is attributed to altered endo-

thelium function. Relaxation of the basilar artery in

humans [69] and cerebral arterioles [70] and the

Fig. 4. Polysomnographic recordings during REM sleep and in the waking state in a patient with OSA. The nasal oral flow,

thoracic and abdominal wall movements, cerebrospinal fluid pressure at the lumbar level, and SaO2 (%) were recorded and are

presented simultaneously. Apneic events are indicated by diamond marks in the upper part of the recording. A 60-second time

scale is indicated in the upper left part of the figure. The SaO2 (%) scale is indicated by the left vertical axis, and the

cerebrospinal fluid pressure scale is indicated by the right vertical axis. As evident in the recording, each apneic event is

accompanied by marked cerebrospinal fluid pressure changes. (Adapted from Sugita Y, Susami I, Yoshio T, et al. Marked

episodic elevation of cerebral spinal fluid pressure during nocturnal sleep in patients with sleep apnea hypersomnia syndrome.

Electroencephalogr Clin Neurophysiol 1985;60:214–9; with permission.)

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237 233

carotid artery in rats [71] in response to endothelium-

dependent agonists is impaired with aging. Deposits

of b-amyloid in brain and cerebral vessels are seen in

aging individuals. Recent data suggest that b-amyloid

may impair endothelium-dependent relaxation by

generation of superoxide anion. This impaired endo-

thelium-dependent relaxation has been attributed to

degradation of nitric oxide by generation of reactive

oxygen species in the vessel wall [71]. Similar impair-

ment of vasoconstrictor responses to several stimuli

has been reported in the human basilar artery [69].

These age-related changes in cerebral blood flow and

the alterations during normal sleep may predispose the

brain to compromised blood supply during sleep.

Humoral mechanisms

In addition to physiologic mechanisms that alter

cerebral blood flow and contribute to hypoperfusion,

several humoral mechanisms may contribute to

increased hypercoagulability in patients with sleep-

disordered breathing and predispose to ischemic and

thromboembolic stroke. Elevated plasma fibrinogen

levels are believed to be associated with increased risk

of stroke and other cardiovascular events [72–77].

Plasma fibrinogen is an acute-phase protein that is

synthesized in the liver and is intrinsically involved in

coagulation. It enhances thrombosis and atheroscle-

rosis by effects on platelet aggregation, blood vessel

wall, and endothelial cell injury [78,79].

Patients with OSA have been shown to have

increased morning levels of fibrinogen [80]; there-

fore, elevated fibrinogen levels may be one mech-

anism that links OSA to stroke. Further evidence of

the association between OSA and increased fibrino-

gen levels and stroke comes from a cross-sectional

study of 113 stroke patients who underwent neuro-

logic rehabilitation. Fibrinogen level was positively

correlated to RDI and length of respiratory events

and negatively correlated with oxygen desaturation

during sleep [81]. As suggested elsewhere [82],

given the cross-sectional nature of this latter study,

it is not clear whether the higher fibrinogen levels

are a reflection of the acute-phase reaction to the

stroke insult, with stroke being worse and fibrinogen

levels being higher in patients with preexisting

OSA. Alternatively, could airway inflammation

associated with OSA induce increases in plasma

fibrinogen? Although it is widely held that BMI

and other measures of obesity may be determinants

of fibrinogen [83], this study showed that OSA, not

BMI, was independently associated with increased

fibrinogen. Whether fibrinogen is simply a marker

for stroke is yet to be determined, but it is provoc-

ative to consider it as a potential intermediate step

of the pathophysiologic pathway between OSA and

stroke. Further studies that explore the effects of

fibrinogen with treatment for sleep apnea should

prove informative.

Increases in platelet reactivity have been associ-

ated with increased risk of cardiovascular event and

death [84–87]. The ability of aspirin, a recognized

inhibitor of platelet function, to prevent stroke, myo-

cardial infarction, and death can be interpreted as

additional evidence linking platelets to these disor-

ders. It also has been demonstrated that platelet

aggregability increases significantly during the pe-

riod from 6:00 AM to 9:00 AM, which is temporally

related to rising plasma catecholamine levels and the

circadian period. This period has the highest risk for

cardiovascular/cerebrovascular events and sudden

cardiac death [49] (see Fig. 3). A small prospective

study of men who underwent polysomnography for

suspected sleep apnea demonstrated significantly

increased spontaneous platelet activation and ag-

gregation in patients with OSA compared with con-

trols without OSA [88]. Although no relationship

could be established between the level of sponta-

neous platelet activation and specific markers of sleep-

disordered breathing, a second important finding

of the study was a reduction of platelet reactivity

after the application of CPAP. The authors speculated

that the mechanisms for increased platelet reactivity

in patients with OSA are possibly the cyclic hypo-

xemia, hypercarbia, and catecholamine surges that

are part of OSA, which also have been reported to

cause enhanced platelet reactivity [89–91]. Several

other humoral factors associated with cardiovascular

morbidity and mortality have been demonstrated to

be increased in patients with OSA, including plasma

homocysteine [92], circulating endothelin-1 (a potent

vasoconstrictor) [93,94], vascular cell adhesion mol-

ecule-1, intracellular adhesion molecule-1, and L-se-

lectin [95].

A last mechanism whereby OSA may increase the

risk of stroke relates to it being provocative of right-

to-left shunting through a patent foramen ovale [96].

The increased right heart pressure–associated apneic

events may serve to increase the exposure time of

right-to-left shunting through a patent foramen ovale,

which increases the risk of paradoxic embolism.

Patients with sleep apnea may have an increased

prevalence of patent foramen ovale [97].

Taken together, cerebral hypoperfusion, sympa-

thetic activation, hypertension, hypercoagulabity, hy-

poxemia, endothelial impairment, and right-to-left

shunting via patent foramen ovale all likely have a

role in pathogenesis of cerebrovascular disease in

patients with sleep-disordered breathing.

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237234

Summary

Sleep-related breathing disorders are strongly

associated with increased risk of stroke independent

of known risk factors. The direction of causation

favors sleep-disordered breathing leading to stroke

rather than the other way around, although definitive

proof of this awaits the results of prospective cohort

studies. If causal, even a moderately elevated risk of

stroke coupled with the high prevalence of sleep-

disordered breathing could have significant public

health implications. The relationship between sleep-

disordered breathing and stroke risk factors is com-

plex, and likely part of the risk for cerebrovascular

events is because of higher cardiovascular risk factors

in patients with increased RDI. The mechanisms

underlying this increased risk of stroke are multi-

factorial and include reduction in cerebral blood flow,

altered cerebral autoregulation, impaired endothelial

function, accelerated atherogenesis, thrombosis, and

paradoxic embolism. Because of the effects of sleep-

disordered breathing on vascular tone, hypertension

is believed to be a major mechanism by which sleep-

disordered breathing might influence risk of stroke.

Because sleep-related breathing disorders are treat-

able, patients with stroke/TIA should undergo inves-

tigation, with a thorough sleep history interview,

physical examination, and polysomnography. Treat-

ment of sleep apnea has been shown to improve

quality of life, lower blood pressure, improve sleep

quality, improve neurocognitive functioning, and

decrease symptoms of excessive daytime sleepiness

[98]. Further treatment trials are needed to determine

whether treatment improves outcome after stroke and

whether treatment may serve as secondary prophy-

laxis and modify the risk of recurrent stroke or death.

References

[1] American Heart Association. 2002 Heart and Stoke

Statistical Update. Dallas (TX): American Heart Asso-

ciation; 2001.

[2] Young T, Peppard PE. Epidemiology of obstructive

sleep apnea: a population health perspective. Am J

Respir Crit Care Med 2002;165:1217–39.

[3] Leegaard O. Diffuse cerebral symptoms in convales-

cents from cerebral infarction and myocardial infarc-

tion. Acta Neurol Scand 1983;67:348–55.

[4] Kales A, Caldwell AB, Cadieux RJ, et al. Severe ob-

structive sleep apnea: associated psychopathology and

psychosocial consequences. J Chronic Dis 1985;38:

427–34.

[5] Newman AB, Nieto JF, Guidry U, et al. Relation of

sleep-disordered breathing to cardiovascular disease

risk factors. Am J Epidemiol 2001;154:50–9.

[6] Telakivi T, Partinen M, Koskenvuo M, et al. Periodic

breathing and hypoxia in snorers and controls: valida-

tion of snoring history and association with bloodpres-

sureandobesity.ActaNeurolScand1987;76:69–75.

[7] Neau JP. Vascular disorders and obstructive sleep ap-

nea syndrome. Rev Neurol (Paris) 2001;157:S34–7.

[8] Palomaki H. Snoring and the risk of ischemic brain

infarction. Stroke 1991;22:1021–5.

[9] Partinen M, Palomaki H. Snoring and cerebral infarc-

tion. Lancet 1985;2:1325–6.

[10] Smirne S, Palazzi S, Zucconi M, et al. Habitual snoring

as a risk factor for acute vascular disease. Eur Respir J

1993;6:1357–61.

[11] Spriggs D, French J, Murdy J, et al. Historical risk fac-

tors for stroke: a case control study. Age Ageing 1990;

19:280–7.

[12] Spriggs D, French J, Murdy J, et al. Snoring increases

the risk of stroke and adversely affects prognosis. Q J

Med 1992;83:555–62.

[13] Askenasy JJ, Goldhammer I. Sleep apnea as a feature

of bulbar stroke. Stroke 1988;19:637–9.

[14] Koskenvuo M, Kaprio J, Telakivi T, et al. Snoring as a

risk factor for ischaemic heart disease and stroke in

men. BMJ 1987;294:16–9.

[15] Hu FB, Willet WC, Manson JE, et al. Snoring and the

risk of cardiovascular disease in women. J Am Coll

Cardiol 2000;35:308–13.

[16] Waller P. Is snoring a cause of vascular disease? An

epidemiological review. Lancet 1989;1:143–6.

[17] Partinen M. Daytime sleepiness and vascular morbidity

at seven-year follow-up in obstructive sleep apnea Pa-

tients. Chest 1990;97:27–32.

[18] Qureshi AI, Giles WH, Croft JB, et al. Habitual sleep

patterns and the risk for stroke and coronary heart dis-

ease. Neurology 1997;48:904–11.

[19] MohseninV, Valor R. Sleep apnea in patients with hemi-

spheric stroke. Arch Phys Med Rehabil 1995;76:71–6.

[20] Barer DH. The natural history and functional conse-

quences of dysphagia after hemispheric stroke. J Neu-

rol Neurosurg Psychiatry 1989;52:236–41.

[21] Gordon C, Hewer RL, Wade DT. Dysphagia in acute

stroke. BMJ 1987;295:411–4.

[22] Dyken ME, Somers VK, Yamada T, et al. Investigating

the relationship between stroke and obstructive sleep

apnea. Stroke 1996;27:401–7.

[23] Shahar E, Whitney CW, Redline S, et al. Sleep-disor-

dered breathing and cardiovascular disease: cross-sec-

tional results of the Sleep Heart Health Study. Am J

Respir Crit Care Med 2001;163:19–25.

[24] Bassetti C, Aldrich MS. Sleep apnea in acute cerebro-

vascular diseases: final report on 128 patients. Sleep

1999;22:217–23.

[25] Bassetti C, Aldrich MS, Chervin RD, et al. Sleep apnea

in patients with transient ischemic attack and stroke: a

prospective study of 59 patients. Neurology 1996;

47:1167–73.

[26] Ricci S, Celani MG, La Rosa F, et al. Silent brain

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237 235

infarctions in patients with first-ever stroke: a commu-

nity-based study in Umbria, Italy. Stroke 1993;24:

647–51.

[27] Awad I, Modic M, Little JR, et al. Focal parenchymal

lesions in transient ischemic attacks: correlation of

computed tomography and magnetic resonance imag-

ing. Stroke 1986;17:399–403.

[28] Parra O, Arboix A, Bechich S, et al. Time course of

sleep-related breathing disorders in first-ever stroke or

transient ischemic attack. Am J Respir Crit Care Med

2000;161:375–80.

[29] Whisnant JD, Busford JR, Bernstein EF, Cooper ES,

Dyken ML, Easton D, et al. Stroke NIoNDa. Special

report: classification of cerebrovascular disease. Stroke

1990;21:637–76.

[30] Neito FJ, Young TB, Lind BK, et al. Association of

sleep-disordered breathing, sleep apnea, and hyperten-

sion in a large community based study. JAMA 2000;

283:1829–36.

[31] Peppard PE, Young T, Palta M, et al. Prospective study

of the association between sleep-disordered breathing

and hypertension. N Engl J Med 2000;342:1378–84.

[32] Britton M, Roden A. Progression of stroke after arrival

at hospital. Stroke 1985;16:629–32.

[33] Davalos A, Cendra E, Teruel J, et al. Deteriorating

ischemic stroke: risk factors and prognosis. Neurology

1990;40:1865–9.

[34] Good DC, Henkle JQ, Gelber D, et al. Sleep-disor-

dered breathing and poor functional outcome after

stroke. Stroke 1996;27:252–9.

[35] Iranzo A, Santamaria J, Berenguer J, et al. Prevalence

and clinical importance of sleep apnea in the first night

after cerebral infarction. Neurology 2002;58:911–6.

[36] Sandberg O, Franklin KA, Bucht G, et al. Nasal con-

tinuous positive airway pressure in stroke patients with

sleep apnoea: a randomized treatment study. Eur Res-

pir J 2001;18:630–4.

[37] Turkington P, Bamford J, Wanklyn P, et al. Prevalence

and predictors of upper airway obstruction in the first

24 hours after acute stroke. Stroke 2002;33:2037–42.

[38] Sandberg O, Franklin KA, Bucht G, et al. Sleep apnea,

delirium, depressed mood, cognition, and ADL ability

after stroke. J Am Geriatr Soc 2001;49:391–7.

[39] Wessendorf T, Wang Y, Thilmann A, et al. Treatment of

obstructive sleep apnoea with nasal continuous positive

airway pressure in stroke. Eur Respir J 2001;18:623–9.

[40] Parra O. Sleep-disordered breathing and stroke: is there

a rationale for treatment? [letter; comment] Eur Re-

spir J 2001;18:619–22.

[41] Hayashi S, ToyoshimaH, TanabeN, et al. Daily peaks in

the incidence of sudden cardiac death and fatal stroke in

Niigata Prefecture. Jpn Circ J 1996;60:193–200.

[42] Hossman V. Circadian variations of hemodynamics and

stroke. NewYork: Springer-Verlag; 1979. p. 171–80.

[43] Marshall J. Diurnal variation in occurrences of strokes.

Stroke 1977;8:230–1.

[44] Elliott W. Circadian variation in the timing of stroke

onset: a meta-analysis. Stroke 1998;29:992.

[45] Hornyak M, Elam M, Matousek M, Wallin G. Sympa-

thetic muscle nerve activity during sleep in man. Brain

1991;114:1281–95.

[46] Klingelhofer J, Hajak G, Matzander G, et al. Dynamics

of cerebral blood flow velocities during normal human

sleep. Clin Neurol Neurosurg 1995;97:142–8.

[47] Somers VK, Dyken ME, Mark AL, et al. Sympathetic

nerve activity during sleep in normal adults. N Engl J

Med 1993;328:303–7.

[48] Andreotti F, Davies GJ, Hackett DR, et al. Major cir-

cadian fluctuations in fibrinolytic factors and possible

relevance to time of onset of myocardial infarction, sud-

den cardiac death, and stroke. Am J Cardiol 1988;

62:635–7.

[49] Toffler GH, Brezinski D, Shafer AI, et al. Concurrent

morning increase in platelet aggregability and the risk

of myocardial infarction and sudden cardiac death.

N Engl J Med 1987;316:1514–8.

[50] Shiomi T, Guilleminault C, Stoohs R, et al. Leftward

shift of the intraventricular septum and pulsus paradox-

us in obstructive sleep apnea syndrome. Chest 1991;

100:894–902.

[51] Lanfranchi P. Obstructive sleep apnea and vascular

disease. Respiratory Research 2001;2:315–9.

[52] Busja D. Cerebral autoregulation. In: The regulation of

cerebral blood flow. Boca Raton (FL): CRC Press;

1994. p. 45–61.

[53] Paulson OB, Strandgaard S. Cerebral autoregulation.

Brain Metab Rev 1990;2:161–92.

[54] Madsen PL, Vorstrup S, Friberg L, Lassen NA, Wild-

schiodtz G. Human regional cerebral blood flow during

rapid-eye-movement sleep. J Cereb Blood Flow Metab

1991;11:502–7.

[55] Diomedi M, Placidi F, Cupini LM, et al. Cerebral he-

modynamic changes in sleep apnea syndrome and ef-

fect of continuous positive airway pressure treatment.

Neurology 1998;51:1051–6.

[56] Droste DW, Berger W, Schuler E, et al. Middle cerebral

artery blood flow velocity in healthy persons during

wakefulness and sleep: a transcranial Doppler study.

Sleep 1993;16:603–9.

[57] Fischer AQ, Taorimina MA, Akhtar B, et al. The effect

of sleep on intracranial hemodynamics: a transcranial

Doppler study. J Child Neurol 1998;6:155–8.

[58] Hoshi Y, Mizukami S. Dynamic features of hemody-

namic and metabolic changes in human brain during

all-night sleep as revealed by near-infrared spectros-

copy. Brain Res 1994;652:257–62.

[59] Lenzi P, Zoccoli G, Walker AM, et al. Cerebral blood

flow regulation in REM sleep: a model for flow me-

tabolism coupling. Arch Ital Biol 1999;137:165–79.

[60] Madsen PL, Schmidt JF, Holm S, et al. Cerebral oxy-

gen metabolism and cerebral blood flow in man during

light sleep (stage 2). Brain Res 1991;557:217–20.

[61] Sakai F, Meyer JS, Karacan I, et al. Normal human

sleep: regional cerebral hemodynamics. Ann Neurol

1980;7:471–8.

[62] Sugita Y, Susami I, Yoshio T, et al. Marked episodic

elevation of cerebral spinal fluid pressure during noctur-

nal sleep in patients with sleep apnea hypersomnia syn-

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237236

drome. Electroencephalogr Clin Neurophysiol 1985;60:

214–9.

[63] Jennum P. Intracranial pressure and obstructive sleep

apnea. Chest 1989;95:279–83.

[64] Fischer AQ, Chaudry BA, Taormina MA, et al. Intra-

cranial hemodynamics in sleep apnea. Chest 1992;102:

1402–6.

[65] Balfors EM. Impairment in cerebral perfusion during

obstructive sleep apneas. Am J Respir Crit Care Med

1994;150:1587–91.

[66] Netzer N, Werner P, Jochums I, et al. Blood flow of the

middle cerebral artery with sleep-disordered breathing.

Stroke 1998;29:87–93.

[67] Melamed E, Lavy S, Bentin S, et al. Reduction in

regional cerebral blood flow during normal aging in

man. Stroke 1980;11:31–5.

[68] Naritomi H, Meyer JS, Sakai F, et al. Effects of ad-

vancing age on regional cerebral blood flow: studies

in normal subjects and subjects with risk factors for

atherothrombotic stroke. Arch Neurol 1979;36:410–6.

[69] Hatake K, Kakishita E, Wakabayashi I, et al. Effect of

aging on endothelium dependent vascular relaxation of

isolated human basilar artery to thrombin and bradyki-

nin. Stroke 1990;21:1039–43.

[70] MayhanWG, Faraci FM. Effect of aging on responses in

cerebral aterioles. Am J Physiol 1990;258:H1138–43.

[71] Paterno R, Feraci FM. Age-related changes in release

of endothelium derived relaxing factor from the carotid

artery. Stroke 1994;25:2459–62.

[72] Di Minno G. Measuring plasma fibrinogen to predict

stroke and myocardial infarction. Arteriosclerosis 1990;

10:1–7.

[73] Kannel W, Wolf P, Castelli W, et al. Fibrinogen and

risk of cardiovascular disease: the Framingham Study.

JAMA 1987;258:1183–6.

[74] Mead TW, North WR, Chakrabarti R, et al. Hemostatic

function and cardiovascular death: early results of a

prospective study. Lancet 1980;1:1050–9.

[75] Resch KL, Ernst E, Matrai A, et al. Fibrinogen and

viscosity as risk factors for subsequent cardiovascular

events in stroke survivors. Ann Intern Med 1992;117:

371–5.

[76] Toss H, Lindhaul B, Siegbahn A, et al. Prognostic

influence of increased fibrinogen and C-reactive pro-

tein levels in unstable coronary artery disease. Circu-

lation 1997;96:4204–10.

[77] Wilhelmsen L, Svardsudd K, Kristoffer KB, et al. Fi-

brinogen as a risk factor stroke and myocardial infarc-

tion. N Engl J Med 1984;311:501–5.

[78] Eber B. Fibrinogen: its role in atherosclerosis. Semin

Thromb Hemost 1993;19:104–7.

[79] Smith EB, Grant A, Stirk C. Fate of fibrinogen in human

arterial intima. Arteriosclerosis 1990;10:263–75.

[80] Chin K, Ohi M, Kita H, et al. Effects of NCPAP therapy

on fibrinogen levels in obstructive sleep apnea syn-

drome. Am J Respir Crit Care Med 1996;153:1972–6.

[81] Wessendorf T, Thilmann A, Wang Y, et al. Fibrinogen

levels and obstructive sleep apnea in ischemic stroke.

Am J Respir Crit Care Med 2000;162:2039–42.

[82] Shamsuzzaman AS. Fibrinogen, stroke, and obstruc-

tive sleep apnea: an evolving paradigm of cardio-

vascular risk. Am J Respir Crit Care Med 2000;162:

2018–9.

[83] Barasch E, Benderly M, Graff E, et al. Plasma fibrin-

ogen levels and their correlates in 6457 coronary heart

disease patients: the Bezafibrate Infarction Prevention

(BIP) study. J Clin Epidemiology 1995;48:757–65.

[84] Elwood PC, Renaud S, Sharp DS, et al. Ischemic heart

disease and platelet aggregation. Circulation 1990;83:

38–44.

[85] Fitzgerald DJ, Roy L, Catella F, et al. Platelet activa-

tion in unstable coronary artery disease. N Engl J Med

1986;315:983–9.

[86] Thaulow E, Erikssen J, Sandvik L, et al. Blood platelet

count and function are related to total and cardiovascu-

lar death in apparently healthy men. Circulation 1991;

84:613–7.

[87] Trip MD, Cats VM, Van Capelle FJ, et al. Platelet

hyperreactivity and prognosis in survivors of myocar-

dial function. N Engl J Med 1990;322:1549–54.

[88] Bokinsky G, Miller M, Ault K, et al. Spontaneous

platelet activation and aggregation during obstructive

sleep apnea and its response to therapy with nasal

continuous positive airway pressure: a preliminary in-

vestigation. Chest 1995;108:625–30.

[89] Dimsdale JE, Coy T, Ziegler MG, et al. The effect of

sleep apnea on plasma and urinary catecholamines.

Sleep 1995;18:377–81.

[90] Fletcher EC, Miller J, Schaaf JW, et al. Urinary cate-

cholamines before and after tracheostomy in patients

with obstructive sleep apnea and hypertension. Sleep

1987;10:35–44.

[91] Wedzicha JA, Syndercombe-Court D. Increased plate-

let aggregate formation in patients with chronic airflow

obstruction and hypoxemia. Thorax 1991;46:504–7.

[92] Lavie L, Perelamn A. Plasma homocysteine levels in

obstructive sleep apnea: association with cardiovascu-

lar mortality. Chest 2001;120:900–8.

[93] Phillips BG, Narkiewicz K, Pesek CA, et al. Effects of

obstructive sleep apnea on endothelin-1 and blood

pressure. J Hypertens 1999;17:61–6.

[94] Saarelainen S, Seppala E, Laasonen K, et al. Circulat-

ing endothelin-1 in obstructive sleep apnea. Endothe-

lium 1997;5:115–8.

[95] Ohga E, Nagase T, Tomita T, et al. Increased levels of

circulating I–CAM-1, VCAM-1, and L-selectin in ob-

structive sleep apnea syndrome. J Appl Physiol 1999;

87:10–4.

[96] Beelke M, Angeli S, Del Sette M, et al. Obstructive

sleep apnea can be provocative for right-to-left shunting

through a patent foramen ovale. Sleep 2002;25:21–7.

[97] Shanoudy H, Soliman A, Raggi P, et al. Prevalence of

patent foramen ovale and its contribution to hypoxemia

in patients with obstructive sleep apnea. Chest 1998;

113:91–6.

[98] Mohsenin V. Sleep-related breathing disorders and risk

of stroke. Stroke 2001;32:1271–8.

H. Yaggi, V. Mohsenin / Clin Chest Med 24 (2003) 223–237 237

Obstructive sleep apnea in epilepsy

Bradley V. Vaughn, MDa,*,1, O’Neill F. D’Cruz, MDa,b,2

aDivision of Sleep and Epilepsy, Department of Neurology, University of North Carolina School of Medicine,

Chapel Hill, NC 27599-7025, USAbDivision of Pediatric Neurology, Department of Neurology, University of North Carolina School of Medicine,

Chapel Hill, NC 27599-7025, USA

The dramatic interplay of sleep and epilepsy has

been known since antiquity. In the fourth century,

Aristotle noted, ‘‘sleep is similar to epilepsy and in

some way, sleep is epilepsy’’ [1]. Even in the second

century, the importance of sleep in the treatment of

epilepsy was observed. Galen cautioned his patients

with seizures against sleepiness, and Soranus noted

that sleep ‘‘must be undisturbed’’ [2]. These early

observations demonstrated the importance of sleep

quality to patients with seizures.

Although sleep apnea is a common disorder, the

first report of treatment of sleep apnea in a patient

with epilepsy was in 1981 by Wyler and Weymuller

[3]. Their patient underwent tracheotomy and attained

control of the generalized seizures and improvement

in the partial seizures. Subsequent reports suggested

significant benefits of treating sleep apnea in patients

with epilepsy [4–7]. Hypotheses of the mechanism

by which sleep apnea seems to exacerbate epilepsy

rest on the physiologic consequences of sleep apnea.

In this article the authors explore these observations

that have led to many questions underlying the

prevalence, mechanisms, and potential therapeutic

relationships of sleep apnea to epilepsy.

Epilepsy

The term ‘‘epilepsy’’ is derived from the Greek

work epilambanien, which means to seize or to attack

[8]. Although epilepsy patients were believed in this

time to be seized by demons, science has come to the

understanding that epileptic seizures are the clinical

manifestations of excessive hypersynchronus central

neuronal activity. The clinical diagnosis of epilepsy is

defined as the chronic condition of recurrent unpro-

voked epileptic seizures. Epileptic seizures typically

are divided into partial and generalized seizures.

Partial seizures start in one location and potentially

spread to other regions of the brain. This seizure type

may be subdivided into simple partial (retention of

memory and consciousness), complex partial (impair-

ment of memory or consciousness), or secondarily

generalized. Primary generalized seizures begin

simultaneously over both hemispheres and comprise

various types of behavior. Absence seizures are

characterized by brief staring episodes. Atonic sei-

zures erupt as a sudden loss of tone that results in a

patient falling, whereas tonic seizures produce diffuse

stiffening from increase in muscle tone. Clonic sei-

zures are associated with repetitive jerking, and

myoclonic seizures are single rapid jerks. Tonic-

clonic seizures start with generalized tonic posturing

that progresses to clonic activity. A summary is

provided in the Box 1 [9].

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00023-6

* Corresponding author.

E-mail address: vaughnb@glial.med.unc.edu

(B.V. Vaughn).1 Dr. Vaughn has research funding from Cyberonics

Inc., UCB Pharma, and GlaxoSmith Kline and is a member

of the speaker’s bureau for Glaxo Smith Kline, Cyberonics,

Abbott, and Sanofi.2 Dr. D’Cruz has research support from UCB Pharma,

Cyberonics, and GlaxoSmith Kline and is a member of the

speaker’s bureau for OrthoMcNeil and Shire.

Clin Chest Med 24 (2003) 239–248

In 1881, Gower reported on the relationship of

sleep and awake states to epilepsy. He noted that 21%

of patients had seizures solely during sleep [10]. He

also found that 42% patients had seizures only during

the awake state, whereas a third group of 37% had

seizures during the awake and asleep states. Later

investigation by Janz revealed that some individuals

have seizures primarily in the first 2 hours after

awakening [11]. Janz coined the term ‘‘awakening’’

epilepsies for these individuals and referred to seizures

that occur without dependence on the sleep-awake

state as the diffuse epilepsies [11,12]. The state-

dependent types of epilepsy may be more susceptible

to alteration in sleep than the diffuse epilepsies.

Obstructive sleep apnea

Obstructive sleep apnea (OSA) is a common

disorder seen in as many as 9% of adult women,

24% of adult men, and 2% of children [13,14].

Defined by repetitive apneas or hypopneas caused

by increased airway resistance in sleep, this disorder

is associated with nocturnal oxygen desaturation or

frequent arousals. The disorder is arguably a com-

bination of altered central nervous system control

over state-dependent regulation of breathing and

predisposition of airway structure [15]. Regardless

of the underlying cause, this disorder influences the

prevalence of hypertension, diabetes mellitus, and

stroke and produces significant neurologic manifes-

tations of cognitive decline and changes in autonomic

regulation [16,17]. These neurologic manifestations

may be a result of sleep deprivation, oxygen desatu-

ration, or disturbance of other systems, such as

neuroendocrine, required for the maximal perform-

ance of the brain.

Conversely, diseases that alter the central nervous

system increase the likelihood of disturbing regu-

lation over sleep-related respiration and propagating

upper airway obstruction. Disorders such as stroke,

Alzheimer’s disease, and myotonic dystrophy have a

higher prevalence of sleep apnea [18–20]. The cir-

cular argument of central nervous system involve-

ment in OSA and disorders of the central nervous

system having a higher association of sleep apnea is

not surprising. Neurologic disorders are likely to alter

the function of the neurons involved in state-depen-

dent regulation of breathing just as neurons are

susceptible to the deleterious effect of sleep apnea.

We have become conscious of the importance of

sleep for maximal performance of the central nervous

system. Treatment of OSA may improve some of the

central nervous system function but not cure the

underlying neurologic process.

Effect of obstructive sleep apnea on epilepsy

Clinicians have inferred that OSA exacerbates

epilepsy from the beneficial effect of treatment of

OSA in patients with epilepsy [4–7]. Several studies

have shown that for some patients, treatment of OSA

resulted in the reduction of seizures in patients with

focal-onset seizures and generalized seizures. This

was seen in adults and children [4,5,7]. The authors

noted a reduction in number of seizures in patients

with state-dependent seizures, whether focal-onset or

generalized seizures, and Devinsky et al and Vaughn

et al reported a greater reduction in the number of

adult patients with generalized seizures [4,5]. Koh

et al demonstrated that 56% of children with various

neurologic disorders had a reduction in seizure fre-

quency [7]. Oliveira found that patients with focal-

onset epilepsy have fewer epileptiform discharges on

their electroencephalogram after treatment of their

Box 1. Seizure types

I. Generalized seizures of Non-FocalOrigin1. Tonic-Clonic2. Tonic3. Clonic4. Absence5. Atonic/akinetic6. Myoclonic

II. Partial Seizures1. Simple partial (without loss

of consciousness)a. motor symptomsb. sensory symptomsc. autonomicd. psychic symptomse. compound forms

2. Complex partial(impaired consciousness)a. simple partial seizure

followed by lossof consciousness

b. Impairment ofconsciousness at onset

c. Automatisms3. Partial seizures evolving to

secondary generalizationIII. Unclassified Seizures

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248240

OSA [21]. These studies uniformly found that some

patients had significant benefit in seizure control with

the treatment of sleep apnea.

In addition to the effect OSA has on seizure

frequency, sleep apnea can deter from a patient’s

quality of life. As seen in patients with OSA without

epilepsy, patients with epilepsy frequently complain

of excessive daytime sleepiness, unrefreshing sleep,

and low energy [5,22]. These symptoms were

improved after the initiation of therapy for OSA

[4,5,22]. Although no quality-of-life studies have

been performed in patients with epilepsy and OSA,

these patients conveyed subjective improvement in

their sleep.

Prevalence

The prevalence of epilepsy in the general popu-

lation is approximately 1%. Epilepsy most frequently

begins in childhood and later adult years [23]. Middle

age adulthood holds the lowest incidence of epilepsy.

OSA, however, has a peak incidence in middle age and

occurs predominantly in adult men. Patients with

neurologic disorders seem to have a greater prevalence

for sleep disturbance than normal subjects. This

increase in prevalence seems to extend to patients with

epilepsy. Miller showed that more than two thirds of

patients with epilepsy seen at a university center had

complaints regarding sleep [24]. Polysomnographic

investigation by Malow et al showed that nearly one

third of patients with medically refractory epilepsy had

a respiratory disturbance index of more than 5 [22]. In

the authors’ cohort of 25 patients with intractable

epilepsy, they found that 36% had a respiratory dis-

turbance index of more than 10. This may have male

gender predominance. In the three adult studies that

showed the effect of treatment of OSA in patients with

epilepsy, men were strikingly more affected than

women. Nine of the ten patients in the authors’ cohort

were men, eight of the nine in Malow’s series were

men, and six of the seven in Devinsky’s report were

men [4,5,22]. This may be caused in part by selection

bias. These patients also may not be obese or have the

‘‘typical’’ body habitus associated with OSA. Two of

the authors’ ten patients had normal body habitus and

did not have features upon examination that suggested

sleep apnea [5]. Although all of these studies are

compelling, larger cohorts are needed to elucidate

the true prevalence and age and gender distribution

of sleep apnea in patients with epilepsy.

The apparent increased prevalence of sleep apnea

in patients with epilepsy may be from several etiol-

ogies. These factors may be inherent in the epileptic

disorder or result from the treatment of the epilepsy.

Disorders of the central nervous system may affect

the regulation of respiration and increase the risk of

sleep apnea. This is seen in patients with other

neurologic disorders, such as Alzheimer’s disease,

strokes, cerebral palsy, and myotonic dystrophy

[18–20,25]. Therapeutic intervention for epilepsy

also may increase the risk of sleep apnea. Some

anticonvulsant medications promote weight gain

and may alter respiratory regulation. Valproate, viga-

batrin, and gabapentin are well known to accelerate

obesity, which increases the likelihood for sleep

apnea. Vigabatrin has been reported to cause a

significant weight gain, which results in a patient

developing overt signs of OSA [26]. Patients who are

given medications that promote weight gain should

have regular visits to monitor their weight and be

queried regarding symptoms of sleep apnea. Benzo-

diazepines and barbiturates may cause suppression in

responsiveness of carbon dioxide and oxygen desatu-

ration and increase upper airway musculature relaxa-

tion [27]. The changes in regulation of breathing may

be more sensitive to these inhibitory medications and

exacerbate underlying sleep-related breathing distur-

bance during certain stages of sleep. Another form of

therapy for epilepsy, vagus nerve stimulation, has

been reported to increase airway disturbance poten-

tially during sleep in some patients [28]. This therapy

may increase airway resistance from stimulation of

recurrent laryngeal nerve or interfere with the respi-

ratory sensory feedback.

Obstructive sleep apnea also may influence the

prevalence of epilepsy. Seizures as a direct result of

apnea are rare. In one patient, apnea in sleep report-

edly caused a seizure after severe oxygen desatura-

tion and cardiac arrest [29]. In another study of

patients with OSA, Sonka et al found that 4% of

their cohort had epilepsy [30]. This prevalence

exceeds that of the general population. Most

(78.9%) of these patients had seizures only during

sleep, and most of the events were generalized

seizures. This study may be skewed by variances in

referral patterns, but the elevated prevalence raises

interesting question of sleep apnea provoking sei-

zures or unmasking a potential for seizures.

Mechanisms

The treatment of sleep apnea seems to reduce the

recurrence of seizure in some patients. The subsequent

inference is that the presence of sleep apnea increases

the recurrence of seizures in these same patients. The

mechanism by which sleep apnea influences the sei-

zure disorder is unclear, however, some observations

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248 241

may provide clues to this relationship. Many patients

who responded to OSA therapy had a reduction of

seizures away from the period of sleep [5]. The

potential mechanism of this influence must take into

account the effect on seizure induction away from the

individual apnea. Potential mechanisms for sleep

apnea increasing the likelihood of seizures rest on

two general hypotheses, which are derived from the

pathologic consequences of the apnea: sleep depriva-

tion and oxygen desaturation. Sleep deprivation and

sleep fragmentation may increase the vulnerability to

seizure recurrence similarly to other potential distur-

bances of sleep. The second hypothesis is based on

apnea producing hypoxemia, which subsequently dis-

rupts the mechanisms inhibiting seizures.

Sleep and sleep deprivation in epilepsy

Sleep can play an important role in the seizure

discharge. This effect can be seen in the interictal and

ictal discharge. Interictal discharges are the electro-

encephalographic signature of epilepsy away from

the seizure. Sleep may activate interictal activity in

approximately one third of patients with epilepsy and

up to 90% of subjects with state-dependent epilepsies

[31–34]. For focal-onset seizures, the interictal dis-

charge may have little to do with the actual seizure.

The focal discharges may increase with the onset of

light sleep and demonstrate a greater frequency and

spatial and morphologic variability with stage 3 and

4 sleep. Rapid eye movement (REM) sleep is asso-

ciated with relative suppression of interictal dis-

charges. The epileptiform discharges that occur in

REM sleep are briefer and have less variability in

morphology and location than those seen in non-

REM sleep. Overnight studies of interictal activity

demonstrated that the focal interictal activity

increases with the entrance into the deeper stages

of non-REM sleep [35]. These stages of sleep are

physiologically linked to greater thalamocortical

neuronal synchronization, when fewer neurons are

engaged in active membrane depolarization [36].

More neurons are in the resting membrane state

and can be recruited into the discharge. This greater

availability of neuronal recruitment may account for

the greater spatial and morphologic variability of the

focal interictal discharge.

For primary generalized epilepsies, little distin-

guishes ictal from interictal discharges in that in-

terictal discharges are shorter and have no clear

behavioral manifestations. These generalized dis-

charges are more common during light sleep and

after awakening, however. Horita found that the

discharges are longest in stage 1 sleep [37]. The

deeper non-REM stages of sleep, stages 3 and 4,

are less likely to activate the generalized discharges,

and REM sleep has a further suppressive action.

Seizures have a slightly different pattern than

interictal discharges in relation to sleep. Focal and

generalized seizures are more likely to occur out of

light sleep or soon after awakening and rarely occur

in REM sleep [38]. Generalized seizures also occur

frequently near arousals or soon after awakenings.

Potential exists that the thalamocortical relay neu-

rons are more vulnerable at these times to hyper-

excitable synchronization and allow for generalized

discharges to propagate into seizures. Shouse has

postulated that a synchronous bursts-pause pattern of

entrained thalamocortical neurons extends beyond

the normal firing and can generate into a spike-wave

discharge [38].

Sleep deprivation can bring out seizures and

increase the frequency of interictal activity. Sleep

deprivation has been used extensively in long-term

epilepsy monitoring settings to trigger seizure activity.

In some patients, sleep deprivation is a powerful

provocative agent, whereas other patients demonstrate

little change in seizure frequency [31,39–41]. Rajna

and Veres found that in 9 of 14 patients with temporal

lobe epilepsy, seizures occurred on more than half of

the days after sleep deprivation [41]. Sleep deprivation

is also noted to increase the frequency of generalized

seizures and increase the interictal discharges in

patients with generalized epilepsies [42]. Although it

is still debated, activation of interictal activity from

sleep deprivation may be related to the promotion of

the onset of sleep or the disruption of central nervous

system processes that inhibit seizures [43].

Obstructive sleep apnea disrupts sleep and can

cause significant sleep deprivation. Janz noted that

sleep deprivation frequently provokes seizures in

patients with the awakening epilepsies. These epi-

lepsies are frequently characterized by generalized

seizures [11,12]. The first case report of OSA and

epilepsy showed a resolution of the generalized sei-

zures after tracheotomy [3]. Devinsky et al found that

the patients with generalized seizures were more

likely to have a reduction in seizure frequency after

the treatment of the OSA than their patients with

focal-onset seizures [4]. Two of their patients, who

had only generalized tonic-clonic seizures, became

seizure free after appropriate continuous positive

airway pressure (CPAP) therapy. The decrease in

seizure frequency in response to treatment of sleep

apnea also seems to extend to children. In their

cohort of 12 children with primary generalized epi-

lepsy and absence seizures, Carney and Kohrman

reported an average of 92% reduction in seizures

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248242

with appropriate treatment of the OSA [44]. None of

these children had oxygen desaturation below 87%,

which supported the idea that some mechanism other

than oxygen desaturation was playing a role for

seizure provocation and recurrence (Carney, personal

communication, 2002).

Patients with focal-onset epilepsy also seem to

have a reduction in their seizures after treatment of

OSA. In a cohort of patients who had focal-onset

seizures, more than half had a more than 50%

reduction in seizure frequency with no alteration in

medication [5]. Although they had a reduction in

seizure frequency, two patients who became seizure

free did not have oxygen desaturation below 88% on

polysomnographic examination. Pediatric patients

with presumed focal-onset epilepsy also showed

improvement in seizure frequency with treatment of

the OSA. Koh et al reported that one patient with

focal-onset seizures had near complete control of

seizures after tonsillectomy [7]. This patient had

oxygen desaturation only to 96%, which supported

the hypothesis that something other than low oxygen

plays a role for the increase in seizure frequency.

In the context of sleep apnea, the changes in sleep

architecture caused by the repetitive apnea increases

the vulnerability to seizures. The patient with sleep

apnea has more frequent arousals, greater percent of

time awake and light sleep, and less REM sleep. The

increased number of arousals and increased amount

of time awake and in light sleep afford a greater

chance of seizure initiation by increasing the percent

of time in a state that is more vulnerable to seizures.

Sleep fragmentation may allow a greater opportunity

for seizure initiation. Most patients with OSA also

have either disrupted REM sleep or diminished time

in REM sleep. REM sleep seems to have an anticon-

vulsant effect and increases the threshold for seizure

occurrence [45]. Patients with REM sleep disruption

have less of the antiseizure effect of this state. REM

sleep is important in reducing seizure recurrence and

may play a role in the provocative seizure effect of

sleep deprivation.

Shouse showed that the propensity for generaliza-

tion of epileptic discharges increases after sleep

deprivation, as seen by the susceptibility of cats to

penicillin-induced seizures after sleep deprivation

[46]. Sleep deprivation also affects the development

of a seizure focus. Animals can be kindled to develop a

seizure focus by repetitive exposure to epileptigenic

chemicals or electrical stimulation. This model for

epilepsy has been correlated to human focal-onset

epilepsy. Total sleep deprivation causes an increase

in the rate of kindling, and REM sleep deprivation

accelerates kindling of amygdala [47,48]. Conse-

quently, sleep fragmentation may increase seizure

frequency by interfering with seizure inhibitory mech-

anisms, and potentially increasing kindling and sleep

deprivation may accelerate the progression of the

epileptic focus. The clinical application of these find-

ings raises a concerning issue regarding progression of

the epileptic process in humanswith epilepsy and sleep

deprivation. It also suggests that sleep fragmentation

and sleep deprivation may have a differentiating effect

on focal and generalized epilepsies.

Hypoxia

Many of the patients described in reports by

Devinsky et al, Vaughn et al, and Koh et al had

significant oxygen desaturations [4,5,7]. Several of

these patients had dramatic improvement in seizure

frequency after therapeutic intervention of the apnea.

The effect of hypoxia must be considered as one

potential mechanism for the improvement in seizure

frequency. Seizures are not a common manifestation

of brief periods of hypoxia, but they are frequently

seen in individuals who suffered anoxic encephalop-

athy. The effect of hypoxia on lowering the seizure

threshold seems to be most prominent in the devel-

oping brain. Hypoxia induces a hyperexcitable state

in the immature hippocampus [49]. This effect seems

to be most significant and long lasting if the hypoxia

occurs during the perinatal period. Animal studies

have shown that hypoxia produces a profound effect

on glutamate synapses and leads to the cascade of

events that ends in cell death and reorganization that

promotes epileptogenesis [50].

Although these findings may have some applica-

tion to children with nocturnal hypoxia, their

application to adults with OSA is unclear. Adult

mice made hypoxic may be more susceptible to

certain types of seizure induction. In adult mice,

hypoxia induced by breathing a 5% oxygen prepara-

tion lowered the seizure threshold to several con-

vulsant agents [51]. This phenomenon was blocked

by the application of adenosine A1 receptor ant-

agonist. In certain mutant mice that lack the Kir6.2

subunit of the potassium sensitive ATP channels,

brief hypoxia can lead to generalized seizures. These

mice lack the ability to enhance the substantia nigra

pars reticulata’s role in seizure suppression [52].

Emerson et al also found some evidence to suggest

that hypoxia preconditioning may enhance the pro-

tective mechanisms of the brain [53]. These studies

suggested that hypoxia may alter the seizure thresh-

old, but we have limited understanding of how

hypoxia influences seizure induction and recurrence

in humans.

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248 243

Epilepsy causing a breathing disorder

The article has focused thus far on patients who

have epilepsy and developed OSA (Figs. 1, 2). There is

a need to address the possibility that the seizure focus

may cause apneas. Respiratory disturbances related to

seizures are not uncommon. Oxygen desaturation

frequently is found during seizures in individuals

who are in an epilepsy monitoring unit [54]. Snoring

and apneas that occur with seizures may be part of

the ictus or may occur as a postictal phenomenon (see

Figs. 1, 2). Repetitive nightly seizures can be mis-

taken for sleep apnea [55]. Seizures also can cause

nocturnal choking, as seen in rolandic epilepsy and

epileptic operculum syndrome [56]. The clinicianmust

be alert for the occurrence of seizures. This is one reas-

on that adequate electroencephalographic monitoring

should be included in the overnight polysomnogram.

Clinical manifestations and evaluation

The disruption of restorative sleep by OSA results

in excessive daytime sleepiness and other symptoms

reminiscent of sleep apnea. Patients with epilepsy and

sleep apnea frequently complain of excessive daytime

sleepiness, unrefreshing sleep and loud snoring. They

may have witnessed apneas or periods in which they

have awoken themselves from horrific snoring or

gasping. The clinician must ask about hypersomno-

lence, snoring, and other symptoms of increased

upper airway resistance and obstruction. The clinician

also should ask about recurrence of seizure, trends of

seizure frequency especially associated with the

symptoms of sleep disturbance, time of seizures,

and intensity of seizure and look for potential

increase in seizures that may suggest a relationship

of OSA to seizures. Patients must relate their current

medication regimen and changes that have occurred

even before the onset of sleep-related symptoms. The

clinician should note weight changes, pattern and

time of sleep, and concurrent use of herb or food

supplements. The differential diagnosis of hypersom-

nolence includes sedating medications, sleep depriva-

tion, circadian rhythm disorders, and other causes of

sleep disruption. Recurrence of seizures also should

be considered. As with any other complex medical

condition, patients should undergo a thorough sleep

history and physical examination before considera-

tion for polysomnography. The polysomnography

should include more extensive electroencephalo-

graphic coverage of the frontal and temporal head

regions [57]. These patients also should be screened

for thyroid abnormalities and other medical disorders

that may increase the likelihood of disturbed sleep. A

complete review of potential causes for the symptoms

offers a better chance for successful identification and

treatment of the underlying cause.

Therapeutic options

Various therapeutic interventions have been used

in patients with OSA and epilepsy. The sleep

Fig. 1. The patient has an apnea after the initiation of the seizure.

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248244

specialist, epileptologist, and patient should work

together with the goal of determining a treatment

option that corrects the breathing disturbance with-

out exacerbating the underlying seizure disorder.

Optimization of medications to avoid progressive

weight gain or sedation should be considered if it

could be accomplished without exacerbating the

seizures. Making sure the patient is educated

regarding sleep apnea and becomes vested in the

therapy is a fundamental feature to ensure high

patient compliance and satisfaction. Patient involve-

ment is an intricate factor in determining the best

treatment option.

The most common intervention for treating OSA in

adults has been CPAP, which can be used safely in

patients with epilepsy. Close and frequent follow-up

that focuses on identifying issues that interfere with

CPAP use and educating patients have been key to

improving compliance. Patients may experience the

predictable difficulties with CPAP and respond to

similar interventions. The authors have not had any

patients become entangled in the tubing or injured with

the device during a seizure. The authors also have used

positional therapy with success. Two patients from

their original series responded well to positional ther-

apy using a tennis ball in the middle of the back of a

nightshirt. These patients had clear positional-related

sleep apnea defined on an overnight study and were

motivated to continue the therapy.

Other investigators have promoted the use of

medications such as protriptyline or acetazolamide

[4,58]. Devinsky et al reported that two of their

patients noted benefits in seizure frequency and

symptoms of OSAwith use of protriptyline [4]. They

also noted trying acetazolamide in three patients, but

the results of seizure reduction and symptoms of OSA

were mixed. Acetazolamide has the attractive benefit

of being a mild anticonvulsant and may improve

seizure control by more than one mechanism. On

the other hand, theophylline should be used with care

in patients with epilepsy because of its potential for

lowering the seizure threshold. Anticonvulsant ther-

apy should be directed toward complete seizure

control with no side effects. If possible, patients

may improve by avoiding any sedating and weight-

enhancing medications. Alternatively, the epileptolo-

gist may consider the use of anticonvulsant agents

that promote weight loss, such as topiramate. Medi-

cations should be optimized to improve respiration

without impairing seizure control.

Airway surgery also has been used successfully

to treat OSA. Wyler and Weymuller’s first report

used the correction of airway obstruction by means

of tracheotomy [3]. Although tracheotomies are

preformed for only the most severe cases of OSA,

alternative surgeries are available and can improve

the sleep apnea and compliance with CPAP. Ton-

sillectomy can be performed safely and is the

treatment choice for many children with OSA.

Whatever the surgery, the operation should be

tailored to the patient, and close postoperative moni-

toring may be required. A team that includes the

Fig. 2. This figure demonstrates an apnea occurring at the end of a seizure.

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248 245

epileptologist, sleep specialist, surgeon, and anes-

thesiologist should discuss the potential benefits and

risks for the patient with the goals of correction of

the airway disturbance.

Oral devices have been proposed as a viable option

in patients with epilepsy [58]. Although these devices

are an alternative therapy for OSA, patients with

history of mastication during or after a seizure should

be counseled on the potential risk of the device being

fractured during a seizure and the possibility of airway

occlusion. Hard, nonpliable dental devices that have a

high retention of the teeth may have lower likelihood

of becoming dislodged and fractured. Another concern

is for patients who have postictal vomiting. A dental

device may impede the ability for the patient to clear

the airway, which should be especially concerning if

the patient has impaired sensorium from the seizure

and postictal somnolence.

The patient with epilepsy and OSA has many

therapeutic options and considerations to maximize

the potential for improvement. Optimal care for the

epilepsy and OSA reduces the chance of untoward

effects of either disease process. Regardless of the

therapy chosen, the patient must be vested in the

therapy, and close follow-up is crucial to success.

Summary

Obstructive sleep apnea can affect an individual

with epilepsy profoundly. These relatively common

disorders can coexist and potentially exacerbate each

other. The identification and appropriate treatment

of OSA may have far-reaching consequences in

improving a patient’s quality of life and recurrence

of seizures. Clinicians must be aware of the relation-

ship of these disorders and keenly question epilepsy

patients, regardless of their body habitus, regarding

potential symptoms of sleep apnea. Although the

underlying pathogenic mechanisms are unclear, we

can model the information gained from the observa-

tions to further the understanding of the relationship

between sleep and epilepsy.

Acknowledgment

The authors wish to extend their great apprecia-

tion to Michelle Wrightsell for her administrative and

editorial assistance that made this article possible.

The authors also wish to extend their appreciation to

Beth Malow for her thoughtful input regarding issues

discussed in the manuscript.

References

[1] Hett WS. Aristotle, on sleep and waking. On the Soul,

Parva naturalia, On breath. Boston: Harvard University

Press; 1957. p. 8–9.

[2] Temkin O. The falling sickness: a history of epilepsy

from the Greeks to the beginning of modern neurol-

ogy. 2nd edition, revised. Baltimore: Johns Hopkins

Press; 1994.

[3] Wyler AR, Weymuller EA. Epilepsy complicated by

sleep apnea. Ann Neurol 1981;9:403–4.

[4] Devinsky O, Ehrenberg B, Bathlen GM, Abramson

HS, Bathlen GM, Luciano D. Epilepsy and sleep apnea

syndrome. Neurology 1994;44:2060–4.

[5] Vaughn BV, D’Cruz OF, Beach R, Messenheimer JA.

Improvement of epileptic seizure control with treat-

ment of obstructive sleep apnea. Seizure 1996;5:73–8.

[6] Ezpeleta D, Garcia-Penna A, Peraita-Adrados R.

Epilepsia y sindrome de apnea del sueno. Rev Neurol

1998;26:389–92.

[7] Koh S, Ward SL, Lin M, Chen LS. Sleep apnea treat-

ment improves seizure control in children with neuro-

developmental disorders. Pediatr Neurol 2000;22:

36–9.

[8] Grigg-Damberger M, Damberger SJ. Historical aspects

of sleep and epilepsy. In: Bazil CW, Malow BA, Sam-

maritano MR, editors. Sleep and epilepsy: the clinical

spectrum. New York: Elsevier; 2002. p. 3–17.

[9] Commission on Classification and Terminology of the

International League against Epilepsy. Proposal for re-

vised clinical and electroencephalographic classifica-

tion of epileptic seizures. Epilepsia 1981;22:489–501.

[10] Passouant P. Historical aspects of sleep and epilepsy.

Epilepsy Res 1991;2:19–30.

[11] Janz D. The grand mal epilepsies and the sleep waking

cycle. Epilepsia 1962;3:69–109.

[12] Janz D. Epilepsy and the sleep-waking cycle. In: Vin-

ken PJ, Bruyn GW, editors. The epilepsias: handbook

of clinical neurology. Amsterdam: North Holland Pub-

lishing Company; 1974. p. 457–90.

[13] Young T, Palta M, Dempsey J, Skatrud J, Weber S,

Badr S. The occurrence of sleep disordered breathing

among middle aged adults. N Engl J Med 1993;

328:1230–5.

[14] Workshop Summary Group. Cardiorespiratory sleep

studies in children: establishment of normative data

and polysomnographic predictors of morbidity. Am J

Respir Crit Care Med 1999;160:1381–7.

[15] Widdicombe J. Upper airway reflexes. Current Opin

Pulm Med 1998;4:376–82.

[16] Naegele B, Thouvard V, Pepin JL, Levy P, Bonnet C,

Perret JE, et al. Deficits of cognitive executive

functions in patients with sleep apnea syndrome. Sleep

1995;18:43–52.

[17] Khoo MC, Belozeroff V, Berry RB, Sassoon CS. Car-

diac autonomic control in obstructive sleep apnea:

effects of long term CPAP therapy. Am J Respir Crit

Care Med 2001;164:807–12.

[18] Turkington PM, Bamford J, Wanklyn P, Elliott MW.

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248246

Prevalence and predictors of upper airway obstruction

in the first 24 hours after acute stroke. Stroke 2002;33:

2037–42.

[19] Guilleminault C, Stoohs R, Quera-Salva MA. Sleep re-

lated obstructive and nonobstructive apneas and neuro-

logic disorders. Neurology 1992;42(7 Suppl 6):53–60.

[20] Bliwise DL, Yesavage JA, Tinklenberg JR, Dement

WC. Sleep apnea in Alzheimer’s disease. Neurobiol

Aging 1989;10:343–6.

[21] Oliveira AJ, Zamagni M, Dolso P, Bassetti MA, Gigli

GL. Respiratory disorders during sleep in patients with

epilepsy: effect of ventilatory therapy on EEG interictal

epileptiform discharges. Clinical Neurophysiology

2000;111(Suppl 2):141–5.

[22] Malow BA, Fromes GA, Aldrich MS. Usefulness of

polysomnography in epilepsy patients. Neurology

1997;48:1389–94.

[23] Hauser WA, Annegers JF, Kurland LT. Incidence of

epilepsy and unprovoked seizures in Rochester, Min-

nesota, 1935–1984. Epilepsia 1994;34:453–68.

[24] Miller MT, Vaughn BV, Messenheimer JA, Finkel AG,

D’Cruz OF. Subjective sleep quality in patients with

epilepsy. Epilepsia 1996;36(Suppl 4):43.

[25] Kotagal S, Gibbons VP, Stith JA. Sleep abnormalities

in patients with severe cerebral palsy. Dev Med Child

Neurol 1994;36:304–11.

[26] Lambert MV, Bird JM. Obstructive sleep apnea follow-

ing rapid weight gain secondary to treatment with viga-

batrin (Sabril). Seizure 1997;6:233–5.

[27] Takhar J, Bishop J. Influence of chronic barbiturate

administration on sleep apnea after hypersomnia pre-

sentation: case study. J Psychiatry Neurosci 2000;25:

321–4.

[28] MalowBA, Edwards J,MarzeeM, Sagher O, Fromes G.

Effects of vagus nerve stimulation on respiration during

sleep: a pilot study. Neurology 2000;55:1450–4.

[29] Kryer M, Quesney LF, Holder D, Gloor P, MacLeod P.

The sleep deprivation syndrome of the obese patient:

a problem with periodic nocturnal upper airway ob-

struction. Am J Med 1974;56:531–8.

[30] Sonka K, Juklichova M, Pretl M, Dostalova S, Horinek

D, Nevsimlova S. Seizures in sleep apnea patients: oc-

currence and time distribution. Sb Lek 2000;1:229–32.

[31] Degen R, Degen HE. Sleep and sleep deprivation in

epileptology. Epilepsy Res 1991;(Suppl 2):235–60.

[32] Gibbs EL, Gibbs FA. Diagnostic and localizing value

of electroencephalographic studies in sleep. Res Publ

Assoc Res Nerv Ment Disease 1947;26:366–76.

[33] Halasz P. Sleep, arousals and electroclinical manifes-

tations of generalized epilepsy with spike wave pattern.

Epilepsy Res 1991;(Suppl 2):43–8.

[34] Wieser HG. Temporal lobe epilepsy, sleep and arousal:

stereo-EEG findings. Epilepsy Res 1991;(Suppl 2):

97–119.

[35] Malow BA, Kushwaha R, Lin X, Morton KJ, Aldrich

MS. Relationship of interictal epileptiform discharges

to sleep depth in partial epilepsy. Electroencephalogr

Clin Neurophysiol 1997;102:20–6.

[36] Steriade M. Corticothalamic resonance, states of

vigilance and mentation. Neuroscience 2000;101:

243–76.

[37] Horita H, Uchida E, Maekawa K. Circadian rhythm or

regular spike-wave discharges in childhood absence

epilepsy. Brain Dev 1991;13:200–2.

[38] Shouse MN, Martins de Silva A, Sammaritano M. Cir-

cadian rhythm, sleep and epilepsy. J Clin Neurophysiol

1996;13:32–50.

[39] Ellinson RJ, Wilkin K, Bennett DR. Efficacy of sleep

deprivation as an activation procedure in epilepsy pa-

tients. J Clin Neurophysiol 1984;1:83–101.

[40] Thomaides TN, Kerezoudi EP, Chaudhuri LR, Chero-

poulous C. Study of EEGs following 24 hour sleep

deprivation in patients with post traumatic epilepsy.

Eur Neurol 1992;32:79–82.

[41] Rajna P, Veres J. Correlations between night sleep du-

ration and seizure frequency in temporal lobe epilepsy.

Epilepsia 1993;343:574–9.

[42] Arne-Bes MC, Calvet V, Thiberge M, Arbus L. Effects

of sleep deprivation in an EEG study of epileptics.

In: Sternman BM, Shouse MN, Passouant P, editors.

Sleep and epilepsy. New York: Academic Press;

1982. p. 441–52.

[43] Glick TH. The sleep-deprived electroencephalogram:

evidence and practice. Arch Neurol 2002;59:1235–9.

[44] Carney PR, Kohrman MH. Relation between epilepsy

and sleep during infancy and childhood. In: Bazil CW,

Malow BA, Sammaritano MR, editors. Sleep and epi-

lepsy: the clinical spectrum. New York: Elsevier; 2002.

p. 359–72.

[45] Kumar P, Raju TR. Seizure susceptibility decreases

with enhancement of rapid eye movement sleep. Brain

Res 2001;922:299–304.

[46] Shouse MN. Sleep deprivation increases susceptibility

to kindled and penicillin seizure events during all wak-

ing and sleep states in cats. Sleep 1998;119:162–71.

[47] Grahnstedt S. Sleep deprivation and kindled seizures.

Exp Neurol 1986;92:248–60.

[48] Kawahara R, Hamazaki Y, Takeshita H. Effect of REM

sleep deprivation on each seizure stage in feline amyg-

daloid kindling. Seishin Shinkeigaku Zasshi 1994;96:

109–21.

[49] Jensen FE, Wang C, Stafstrom CE, Liu Z, Geary C,

Stevens MC. Acute and chronic increases in excitabil-

ity in rat hippocampal slices after perinatal hypoxia in

vivo. J Neurophysiol 1998;79:73–81.

[50] Johnston MV. Excitotoxicity in neonatal hypoxia.

Mental Retardation and Developmental Disabilities

Research Reviews 2001;7:229–34.

[51] Zgodzinski W, Rubaj A, Kleinrok Z, Sieklucka Dziuba

M. Effect of adenosine A1 and A2 receptor stimulation

on hypoxia induced convulsions in adult mice. Pol J

Pharmacol 2001;53:83–92.

[52] Yamada K, Ji JJ, Yuan J, Miki T, Sato S, Horimoto N,

et al. Protective role of ATP sensitive potassium chan-

nels in hypoxia induced generalized seizure. Science

2001;292:1543–6.

[53] Emerson MR, Samson FE, Pazdernik TL. Effects of

hypoxia preconditioning on expression of metallothio-

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248 247

nein 1,2 and heme oxygenase-1 before and after

kainic acid induced seizures. Cell Mol Biol 2000;46:

619–26.

[54] Blum AS, Ives JR, Goldberger AL, Al-Aweel IC,

Krishnamurthy KB, Drislane FW, et al. Oxygen de-

saturations triggered by partial seizures: implications

for cardiopulmonary instability in epilepsy. Epilepsia

2000;41:536–41.

[55] Sharp S, D’Cruz OF. Seizures as a cause of sleep re-

lated respiratory disturbance. Sleep 1998;21(Suppl 3):

77.

[56] Watson DB, D’Cruz OF. Epileptic operculum syn-

drome presenting as obstructive sleep apnea. Sleep

2001;24(Suppl):A369–70.

[57] Foldvary N, Caruso AC, Mascha E, Perry M, Klem G,

McCarthy V, et al. Identifying montages that best de-

scribe electrographic seizure activity during polysom-

nography. Sleep 2000;23:221–9.

[58] Ehrenberg B. Sleep apnea and epilepsy. In: Bazil CW,

Malow BA, Sammaritano MR, editors. Sleep and epi-

lepsy: the clinical spectrum. New York: Elsevier; 2002.

p. 373–82.

B.V. Vaughn, O.F. D’Cruz / Clin Chest Med 24 (2003) 239–248248

Clin Chest Med 24 (2003) 249–259

Neuropsychological impairment and quality of life in

obstructive sleep apnea

Michael J. Sateia, MDa,b,*

aDepartment of Psychiatry, Dartmouth Medical School, 1 Medical Center Drive, Lebanon, NH 03756, USAbSleep Disorders Center, Dartmouth-Hitchcock Medical Center, 1 Medical Center Drive, Lebanon, NH 03756, USA

Obstructive sleep apnea (OSA) is characterized by

repetitive partial or complete airway obstructions

during sleep, with associated sleep disruption and

varying degrees of transient oxygen desaturation. The

indications for treatment of OSA fall into three broad

categories: (1) social complications, such as spousal

disturbance or patient embarrassment related to snor-

ing, (2) daytime dysfunction, including sleepiness,

psychological disturbance, cognitive impairment, or

quality-of-life issues, and (3) cardiovascular conse-

quences (eg, risk of systemic or pulmonary hyper-

tension, congestive heart failure, or arrhythmia). Of

these three categories, the daytime disturbances, in all

likelihood, are the most frequent motivations for

physicians and patients to pursue definitive treatment

for OSA. Although much attention has been paid to

excessive sleepiness as a complication of this con-

dition, there is less understanding about the relation-

ships between OSA and various cognitive and

psychological disturbances and the relationship of

these disturbances to quality of life.

Although research directed to the issue of cogni-

tive and psychological consequences of OSA has been

ongoing for more than 20 years, a clear picture has yet

to emerge, mainly because the area is complex and

study designs have varied significantly, which makes

comparisons between studies problematic. The pop-

ulations assessed in these investigations have varied

with respect to severity of their respiratory distur-

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All righ

doi:10.1016/S0272-5231(03)00014-5

This work was supported by Grant No. K07-HL03646

from the National Heart, Lung, and Blood Institute.

* Sleep Disorders Center, Dartmouth-Hitchcock Medi-

cal Center, 1 Medical Center Drive, Lebanon, NH 03756.

E-mail address: michael.j.sateia@dartmouth.edu.

bance, age, baseline physical characteristics, coexist-

ing medical illness, and other measures. Hypoxemia,

obesity, medications, and psychiatric state all have

potential bearing on the outcome of neuropsycholog-

ical assessment. To date, the nature of the relation-

ships between these factors and daytime impairments

in obstructive sleep apnea is not clearly defined.

This article summarizes current information re-

garding OSA and quality of life, cognitive disturbance,

and psychological factors, identifies limitations of the

available data, draws tentative conclusions, and dis-

cusses future research considerations.

Neuropsychological function

Background considerations

Cognitive function and neuropsychological testing

have been assessed in numerous studies over the past

20 years. In evaluating the results of these investi-

gations, differences in sampling and study design

must be considered. The characteristics of study

populations vary significantly across these studies.

The most obvious source of variation is the severity

of respiratory disturbance, as dictated by defined

inclusion criteria and sampling pool. A recent review

[1] noted that the effect size of cognitive impairment

in OSA correlated highly with the severity of breath-

ing disturbance, with effect size ranging from 0.2 to

0.3 standard deviations in samples with milder apnea

[2,3] to 2 to 3 standard deviations in clinical pop-

ulations with moderate to severe OSA [4,5]. As

detailed later in this section, severity also influenced

the type of cognitive impairment observed [1,6].

ts reserved.

M.J. Sateia / Clin Chest Med 24 (2003) 249–259250

Assessment of cognitive function also may be

influenced by other subject characteristics that vary

across studies. For example, early work by Findley et

al demonstrated significant cognitive impairment in a

population of 26 patients with OSA [7]. These

patients demonstrated daytime CO2 retention and

hypoxemia, which may play a role in the genesis of

cognitive dysfunction independent of sleep apnea.

Many studies have failed to assess or control for

other pertinent variables that may impact cognitive

performance, such as the psychological state of sub-

jects, particularly the degree of depression. As noted

by Bliwise [8] and Telakivi et al, [9] age and baseline

cognitive function of subjects also may play an

important role in determining degree of identified

neuropsychological impairment in OSA. Individuals

with high baseline function may have the ability to

compensate for the effects of mild to moderate OSA

on standard evaluation instruments that have been

designed primarily to detect dementia.

Box 1. Neuropsychological instruments in assessm

Instrument FunctiWechsler Adult Intelligence

Scale-Revised (WAIS-R)Gener

Simple Reaction Time (SRT) AlertnFour Choice Reaction Time (FCRTT) AlertnContinuous Performance Task (CPT) AlertnCritical Flicker Fusion (CFF) AlertnDigit Span (DS) AttentDigit Symbol Substitution (DSS) AttentTrail-Making (TM) Attent

(sequeStroop Color-Word AttentPaced Auditory SerialAddition (PASAT)

Attent

Wisconsin Card Sorting (WCS) ExecuTower of Toronto / London ExecuVerbal Fluency ExecuPicture Completion ExecuBlock Design (BD) ExecuPicture Arrangement Execu

conceObject Assembly Execu

constrWechsler Memory Scale (WMS) ImmedBenton Visual Retention Short-Rey Auditory Verbal Learning Memo

ExecuPurdue Pegboard PsychFinger Tapping Psych

A range of comparison groups has been used in

these investigations. Comparisons include use of

published normative data [7], healthy controls

[4,5,10,11], and other groups, such as insomniacs

[12,13], persons with other hypersomnolence disor-

ders [5], and patients with treated chronic obstructive

pulmonary disease [14].

An array of neuropsychological instruments has

been used in the assessment of patients with OSA.

Although there has been some overlap in the particu-

lar performance batteries used in these investigations,

the inevitable differences in such batteries, coupled

with other design variations, make comparisons

among studies difficult. In developing a specific

battery for research purposes, the neuropsychologist

chooses from a large number of individual tests, each

of which is purported to measure a specific domain or

domains of performance. Commonly used tests and

the primary domain(s) that they assess are listed in

Box 1. Decary et al have reviewed the subject of

ent of obstructive sleep apnea patients

onal intelligence

ess/vigilanceess/vigilanceess/vigilanceess/vigilance; visual motor functionion/short-term memoryion/concentration/psychomotorion/executive functionncing/visual search)ion/executive function (response inhibition)ion/executive function

tive functiontive function (problem solving)tive functiontive function (concept formation/reasoning)tive function (constructional ability)tive function (planning/organization/pt formation)tive function (concept formation/uctional ability)iate/delayed memory (logical/figure [visual])term memory (figure [visual] retention)ry (Immediate/delayed verbal learning)tive function (planning)omotor coordinationomotor coordination

M.J. Sateia / Clin Chest Med 24 (2003) 249–259 251

neurocognitive function in OSA and note the most

relevant areas of assessment: (1) general intellectual

function, (2) attention/vigilance/concentration, (3)

memory (working/episodic/procedural) and learning,

and (4) executive and motor function [15]. It is

important to note that because many of these tools

involve multiple functions, there is not always uni-

form agreement as to the significance and meaning of

impaired performance on a given test. The authors

have suggested a standard battery for neuropsycho-

logical assessment of sleep apnea patients that

includes measures in each of these areas.

The following issues are most commonly ad-

dressed in studies of neuropsychological impairment

in obstructive sleep apnea: (1) Is significant impair-

ment of neuropsychological function associated with

obstructive sleep apnea and what is the nature of

that impairment? (2) What is the relationship

between the severity of apnea and the type/degree

of dysfunction? (3) Which aspects of OSA (eg,

hypoxemia, frequency of events, disturbances of

sleep architecture) are most predictive of dysfunc-

tion? (4) Do different types of dysfunction correlate

with different characteristics of the sleep apnea? (5)

Is functional impairment reversible with treatment

and, if so, what is the pattern and time course

associated with that reversal?

General intellectual function

Numerous studies have identified a significant

degree of neuropsychological impairment with OSA

[4,5,7,11,16,17]. Deficits in global intellectual func-

tioning, typically assessed by IQ scores of the Wechs-

ler Adult Intelligence Scale (WAIS-R), have been

demonstrated [4,5,7,16]. For the most part, these

studies have suggested that deterioration in general

intellectual function in OSA is a function of hypox-

emia [4,7,16], although not all analyses have demon-

strated such an association or explored possible

correlation with other variables.

Attention and concentration

Measures of attention and concentration likewise

have yielded indication of impairment associated

with sleep apnea. Significant differences between

apnea subjects and controls have been described on

Trail Making [4,5,7,16], Stroop Color Test [11], Digit

Symbol [4,16], Paced Auditory Serial Addition Test

[7], and Letter Cancellation [4,5]. Similarly, mea-

sures of vigilance, such as Choice Reaction Time

[4,7,18], Continuous Performance Test [3,14], and

Steer Clear [18,19], document impairment, which is

consistent with the sleepiness manifested in most

sleep apnea patients.

Memory

Assessment of memory function in OSA has

suggested deficits in short- and long-term memory.

Bedard et al described diminished performance on a

short-term memory measure in patients with mod-

erate and severe sleep apnea, although only the

severe group demonstrated evidence of impairment

in delayed recall [4]. These disturbances were pri-

marily associated with decrease in vigilance. Green-

berg et al found no differences between patients with

OSA and controls on subscales of the Wechsler

Memory Scale (in either immediate or delayed con-

ditions) but did demonstrate a modest impairment in

immediate recall on digit span [5]. Short-term mem-

ory deficit has been noted by others [7,11]. Naegele

et al also demonstrated significant abnormalities in

long-term memory [11]. They argued, however, that

the apparent long-term disturbance reflects the defi-

cit in initial learning and that, in this respect, patients

with OSA more closely resemble persons with

frontal lobe lesions than persons with temporal

lesions, for whom true forgetting is a core feature.

Others have reported short- and long-term memory

problems [10,14,20].

Executive function

From a neuroanatomic standpoint, disturbances in

executive function and problem-solving areas are

associated with frontal lobe dysfunction. Although

the available data do reveal disturbance in executive

function in some cases, results are mixed and suggest

that these abnormalities are most evident in patients

with more severe sleep apnea. One group reported

mild performance deficits on standard measures of

executive function (Wisconsin Card Sorting [WCS]/

Tower of Toronto) [11]. Others found impaired per-

formance on Paced Auditory Serial Addition Task

(PASAT) in hypoxemic sleep apnea patients com-

pared with a nonhypoxemic group [7]. Bedard et al

described widespread deficits in various executive

functions (verbal fluency, planning, sequential think-

ing, and constructional ability), with extent and

severity of impairment apparently advancing in asso-

ciation with severity of the breathing abnormality [4].

In a more recent study, Salorio et al found no

difference in performance in WCS between patients

with OSA and controls, although results on verbal

fluency measures were mixed [10]. Studies of older

subjects with mild apnea have not revealed signifi-

M.J. Sateia / Clin Chest Med 24 (2003) 249–259252

cant disturbance in executive function, memory, or

other aspects of cognitive performance [21,22].

Cognitive impairment and severity

For purposes of comparison, OSA severity data,

as addressed in studies of cognitive impairment, are

usually based on frequency of respiratory events

(apnea-hypopnea index [AHI]). One should note that

this measure might best serve as a proxy for sleep

disruption and does not consistently reflect severity

of hypoxemia. Many of these studies, however, do

include analyses of the relative effects of hypoxemia

versus sleep disturbance on cognitive function. Eng-

leman et al, in reviewing case-control studies, noted a

distinct trend toward increasing effect size of cogni-

tive impairment with increasing AHI [1]. Other

researchers found only limited and relatively mild

differences between patients with moderate OSA and

controls, whereas the severe group had more wide-

spread and obvious dysfunction [4].

Analysis of correlations between specific physio-

logic parameters associated with OSA and disturbed

neuropsychological function is complicated by the

choice of physiologic parameters for analysis and the

definitions of those parameters. For example, an

assessment of the role of hypoxemia in generation

of impairment may use the number of desaturations

of 4% or more, the percentage of time spent below

specific thresholds, or the minimum saturations per

event. Likewise, exploration of the impact of sleep

disturbance on neuropsychological function may use

AHI, numbers of arousals (of varying definitions), or

direct measures of daytime vigilance/sleepiness, such

as reaction times or multiple sleep latency tests. The

outcome of these investigations depends—at least to

some extent—on the choice of measures. Which

measures produce the most significant and accurate

correlations remain unclear.

Not surprisingly, the investigations of correlations

among sleep, respiratory variables, hypoxemia, and

various neuropsychological tests have demonstrated

mixed results (Table 1). Correlations between global

intellectual impairment and hypoxemia have been

noted in several investigations. Other investigations

have described an association between executive func-

tion (eg, WCS or Block Design) and oxygen desatura-

tion in OSA. Measures of sleep disruption or direct

assessments of daytime sleepiness/alertness (eg,

arousals or Multiple Sleep Latency Test [MSLT]) have

been noted to correlate most closely with memory

disturbance (eg, digit span, Wechsler Memory Scale)

and tests of vigilance/alertness/ concentration, such as

Four Choice Reaction Time, Critical Flicker Fusion,

Digit Symbol Substitution, or Simple Reaction Time.

These relationships, however, are variable, overlap-

ping, relatively weak, and, as Englemann et al point

out, not strongly predictive of specific dysfunction [1].

The evidence of general intellectual slowing and,

more specifically, disturbance in executive functions

has led to the suggestion that OSA (and specifically

hypoxemia) may be associated with frontal lobe

dysfunction [4,23]. Others have postulated distur-

bance in neurotransmitter synthesis as the basis for

cognitive dysfunction [4], but the precise mecha-

nisms of these disturbances are not known.

Treatment and reversibility

Trials that examine the reversibility of cognitive

dysfunction in patients with OSA have used various

treatments (primarily continuous positive airway pres-

sure [CPAP]) and study designs (including normal

controls, placebo CPAP groups, and cross-over

designs). Bedard et al assessed ten patients with

moderate to severe OSA at baseline and 6 months after

CPAP treatment and compared them to ten control

subjects [24]. Significant baseline deficits in function

normalized to near control levels in most cases, but

tests of executive function (Trail Making Test [TMT]/

verbal fluency) did not significantly improve. The

investigators suggested that this continued impair-

ment might reflect irreversible hypoxic damage.

Naegele et al studied ten patients treated with

CPAP for 4 to 6 months and compared them to ten

controls [25]. At baseline, subjects demonstrated

significant differences from controls in multiple areas

of cognitive function. After treatment, these subjects

differed from controls only in persistence of short-

term memory deficits. The investigators argued that

persistent frontal lobe disturbance may be the basis of

the ongoing memory problems. Lojander et al eval-

uated the impact of surgical treatment (23 patients)

and nasal CPAP (27 patients) on cognitive function

[26]. In this group of patients with moderate sleep

apnea, CPAP treatment at 3 and 12 months was

associated with significantly greater improvement

(versus conservative management) on the Benton

Visual retention Test only. Patients who were treated

surgically did not differ from the conservative treat-

ment group on any psychometrics. These patients

showed only mild impairment at baseline, however.

Other researchers have found varying degrees of

improvement with CPAP in uncontrolled investi-

gations [23,27]. Only two recent studies have used

a credible placebo (subtherapeutic or sham CPAP) in

assessing effect of treatment on cognitive function. A

subtherapeutic CPAP-controlled, randomized cross-

Table 1

Correlation of neuropsychological performance with obstructive sleep apnea variables

Source Hypoxemia Other respiratory variables Sleep variables Alertness variables

Kingshott et al [58] Intellectual ability component

score correlates with min.

O2 (0.15)

Intellectual ability component

score correlates with AHI

(� 0.14)

Response slowing component

score correlates with wakefulness

component score (� 0.34)

Greenberg et al [5] Perceptual organization/motor

speed correlate with min. O2; global

performance shows no correlation

NA

Telakivi et al [20] Logical memory correlates

with DESA4 (snorers)

Spatial skills/memory retention/

WAIS-VS correlate with sleepiness

Berry et al [59] Logical memory/WMS/WAIS-PS/

verbal fluency correlates with

number of desaturations

Logical memory/visual memory/

WMS correlates with AI

Yesavage et al [60] NA Concentration/response inhibition/

eye-hand coordination/ (Raven)/

(Peabody) correlate with RDI

Findley et al [7] Short-term memory/problem solving/

attention impaired in hypoxemic vs.

nonhypoxemic group; global

performance impairment correlates

with median sleep SaO2

Sleep variables not correlated

with global cognitive function

Verstraeten et al [12] Psychomotor function/attention

correlates with alertness (FCRTT)

Bedard et al [4] General intellectual function

(WAIS) and executive functions

may be associated with hypoxemia

(cumulative % below threshold)

Cheshire et al [16] Attention/executive function

(TMT)/global IQ correlates with min.

O2; executive function (BD)/attention

(SRT)/global IQ correlates with DESA4

Vigilance (MSLT/FCRTT) associated

with attention/verbal memory

Telakivi et al [9] No cognitive function correlates

with DESA4 or median SaO2

Attention/executive function/

global IQ correlates with AHI

Executive function correlates

with arousals (� 0.41)

Redline et al [3] Subjective sleepiness estimates show

no correlation with any impairment

Naegele et al [11] Executive function associated with

hypoxemia in logistic regression

Short-term memory impairment

associated with AHI in

logistic regression

Upward sleep stage shifts correlate

with executive function

Sleepiness (MSLT) correlated with

executive function (WCS)/

short-term memory

M.J.

Sateia

/Clin

Chest

Med

24(2003)249–259

253

M.J. Sateia / Clin Chest Med 24 (2003) 249–259254

over study of 46 patients demonstrated significant

improvements in multiple areas of cognitive function

from baseline to treatment condition, as had been

demonstrated in previous studies [28]. No difference

in improvement was observed between CPAP and

placebo groups, however. Similarly, Bardwell et al

found that only 1 of 22 neuropsychological test

scores showed a difference between CPAP and sham

groups, a result attributable to chance alone, although

rank-sum analysis did suggest a greater overall

improvement in general cognitive function in the

treatment group [29].

The mean treatment periods for these studies were

only approximately 2 weeks and 1 week, respec-

tively, which leaves open the possibility that cogni-

tive functional improvement in response to CPAP

may require more extended periods of treatment. In

the Bardwell study, AHI on the sham CPAP was

reduced from 44 to 28, which raised the consideration

that this is actually ‘‘partially therapeutic’’ as opposed

to subtherapeutic. No reduction was observed in

another CPAP placebo investigation [28]. Although

numerous uncontrolled studies show improvement in

cognitive function after initiation of CPAP, the results

of placebo-controlled investigations do not provide

unequivocal support for the hypothesis that this

change is directly attributable to CPAP.

Psychological factors

Most of the studies that examine psychological

variables and OSA have focused on depression,

including baseline and posttreatment analyses. The

results have been mixed. Early studies identified

evidence of increased depression in patients with

OSA on Minnesota Multiphasic Personality Inventory

(MMPI) [30,31]. Reynolds noted that 20% of patients

with OSA met criteria for a past or present episode of

depression and noted increased likelihood of depres-

sion in the sleepier group [32]. Millman found that

45% of 55 patients with OSA generated scores of more

than 50 on the Zung Depression Self-Rating Scale

[33]. Statistically significant correlation between AHI

and depression was not identified, although a trend

toward higher Respiratory Disturbance Index (RDI) in

the depressed group was noted. In a controlled MMPI

study, investigators noted elevation on multiple scales,

including depression, in patients with moderate to high

severity OSA compared with controls [34].

Mosko et al used a questionnaire based on Diag-

nostic and Statistical Manual of Mental Disorders

(DSM-III) to assess depression [35]. They reported

that 58% of the OSA group met DSM-III criteria for

major depression of four or more depressive symp-

toms. Similar results were noted for narcolepsy and

periodic limb movement patients. Only 26% of

patients described themselves as currently depressed,

however. Dahlof et al found that 34% met criteria for

depression using the Comprehensive Psychiatric Rat-

ing Scale and clinical interview [36].

Several other investigations have failed to find

significant increases in depression associated with

sleep apnea. In a 5-year longitudinal study, Phillips

found no evidence of significant psychopathology in

a population of older adults, although this population

had relatively mild OSA [22]. An investigation of

2271 patients screened for sleep apnea found no

significant association between breathing disturbance

and depression or other psychopathology, as assessed

by the Symptom Check List-90 [37]. Although

women with simple snoring and patients with severe

sleep apnea were noted to have elevated depression

scores, the investigators did not find a consistent

relationship between depression and apnea in this

group and concluded that the higher rates of depres-

sion are related to gender and personality differences

rather than sleep apnea. Likewise, other investigators,

using various assessment instruments, discovered no

association between OSA and depression [38,39].

Treatment and reversibility

If sleep apnea does significantly predispose to

depression, one would expect at least some degree of

improvement in mood as a result of treatment for OSA.

Several studies have demonstrated this. Most treat-

ment response studies unfortunately are not well

controlled. Derderian described a reduction of Profile

of Mood States (POMS) depression scores in seven

patients with moderate to severe apnea and noted that

improvement in depression correlated with increase in

slow-wave sleep [40]. In the Dahlof study [36], the

percentage of persons who meet criteria for depression

fell from 34% to 10% after uvulopalatopharyngo-

plasty. Similarly, Mosko’s patient group had signifi-

cant reduction of depression (and fatigue and anger)

POMS scores 2 to 3 months after various corrective

upper airway surgeries [35]. Millman also found

substantial reduction in Zung Self-Rating Scale scores

in his sample after initiation of nasal CPAP [33].

Positive airway pressure resulted in modest but

progressive improvement of the MMPI depression

scale and several other scales in 23 apnea patients

described by Platon and Sierra [34]. Of note, the

improvement did not reach statistical significance until

the third follow-up, which occurred at 11 to 14 months

after initiation of nasal CPAP. More recently, others

M.J. Sateia / Clin Chest Med 24 (2003) 249–259 255

have reported improvement at 1 and 3 months on nasal

CPAP using the Beck Depression Inventory [41].

Recent placebo-controlled (oral placebo or subthera-

peutic nasal CPAP) studies that examined change in

depression in response to treatment have cast some

doubt on the earlier results. Engleman et al did find

significant improvement in Hospital Anxiety and

Depression Scale depression ratings in a population

of patients with mild OSA treated with CPAP for

4 weeks versus a group treated with oral placebo [6].

Barnes et al noted no difference between patients

who received CPAP and the oral placebo group for

POMS or Beck depression scores [42]. Henke et al

saw no difference between CPAP and subtherapeutic

(0–1 cm H2O) CPAP groups using the Geriatric

Depression Scale [28]. Yu et al also used subther-

apeutic CPAP as a placebo control [43]. They found

improvement in the active treatment and placebo

groups, which suggested that the improvement is

primarily placebo response. Several design limita-

tions of this study, as noted by the authors, suggest

a need for caution in interpretation of these results.

Patients with more severe depression were excluded

from the study, which produced a study population

with relatively low depression POMS scores. Post-

treatment assessment was conducted at 1 week,

whereas most treatment studies have used follow-up

evaluation at 1 to 14 months. The placebo group did

show an almost 30% reduction in RDI, which raised

the possibility that there may have been a partial

therapeutic response to the ‘‘placebo.’’

Although the standard clinical perspective regard-

ing psychological function and OSA suggests that the

disorder is commonly associated with some degree

of depression that typically remits with treatment,

the evidence, particularly from placebo-controlled

treatment trials, is mixed. Numerous variables must

be considered in assessing the methodology of

these investigations. The length of treatment period

for these studies varied from 1 week to more than

1 year. At least one study suggested that full treatment

response may not occur for months, which raised

some question about those studies with short follow-

up periods. The baseline severity of apnea and the

severity of depression may impact outcomes assess-

ment. Although effect may be more apparent in

populations with more severe breathing disturbance

and associated symptoms, it is important to define the

lowest level of severity at which treatment interven-

tion effects significant change. Not all studies report

compliance with CPAP (which typically is moderate,

at best). Any assessment of outcome clearly must

include determination of compliance. Finally, it has

been demonstrated that these patients exhibit substan-

tial placebo response, which underscores the necessity

of placebo controls in these investigations. Subther-

apeutic CPAP seems the most appropriate placebo

intervention, although further understanding of the

extent to which a partial therapeutic response to this

‘‘placebo’’ may compromise distinctions between

treatment and control groups would be helpful.

Quality of life

Patents who are treated for moderate to severe

obstructive apnea typically note marked subjective

improvement in quality of life. The baseline disturb-

ance and the treatment response have been well

documented in numerous quality-of-life assessment

studies.Most of these studies have used the Short Form

36 (SF-36), a 36-item subscale of the Medical Out-

comes Survey (MOS), which measures physical func-

tioning, role limitations caused by physical and

emotional difficulties, mental health, physical

pain, vitality/energy, and general health perception.

Although there is variance across studies with respect

to the particular areas of disturbance, almost all studies

demonstrated some impairment in one or more areas.

Some studies suggested a linear relationship between

the severity of apnea and breadth and degree of

functional disturbance. Most of the studies also

revealed marked improvement—if not complete res-

olution—of the dysfunction with effective treatment.

The Wisconsin Sleep Cohort Study evaluated

738 patents with the SF-36 [44]. Although the qual-

ity-of-life assessment was not conducted until some

time after the sleep study, diminished general health

was correlated with apnea in dose-response fashion,

even after controlling for age, body mass index, and

other health factors. Increasing impairments in phys-

ical function, mental health, role function associated

with physical problems, social role, and energy were

associated with increasing severity of OSA. A trend

toward diminished life satisfaction correlated with

breathing disturbance also was noted. An investigation

of 5816 patents from the Sleep Heart Health Study

found that energy/vitality was the only scale that

demonstrated a linear relationship with apnea [45].

Severe sleep apnea, however, was associated with

significant abnormalities in multiple SF-36 scales,

including physical and social function, vitality, and

general health. The investigators also identified asso-

ciations between insomnia and sleepiness complaints

and disturbance on all SF-36 scales. Both of these

groups point out that the degree of impairment noted in

the samples is on the same order as that noted in other

populations of patents with significant medical illness,

M.J. Sateia / Clin Chest Me256

such as diabetes, heart disease, arthritis, or clinical

depression. Several other investigations have de-

scribed decrements in various domains of the SF-36

[46–50]. These disturbances have been seen in cohorts

of persons with mild sleep apnea [48] and in persons

with more severe apnea. Some of these studies iden-

tified a relationship between severity and degree

of disturbance.

Analysis of predictors of daytime dysfunction does

not provide a clear conclusion. The Finn et al study

found that AHI correlated significantly with multiple

SF-36 scales [44]. Baldwin et al analyzed data by RDI

4% and by clinical categories of severity and found

only a linear relationship between the latter and vitality

[45]. Reports of difficulty initiating or maintaining

sleep and excessive sleepiness did predict widespread

disturbance in quality-of-life measures. Moore et al

found that RDI did correlate with health distress,

energy/fatigue, mobility, and social function when

age and gender were controlled [51]. Finally, Bennett

noted only a weak relationship between pretreatment

SF-36 scores and sleep fragmentation indices [50].

Treatment and reversibility

Studies of the impact of treatment on quality of

life have focused primarily on CPAP. Bolitschek et al

and Bennett et al described normalization of daytime

function after 3 months and 4 weeks of nasal CPAP,

respectively [50,52]. Others found broad improve-

ment in quality-of-life measures after 6 months on

CPAP [47]. No relationship between arousals and

change in daytime function was identified, but a

correlation between hypoxemia indices and quality-

of-life improvement was noted. Two studies have

examined the issue of CPAP response in placebo-

controlled trials. Engleman et al, in an oral placebo

investigation of subjects with mild sleep apnea,

reported improvement in Nottingham Health Profile

total score for the CPAP versus placebo group,

although this reached statistical significance only

for the better CPAP compliers [6]. An earlier inves-

tigation by the same group revealed significant

improvement versus placebo in patients with mod-

erate to severe OSA [53]. Jenkinson et al adminis-

tered CPAP and subtherapeutic CPAP in randomized

fashion to a total of 107 patients with moderate OSA

[54]. They found significantly greater improvement

in the CPAP group for numerous SF-36 scales, with

effect sizes of 1.02 for mental component summary

and 1.68 for energy/vitality. Other researchers have

described an association between degree of improve-

ment in quality-of-life measures and severity of

baseline impairment [46].

Other instruments

Flemons et al, noting that generic instruments

such as the SF-36 or Nottingham Health Profile

may not be optimal instruments for identifying and

tracing symptoms of sleep apnea, have developed the

Calgary Sleep Apnea Quality of Life Index [55]. This

35-question instrument addresses four domains,

including daily function, social interaction, emotional

function, and symptoms. Correlations between the

Sleep Apnea Quality of Life Index and SF-36 total

scores at baseline were relatively low (0.21). Change

scores for the two instruments showed significant

correlations for total scores and for several SF-36

subscales, however. In an expanded study they

described assessment of 90 patients before and after

CPAP [56]. The Sleep Apnea Quality of Life Index

did not correlate with severity of OSA but did show

moderate (0.36 –0.71) correlations with various

SF-36 scales. Based on 62 subjects who completed

at least 4 weeks of CPAP, they found that changes in

Sleep Apnea Quality of Life Index were most

strongly associated with change in RDI, global qual-

ity of life rating, and vitality and social function

scales of SF-36. Another sleep apnea-specific instru-

ment, the Obstructive Sleep Apnea Patient-Oriented

Severity Index, assesses 32 items and demonstrates

significant correlation with patients’ subjective global

assessment of quality of life [57].

The available data strongly suggest that even mild

sleep apnea is associated with some degree of impair-

ment in quality of life. Although the exact nature of

the impairment may vary from study to study depend-

ing on the characteristics of the patient sample, the

severity of apnea, and the specific instrument(s) used

to measure quality of life, the weight of evidence

supports significant dysfunction, possibly on the order

of that observed in common chronic illnesses. Several

considerations must be weighed in interpreting these

data, however. The specific evaluation tool may

influence substantially the outcome of such studies.

Although instruments such as the SF-36 or Notting-

ham Health Profile are well-validated and widely used

devices, they may not provide the most accurate

evaluation of quality of life in patients with sleep

apnea, as Flemons and others [55] have pointed out.

Not only do they fail to assess directly many symp-

toms of potential relevance but they also demonstrate

a ceiling effect in healthy controls and treatment

responders that may obscure significant differences

between controls/responders and untreated apnea

patients. Other instruments designed to measure more

specific symptoms in OSA show promise, but further

assessment is required.

d 24 (2003) 249–259

M.J. Sateia / Clin Chest Med 24 (2003) 249–259 257

Quality-of-life data do not consistently dem-

onstrate a strong association between impairment

and severity of sleep apnea, as measured by AHI or

degree of sleep fragmentation. As several investiga-

tors pointed out, the absence of a strong correlation

between a particular OSA symptom or symptom

cluster and a specific index of respiratory or sleep

disturbance is hardly unprecedented. The same holds

true for other symptoms, such as sleepiness. Numer-

ous factors contribute to these symptom presenta-

tions, and perhaps it is unrealistic to expect high

degrees of association between these symptoms and

any one variable.

Future studies of quality of life in OSA must focus

on several issues to strengthen further a demonstra-

tion of dysfunction. A single, well-recognized assess-

ment tool would allow ready comparison of results

among investigators. Although the SF-36 has played

that role to some extent thus far, an instrument more

specific for OSA may be conducive to more accurate

and efficient identification of differences. Any anal-

ysis of this sort must control for the variety of

confounding variables, such as age, gender, body

mass index, smoking, alcohol consumption, and the

potential contribution of co-varying medical or psy-

chiatric disorders that may, in their own right, be

associated with significant quality of life impairment.

Studies of treatment response must use adequate

placebo controls, such as subtherapeutic CPAP.

Cross-over designs should use adequate washout

periods to reduce the risk of carry-over effects that

might contaminate results.

Summary

Although clinical experience has suggested for

more than two decades that OSA is associated with

impairment of cognition, emotional state, and quality

of life and that treatment with nasal CPAP produces

significant improvements in these areas, sound empir-

ical evidence to support this view, especially regard-

ing treatment outcome, has been lacking. More recent

investigations have begun to provide this support

from randomized, adequately controlled studies.

These assessments suggest that some degree of cog-

nitive dysfunction is associated with OSA. The

effects are most apparent in the severe cases, whereas

results in mild cases are more equivocal. Reported

impairments include global intellectual dysfunction

and deficits in vigilance, alertness, concentration,

short- and long-term memory, and executive and

motor function. Considerable discrepancy exists

across studies with respect to type and degree of

dysfunction, however. Disturbances in general intel-

lectual function and executive function show stron-

gest correlations with measures of hypoxemia. Not

unexpectedly, alterations in vigilance, alertness, and,

to some extent, memory seem to correlate more with

measures of sleep disruption. Although many inad-

equately controlled investigations have demonstrated

reversibility of most or all of these deficits with

effective treatment, more recent placebo-controlled

studies have raised doubts regarding whether the

observed changes are truly a function of treatment.

This issue requires further systematic exploration

with adequate controls and step-wise analysis of

treatment duration effects.

A similar set of considerations exists with respect

to the relationship between psychological distur-

bance, primarily depression, and OSA. Although

several studies suggest significant depression in these

patients, the results are mixed. Placebo-controlled

treatment trials fail to demonstrate consistently a

difference in mood improvement between active

treatment groups and controls, although several meth-

odologic considerations suggest that these results

should be interpreted with caution.

Numerous investigations leave little doubt about

the issue of quality of life impairment among persons

with OSA. Further characterization of impairment,

particularly in areas specific to this population, will

provide clearer understanding of the problem.

Preliminary investigations of treatment response in

controlled studies indicate significantly greater

improvement of quality of life in response to CPAP.

Although patients with OSA commonly report

disturbances in cognitive and psychological function

and general quality of life, the increased rates of

obesity, hypertension, diabetes, cardiovascular dis-

ease, medication use, and related psychosocial com-

plications present a host of potential etiologies that

might explain the impairments noted. There can be

little doubt that these covariants do, in some cases,

contribute to neuropsychological dysfunctions. It is

essential that future studies continue to define those

disturbances that are specific to OSA, the relationship

between levels of severity and impairment, the role of

treatment in reversing these dysfunctions, and the

correlation between test results and significant day-

to-day social and occupational functional impairment.

References

[1] Engleman HM, Kingshott RN, Martin SE, Douglas NJ.

Cognitive function in the sleep apnea/hypopnea syn-

drome (SAHS). Sleep 2000;23:S102–8.

M.J. Sateia / Clin Chest Med 24 (2003) 249–259258

[2] Kim HC, Young T, Matthews CG, Weber SM, Wood-

ward AR, Palta M. Sleep-disordered breathing and

neuropsychological deficits: a population-based study.

Am J Respir Crit Care Med 1997;156:1813–9.

[3] Redline S, Strauss ME, Adams N, Winters M, Roebuck

T, Spry K, et al. Neuropsychological function in mild

sleep-disordered breathing. Sleep 1997;20:160–7.

[4] Bedard MA, Montplaisir J, Richer F, Rouleau I, Malo J.

Obstructive sleep apnea syndrome: pathogenesis of

neuropsychological deficits. J Clin Exp Neuropsychol

1991;13:950–64.

[5] Greenberg GD, Watson RK, Deptula D. Neuropsy-

chological dysfunction in sleep apnea. Sleep 1987;

10:254–62.

[6] Engleman HM, Kingshott RN, Wraith PK, Mackay

TW, Deary IJ, Douglas NJ. Randomized placebo-con-

trolled crossover trial of continuous positive airway

pressure for mild sleep apnea/hypopnea syndrome.

Am J Respir Crit Care Med 1999;159:461–7.

[7] Findley LJ, Barth JT, Powers DC, Wilhoit SC, Boyd

DG, Suratt PM. Cognitive impairment in patients with

obstructive sleep apnea and associated hypoxemia.

Chest 1986;90:686–90.

[8] Bliwise DL. Sleep apnea and cognitive function: where

do we stand now? Sleep 1993;16:S72–3.

[9] Telakivi T, Kajaste S, PartinenM, Brander P, NyholmA.

Cognitive function in obstructive sleep apnea. Sleep

1993;16:S74–5.

[10] Salorio CF, White DA, Piccirillo J, Duntley SP, Uhles

ML. Learning, memory, and executive control in indi-

viduals with obstructive sleep apnea syndrome. J Clin

Exp Neuropsychol 2002;24:93–100.

[11] Naegele B, Thouvard V, Pepin JL, Levy P, Bonnet C,

Perret JE, et al. Deficits of cognitive executive functions

in patients with sleep apnea syndrome. Sleep 1995;

18:43–52.

[12] Verstraeten E, Cluydts R, Verbraecken J, De Roeck J.

Neuropsychological functioning and determinants of

morning alertness in patients with obstructive sleep

apnea syndrome. J Int Neuropsychol Soc 1996;2:

306–14.

[13] Stone J, Morin CM, Hart RP, Remsberg S, Mercer J.

Neuropsychological functioning in older insomniacs

with or without obstructive sleep apnea. Psychol Aging

1994;9:231–6.

[14] Roehrs T, Merrion M, Pedrosi B, Stepanski E, Zorick

F, Roth T. Neuropsychological function in obstructive

sleep apnea syndrome (OSAS) compared to chronic

obstructive pulmonary disease (COPD). Sleep 1995;

18:382–8.

[15] Decary A, Rouleau I, Montplaisir J. Cognitive defi-

cits associated with sleep apnea syndrome: a pro-

posed neuropsychological test battery. Sleep 2000;

23:369–81.

[16] Cheshire K, Engleman H, Deary I, Shapiro C, Douglas

NJ. Factors impairing daytime performance in patients

with sleep apnea/hypopnea syndrome. Arch Intern

Med 1992;152:538–41.

[17] Dealberto MJ, Pajot N, Courbon D, Alperovitch A.

Breathing disorders during sleep and cognitive per-

formance in an older community sample: the EVA

Study. J Am Geriatr Soc 1996;44:1287–94.

[18] Nykamp K, Rosenthal L, Guido P, Roehrs T, Rice FM,

Syron ML, et al. The effects of sleepiness on perform-

ance among patients with OSA. Sleep Research 1997;

26:450.

[19] Findley L, Unverzagt M, Guchu R, Fabrizio M, Buck-

ner J, Suratt P. Vigilance and automobile accidents in

patients with sleep apnea or narcolepsy. Chest 1995;

108:619–24.

[20] Telakivi T, Kajaste S, Partinen M, Koskenvuo M, Sal-

mi T, Kaprio J. Cognitive function in middle-aged

snorers and controls: role of excessive daytime somno-

lence and sleep-related hypoxic events. Sleep 1988;11:

454–62.

[21] Knight H, Millman RP, Gur RC, Saykin AJ, Doherty

JU, Pack AI. Clinical significance of sleep apnea in the

elderly. Am Rev Respir Dis 1987;136:845–50.

[22] Phillips BA, Berry DT, Lipke-Molby TC. Sleep-disor-

dered breathing in healthy, aged persons: fifth and final

year follow-up. Chest 1996;110:654–8.

[23] Feuerstein C, Naegele B, Pepin JL, Levy P. Frontal

lobe-related cognitive functions in patients with sleep

apnea syndrome before and after treatment. Acta Neu-

rol Belg 1997;97:96–107.

[24] Bedard MA, Montplaisir J, Malo J, Richer F, Rouleau I.

Persistent neuropsychological deficits and vigilance

impairment in sleep apnea syndrome after treatment

with continuous positive airways pressure (CPAP).

J Clin Exp Neuropsychol 1993;15:330–41.

[25] Naegele B, Pepin JL, Levy P, Bonnet C, Pellat J,

Feuerstein C. Cognitive executive dysfunction in pa-

tients with obstructive sleep apnea syndrome (OSAS)

after CPAP treatment. Sleep 1998;21:392–7.

[26] Lojander J, Kajaste S, Maasilta P, Partinen M. Cogni-

tive function and treatment of obstructive sleep apnea

syndrome. J Sleep Res 1999;8:71–6.

[27] Valencia-Flores M, Bliwise DL, Guilleminault C, Cil-

veti R, Clerk A. Cognitive function in patients with

sleep apnea after acute nocturnal nasal continuous pos-

itive airway pressure (CPAP) treatment: sleepiness and

hypoxemia effects. J Clin Exp Neuropsychol 1996;18:

197–210.

[28] Henke KG, Grady JJ, Kuna ST. Effect of nasal contin-

uous positive airway pressure on neuropsychological

function in sleep apnea-hypopnea syndrome: a random-

ized, placebo-controlled trial. Am J Respir Crit Care

Med 2001;163:911–7.

[29] Bardwell WA, Ancoli-Israel S, Berry CC, Dimsdale

JE. Neuropsychological effects of one-week continu-

ous positive airway pressure treatment in patients with

obstructive sleep apnea: a placebo-controlled study.

Psychosom Med 2001;63:579–84.

[30] Beutler LE, Ware JC, Karacan I, Thornby JI. Differ-

entiating psychological characteristics of patients with

sleep apnea and narcolepsy. Sleep 1981;4:39–47.

[31] Kales A, Caldwell AB, Cadieux RJ, Vela-Bueno A,

Ruch LG, Mayes SD. Severe obstructive sleep apnea.

M.J. Sateia / Clin Chest Med 24 (2003) 249–259 259

II: Associated psychopathology and psychosocial con-

sequences. J Chronic Dis 1985;38:427–34.

[32] Reynolds III CF, Kupfer DJ, McEachran AB, Taska LS,

Sewitch DE, Coble PA. Depressive psychopathology in

male sleep apneics. J Clin Psychiatry 1984;45:287–90.

[33] Millman RP, Fogel BS, McNamara ME, Carlisle CC.

Depression as a manifestation of obstructive sleep ap-

nea: reversal with nasal continuous positive airway

pressure. J Clin Psychiatry 1989;50:348–51.

[34] Ramos Platon MJ, Espinar Sierra J. Changes in psy-

chopathological symptoms in sleep apnea patients after

treatment with nasal continuous positive airway pres-

sure. Int J Neurosci 1992;62:173–95.

[35] Mosko S, Zetin M, Glen S, Garber D, DeAntonio M,

Sassin J, et al. Self-reported depressive symptomatol-

ogy, mood ratings, and treatment outcome in sleep

disorders patients. J Clin Psychol 1989;45:51–60.

[36] Dahlof P, Norlin-Bagge E, Hedner J, Ejnell H, Hetta J,

Hallstrom T. Improvement in neuropsychological per-

formance following surgical treatment for obstructive

sleep apnea syndrome. Acta Otolaryngol (Stockh)

2002;122:86–91.

[37] Pillar G, Lavie P. Psychiatric symptoms in sleep apnea

syndrome: effects of gender and respiratory distur-

bance index. Chest 1998;114:697–703.

[38] Bliwise DL, Yesavage JA, Sink J, Widrow L, Dement

WC. Depressive symptoms and impaired respiration in

sleep. J Consult Clin Psychol 1986;54:734–5.

[39] Cassel W. Sleep apnea and personality. Sleep 1993;16:

S56–8.

[40] Derderian SS, Bridenbaugh RH, Rajagopal KR. Neu-

ropsychologic symptoms in obstructive sleep apnea

improve after treatment with nasal continuous positive

airway pressure. Chest 1988;94:1023–7.

[41] Sanchez AI, Buela-Casal G, Bermudez MP, Casas-

Maldonado F. The effects of continuous positive air

pressure treatment on anxiety and depression levels

in apnea patients. Psychiatry Clin Neurosci 2001;55:

641–6.

[42] Barnes M, Houston D, Worsnop CJ, Neill AM, Myky-

tyn IJ, Kay A, et al. A randomized controlled trial of

continuous positive airway pressure in mild obstructive

sleep apnea. Am J Respir Crit Care Med 2002;165:

773–80.

[43] Yu BH, Ancoli-Israel S, Dimsdale JE. Effect of CPAP

treatment on mood states in patients with sleep apnea.

J Psychiatr Res 1999;33:427–32.

[44] Finn L, Young T, Palta M, Fryback DG. Sleep-disor-

dered breathing and self-reported general health status

in the Wisconsin Sleep Cohort Study. Sleep 1998;21:

701–6.

[45] Baldwin CM, Griffith KA, Nieto FJ, O’Connor GT,

Walsleben JA, Redline S. The association of sleep-dis-

ordered breathing and sleep symptoms with quality of

life in the Sleep Heart Health Study. Sleep 2001;24:

96–105.

[46] D’Ambrosio C, Bowman T, Mohsenin V. Quality of

life in patients with obstructive sleep apnea: effect of

nasal continuous positive airway pressure. A prospec-

tive study. Chest 1999;115:123–9.

[47] Kingshott RN, Vennelle M, Hoy CJ, Engleman HM,

Deary IJ, Douglas NJ. Predictors of improvements in

daytime function outcomes with CPAP therapy. Am J

Respir Crit Care Med 2000;161:866–71.

[48] Gall R, Isaac L, Kryger M. Quality of life in mild

obstructive sleep apnea. Sleep 1993;16:S59–61.

[49] Yang EH, Hla KM, McHorney CA, Havighurst T, Badr

MS, Weber S. Sleep apnea and quality of life. Sleep

2000;23:535–41.

[50] Bennett LS, Barbour C, Langford B, Stradling JR, Da-

vies RJ. Health status in obstructive sleep apnea: rela-

tionship with sleep fragmentation and daytime

sleepiness, and effects of continuous positive airway

pressure treatment. Am J Respir Crit Care Med 1999;

159:1884–90.

[51] Moore P, Bardwell WA, Ancoli-Israel S, Dimsdale JE.

Association between polysomnographic sleep mea-

sures and health-related quality of life in obstructive

sleep apnea. J Sleep Res 2001;10:303–8.

[52] Bolitschek J, Schmeiser-Rieder A, Schobersberger R,

Rosenberger A, Kunze M, Aigner K. Impact of nasal

continuous positive airway pressure treatment on qual-

ity of life in patients with obstructive sleep apnoea. Eur

Respir J 1998;11:890–4.

[53] Engleman HM, Martin SE, Deary IJ, Douglas NJ. Ef-

fect of continuous positive airway pressure treatment

on daytime function in sleep apnoea/hypopnoea syn-

drome. Lancet 1994;343:572–5.

[54] Jenkinson C, Davies RJ, Mullins R, Stradling JR.

Comparison of therapeutic and subtherapeutic nasal

continuous positive airway pressure for obstructive

sleep apnoea: a randomised prospective parallel trial.

Lancet 1999;353:2100–5.

[55] Flemons WW, Reimer MA. Development of a disease-

specific health-related quality of life questionnaire for

sleep apnea. Am J Respir Crit Care Med 1998;158:

494–503.

[56] Flemons WW, Reimer MA. Measurement properties of

the Calgary sleep apnea quality of life index. Am J

Respir Crit Care Med 2002;165:159–64.

[57] Piccirillo JF. Outcomes research and obstructive sleep

apnea. Laryngoscope 2000;110:16–20.

[58] Kingshott RN, Engleman HM, Deary IJ, Douglas NJ.

Does arousal frequency predict daytime function? Eur

Respir J 1998;12(6):1264–70.

[59] Berry DT, Webb WB, Block AJ, Bauer RM, Switzer

DA. Nocturnal hypoxia and neuropsychological varia-

bles. J Clin Exp Neuropsych 1986;8(3):229–38.

[60] Yesavage J, Bliwise D, Guilleminault C, Carskadon M,

Dement W. Preliminary communication: intellectual

deficit and sleep-related respiratory disturbance in the

elderly. Sleep 1985;8(1):30–3.

Obstructive sleep-disordered breathing in children:

new controversies, new directions

John L. Carroll, MD

Pediatric Sleep Disorders Center, Division of Pediatric Pulmonary Medicine, University of Arkansas for Medical Sciences,

Arkansas Children’s Hospital, 800 Marshall Street, Little Rock, AR 72202, USA

Obstructive sleep-disordered breathing (SDB) in

children was characterized until recently as obstruc-

tive sleep apnea syndrome (OSAS), usually treated by

surgical adenotonsillectomy, versus primary snoring,

which was believed to be of no clinical significance

and did not require treatment [1].

Although classic childhood OSAS is estimated to

occur in approximately 2% of children, the reported

prevalence of loud nightly snoring is much higher,

ranging to more than 20% in children [2–10].

Nightly snoring is common in children and, if asso-

ciated with significant morbidity, could represent an

enormous public health problem.

In the past, the pathophysiology of childhood ob-

structive SDB was believed to be relatively straight-

forward; sleep disruption was the likely cause of

daytime sleepiness and hypoxemia was believed to

result in growth impairment and cardiovascular com-

plications [1]. Other daytime symptoms of childhood

SDB were not widely recognized. Diagnosis of ob-

structive SDB in children was also believed to be

straightforward until recently, with polysomnography

touted as the ‘‘gold standard’’ for neatly separating

snoring children into categories of childhood OSAS

versus clinically benign ‘‘primary snoring.’’ Al-

though this approach was simple and straightforward,

recent advances suggest that it was incorrect or, at

best, incomplete.

The clinical picture of childhood SDB was com-

plicated in the late 1990s by general acknowledgment

that upper airway resistance syndrome (UARS) occurs

in children. That is, snoring children without classic

OSAS could exhibit significant daytime symptoms

related to increased upper airway resistance during

sleep. In 1999, an American Thoracic Society (ATS)

workshop summary on sleep studies in children

included childhood UARS and obstructive hypoven-

tilation in the classification of childhood SDB but

retained the concept of ‘‘primary snoring’’ [11].

Recent evidence indicates that childhood

obstructive SDB is not easily categorized into simple

clinical entities and that symptoms in children may

be varied, subtle, and difficult to detect. Far from

being ‘‘straightforward,’’ the area of childhood SDB

is currently characterized by a lack of consensus on

definitions, lack of diagnostic criteria, numerous

unanswered mechanistic questions, and several excit-

ing new directions. Because many reviews of child-

hood OSAS have been published [12–18], this article

focuses on new developments and controversies.

Clinical picture of childhood obstructive sleep-

disordered breathing: then and now

Snoring always indicates some degree of partial

airway obstruction. Although once believed to be

‘‘benign,’’ it is currently recognized that snoring, in

the absence of obstructive sleep apnea (OSA) or

hypoxemia during sleep, may be associated with

sleep disruption and daytime symptoms as severe or

worse than symptoms associated with full-blown

‘‘classic’’ childhood OSAS. From a respiratory per-

spective, childhood obstructive SDB is continuum,

with snoring on one end and complete upper airway

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00024-8

E-mail address: carrolljohnl@uams.edu

Clin Chest Med 24 (2003) 261–282

obstruction, hypoxemia, and obstructive hypoventila-

tion on the other (Fig. 1). The relationship among

daytime symptoms, nighttime breathing patterns, and

physiologic abnormalities is not simple. In the

absence of apnea, hypopnea, hypoxemia, or hyper-

capnia, a child with snoring may have disturbed sleep

and severe daytime symptoms, whereas a child with

severe sleep-related upper airway obstruction may

have minimal or no discernible daytime symptoms.

Even the absence of discernable daytime symptoms

(see Fig. 1B) does not rule out a pathologic condition

associated with snoring. The snoring ‘‘C student’’

may have been an ‘‘A student’’ were it not for subtle

sleep disruption associated with ‘‘apparently’’ benign

snoring [19–21]. The child who grows along the fifth

percentile for height and weight may have been in the

fiftieth percentile were it not for ‘‘apparently’’ benign

snoring [22,23]. The behavior of the ‘‘difficult’’ child

may have been better were it not for ‘‘apparently’’

benign snoring [6,24,25].

Classically, childhood OSAS was defined as par-

tial or complete upper airway obstruction during

sleep, usually associated with some combination of

sleep disruption, hypoxemia, hypercapnia, or daytime

symptoms attributable to the sleep-related airway

obstruction (see Fig. 1D, E). Normative polysomno-

graphic data for children were published in the 1980s

and 1990s [26], and as in adults, the diagnosis of

‘‘childhood OSAS’’ was based on threshold criteria

such as apnea index and degree of oxygen desatura-

tion. Children who snored but did not meet the

threshold criteria for childhood OSAS were classified

as ‘‘primary snorers,’’ which was believed to be

clinically insignificant (see Fig. 1B).

In recent years UARS has been used to describe

daytime symptoms caused by nighttime breathing-

related sleep disruption but without OSA or hypopnea

(see Fig. 1C) [13,27–30]. Guilleminault et al de-

scribed the clinical picture of childhood UARS as

early as 1982 [31], although it was not termed

‘‘UARS’’ until the 1990s [29]. It is currently widely

accepted that snoring children may exhibit a range of

daytime symptoms from subtle to disabling, regard-

less of whether they meet criteria for classic OSAS.

Despite these advances, the use of unvalidated thresh-

old indices (eg, apnea-index) and other unvalidated

‘‘diagnostic criteria’’ for childhood SDB continues.

‘‘New’’ perspective on daytime symptoms

More than 25 years ago it was recognized that

childhood obstructive SDB was associated with im-

paired daytime neurocognitive function and behavior

[31–35]. These observations had little effect on

diagnostic testing, however, which continued to focus

almost entirely on nighttime breathing measurements

and unvalidated threshold criteria to ‘‘score’’ the

degree of sleep disruption and abnormal breathing.

Although it was has been known for decades that

severe childhood OSAS could cause developmental

delay and cognitive impairment [35], little attention

Fig. 1. Continuum of upper airway resistance and airway obstruction.

J.L. Carroll / Clin Chest Med 24 (2003) 261–282262

was given to mild OSAS or snoring children without

OSAS. Despite several additional studies in the 1990s

that explored behavioral effects associated with snor-

ing [6,23–25], diagnosis and treatment practices

remained unchanged.

In 1998, Gozal reported that first grade children

with poor school performance had a higher-than-

expected prevalence of snoring and sleep-related

hypoxemia. Children treated with adenotonsillectomy

showed a statistically significant improvement in

school grades, whereas untreated children with SDB

showed no improvement [21]. This study marked a

turning point, with the full realization that ‘‘classic’’

childhood OSAS (see Fig. 1D, E) probably represents

only the ‘‘tip of the iceberg’’ [15]. Recently, the

major focus has shifted to the other end and middle

of the spectrum: children with snoring and important

but subtle and nonspecific behavioral and neurocog-

nitive daytime symptoms (see Fig. 1). A clear result

of this shift is the recognition that current approaches

to the identification and diagnosis children with

obstructive SDB are inadequate and much in need

of revision and standardization.

Childhood obstructive sleep-disordered breathing

A growing body of evidence suggests that the

traditional diagnosis of childhood OSAS encom-

passed only a small proportion of children with SDB

[13,15,36]. Nearly all of the existing literature on

childhood obstructive SDB is based on arbitrary, non-

validated criteria for classic OSAS and primary snor-

ing that were borrowed from the adult medical

literature decades ago, however. To complicate matters

further, even in the ‘‘classic’’ childhood OSA litera-

ture, data are highly variable because of lack of stan-

dardized definitions and diagnostic criteria [37,38].

Definitions

Discussion of childhood obstructive SDB should

start with the definition. Currently, however, there is

no standard, widely accepted definition. Given that

clinical symptoms can result from the entire spectrum

of childhood obstructive SDB, it seems reasonable to

consider classical OSAS, obstructive hypoventilation,

and snoring with daytime symptoms (UARS) as

manifestations of the same underlying pathophysiol-

ogy, under the heading ‘‘childhood obstructive SDB.’’

Currently, there are no officially endorsed diagnostic

criteria for childhood SDB similar to those published

for adult OSAS [39]. Although a ‘‘consensus confer-

ence’’ on childhood OSAS was held in the early 1990s

[1], an evidence-based ‘‘definitions conference,’’ sim-

ilar to the one convened to clarify definitions of adult

SDB [40], has never been organized.

Childhood obstructive SDB may be defined as a

disorder of breathing during sleep characterized by

prolonged increased upper airway resistance, partial

upper airway obstruction, or complete obstruction

that disrupts pulmonary ventilation, oxygenation, or

sleep quality. Nighttime manifestations include some

combination of snoring, increased respiratory effort,

episodic hypoxemia, CO2 retention, restless sleep,

and increased numbers of arousals and awakenings

from sleep. Daytime symptoms include excessive

daytime sleepiness, daytime tiredness, fatigue, poor

school performance, inattention, hyperactivity, oppo-

sitional behavior, and other subtle behavioral distur-

bances. This definition, modified from the American

Thoracic Society definition [1], encompasses all

childhood SDB diagnoses, including childhood

OSAS, obstructive hypoventilation, and UARS.

Because there is no consensus on diagnostic

criteria, practitioners are still faced with basic, fun-

damental questions, such as ‘‘What are the diagnostic

criteria for obstructive SDB in children?’’ ‘‘How to

identify children with obstructive SDB?’’ ‘‘What are

the indications for testing?’’ ‘‘What are the appropri-

ate methods of testing?’’ ‘‘What are the indications

for treatment (including avoidance of future morbid-

ity)?’’ ‘‘What are the short- and long-term outcomes

of treatment versus no treatment?’’ It is particularly

important for parents, teachers, family practitioners,

pediatricians, and third-party payors to discard the old

‘‘mindset’’ of childhood obstructive SDB manifesting

only as severe nighttime airway obstruction or overt

daytime sleepiness [41,42].

Epidemiology

The prevalence of snoring ‘‘often’’ or ‘‘nightly’’

(so-called ‘‘habitual snorers’’) ranges from 3.2% to

21% in children [2–4,6–9,43,44], and little is known

about the natural history of snoring in children. Ali et

al studied the natural history of snoring in a group of

children from age 4 to 7 years and found that the

overall prevalence of snoring did not change (12.1%

in 1989–1990 versus 11.4% in 1992) [5]. More than

half of the children who snored habitually at age 4 to

5 no longer did so by age 7, however. Although the

overall prevalence of daytime sleepiness decreased

with age, hyperactivity, excessive daytime sleepiness,

and restless sleep were more common in snoring

children compared with children who reported never

snoring. The prevalence of snoring in adolescents

and adults is higher than that reported for preadoles-

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 263

cent children, which suggests that snoring increases

with age.

Nothing is known about the prevalence of UARS

in children. The prevalence of ‘‘classic’’ childhood

OSAS is believed to be approximately 1% to 3% and

occurs in children of all ages. In children with normal

craniofacial structure, the peak incidence occurs

between approximately 3 and 6 years of age, which

corresponds to the age range when upper airway

lymphoid tissue enlargement, relative to craniofacial

size, is greatest [45]. OSAS seems to be more

frequent in African-American children, children with

respiratory disease, obese children, and children with

a family history of OSAS [46]. It is unclear whether

gender is a predisposing factor for OSAS in children.

Clinical features

Nighttime symptoms

Snoring is the most common nighttime symptom

of OSAS in children. The snoring sounds made by

children may have a higher pitched, more guttural, or

harsh sound than classic ‘‘nasal’’ snoring, however,

and some parents may not identify their child’s noisy

breathing during sleep as snoring. Simple questions

such as ‘‘Does your child snore?’’ may fail to identify

children with significant SDB. Parents usually do not

sleep in the same room as the child and may be

unaware of the child’s breathing sounds and pattern

during sleep.

Children may exhibit the classic adult pattern of

continuous snoring interrupted by pauses. In children,

however, OSAS tends to occur mainly in rapid eye

movement (REM) sleep; therefore, the snoring or

pauses may be absent for significant periods of the

night. Children with SDB also tend to exhibit a pattern

of prolonged partial upper airway obstruction and may

have few or even no complete obstructive apneas

[42,47]. Children may manifest SDB by making other

sounds, such as stridor, snorting, gasping, or grunting.

Loud gasping often accompanies arousals after

obstructive episodes. Children with SDB may have

obviously increased respiratory effort, which is often

manifested as paradoxical inward rib cage motion, and

some parents may describe this as ‘‘struggling’’ to

breathe during sleep [30]. Paradoxical inward rib cage

motion is normal in children during REM sleep until

age 3. Cyanosis is rarely observed by parents, even in

cases of severe childhood OSAS.

Sleep disturbances caused by SDB may be man-

ifested as restless sleep, increased movement during

sleep, ‘‘bed thrashing,’’ frequent arousals (sometimes

accompanied by gasping noises), frequent awakening,

and unusual sleeping positions (eg, sitting, propped up

on pillows). Other signs may be observed, such as

increased sweating during sleep or sleeping with the

neck hyperextended. Although enuresis has been

associated with OSAS in children [48], subsequent

studies have not confirmed the association [8].

Daytime symptoms

The most prominent daytime symptom of SDB in

adults—excessive daytime sleepiness—is absent in

most children with polysomnography-proven OSAS

[30,49]. A recent study using multiple sleep latency

testing confirmed that most children with OSAS

do not exhibit excessive daytime sleepiness [50]. This

is a major difference between children and adults

with SDB.

If children with SDB are not overtly sleepy during

the day, what are their daytime symptoms? The ef-

fects of obstructive SDB on mental development in

children were well recognized more than 100 years

ago [51–53]. Studies from the early 1980s showed

that children with OSAS may exhibit daytime behav-

iors, such as pathologic shyness, social withdrawal,

hyperactivity, aggressiveness, tiredness, and fatigue

[31,33,35,49,54]. Older children were reported to

exhibit lethargy, excessive ‘‘daydreaming,’’ rebellious

behavior in school, ‘‘phasing out,’’ ‘‘lapses’’ in aware-

ness, or being unresponsive to questions [31,54].

Numerous more recent studies of symptoms [6,24,

25,55–62] and objective measures [20,21,24,56,63]

have confirmed and expanded early observations that

SDB in children is associated with behavioral symp-

toms or impaired cognitive or school performance (eg,

the 1998 study by Gozal [21]). In children with classic

OSAS, there is clearly an important association be-

tween SDB, poor school performance, and other man-

ifestations of impaired daytime cognitive function.

The larger question is whether similar daytime

neurocognitive impairment occurs in children with

SDB who do not meet the criteria for classic OSAS.

Twenty years ago, Guilleminault et al reported on 25

children with heavy snoring and daytime symptoms,

including abnormal behavior and excessive daytime

sleepiness, but without OSA or oxygen desaturation

on polysomnography [31]. In every case, tonsillec-

tomy or adenoidectomy resulted in improvement or

complete disappearance of daytime symptoms [31].

Other studies also suggested that snoring children who

did not fit criteria for classic OSAS, may have clini-

cally significant daytime dysfunction (UARS) [27,36].

Morning headaches have been reported by several

authors to be a symptom of childhood OSAS, al-

J.L. Carroll / Clin Chest Med 24 (2003) 261–282264

though one of the few controlled studies of childhood

OSAS did not confirm this association [49]. Compar-

ison of children who underwent adenoidectomy with

‘‘normal’’ controls also revealed no difference in the

incidence of morning headaches [64]. Daytime mouth

breathing is a common finding in children with

adenotonsillar hypertrophy and is a common finding

in OSAS.

Because children may have significant daytime

symptoms (eg, neurocognitive impairment, behavioral

abnormalities, poor school performance, poor growth)

even if the polysomnography does not indicate OSAS,

is it unfortunate that the 2002 Academy of Pediatrics

(AAP) Clinical Practice Guideline for Diagnosis and

Management of Childhood Obstructive Sleep Apnea

simply recommends ‘‘further clinical evaluation and

treatment as warranted’’ for such a child [65].

Although the prevalence of UARS in children is

unknown, data from a large study of children referred

to a pediatric sleep center suggested that UARSmay be

present in most children referred for snoring [36].

Presentation of childhood SDB as UARS seemed to

be much more common than classic OSAS [36].

Predisposing factors for childhood obstructive

sleep apnea syndrome

Obesity

Although obesity predisposes to sleep-related

upper airway obstruction in children, most children

with OSAS are not obese. A study of OSAS in obese

Singapore children estimated the prevalence to be

5.7% overall and 13.3% in morbidly obese children

(percent ideal body weight >180) [66]. Redline et al

found obesity to be a significant risk factor (odds

ratio, 4.59; 95% confidence interval 1.58 to 13.33)

for OSAS in children and adolescents [46], and

numerous studies have found that obese children

are overrepresented in groups of children referred

for suspected OSAS [42].

Other factors

Snoring increases during upper respiratory tract

infection in children. Various nasal, oropharyngeal,

laryngeal, and neurologic problems also may predis-

pose to sleep-related airway obstruction (Box 1). A

long list of syndromes and other medical conditions

are known to be associated with an increased inci-

dence of childhood OSAS. Major genetic syndromes

and disorders associated with SDB in children

include Down syndrome, Prader-Willi syndrome,

achondroplasia, Arnold-Chiari syndrome, and myelo-

meningocele. SDB is common in children with

cerebral palsy [67]. Any syndrome or disorder that

affects one or some combination of upper airway

structure, airway muscle tone, upper airway muscle

control, or sleep may predispose to OSAS in children.

Obstructive sleep apnea syndrome in children with

Down syndrome is especially noteworthy. Marcus et al

found a high incidence of OSAS in patients with Down

syndrome 2 weeks to 52 years of age, even in persons

in whom it was not clinically suspected [68]. Children

with Down syndrome also tend to have significant

sleep fragmentation that is only partly explained by

SDB [69]. Practitioners should have a low threshold

Box 1. Predisposing factors for childhoodobstructive sleep apnea syndrome

Nasal RhinitisNasal polypsAdenoid hypertrophyPharyngeal flap surgeryNasal stenosisChoanal atresia

Pharyngeal Tonsil enlargementMicrognathiaRetrognathiaLingual tonsil

enlargementCleft palate repairAirway narrowing

caused by obesityTissue infiltration (eg,mucopolysaccharidoses)

Laryngeal Laryngeal webSubglottic stenosisVocal cord paralysisLaryngomalaciaLaryngeal masses

and tumorsInflammation caused

by gastroesophagealreflux

Neurologic Cerebral palsyArnold-Chiari

malformationPharmacologic Sedation

AnesthesiaOther Allergy/atopy

Cigarette smokeexposure

Sleep deprivation

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 265

for performing a detailed sleep assessment in patients

with Down syndrome.

Complications of obstructive sleep apnea

syndrome in children

Growth

Growth impairment is a well-documented com-

plication of OSAS in children. OSAS may result in

failure to thrive. Not all children with growth impair-

ment caused by SDB are below the fifth percentile on

growth charts for height or weight, however. Some

children with SDB demonstrate increased growth

velocity after adenotonsillectomy, even if they were

not less than fifth percentile before surgery [22,23,70].

From a practical point of view, SDB should be con-

sidered in any child with questionable weight gain or

stature and snoring.

Cardiovascular complications

It has been known for five decades that severe

OSAS in children may lead to congestive heart failure

and cor pulmonale. In the 1950s, childhood OSAS

was diagnosed mainly by cardiologists and endocri-

nologists when children presented in heart failure or

severe growth impairment. The more relevant ques-

tion currently is whether milder forms of childhood

SDB are associated with cardiovascular morbidity.

Tal et al, using radionuclide ventriculography to

study children with OSAS, found significant reduc-

tions in right ventricular ejection fraction that were

reversible after adenotonsillectomy [71]. Children

with polysomnographic-proven OSAS have been

shown as a group to have higher diastolic blood

pressures compared with children with snoring but

without OSAS [72]. Amin et al recently reported

abnormal left ventricular geometry in approximately

40% of children with OSA and approximately 15% of

children with snoring alone [73]. Whether such

changes are a precursor for cardiovascular disease

in adults remains to be determined.

Mortality

The mortality rate for childhood SDB or OSAS is

unknown. Death during sleep caused by OSAS in

children is apparently rare, and most deaths are

believed to be perioperative after adenotonsillectomy.

Children with unrecognized OSAS and cardiovascu-

lar compromise may decompensate during general

anesthesia [74,75]. Death caused by OSAS may occur

after surgical correction of velopharyngeal incompe-

tence [76].

Pathophysiology of childhood obstructive sleep

apnea syndrome

Sleep-related airway obstruction

Given the wide variety of predisposing factors for

childhood SDB, no single pathophysiology accounts

for all cases (Fig. 2). Sleep-related upper airway

Fig. 2. Pathophysiology of childhood SDB. Any one factor alone (eg, adenotonsillar hypertrophy) may not be sufficient to cause

obstructive SDB. The same degree of hypertrophy may cause SDB when combined with predisposing factors such as abnormal

arousal mechanisms, decreased neural drive to upper airway muscles, or abnormal load compensation mechanisms, however.

J.L. Carroll / Clin Chest Med 24 (2003) 261–282266

obstruction in children is not simply a matter of big

tonsils and adenoids; it is dynamic airway collapse

related to muscle tone, motor control, and structure.

Children with SDB do not exhibit upper airway

obstruction while awake, which indicates that sleep-

related dynamic factors responsible for maintaining

airway patency are involved. The other clear evidence

in support of this view is that OSAS in children is

highly state related and occurs largely during REM

sleep [77]. A child with severe REM-related OSAS

may have minimal upper airway obstruction during

non-REM sleep, which indicates that state-related

upper airway control plays a major role.

Numerous studies have failed to find a simple

relationship between adenotonsillar size (or volume)

and the occurrence of OSAS in children. This has led

to the speculation that children who develop OSAS

must have an underlying abnormality of upper airway

structure, muscle tone, or upper airway reflex muscle

control [13,78,79]. In otherwise normal children with

OSAS the current view is that adenotonsillar hyper-

trophy causes airway narrowing that, when super-

imposed on subtle abnormalities of upper airway

motor control or tone (neural drive), leads to clin-

ically significant dynamic airway obstruction during

sleep (see Fig. 2) [79].

Daytime symptoms and complications

The pathologic mechanisms underlying daytime

symptoms of childhood OSAS are unknown,

although intermittent hypoxia and sleep fragmenta-

tion likely play a role [19]. Snoring children without

OSAS may have debilitating daytime symptoms.

Conversely, children with severe OSAS and severe

hypoxemia during sleep may have minimal daytime

symptoms. Proposed mechanisms have included

sleep disruption or fragmentation, nighttime hypox-

emia, hypoxia, or sleep fragmentation – induced

alterations in brain neurochemistry, inflammation,

hormonal changes caused by sleep fragmentation or

deprivation, and changes in cerebral blood flow

caused by blood gas changes or altered cerebral

perfusion pressure [15,19]. In reality, how childhood

SDB leads to complex behavioral and neurocognitive

derangements remains unknown. Additional research

in this area is critically important to determine appro-

priate thresholds for the treatment of various aspects

of childhood SDB.

Complications of SDB, such as cardiovascular

compromise, hypertension, and growth failure or

impairment, are likely caused partly by known effects

of intermittent hypoxia during sleep [73,80]. Large

swings in intrathoracic pressure may affect cardiac

afterload directly, hypoxemia or sleep fragmentation

may affect brain neurochemistry in cardiovascular

control areas, and growth hormone secretion may be

affected by sleep fragmentation. Potential mecha-

nisms by which intermittent hypoxia may lead to such

derangements recently were reviewed in detail [81].

Relationship between childhood sleep-disordered

breathing and attention deficit disorders

Children with attention deficit hyperactivity dis-

order (ADHD) have difficulty sustaining attention,

attending to details, finishing tasks, listening to

others, and organizing behaviors. These children are

easily distracted, forgetful, and impulsive and have

difficulty sitting still. Reported symptoms of child-

hood SDB include hyperactivity, inattention, im-

pulsive behavior, and oppositional behavior. It is

reasonable to assume that SDB in some children

may exacerbate ADHD or that some children with

hyperactivity caused by SDB may be misdiagnosed as

having ADHD. The possible relationship is strength-

ened by the observation that children with ADHD

have high rates of sleep complaints and disturbances.

The medications used to treat ADHD also can inter-

fere with sleep, and the behavior problems manifested

by these children may interfere with sleep hygiene.

There is evidence that children diagnosed with

ADHD have increased rates of snoring or sleep

disturbances, such as periodic limb movement dis-

order [55–57,82–84]. Although the precise relation-

ship between SDB and ADHD is unknown, because

of the symptom overlap, snoring children with a

diagnosis of ADHD are commonly evaluated for the

possibility that SDB is causing or exacerbating their

behavioral symptoms.

Polysomnographic findings in childhood

obstructive sleep-disordered breathing

Procedure and limitations

Polysomnography originally was developed for

adults and later adapted for use in children. As a

diagnostic test for childhood SDB, polysomnography

has numerous shortcomings. Polysomnography

focuses heavily on breathing during sleep, with only

a few crude measures of sleep quality. More impor-

tantly, no studies have documented the relationship

between anything measured by polysomnography

and daytime sleepiness, impaired neurocognitive

function, behavioral abnormalities, or other adverse

outcomes related to SDB in children. Finally, no

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 267

studies have validated whether polysomnography has

any ability to predict which children need (or do

not need) treatment to avoid adverse clinical conse-

quences. One of the few studies that examined this

question reported that conventional polysomno-

graphic scoring criteria failed to identify children

with significant sleep-related upper airway obstruc-

tion [47]. Classic scoring and interpretation of poly-

somnography for children does not identify children

with UARS.

Polysomnography, although considered by many

to be the ‘‘gold standard,’’ is one of several poorly

validated tests for childhood SDB. The ability of

polysomnography to identify children at risk (or

not) for significant adverse clinical outcomes is

unknown. In reality, most children who undergo

adenotonsillectomy for apparent symptoms of child-

hood SDB do not receive any diagnostic studies. For

the small fraction of children who are tested before

adenotonsillectomy for ‘‘obstructive symptoms,’’

practitioners in pediatric sleep laboratories worldwide

use different clinical and laboratory testing proce-

dures and diagnostic criteria.

Guidelines for polysomnography in children were

published by the American Thoracic Society based

on a consensus of opinion in the early 1990s [1].

Because of a lack of data on polysomnography in

children at that time, however, the guidelines are not

definitive. After an exhaustive review of the litera-

ture, the 2002 AAP Technical Report on the Diagnosis

and Management of Childhood Obstructive Sleep

Apnea Syndrome concluded that the ‘‘gold standard’’

is poorly validated and that normative standards for

polysomnography in children have not been shown to

have any validity as predictors of the occurrence

complications [37]. Because the scope of childhood

SDB has been expanded beyond OSAS to include

UARS, the use of polysomnography as a ‘‘gold

standard’’ becomes even more dubious because of

its heavy focus on breathing and minimal measures of

sleep quality.

Most pediatric laboratories record standard elec-

troencephalogram leads for sleep staging, chest/abdo-

men motion by strain gauges or respiratory inductance

plethysmography, extraoccular muscle electromyo-

gram, submental and limb electromyogram, electro-

cardiogram, a measure of nasal/oral airflow (eg,

thermistor), pulse oximetry, and a method of detecting

CO2 retention (end-tidal or transcutaneous CO2).

Some laboratories include esophageal pressure mon-

itoring for detection of UARS [36]. With the

expanded scope of clinically significant SDB, much

more research is needed on methods for identification

and diagnosis of children with all forms of SDB,

particularly methods that do not meet criteria for

classic childhood OSAS.

What is normal?

The field of childhood obstructive SDB includes

the following major problems: (1) ‘‘Normal’’ has

never been defined with respect to childhood obstruc-

tive SDB. (2) Normative data are not available for

many polysomnographic measures. (3) Existing nor-

mal values are limited to classic OSAS. During the last

two decades, when most studies on childhood SDB

were conducted, snoring was believed to be ‘‘benign’’

and the behavioral symptoms of SDB were unrecog-

nized. Although normal polysomnographic values for

children and adolescents were published in a landmark

paper byMarcus et al [26], that study included snoring

children and possibly included children with UARS.

Currently, polysomnographic diagnostic criteria for

childhood UARS have not been developed, and nor-

mal polysomnographic values for asymptomatic, non-

snoring children are lacking. In the discussion that

follows, where possible, normative polysomnographic

values were extracted from the asymptomatic control

groups of several studies.

Sleep

Pediatric sleep laboratories analyze polysomno-

graphic data to derive arousal index (arousals/hour of

sleep time), sleep efficiency (time asleep/time in bed),

number of awakenings per hour, and time spent in

stages 1, 2, 3, 4 non-REM sleep and REM sleep. No

data exist on the relationship between sleep architec-

ture variables and daytime symptoms or other adverse

outcomes of childhood SDB, however. The positive

and negative predictive values of polysomnographic

sleep data are simply unknown.

Arousal index was reported by Goh et al to be

5/hour F 2/hour (meanF SD) in ten nonsnoring,

prepubertal children [77]. Guilleminault et al, in 36

asymptomatic prepubertal children with no evidence

of SDB, reported an electroencephalogram arousal

index of 2.7/hourF 1.9/hour [85]. These data suggest

that an arousal index of ten or more arousals/hour is

clearly outside of the normal range for asymptomatic

children (excluding infants). Goh el al found that

sleep efficiency was 84%F 13% (meanF SD) in

nonsnoring control children and that sleep architec-

ture, with respect to sleep stages and sleep efficiency,

was the same in nonsnoring controls versus children

with polysomnographically proven OSAS [77]. Mean

arousal index reported for children with classic OSA

was 11/hourF 4/hour in the study of Goh et al but

J.L. Carroll / Clin Chest Med 24 (2003) 261–282268

ranged from 2.5/hour up to 17/hour in other studies

[27,30,77,86]. Although it seems that normal children

should have an arousal index of less than ten

arousals/hour, some children with classic OSAS have

an arousal index within the normal range. Because

excessive daytime sleepiness and various other

daytime symptoms are known to occur in children

with UARS and OSAS, either current polysomno-

graphic techniques fail to detect significant sleep

disruption in children or other mechanisms underlie

these daytime symptoms.

Breathing pattern

Respiratory rate is usually normal in children with

OSAS unless they have lung disease or breathing

control abnormalities. Children with classic OSAS

exhibit obstructive apnea and hypopnea, defined

essentially as they are for adults except for duration.

In children, artificial time limits (eg, 10 seconds) are

usually not placed on obstructive apnea or hypopnea.

A standard approach is to consider significant any

obstructive episode that lasts longer than two respi-

ratory cycle times [87]. Young children may exhibit

oxygen desaturation with apnea as brief as 3 to

5 seconds. Fig. 3 shows classic OSA in a 4-year-

old boy that consisted of no airflow for 20 seconds

while paradoxical respiratory efforts continued. This

event was accompanied by a fall in oxygen saturation

from 96% to less than 80%. Fig. 4 shows a 2-minute

sample from the same child’s polysomnograph,

which indicates repetitive obstructive apnea associ-

ated with a ‘‘saw tooth’’ pattern of oxygen desatura-

tion and five arousals in 2 minutes. Children with

OSAS may exhibit obvious patterns of obstruction

similar to that observed in adults, and in such cases,

the diagnosis is not difficult.

One of the most remarkable findings in childhood

OSAS is the clustering of events in REM sleep [77]. It

is common in childhood OSAS to find most obstruc-

tive apnea, hypopnea, hypoxemia, or hypercarbia

occurring during REM sleep. Fig. 5 shows a 1-minute

sample from an 8-year-old child with obstructive

hypoventilation. This child had a normal apnea-

hypopnea index, with only two obstructive apneas

the entire night, yet during each REM period he

exhibited severe obstructive hypoventilation without

any complete obstructive apnea. As shown in the

figure, end-tidal CO2 exceeded 76 mmHg and oxygen

desaturation was moderately severe despite continued

Fig. 3. Typical obstructive apnea in a 4-year-old boy. (A) Absence of flow in end-tidal CO2 tracing. (B) Paradoxical inward rib

cage motion during period of airway obstruction. (C) Oxygen desaturation from 96% at the beginning to approximately 75% by

the end of the obstructive apnea. (D) Arousal from sleep at end of obstructive apnea.

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 269

Fig. 5. Continuous obstructive hypoventilation. (A) No interruption of oronasal airflow. (B) Continuous paradoxical inward rib

cage motion. (C) End-tidal CO2 between 72 and 76 mm Hg. (D) Oxygen saturation 84% to 88% without recovery to normal.

This child, despite severe hypoventilation and persistent hypoxemia, simply would seem to be snoring if observed by parents

during sleep.

Fig. 4. Repetitive obstructive apnea in a 4-year-old boy. (A) Obstructive apnea with absent airflow. (B) Paradoxical inward

rib cage motion during airway obstruction. (C) In phase rib cage and abdomen motion during nonobstructed breathing.

(D) Oxygen desaturation from 98% at the beginning to less than 75%. (E) Arousal from sleep at end of obstructive apnea.

There were five arousals during the 2-minute period. Note ‘‘sawtooth’’ pattern of severe oxygen desaturation after each ob-

structive apnea.

J.L. Carroll / Clin Chest Med 24 (2003) 261–282270

airflow with each breath. It is generally accepted that

some measure of CO2 is necessary to detect obstruc-

tive hypoventilation in children, and many pediatric

sleep laboratories measure end-tidal CO2.

Sleep-related upper airway obstruction may be

worse during the second half of the night [77]. A

significant proportion of children with classic OSAS

exhibit most of their obstructive episodes during the

second half of the polysomnograph. This has several

important implications. First, brief studies (eg, nap

studies) are relatively insensitive for detecting OSAS

in children [88]. Second, unless parents stay up or

wake up to observe their child sleeping during the

second half of the night, they may be unaware of the

severity of their child’s SDB. Finally, any attempt to

perform ‘‘split night’’ studies (eg, continuous positive

airway pressure [CPAP] titration during the second

half of the polysomnograph) results in a high prob-

ability of missing the child’s worst SDB.

By definition, children with UARS do not exhibit

hypoxemia, hypercapnia, or obstructive apnea during

sleep, although they may snore and may seem to have

increased respiratory effort or disrupted sleep. The

diagnosis of childhood UARS is controversial cur-

rently, and there are no diagnostic standards. The most

comprehensive study to date indicated that children

with UARS exhibit several patterns of sleep-related

increased respiratory effort that are best detected using

esophageal pressure monitoring [13,36,78]. Children

may not tolerate esophageal pressure monitoring, and

more research is needed to determine the best methods

and criteria for diagnosis of UARS in children.

Hypoxemia

Oxygen saturation measured by pulse oximetry

during sleep generally remains approximately 95% or

more in children [26,77], and numerous studies have

reported oxygen desaturation in children with classic

OSAS. Children with OSAS or obstructive hypoven-

tilation may experience episodic or continuous hypox-

emia during sleep that can range from minimal to

severe. Most pediatric sleep laboratories record oxy-

gen saturation continuously all night using pulse

oximetry (SpO2) and report nadir SpO2, respiratory

events associated with oxygen desaturation more than

4%, and percent of total sleep time spent with SpO2

less than 90%, 92%, or other threshold values. It is

generally assumed that hypoxemia is bad for children

and oxygen saturations less than 90% or 92% are

considered harmful. The positive or negative predic-

tive values of polysomnographic oximetry data in

children are unknown, however. Oxygen saturation

during sleep is normal in children with UARS.

Hypercapnia

Normal children maintain end-tidal CO2 less than

53 mm Hg and do not spend more than 10% of total

sleep time with an end-tidal CO2 more than 50 mm

Hg according Marcus et al [26,77]. As seen in Fig. 5,

some children exhibit obstructive hypoventilation

without apnea. Other children present a mixed picture

of obstructive apnea, hypopnea, and elevated end-

tidal CO2. As with hypoxemia, numerous studies

have reported hypercapnia in children referred for

suspected OSAS [30,35,47]. These groups are highly

selected, and the actual prevalence of hypercapnia in

children with obstructive SDB is unknown. Obese

children and children with genetic abnormalities may

be more likely to exhibit sleep-related hypercapnia.

Silvestri found that three quarters of obese children

with OSAS were hypercapneic during sleep [89].

These authors reported that OSAS with hypercapnia

was significantly more likely if weight was 200% or

more than ideal body weight.

Scoring polysomnography for childhood obstructive

sleep apnea syndrome

There are no widely accepted standardized guide-

lines or diagnostic criteria for classic OSAS in

children. The 2002 AAP Clinical Practice Guideline

for the Diagnosis and Management of Childhood Ob-

structive Sleep Apnea Syndrome [37,65] acknowledg-

ed that polysomnography remains unvalidated. The

AAP technical report on childhood SDB [37] states:

‘‘It is assumed that PS is a benign condition and

OSAS is associated with undesirable complications.

Normative standards for their polysomnographic

determination have been chosen on the basis of

statistical distribution of data, but it has not been

established that those standards have any validity as

predictors of the occurrence of complications.’’

The AAP Clinical Practice Guideline [65] sum-

marizes their findings by stating that in children:

‘‘Although we know which polysomnographic

parameters are statistically abnormal, studies have

not definitively evaluated which polysomnographic

criteria predict morbidity.’’

Pediatric sleep laboratories choose threshold val-

ues, usually based on the ATS standards for cardio-

respiratory sleep studies in children [1], that they

consider to be diagnostic or strongly suggestive of

significant childhood SDB (typical values shown in

Box 2). Polysomnography also yields data on the se-

verity of the sleep-related airway obstruction, hypox-

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 271

emia, hypercapnia, and degree of sleep disruption.

Severity criteria have been shown to correlate with the

probability of postoperative complications [90–95]

and response to treatment [96].

Scoring polysomnography for childhood upper

airway resistance syndrome

Currently there are no polysomnographic criteria

or guidelines for diagnosing UARS in children. By

definition, children with UARS do not meet criteria

for classic OSAS (see Box 2). Guilleminault et al

recommend esophageal pressure monitoring during

polysomnography to diagnose several distinct patterns

of increased respiratory effort during sleep in children

with UARS [13,36,78]. There are no widely accepted,

validated standards for interpretation of esophageal

pressure measurements on polysomnography, how-

ever, and their diagnostic and predictive value (of

adverse outcome) has not yet been determined. Some

adult sleep laboratories measure respiratory effort-

related arousal index [97]. There are no evidence-

based guidelines for respiratory effort-related arousal

index use in children. Some pediatric sleep labora-

tories, based on experience, determine a threshold for

nocturnal awakenings (eg, <1/hour), arousal index

(eg, <10/hour), and sleep efficiency (eg, >80%) and

consider these ‘‘suggestive’’ of UARS when

exceeded. Finally, even if polysomnography is ‘‘nega-

tive’’ for classic OSAS, the interpreter often can get a

strong impression of increased upper airway resist-

ance from viewing the video/audio tape, reviewing the

technician’s comments, and reviewing the tracings.

Such values and impressions are not evidence-based,

however, and in reality, UARS in children remains an

individualized clinical diagnosis based on judgment

and experience.

Other diagnostic tools for childhood obstructive

sleep apnea syndrome detection

Various alternatives to polysomnography for diag-

nosis of classic OSAS in children have been pro-

posed, including simple video or audiotaping and

analysis of snoring patterns. Videotaping a child at

home can be useful if it shows obvious sleep-related

upper airway obstruction. Simple videotaping by

parents does not allow assessment of severity and

provides no data on degree of hypoxemia. Unless the

parents happen to film the child during REM sleep,

significant SDB easily can be missed. Sivan et al

studied the predictive value of video, taken by parents

of their sleeping child, for diagnosing childhood

OSAS [98]. The authors analyzed a 30-minute video-

tape for noisy breathing, movement, arousals, and

other signs of OSAS. The results of the videotape

analysis correlated with polysomnography diagnosis

of classic OSAS in 84% of cases. This study did not

address the important issue of UARS, however.

Simple audio recordings, although touted by some

as useful for detecting SDB in children, are probably

not useful and may be misleading. A recent study that

compared home audiotape analysis with polysomno-

graphy found that the sensitivity rate of audiotape for

diagnosis of OSAS was only 46% [99].

A more sophisticated video-based home sleep

study methodology was described by Brouillette

et al [100–103]. These authors developed a home

sleep study system that uses a simple cardiorespira-

tory montage (EKG, respiratory inductance plethys-

mography, SpO2) combined with videotaping. The

videotapes are analyzed using a computerized move-

ment detection system. This system’s ability to detect

OSAS in children with adenotonsillar hypertrophy

has been validated relative to polysomnography and

it has several advantages. The child can be studied in

his or her natural sleeping environment at home and

there are no leads or sensors on the face. The utility of

this system for diagnosis of UARS is unknown, and

the authors are careful to point out that this system is

not appropriate when detailed information on sleep

staging, ventilation, or respiratory muscle function is

required. Once fully validated with respect to daytime

symptoms of UARS and OSAS, it may prove to be an

alternative to polysomnography.

Other home study approaches, which range from

overnight oximetry to complex multichannel record-

ings, recently were reviewed by the AAP subcom-

mittee on OSAS [37]. Oximetry alone should be used

with caution, although it may provide useful screen-

ing information [90,104,105]. A ‘‘negative’’ over-

night oximetry study does not rule out significant

sleep disturbance, hypoventilation, or significant

increased upper airway resistance. All of these meth-

ods currently suffer from the same shortcomings as

full polysomnography; that is, they lack ability to

predict daytime symptoms, complications, and other

Box 2. Abnormal values on pediatricpolysomnography (example)

Obstructive apnea index (AI) >1/hApnea-hypopnea index >5/hPeak end-tidal CO2 >53 mm HgEnd-tidal CO2 >50 mm Hg for >10%

of total sleep timeMinimum SpO2<92%

J.L. Carroll / Clin Chest Med 24 (2003) 261–282272

adverse outcomes, and threshold levels of abnormal-

ity that merit treatment remain unknown.

In summary, there is no ‘‘gold standard’’ for the

diagnosis of childhood SDB (UARS and OSAS). As

Ali and Stradling recently observed, polysomnogra-

phy is not the ‘‘gold standard’’ methodology against

which other techniques must be compared, it is

simply the oldest [106]. Polysomnography can

identify statistically abnormal breathing that suggests

classic OSAS, but it is certainly not a ‘‘gold stan-

dard’’ for diagnosis of UARS in children. As new

approaches to the diagnosis of childhood SDB are

developed, critical evaluation and validation—par-

ticularly with respect to their ability to predict clinical

symptoms, adverse outcomes, and response to treat-

ment—will be essential.

Tests for daytime symptoms of

sleep-disordered breathing

None of the subjective scales commonly used for

adults with SDB (eg, Epworth Sleepiness Scale) has

been validated for children. Quality-of-life assess-

ment tools, such as the Child Behavior Checklist,

OSA-18, and CHQ-PF50, may be valid for detecting

signs of impaired health or improvement in symp-

toms after adenotonsillectomy [107–111]. The diag-

nostic value of such tools is unknown for children.

Similarly, the Maintenance of Wakefulness Test also

has not been validated for children with SDB. Cur-

rently, the only standardized test for daytime sleepi-

ness in children is the multiple sleep latency test.

Diagnosis of childhood obstructive

sleep-disordered breathing

Presenting symptoms and signs

Children with obstructive SDB may present with

any combination of snoring, noisy breathing during

sleep, restless sleep, daytime fatigue, excessive day-

time sleepiness, abnormal or difficult behavior,

impaired school performance, attention problems,

developmental delay, and impaired growth. Children

with UARS may not even snore. The diagnosis of

obstructive SDB in children often requires a high level

of suspicion and detailed clinical history. The symp-

toms of childhood SDB clearly overlap with numer-

ous other potential causes, and they usually cannot be

attributed to SDB on the basis of history alone.

Excessive daytime sleepiness, the hallmark of

OSAS in adults, occurs in only a small proportion of

children with obstructive SDB. Children do present

with excessive daytime sleepiness as the chief com-

plaint, however, and may turn out to have UARS,

OSAS, idiopathic hypersomnia, narcolepsy, poor sleep

hygiene, some combination of the above diagnoses, or

various other causes of excessive daytime sleepiness.

History and physical examination

The value of clinical history for diagnosing classic

childhood OSAS has been questioned by numerous

studies [30,96,112–115], all of which were per-

formed before childhood UARS became widely

acknowledged. These studies, including one from

the author’s laboratory [30], examined the ability of

limited clinical history (focused mostly on breathing

symptoms and excessive daytime sleepiness) to dis-

tinguish classic OSAS from snoring without OSAS,

whereas the question of UARS was not addressed. In

retrospect it is likely that the ‘‘primary snoring’’

groups in such studies included children with UARS.

The relevant question is whether clinical history (or a

clinical ’’score’’ based on history/examination) has

predictive validity with respect to symptoms or com-

plications of childhood obstructive SDB (including

UARS). The answer to this question remains

unknown, and research in this area is critically

important for the field to advance toward a definitive

diagnostic approach.

Despite the limitations and controversy, the

evaluation for suspected SDB should begin with a

detailed history of the child’s sleep, breathing during

sleep, and daytime symptoms. Sleep history should

start by defining where the child sleeps in relation to

the caregiver being interviewed and the degree to

which the caregiver is aware of the child’s sleep

problems. This is not trivial. Parents may be unaware

of the child’s nightly sleep/breathing patterns (eg, the

child lives with grandmother, is brought by the

mother but lives five nights/week with the father)

or daytime symptoms (eg, at school). The same

parent will answer ‘‘no’’ to ‘‘does your child snore’’

and similar questions rather than reveal that he or she

simply does not know. Taking a detailed sleep/

breathing history of a child from adult caregivers is

fraught with pitfalls for the unwary. Suggested points

to cover in the history are outlined in Table 1.

Research is badly needed to develop validated,

age-specific, standardized questionnaire tools ca-

pable of identifying neurobehavioral abnormalities

and other symptoms or sequelae in children with

obstructive SDB.

Physical examination is also important for assess-

ing airway structure and exacerbating factors (Table 2).

The possible significance of abnormal craniofacial

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 273

morphology and airway anatomy in childhood SDB

recently was reviewed [13,78]. Some children fall

asleep during the office visit and exhibit overt OSAS

or obviously abnormal upper airway resistance. Most

of the time, however, even children with severe OSAS

appear normal while awake. The visit is an opportu-

nity to observe the child for subtle signs of excessive

daytime sleepiness, developmental impairment, or

behavior problems.

Approaches to diagnosis of childhood obstructive

sleep-disordered breathing

Despite the confusion, lack of data, and lack of

validated diagnostic approaches, physicians still must

evaluate snoring children and recommend treatment.

The 2002 AAP Clinical Practice Guideline for Dia-

gnosis and Management of Childhood Obstructive

Sleep Apnea proposes an algorithm for diagnosis and

management of uncomplicated childhood OSAS [65].

The reader is referred to the AAP guidelines [65] and

the accompanying technical report [37] for extensive

review of evaluation options for snoring children.

The main highlights of those guidelines for diagnosis

are as follows: (1) all children should be screened for

snoring; (2) complex patients should be referred to a

specialist; (3) patients with cardiorespiratory failure

cannot await elective evaluation; (4) diagnostic evalu-

ation is useful to distinguish ‘‘primary snoring’’ from

OSAS; and (5) polysomnography is the ‘‘gold stan-

dard’’ [65].

Unfortunately, although the 2002 AAP technical

report [37] provided an outstanding review of the

research literature on childhood SDB, there are sig-

nificant gaps and limitations in the resulting AAP

Clinical Practice Guideline for Diagnosis and Man-

agement of Childhood Obstructive Sleep Apnea [65].

The guidelines acknowledge repeatedly that neuro-

Table 1

Clinical history of the child with snoring and suspected sleep-disordered breathing

Sleeping environment Usual sleeping location; does child sleep in bed? age of mattress, type of pillow(s),

age of pillows, pillow/mattress covers? Bed sharing, room sharing, bed location,

distracting factors in sleeping environment (eg, television, outside noises, lights),

smoke exposure, pets in home

Sleep history Usual bedtime, bedtime behavior (eg, resistance), usual sleep onset time, nighttime

awakenings, parasomnias (sleep talking, walking, nightmares), usual sleeping

position, unusual sleeping positions, movement during sleep, enuresis (primary or

secondary), usual time of awakening, problems with awakening in morning

Snoring/breathing history Age at onset of snoring, frequency (nightly, most nights of week, only with upper

respiratory infection), proportion of night spent snoring, quality (pitch, harshness,

loudness, whether it disturbs others), pauses in snoring, observed struggle to breathe

or increased breathing effort during sleep, observed paradoxical inward rib cage mo-

tion, neck position (eg, hyperextended), parental interventions to improve breathing

(eg, change head position, prop up on pillows, awaken child)

Daytime symptoms Excessive daytime sleepiness: daytime sleepiness, inappropriate naps (for age),

falling asleep in school, inappropriately early bedtime (for age)

Behavioral/functional: cranky, irritable, oppositional, inattentive, hyperactive, poor

school performance, morning headaches, difficulty awakening in morning

Neurocognitive: loss of developmental milestones, poor school performance, mem-

ory problems, ‘‘blank’’ periods during day, oppositional behavior

Other: daytime mouth breathing, nasal obstruction, constant runny nose, frequent

sore throats, poor eating (likely related to tonsil/adenoid hypertrophy), poor growth,

allergies, nasal congestion

Medications Current medications, with focus on medications that may affect nasal resistance,

upper airway tone, or sleep quality; also important for planning polysomnography

(eg, medications that interfere with sleep)

Past medical and surgical history Previous airway manipulation (eg, intubation in neonatal intensive care unit),

previous airway surgery (adenoidectomy, tonsillectomy, uvulopalatopharyngoplasty),

previous cleft lip and/or palate repair, previous nasal surgery, recent weight gain,

thyroid or other metabolic problems

Family history Snoring, OSAS, UARS, obesity, family members on CPAP

Review of systems Thorough review of systems to elucidate any possible exacerbating factors (eg,

smoke exposure) or complications (eg, signs of cor pulmonale, congestive heart fail-

ure, seizures)

J.L. Carroll / Clin Chest Med 24 (2003) 261–282274

cognitive impairment and behavior problems may be

a symptom or complication of childhood SDB, yet

they never explicitly acknowledge the existence of

UARS, nor do they provide any guidance in cases in

which the child is symptomatic but the ‘‘gold stand-

ard’’ polysomnography is ‘‘negative’’ for OSAS. The

current polysomnographic diagnostic criteria for

childhood OSAS are based on statistical norms (from

limited, small studies) and never have been shown to

have any diagnostic validity with respect to symptoms

or complications. Symptomatic snoring in children

who do not meet the current diagnostic criteria for

OSAS may be a common presentation of childhood

SDB [13,19,27,36,85]. Diagnosing classic OSAS

by polysomnography is easy when the polysomno-

gram is abnormal. When a symptomatic child’s poly-

somnogram does not demonstrate classic OSAS,

the AAP guidelines simply recommend ‘‘further clin-

ical evaluation and treatment as warranted,’’ which

leaves the practitioner without guidance for the most

difficult cases.

In reality, most children with snoring and daytime

symptoms of childhood SDB never see a sleep spe-

cialist; they are either referred to an otolaryngologist

or remain unidentified. For the snoring child with

daytime symptoms and enlarged tonsils or adenoids, it

has been argued that the diagnosis is likely UARS or

OSAS and polysomnography is not indicated [116].

The same author further suggested that polysomno-

graphy, as currently performed for children (without

esophageal pressure monitoring or detailed analysis of

sleep microarchitecture), does not detect UARS

anyway. The latter assertion has merit and should be

resolved with appropriate research studies. Adenoton-

sillectomy is a procedure with risk of morbidity and

mortality, however, and should not be undertaken

without the clearest diagnosis possible. The daytime

symptoms of childhood SDB, particularly neurobe-

havioral symptoms, are all nonspecific and possibly

the result of various causes. The major dilemma in

this field currently is that the so-called ‘‘gold stan-

dard’’ diagnostic test, traditional polysomnography,

fails to identify children with significant morbidity

caused by SDB (UARS).

Approaches to this diagnostic dilemma to date

have included measurement of esophageal pressure

during polysomnography to detect UARS [13,36],

attempts to detect increased ‘‘airway resistance’’

using nasal pressure measurements [117,118], use of

unvalidated locally derived criteria to diagnose

UARS, and surgical adenotonsillectomy without di-

agnostic testing. As acknowledged by the AAP

Clinical Practice Guideline for Diagnosis and Man-

agement of Childhood Obstructive Sleep Apnea [65],

there is currently a shortage of pediatric sleep labo-

ratories to perform polysomnography. For evaluation

of the snoring child, many if not most otolaryngolo-

gists only use polysomnography for children deemed

to be borderline or high risk for adenotonsillectomy

[18,116]. In the absence of guidelines, some pediatric

sleep laboratories make up diagnostic standards for

childhood UARS based on experience. For example,

in a symptomatic snoring child who does not meet

ATS criteria for classic OSAS, a diagnosis of UARS

may be made based on arousal index, number of awak-

enings, sleep efficiency, number of sleep stage shifts,

severity of snoring, technician observations of in-

creased respiratory effort, and ‘‘gut feeling’’ of the

interpreting physician. It is critically important to

develop unambiguous definitions, effective diagnostic

tools (including quality-of-life assessment, symptom

questionnaires, clinical scores), and validated guide-

lines for diagnosis of UARS in children.

Management of obstructive sleep-disordered

breathing in children

Despite the lack of diagnostic criteria for child-

hood SDB, practitioners must make difficult manage-

ment decisions. Which snoring child needs treatment

and which treatments are indicated? How should

polysomnographic data be used to guide treatment

Table 2

Physical examination of the child with snoring and suspected

sleep-disordered breathing

Examination Focus of examination

Vital signs Include height, weight, growth curve,

blood pressure

Body habitus Obesity, neck anatomy (eg, short neck)

Ear, nose, throat Emphasis on oropharyngeal size, tonsil

size (0–4+), adenoid enlargement,

nasal patency, evidence for chronic

nasal congestion, neck masses,

thyroid examination

Craniofacial Facial shape/features (eg, ‘‘adenoid’’

facies, long face), mid-face hypoplasia,

micrognathia, retrognathia, elongated

soft palate, small triangular chin,

steep mandibular plane, narrow

intermolar width

Cardiovascular Emphasis on signs of cor pulmonale

Other The remainder of the examination may

focus on features associated with SDB,

such as neuromuscular weakness,

spasticity, cerebral palsy, and other

associated conditions (eg, genetic)

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 275

of obstructive SDB in children? What follow-up

measures are appropriate for the child with obstruc-

tive SDB?

The 1996 ATS Consensus Statement on Standards

and Indications for Cardiopulmonary Sleep Studies in

Children outlined a few polysomnographic findings

that ‘‘should be considered abnormal’’ but stopped

short of treatment recommendations. The 1996 con-

sensus statement concludes with the following

‘‘research question’’ [1]:

‘‘Which PSG abnormalities (number of respiratory

events, cumulative hypercapnia, severity of desatu-

ration, and degree of sleep disruption) in infants and

children with OSAS correlate with morbidity?’’

The relationship of polysomnographic findings to

treatment was left as an open question because of a

lack of data in 1992, when the conference actually

took place. 10 years later, on the question of poly-

somnography interpretation, the 2002 AAP Clinical

Practice Guideline for Diagnosis and Management of

Childhood Obstructive Sleep Apnea [65] recom-

mends following the ATS Consensus Statement [1]

which did not provide guidelines for polysomno-

graphic interpretation with respect to treatment. In

other words, there are no accepted guidelines on how

to relate polysomnographic results to treatment deci-

sions for children with obstructive SDB.

Generally speaking, diagnostic criteria such as

those outlined in the previous box are used to deter-

mine that polysomnographic results fall outside of

the normal range. In reality, practitioners combine

‘‘abnormal polysomnographic results’’ with data

gleaned from history, physical examination, and other

testing (Table 3) and make a decision based on clinical

judgment. On choice of treatment, the AAP 2002

Clinical Practice Guideline recommends the follow-

ing: (1) Adenotonsillectomy is the first line of treat-

ment for most children, and CPAP is an option for

children who are not candidates for surgery or do not

respond to surgery. (2) High-risk patients should be

monitored as inpatients postoperatively. (3) Patients

should be reevaluated postoperatively to determine

whether additional treatment is required [65].

Medical treatment

Any child with obstructive SDB may show some

degree of improvement with nonsedating deconges-

tants or nasal steroid sprays. A recent study by

Brouillette et al showed significant improvement in

obstructive event indices but not resolution of OSAS

in children after treatment with nasal fluticasone

[119]. The significant improvement in upper airway

obstruction with fluticasone is promising, which

suggests that effective nasal steroid therapy may

suffice for some children with mild UARS or mild

OSAS. Further study of this approach is needed

before it can be recommended, however. Obese

children with OSAS will benefit from weight loss

and adenotonsillectomy [120].

Surgical treatment

For otherwise normal children with adenotonsillar

hypertrophy and OSAS or UARS, the current surgical

treatment of choice is tonsillectomy and adenoidec-

tomy [65,96,121–125]. It is important to note, how-

ever, that adenotonsillectomy does not resolve

obstructive SDB fully in all children, particularly

children with severe preoperative symptoms. Some

children with persistent SDB after adenotonsillectomy

may benefit from uvulopalatopharyngoplasty, lingual

tonsillectomy, maxillary or mandibular surgery, or

tracheostomy. Alternative surgical procedures for

Table 3

Laboratory evaluation of the child with suspected sleep-

disordered breathing

Type of test Test

To identify Lateral neck radiographs

predisposing Laryngoscopy/bronchoscopy

conditions Upper airway fluoroscopy

Sleeping MR cine-fluoroscopy

Cephalometric assessment

of radiographs

To identify daytimeNeuropsychological testing

symptoms or Multiple sleep latency testing

complications Actigraphy

Maintenance of wakefulness testing

Electrocardiogram

Echocardiogram

Hematocrit

To determine

diagnosis

Studies for screening or to provide

complementary information

Questionnaire or

history-based scores

Videotaping by parents

Audiotaping by parents

Overnight oximetry

Daytime nap polysomnography

Other: combinations of oximetry,

videotaping, other channels

Diagnostic studies

Conventional fully polysomnography

Cardiorespiratory video system

(see text)

Some multichannel home

study methodologies

J.L. Carroll / Clin Chest Med 24 (2003) 261–282276

obstructive SDB in children were reviewed recently

[18,126].

Children with SDB and genetic craniofacial

anomalies, cerebral palsy, very young age, lung dis-

ease, and other medical conditions present special

problems with respect to treatment. When the tonsils

and adenoids are enlarged, simple adenotonsillectomy

or other procedures can be beneficial or even curative

of SDB in a substantial proportion of these complex

patients without resorting to long-term tracheostomy

[123,127–132].

Mechanical treatment

Obstructive SDB in children is not always cor-

rectable with medical or surgical treatment. In such

cases, CPAP or bi-level positive airway pressure

(BiPAP) may be indicated and can be used success-

fully by children of all ages, including infants [133].

CPAP provides positive pressure to the lumen of the

airway, which supports soft tissues and decreases

airway collapsibility. In most children, CPAP by

nasal mask is tolerated and effective [134–136]. It

is important that the initial approach to the family and

child be performed correctly and successfully by

practitioners experienced in techniques of desen-

sitization, parent training, and modeling [135]. CPAP

therapy should be titrated during polysomnography to

determine effective pressures, and children on CPAP

therapy should be followed regularly to ensure com-

pliance and proper fit of masks, headbands, straps,

and other equipment as the child grows.

Bi-level positive airway pressure is more comfort-

able to use, especially with higher pressures, and

children may tolerate it better. BiPAP also allows

higher inspiratory pressures to be used, allows setting

of a backup rate, and provides some ventilatory

assistance. BiPAP therapy is particularly appropriate

for the child who will not use CPAP and patients with

sleep-related hypoventilation caused by muscle

weakness, neurologic disease, or obesity. One poten-

tial complication of long-term nasal mask CPAP or

BiPAP is mid-face hypoplasia. Li et al recently

reported the case of a 15-year-old boy who received

face-mask CPAP for 10 years and developed severe

mid-face hypoplasia [137]. In children on long-term

nasal mask CPAP or BiPAP, maxillomandibular

growth should be monitored carefully.

Supplemental oxygen

There are no widely accepted guidelines or stan-

dards for the use of supplemental oxygen in children

with obstructive SDB. Supplemental oxygen may be

used as a temporary treatment for children with SDB

who are awaiting surgery or may be used postoper-

atively if sleep-related hypoxemia persists after thor-

ough evaluation and treatment. Oxygen also can be

used in combination with CPAP and BiPAP when

needed in children with nonobstructive causes of

hypoxemia (eg, lung disease). Caution is strongly

advised when starting supplemental oxygen in patients

with SDB. Although most patients tolerate supple-

mental oxygen well [138,139], some children develop

hypoventilation, and a small subgroup of children are

at risk for developing frank respiratory failure when

placed on supplemental oxygen [140]. Experience

suggests that the children at highest risk for hypoven-

tilation with supplemental oxygen tend to be children

with the most severe, long-standing SDB. The safest

approach is to start oxygen therapy during polysomno-

graphy (or at least while monitoring PCO2).

Follow-up

Some children continue to have upper airway

obstruction, increased upper airway resistance, hyper-

capnia, hypoxemia, and daytime symptoms after

surgery [141,142]. Follow-up is critically important

when the SDB is moderate to severe [96,143] or

when the risk of surgical treatment failure is high (eg,

Down syndrome, cerebral palsy, severe obesity).

Even when therapy of SDB is successful, the original

presenting symptoms may not resolve. Children with

excessive daytime sleepiness may have narcolepsy,

idiopathic hypersomnia, or other sleep disorders.

Follow-up is important, regardless of the treatment

used for childhood SDB.

Summary

Although it may seem that confusion and uncer-

tainty reign in the field of pediatric sleep medicine,

the recent realizations that the scope of childhood

SDB is wider, the symptomatology is broader, and

the prevalence is higher than previously believed are

major advances. Likewise, recent acknowledgment of

the lack of true ‘‘gold standards’’ for diagnosing

UARS and OSAS in children is also a major advance-

ment in this field. Critical assessment of the current

‘‘state of the art’’ by the 2002 AAP Technical Report

on the Diagnosis and Management of Childhood

Obstructive Sleep Apnea Syndrome [37] is another

major advance that sets the stage for the next steps.

The field needs an evidence-based definitions con-

ference, standardization of definitions across all

research studies, and much more research on clinical

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 277

features, pathophysiology, diagnosis, and treatment of

the ‘‘new’’ obstructive SDB, including the full range

of morbidity caused by increased upper airway re-

sistance. This should include further inquiry into the

origins of adult morbidity that resulted from child-

hood SDB and how it can be prevented.

References

[1] American Thoracic Society. Standards and indications

for cardiopulmonary sleep studies in children. Am J

Respir Crit Care Med 1996;153:866–78.

[2] CorboGM, Forastiere F, Agabiti N, Pistelli R, Dell’Or-

co V, Perucci CA, et al. Snoring in 9- to 15-year-old

children: risk factors and clinical relevance. Pediatrics

2001;108:1149–54.

[3] Corbo GM, Fuciarelli F, Foresi A, De Benedetto F.

Snoring in children: association with respiratory

symptoms and passive smoking. BMJ 1989;299:

1491–4.

[4] Teculescu DB, Caillier I, Perrin P, Rebstock E, Rauch

A. Snoring in French preschool children. Pediatr Pul-

monol 1992;13:239–44.

[5] Owen GO, Canter RJ, Robinson A. Snoring, apnoea

and ENT symptoms in the paediatric community. Clin

Otolaryngol 1996;21:130–4.

[6] Ali NJ, Pitson DJ, Stradling JR. Snoring, sleep dis-

turbance, and behaviour in 4–5 year olds. Arch Dis

Child 1993;68:360–6.

[7] Smedje H, Broman JE, Hetta J. Parents’ reports of

disturbed sleep in 5–7-year-old Swedish children.

Acta Paediatr 1999;88:858–65.

[8] Ferreira AM, Clemente V, Gozal D, Gomes A, Pis-

sarra C, Cesar H, et al. Snoring in Portuguese primary

school children. Pediatrics 2000;106:E64.

[9] Ipsiroglu OS, Fatemi A, Werner I, Tiefenthaler M,

Urschitz MS, Schwarz B. Prevalence of sleep disor-

ders in school children between 11 and 15 years of

age. Wien Klin Wochenschr 2001;113:235–44.

[10] Anuntaseree W, Rookkapan K, Kuasirikul S, Thong-

suksai P. Snoring and obstructive sleep apnea in Thai

school-age children: prevalence and predisposing fac-

tors. Pediatr Pulmonol 2001;32:222–7.

[11] American Thoracic Society. Cardiorespiratory sleep

studies in children: establishment of normative data

and polysomnographic predictors of morbidity. Am J

Respir Crit Care Med 1999;160:1381–7.

[12] Carroll JL. Sleep-related upper-airway obstruction in

children and adolescents. Child Adolesc Psychiatr

Clin N Am 1996;5:617–47.

[13] Guilleminault C, Khramtsov A. Upper airway resist-

ance syndrome in children: a clinical review. Semin

Pediatr Neurol 2001;8:207–15.

[14] Guilleminault C, Pelayo R. Sleep-disordered breath-

ing in children. Ann Med 1998;30:350–6.

[15] Gozal D.Morbidity of obstructive sleep apnea in child-

ren: facts and theory. Sleep Breath 2001;5:35–42.

[16] Marcus CL. Sleep-disordered breathing in children.

Am J Respir Crit Care Med 2001;164:16–30.

[17] Marcus CL. Sleep-disordered breathing in children.

Curr Opin Pediatr 2000;12:208–12.

[18] Bower CM, Gungor A. Pediatric obstructive sleep

apnea syndrome. Otolaryngol Clin North Am 2000;

33:49–75.

[19] O’Brien LM, Gozal D. Behavioural and neurocogni-

tive implications of snoring and obstructive sleep ap-

noea in children: facts and theory. Paediatr Respir Rev

2002;3:3–9.

[20] Gozal D, Pope Jr DW. Snoring during early childhood

and academic performance at ages thirteen to fourteen

years. Pediatrics 2001;107:1394–9.

[21] Gozal D. Sleep-disordered breathing and school per-

formance in children. Pediatrics 1998;102(3 Pt 1):

616–20.

[22] Williams III EF, Woo P, Miller R, Kellman RM. The

effects of adenotonsillectomy on growth in young

children. Otolaryngol Head Neck Surg 1991;104:

509–16.

[23] Stradling JR, Thomas G, Warley AR, Williams P,

Freeland A. Effect of adenotonsillectomy on noctur-

nal hypoxaemia, sleep disturbance, and symptoms in

snoring children. Lancet 1990;335:249–53.

[24] Ali NJ, Pitson D, Stradling JR. Sleep disordered

breathing: effects of adenotonsillectomy on behaviour

and psychological functioning. Eur J Pediatr 1996;

155:56–62.

[25] Ali NJ, Pitson D, Stradling JR. Natural history of

snoring and related behaviour problems between the

ages of 4 and 7 years. Arch Dis Child 1994;71:74–6.

[26] Marcus CL, Omlin KJ, Basinki DJ, Bailey SL, Rachal

AB, Von Pechmann WS, et al. Normal polysomno-

graphic values for children and adolescents. Am Rev

Respir Dis 1992;146(5 Pt 1):1235–9.

[27] Downey III R, Perkin RM, Macquarrie J. Upper air-

way resistance syndrome: sick, symptomatic but

underrecognized. Sleep 1993;16:620–3.

[28] Guilleminault C, Black JE, Palombini L, Ohayon M.

A clinical investigation of obstructive sleep apnea

syndrome (OSAS) and upper airway resistance syn-

drome (UARS) patients. Sleep Med 2000;1:51–6.

[29] Guilleminault C, Stoohs R, Clerk A, Cetel M,Maistros

P. A cause of excessive daytime sleepiness: the upper

airway resistance syndrome. Chest 1993;104:781–7.

[30] Carroll JL, McColley SA, Marcus CL, Curtis S,

Loughlin GM. Inability of clinical history to distin-

guish primary snoring from obstructive sleep apnea

syndrome in children. Chest 1995;108:610–8.

[31] Guilleminault C, Winkle R, Korobkin R, Simmons B.

Children and nocturnal snoring: evaluation of the ef-

fects of sleep related respiratory resistive load and

daytime functioning. Eur J Pediatr 1982;139:165–71.

[32] Weissbluth M, Davis AT, Poncher J, Reiff J. Signs of

airway obstruction during sleep and behavioral, devel-

opmental, and academic problems. J Dev Behav Pe-

diatr 1983;4:119–21.

[33] Guilleminault C, Eldridge FL, Simmons FB, Dement

J.L. Carroll / Clin Chest Med 24 (2003) 261–282278

WC. Sleep apnea in eight children. Pediatrics 1976;

58:23–30.

[34] Carskadon MA, Pueschel SM, Millman RP. Sleep-dis-

ordered breathing and behavior in three risk groups:

preliminary findings from parental reports. Childs

Nerv Syst 1993;9:452–7.

[35] Brouillette RT, Fernbach SK, Hunt CE. Obstructive

sleep apnea in infants and children. J Pediatr 1982;

100:31–40.

[36] Guilleminault C, Pelayo R, Leger D, Clerk A, Bocian

RC. Recognition of sleep-disordered breathing in

children. Pediatrics 1996;98:871–82.

[37] Schechter MS. Technical report: diagnosis and man-

agement of childhood obstructive sleep apnea syn-

drome. Pediatrics 2002;109:E69.

[38] Tang JP, Rosen CL, Larkin EK, DiFiore JM, Arnold

JL, Surovec SA, et al. Identification of sleep-disor-

dered breathing in children: variation with event def-

inition. Sleep 2002;25:72–9.

[39] American Academy of Sleep Medicine. International

classification of sleep disorders: diagnostic and coding

manual. Rochester (MN): AmericanAcademy of Sleep

Medicine; 2001.

[40] American Academy of Sleep Medicine Task Force.

Sleep-related breathing disorders in adults: recom-

mendations for syndrome definition and measure-

ment techniques in clinical research. Sleep 1999;22:

667–89.

[41] Richards W, Ferdman RM. Prolonged morbidity due

to delays in the diagnosis and treatment of obstructive

sleep apnea in children. Clin Pediatr (Phila) 2000;39:

103–8.

[42] Rosen CL. Clinical features of obstructive sleep ap-

nea hypoventilation syndrome in otherwise healthy

children. Pediatr Pulmonol 1999;27:403–9.

[43] Gislason T, Benediktsdottir B. Snoring, apneic epi-

sodes, and nocturnal hypoxemia among children

6 months to 6 years old: an epidemiologic study of

lower limit of prevalence. Chest 1995;107:963–6.

[44] Hultcrantz E, Lofstrand-Tidestrom B, Ahlquist-Rastad

J. The epidemiology of sleep related breathing disorder

in children. Int J Pediatr Otorhinolaryngol 1995;

32(Suppl):S63–6.

[45] Jeans WD, Fernando DC, Maw AR, Leighton BC. A

longitudinal study of the growth of the nasopharynx

and its contents in normal children. Br J Radiol 1981;

54:117–21.

[46] Redline S, Tishler PV, Schluchter M, Aylor J, Clark K,

Graham G. Risk factors for sleep-disordered breathing

in children: associations with obesity, race, and respi-

ratory problems. Am J Respir Crit Care Med 1999;159

(5 Pt 1):1527–32.

[47] Rosen CL, D’Andrea L, Haddad GG. Adult criteria

for obstructive sleep apnea do not identify children

with serious obstruction. Am Rev Respir Dis 1992;

146(5 Pt 1):1231–4.

[48] Weider DJ, Hauri PJ. Nocturnal enuresis in children

with upper airway obstruction. Int J Pediatr Otorhi-

nolaryngol 1985;9:173–82.

[49] Brouillette RT, Hanson D, David R, Klemka L, Szat-

kowski A, Fernbach S, et al. A diagnostic approach to

suspected obstructive sleep apnea in children. J Pediatr

1984;105:10–4.

[50] Gozal D, Wang M, Pope Jr DW. Objective sleepiness

measures in pediatric obstructive sleep apnea. Pedia-

trics 2001;108:693–7.

[51] Osler W. Chronic tonsillitis: principles and practice

of medicine. New York: Appleton and Co.; 1892.

p. 335–9.

[52] McKenzie M. A manual of diseases of the throat and

nose, including the pharynx, larynx, trachea, oesopha-

gus, nasal cavities and neck. London: Churchill; 1880.

[53] Hill W. On some causes of backwardness and stupidity

in children: and the relief of the symptoms in some

instances by nasopharyngeal scarifications. BMJ

1889;2:711–2.

[54] Guilleminault C, Korobkin R, Winkle R. A review

of 50 children with obstructive sleep apnea syndrome.

Lung 1981;159:275–87.

[55] Chervin RD, Dillon JE, Bassetti C, Ganoczy DA, Pi-

tuch KJ. Symptoms of sleep disorders, inattention, and

hyperactivity in children. Sleep 1997;20:1185–92.

[56] Chervin RD, Archbold KH. Hyperactivity and

polysomnographic findings in children evaluated

for sleep-disordered breathing. Sleep 2001;24:

313–20.

[57] Chervin RD, Archbold KH, Dillon JE, Panahi P, Pi-

tuch KJ, Dahl RE, et al. Inattention, hyperactivity,

and symptoms of sleep-disordered breathing. Pedia-

trics 2002;109:449–56.

[58] Hansen DE, Vandenberg B. Neuropsychological fea-

tures and differential diagnosis of sleep apnea syn-

drome in children. J Clin Child Psychol 1997;26:

304–10.

[59] Owens J, Opipari L, Nobile C, Spirito A. Sleep and

daytime behavior in children with obstructive sleep ap-

nea and behavioral sleep disorders. Pediatrics 1998;

102:1178–84.

[60] Kaplan BJ, McNicol J, Conte RA, Moghadam HK.

Sleep disturbance in preschool-aged hyperactive and

nonhyperactive children. Pediatrics 1987;80:839–44.

[61] Corkum P, Tannock R, Moldofsky H. Sleep distur-

bances in children with attention-deficit/hyperactivity

disorder. J Am Acad Child Adolesc Psychiatry 1998;

37:637–46.

[62] Corkum P, Tannock R, Moldofsky H, Hogg-Johnson

S, Humphries T. Actigraphy and parental ratings of

sleep in children with attention-deficit/hyperactivity

disorder (ADHD). Sleep 2001;24:303–12.

[63] Rhodes SK, Shimoda KC, Waid LR, O’Neil PM, Oex-

mann MJ, Collop NA, et al. Neurocognitive deficits in

morbidly obese children with obstructive sleep apnea.

J Pediatr 1995;127:741–4.

[64] Hibbert J. The occurrence of adenoidal signs and symp-

toms in normal children. Clin Otolaryngol 1981;6:

97–100.

[65] Section on Pediatric Pulmonology, Subcommittee on

Obstructive Sleep Apnea Syndrome. American Acad-

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 279

emy of Pediatrics. Clinical practice guideline: diag-

nosis and management of childhood obstructive sleep

apnea syndrome. Pediatrics 2002;109:704–12.

[66] Chay OM, Goh A, Abisheganaden J, Tang J, LimWH,

Chan YH, et al. Obstructive sleep apnea syndrome

in obese Singapore children. Pediatr Pulmonol 2000;

29:284–90.

[67] Kotagal S, Gibbons VP, Stith JA. Sleep abnormalities

in patients with severe cerebral palsy. Dev Med Child

Neurol 1994;36:304–11.

[68] Marcus CL, Keens TG, Bautista DB, Von Pechmann

WS, Ward SL. Obstructive sleep apnea in children

with Down syndrome. Pediatrics 1991;88:132–9.

[69] Levanon A, Tarasiuk A, Tal A. Sleep characteristics

in children with Down syndrome. J Pediatr 1999;134:

755–60.

[70] Marcus CL, Carroll JL, Koerner CB, Hamer A, Lutz J,

Loughlin GM. Determinants of growth in children

with the obstructive sleep apnea syndrome. J Pediatr

1994;125:556–62.

[71] Tal A, Leiberman A, Margulis G, Sofer S. Ventricular

dysfunction in children with obstructive sleep apnea:

radionuclide assessment. Pediatr Pulmonol 1988;4:

139–43.

[72] Marcus CL, Greene MG, Carroll JL. Blood pressure

in children with obstructive sleep apnea. Am J Respir

Crit Care Med 1998;157(4 Pt 1):1098–103.

[73] Amin RS, Kimball TR, Bean JA, Jeffries JL, Willging

JP, Cotton RT, et al. Left ventricular hypertrophy and

abnormal ventricular geometry in children and ado-

lescents with obstructive sleep apnea. Am J Respir

Crit Care Med 2002;165:1395–9.

[74] Sie KC, Perkins JA, Clarke WR. Acute right heart

failure due to adenotonsillar hypertrophy. Int J Pediatr

Otorhinolaryngol 1997;41:53–8.

[75] Warwick JP, Mason DG. Obstructive sleep apnoea

syndrome in children. Anaesthesia 1998;53:571–9.

[76] Kravath RE, Pollak CP, Borowiecki B, Weitzman ED.

Obstructive sleep apnea and death associated with sur-

gical correction of velopharyngeal incompetence. J Pe-

diatr 1980;96:645–8.

[77] Goh DY, Galster P, Marcus CL. Sleep architecture and

respiratory disturbances in children with obstructive

sleep apnea. Am J Respir Crit Care Med 2000;162

(2 Pt 1):682–6.

[78] Guilleminault C, Quo SD. Sleep-disordered breath-

ing: a view at the beginning of the new millennium.

Dent Clin North Am 2001;45:643–56.

[79] Marcus CL. Pathophysiology of childhood obstructive

sleep apnea: current concepts. Respir Physiol 2000;

119(2–3):143–54.

[80] Gozal D, Daniel JM, Dohanich GP. Behavioral and

anatomical correlates of chronic episodic hypoxia

during sleep in the rat. J Neurosci 2001;21:2442–50.

[81] Gozal E, Gozal D. Respiratory plasticity following

intermittent hypoxia: developmental interactions.

J Appl Physiol 2001;90:1995–9.

[82] Ball JD, Tiernan M, Janusz J, Furr A. Sleep patterns

among children with attention-deficit hyperactivity

disorder: a reexamination of parent perceptions. J Pe-

diatr Psychol 1997;22:389–98.

[83] Chervin RD, Archbold KH, Dillon JE, Pituch KJ,

Panahi P, Dahl RE, et al. Associations between symp-

toms of inattention, hyperactivity, restless legs, and

periodic leg movements. Sleep 2002;25:213–8.

[84] Picchietti DL, England SJ, Walters AS, Willis K, Ver-

rico T. Periodic limb movement disorder and restless

legs syndrome in children with attention-deficit hyper-

activity disorder. J Child Neurol 1998;13:588–94.

[85] Guilleminault C, Palombini L, Pelayo R, Chervin RD.

Sleepwalking and sleep terrors in prepubertal children:

what triggers them? Pediatrics 2003;111:E17–25.

[86] Stepanski E, Zayyad A, Nigro C, Lopata M, Basner R.

Sleep-disorderedbreathing in apredominantlyAfrican-

American pediatric population. J Sleep Res 1999;8:

65–70.

[87] Carroll JL, Loughlin GM. Diagnostic criteria for ob-

structive sleep apnea syndrome in children. Pediatr

Pulmonol 1992;14:71–4.

[88] Saeed MM, Keens TG, Stabile MW, Bolokowicz J,

Davidson Ward SL. Should children with suspected

obstructive sleep apnea syndrome and normal nap

sleep studies have overnight sleep studies? Chest

2000;118:360–5.

[89] Silvestri JM, Weese-Mayer DE, Bass MT, Kenny AS,

Hauptman SA, Pearsall SM. Polysomnography in

obese children with a history of sleep-associated

breathing disorders. Pediatr Pulmonol 1993;16:

124–9.

[90] Wilson K, Lakheeram I, Morielli A, Brouillette R,

Brown K. Can assessment for obstructive sleep apnea

help predict postadenotonsillectomy respiratory com-

plications? Anesthesiology 2002;96:313–22.

[91] Januszkiewicz JS, Cohen SR, Burstein FD, Simms C.

Age-related outcomes of sleep apnea surgery in in-

fants and children. Ann Plast Surg 1997;38:465–77.

[92] Riley RW, Powell NB, Guilleminault C, Pelayo R,

Troell RJ, Li KK. Obstructive sleep apnea surgery: risk

management and complications. Otolaryngol Head

Neck Surg 1997;117:648–52.

[93] Ruboyianes JM, Cruz RM. Pediatric adenotonsillec-

tomy for obstructive sleep apnea. Ear Nose Throat J

1996;75:430–3.

[94] Rosen GM, Muckle RP, Mahowald MW, Goding GS,

Ullevig C. Postoperative respiratory compromise in

children with obstructive sleep apnea syndrome: can

it be anticipated? Pediatrics 1994;93:784–8.

[95] McColley SA, April MM, Carroll JL, Naclerio RM,

Loughlin GM. Respiratory compromise after adeno-

tonsillectomy in children with obstructive sleep apnea.

Arch Otolaryngol Head Neck Surg 1992;118:940–3.

[96] Suen JS, Arnold JE, Brooks LJ. Adenotonsillectomy

for treatment of obstructive sleep apnea in children.

Arch Otolaryngol Head Neck Surg 1995;121:525–30.

[97] Loube DI, Andrada TF. Comparison of respiratory

polysomnographic parameters in matched cohorts of

upper airway resistance and obstructive sleep apnea

syndrome patients. Chest 1999;115:1519–24.

J.L. Carroll / Clin Chest Med 24 (2003) 261–282280

[98] Sivan Y, Kornecki A, Schonfeld T. Screening obstruc-

tive sleep apnoea syndrome by home videotape record-

ing in children. Eur Respir J 1996;9:2127–31.

[99] Lamm C, Mandeli J, Kattan M. Evaluation of home

audiotapes as an abbreviated test for obstructive sleep

apnea syndrome (OSAS) in children. Pediatr Pulmo-

nol 1999;27:267–72.

[100] Brouillette RT, Jacob SV, Waters KA, Morielli A,

Mograss M, Ducharme FM. Cardiorespiratory sleep

studies for children can often be performed in the

home. Sleep 1996;19(10 Suppl):S278–80.

[101] Morielli A, Ladan S, Ducharme FM, Brouillette RT.

Can sleep and wakefulness be distinguished in child-

ren by cardiorespiratory and videotape recordings?

Chest 1996;109:680–7.

[102] Jacob SV, Morielli A, Mograss MA, Ducharme FM,

SchlossMD, Brouillette RT. Home testing for pediatric

obstructive sleep apnea syndrome secondary to adeno-

tonsillar hypertrophy. Pediatr Pulmonol 1995;20:

241–52.

[103] Mograss MA, Ducharme FM, Brouillette RT. Move-

ment/arousals: description, classification, and relation-

ship to sleep apnea in children. Am J Respir Crit Care

Med 1994;150(6 Pt 1):1690–6.

[104] Nixon GM, Brouillette RT. Diagnostic techniques for

obstructive sleep apnoea: is polysomnography neces-

sary? Paediatr Respir Rev 2002;3:18–24.

[105] Brouillette RT, Lavergne J, Leimanis A, Nixon GM,

Ladan S, McGregor CD. Differences in pulse oximetry

technology can affect detection of sleep-disordered

breathing in children. Anesth Analg 2002;94

(1 Suppl):S47–53.

[106] Ali NJ, Stradling JR. Epidemiology and natural his-

tory of snoring and sleep-disordered breathing in

children. In: Loughlin GM, Carroll JL, Marcus CL,

editors. Sleep and breathing in children: a develop-

mental approach. New York: Marcel Dekker, Inc.;

2000. p. 555–74.

[107] Rosen CL, Palermo TM, Larkin EK, Redline S.

Health-related quality of life and sleep-disordered

breathing in children. Sleep 2002;25:657–66.

[108] Goldstein NA, Fatima M, Campbell TF, Rosenfeld

RM. Child behavior and quality of life before and

after tonsillectomy and adenoidectomy. Arch Otola-

ryngol Head Neck Surg 2002;128:770–5.

[109] Goldstein NA, Post JC, Rosenfeld RM, Campbell TF.

Impact of tonsillectomy and adenoidectomy on child

behavior. Arch Otolaryngol Head Neck Surg 2000;

126:494–8.

[110] Franco Jr RA, Rosenfeld RM, Rao M. First place,

resident clinical science award 1999: quality of life

for children with obstructive sleep apnea. Otolaryngol

Head Neck Surg 2000;123(1 Pt 1):9–16.

[111] De Serres LM, Derkay C, Astley S, Deyo RA, Rose-

nfeld RM, Gates GA. Measuring quality of life in

children with obstructive sleep disorders. Arch Oto-

laryngol Head Neck Surg 2000;126:1423–9.

[112] Nieminen P, Tolonen U, Lopponen H, Lopponen T,

Luotonen J, Jokinen K. Snoring children: factors pre-

dicting sleep apnea. Acta Otolaryngol Suppl 1997;529:

190–4.

[113] Leach J, Olson J, Hermann J, Manning S. Polysomno-

graphic and clinical findings in children with obstruc-

tive sleep apnea. Arch Otolaryngol Head Neck Surg

1992;118:741–4.

[114] Wang RC, Elkins TP, Keech D, Wauquier A, Hubbard

D. Accuracy of clinical evaluation in pediatric ob-

structive sleep apnea. Otolaryngol Head Neck Surg

1998;118:69–73.

[115] Goldstein NA, Sculerati N, Walsleben JA, Bhatia N,

Friedman DM, Rapoport DM. Clinical diagnosis of

pediatric obstructive sleep apnea validated by poly-

somnography. Otolaryngol Head Neck Surg 1994;

111:611–7.

[116] Messner AH. Evaluation of obstructive sleep apnea by

polysomnography prior to pediatric adenotonsillec-

tomy. Arch Otolaryngol Head Neck Surg 1999;125:

353–6.

[117] Serebrisky D, Cordero R, Mandeli J, Kattan M,

Lamm C. Assessment of inspiratory flow limitation

in children with sleep-disordered breathing by a nasal

cannula pressure transducer system. Pediatr Pulmonol

2002;33:380–7.

[118] Trang H, Leske V, Gaultier C. Use of nasal cannula

for detecting sleep apneas and hypopneas in infants

and children. Am J Respir Crit Care Med 2002; 166:

464–8.

[119] Brouillette RT, Manoukian JJ, Ducharme FM, Oud-

jhane K, Earle LG, Ladan S, et al. Efficacy of fluti-

casone nasal spray for pediatric obstructive sleep

apnea. J Pediatr 2001;138:838–44.

[120] Kudoh F, Sanai A. Effect of tonsillectomy and adenoi-

dectomy on obese children with sleep-associated

breathing disorders. Acta Otolaryngol Suppl 1996;

523:216–8.

[121] Nieminen P, Tolonen U, Lopponen H. Snoring and

obstructive sleep apnea in children: a 6-month follow-

up study. Arch Otolaryngol Head Neck Surg 2000;

126:481–6.

[122] Shintani T, Asakura K, Kataura A. The effect of ad-

enotonsillectomy in children with OSA. Int J Pediatr

Otorhinolaryngol 1998;44:51–8.

[123] Wiet GJ, Bower C, Seibert R, Griebel M. Surgical

correction of obstructive sleep apnea in the compli-

cated pediatric patient documented by polysomnogra-

phy. Int J Pediatr Otorhinolaryngol 1997;41:133–43.

[124] Nishimura T, Morishima N, Hasegawa S, Shibata N,

Iwanaga K, Yagisawa M. Effect of surgery on obstruc-

tive sleep apnea. Acta Otolaryngol Suppl 1996;523:

231–3.

[125] Nuyens MR, Vella S, Bassetti C, Caversaccio M,

Hausler R. Objective value of adenotonsillectomy in

the child: a prospective study of incidence of tonsil-

litis, snoring, pulse oximetry and polysomnography

and general development before and after adenoton-

sillectomy. Schweiz Rundsch Med Prax 1999;88:

893–9.

[126] Cohen SR, Holmes RE, Machado L, Magit A. Surgi-

J.L. Carroll / Clin Chest Med 24 (2003) 261–282 281

cal strategies in the treatment of complex obstructive

sleep apnoea in children. Paediatr Respir Rev 2002;

3:25–35.

[127] Bower CM, Richmond D. Tonsillectomy and adenoid-

ectomy in patients with Down syndrome. Int J Pediatr

Otorhinolaryngol 1995;33:141–8.

[128] Cohen SR, Simms C, Burstein FD, Thomsen J. Alter-

natives to tracheostomy in infants and children with

obstructive sleep apnea. J Pediatr Surg 1999;34:

182–6.

[129] Cohen SR, Lefaivre JF, Burstein FD, Simms C, Kattos

PH, Scott PH, et al. Surgical treatment of obstructive

sleep apnea in neurologically compromised patients.

Plast Reconstr Surg 1997;99:638–46.

[130] Magardino TM, Tom LW. Surgical management of

obstructive sleep apnea in children with cerebral palsy.

Laryngoscope 1999;109:1611–5.

[131] Kosko JR, Derkay CS. Uvulopalatopharyngoplasty:

treatment of obstructive sleep apnea in neurologically

impaired pediatric patients. Int J Pediatr Otorhinola-

ryngol 1995;32:241–6.

[132] Seid AB, Martin PJ, Pransky SM, Kearns DB. Surgi-

cal therapy of obstructive sleep apnea in children with

severe mental insufficiency. Laryngoscope 1990;100:

507–10.

[133] McNamara F, Sullivan CE. Obstructive sleep apnea in

infants and its management with nasal continuous

positive airway pressure. Chest 1999;116:10–6.

[134] Marcus CL,Ward SL,Mallory GB, Rosen CL, Becker-

man RC, Weese-Mayer DE, et al. Use of nasal contin-

uous positive airway pressure as treatment of

childhood obstructive sleep apnea. J Pediatr 1995;

127:88–94.

[135] Rains JC. Treatment of obstructive sleep apnea in

pediatric patients: behavioral intervention for compli-

ance with nasal continuous positive airway pressure.

Clin Pediatr (Phila) 1995;34:535–41.

[136] Waters KA, Everett FM, Bruderer JW, Sullivan CE.

Obstructive sleep apnea: the use of nasal CPAP in

80 children. Am J Respir Crit Care Med 1995;152:

780–5.

[137] Li KK, Riley RW, Guilleminault C. An unreported

risk in the use of home nasal continuous positive air-

way pressure and home nasal ventilation in children:

mid-face hypoplasia. Chest 2000;117:916–8.

[138] Marcus CL, Carroll JL, Bamford O, Pyzik P, Lough-

lin GM. Supplemental oxygen during sleep in child-

ren with sleep-disordered breathing. Am J Respir Crit

Care Med 1995;152(4 Pt 1):1297–301.

[139] Aljadeff G, Gozal D, Bailey-Wahl SL, Burrell B,

Keens TG, Ward SL. Effects of overnight supplemen-

tal oxygen in obstructive sleep apnea in children. Am

J Respir Crit Care Med 1996;153:51–5.

[140] Brouillette RT, Waters K. Oxygen therapy for pedia-

tric obstructive sleep apnea syndrome: how safe?

How effective? Am J Respir Crit Care Med 1996;

153:1–2.

[141] Kerbl R, Zotter H, Schenkeli R, Hoffmann E, Perro-

gon A, Zotsch W, et al. Persistent hypercapnia in

children after treatment of obstructive sleep apnea

syndrome by adenotonsillectomy. Wien Klin Wo-

chenschr 2001;113:229–34.

[142] Nandapalan V, McCormick MS, Jones TM, Gibson H.

Does adenotonsillectomy cure hypoxaemia in children

with sleep apnoea and congenital cardiac pathology?

Int J Pediatr Otorhinolaryngol 1999;50:55–62.

[143] Morton S, Rosen C, Larkin E, Tishler P, Aylor J,

Redline S. Predictors of sleep-disordered breathing

in children with a history of tonsillectomy and/or

adenoidectomy. Sleep 2001;24:823–9.

J.L. Carroll / Clin Chest Med 24 (2003) 261–282282

State of home sleep studies

Christopher K. Li, MD, W. Ward Flemons, MD*

Division of Respiratory Medicine, Department of Medicine, Foothills Hospital, University of Calgary, #211,

108 Edgeridge Terrace Northwest, 1403 29th Street Northwest, Calgary, Alberta, T3A 6C4 Canada

‘‘Sleep is a reward for some, a punishment for

others.’’

Isidore Ducasse

Monsieur Ducasse, a nineteenth century French

poet, recognized a subset of the population with

badly troubled sleep. Some of these people must

have had obstructive sleep apnea, a common dis-

order defined by recurrent apneas or near-apneas

(hypopneas) during sleep. Obstructive sleep apnea is

suspected especially in obese patients who snore,

have systemic or pulmonary hypertension, or are

hypersomnolent [1]. In the Wisconsin Sleep Cohort

study, a working population aged 30 to 60 years was

surveyed to determine the prevalence of sleep apnea

and commonly associated symptoms. Sleep apnea,

defined as an apnea-hypopnea index (AHI, the

number of apneas and hypopneas per hour of sleep)

more than 5, was present in 24% of male subjects

and 9% of female subjects [2]. The prevalence of

symptomatic sleep apnea (AHI > 5 with excessive

daytime somnolence) was 4% and 2%, respectively;

habitual snoring, 44% and 28%, respectively; and

self-reported hypersomnolence, 16% and 23%,

respectively [2]. The prevalence of hypertension in

this study group was 34% [3], whereas the preva-

ence of obesity (body mass index >30 kg/m2) in

the general population aged 20 to 74 is approxi-

mately 27% [4]. The percentage of the population

who are ‘‘at risk’’ of having sleep apnea is high.

Because it is expected that treatment would make a

significant difference in quality of life for many of

these people, there is a steadily increasing demand

for investigation.

The widely accepted reference standard for the

diagnosis of sleep apnea is the polysomnogram [5];

however, this labor-intensive test is time consuming

and requires considerable technical expertise to per-

form and interpret. As a result, most health care

jurisdictions have unacceptably long waiting times

for sleep studies, which causes many clinicians to

seek simpler, more accessible tests. In 1992, Douglas

et al reported in a sample of 200 consecutive patients

who underwent diagnostic polysomnography that the

omission of the electroencephalogram, electromyo-

gram, and electrooculogram, which allow staging of

sleep and detection of arousals, had little or no

influence on their diagnostic conclusions [6]. This

strongly suggested that devices that monitor only

respiration might well prove to be satisfactory for

investigating many cases of suspected sleep apnea.

Diagnosis could be more accessible in simpler cases,

and waiting times for polysomnography might be

reduced in more complicated cases. With this in

mind, numerous devices designed to monitor respira-

tion at home have been developed.

Compared with polysomnography, portable mon-

itors are less costly, do not require a technician in

attendance, and record patients in the natural envi-

ronment of their own beds. Most of the devices are

more prone to technical failures, give no information

about sleep state or even whether the patient was

asleep, fail to detect problems other than sleep apnea,

and have not been shown to distinguish central from

obstructive apneas. Use of portable monitors at home

for managing sleep apnea patients remains controver-

sial and is not currently considered accepted practice

by any specialty group.

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00018-2

* Corresponding author.

E-mail address: flemons@ucalgary.ca (W.W. Flemons).

Clin Chest Med 24 (2003) 283–295

Classification of portable monitors

The technology for data acquisition and analysis

for home monitors has evolved rapidly, and several

devices have been modified repeatedly over the

years. The American Academy of Sleep Medicine,

formerly the American Sleep Disorders Association,

developed a classification system for portable mon-

itors based on the number and type of parameters

recorded (Table 1) [7].

Defining breathing-disordered events

Although polysomnography is widely recognized

as the reference standard for evaluating patients who

are suspected of sleep apnea, the methods and criteria

for defining events are not standardized across sleep

laboratories or research studies [5]. In general, breath-

ing disturbances are identified during polysomnogra-

phy by a clear reduction in a measurement of

breathing, with or without an accompanying decrease

in oxygen saturation or arousal [5]. The most com-

mon method for detecting reductions in airflow dur-

ing polysomnography is a nasal thermistor, which

detects changes in air temperature. Thermistors are

nonquantitative, however, and some experts recom-

mend that they not be used [5].

Type 2 monitors use the same bioelectric signals

as standard polysomnography, which allows quan-

tification of total sleep time and calculation of the

AHI. Type 3 monitors use similar channels and

definitions for detecting breathing events as type 1

and 2 monitors but lack the bioelectric signals for

sleep staging. Because electroencephalogram, elec-

trooculogram, and electromyogram are not recorded,

arousals cannot be used to identify respiratory dis-

turbances and total sleep time cannot be determined.

Types 3 and 4 monitors most commonly divide the

number of events by total monitoring time to derive a

respiratory disturbances index (RDI), which neces-

sarily underestimates AHI to some degree. To address

this potential problem, White et al used a combina-

tion of electrooculogram channels and anterior tibialis

electromyogram channels to estimate total sleep time;

the correlation with electroencephalogram-based

scoring of sleep time was 0.72 [8]. Others have used

leg movements on electromyogram alone to estimate

periods of wakefulness and subtracted these from the

total monitoring time [9]. It is not clear how such

estimates of total sleep time affect the diagnostic

performance of portable monitors.

Type 4 monitors have used several methods to

define breathing disturbances. Most methods use

oxygen saturation as the primary parameter, but there

are many different techniques for analyzing the data.

The various oximeters use different algorithms for

calculating oxygen saturation, have different sam-

pling frequencies, and store or display the signal at

different intervals. Some oximeters take multiple

readings, store them in memory, average them, and

report a value every 21 seconds [10]; others sample

Table 1

American Academy of Sleep Medicine classification system for sleep apnea evaluation studies

Type 1

Standard

polysomnography

Type 2

Comprehensive

portable polysomnography

Type 3

Modified portable

sleep apnea testing

Type 4

Continuous single

or dual parameter

recording

Parameters Minimum of 7,

including EEG,

EOG, chin EMG,

ECG, airflow,

respiratory effort,

oxygen saturation

Minimum of 7, including

EEG, EOG, chin EMG,

ECG, airflow, respiratory

effort, oxygen saturation

Minimum of 4,

including ventilation

(at least 2 channels

of respiratory movement,

or respiratory movement

and airflow), heart rate or

ECG, oxygen saturation

Minimum of 1:

oxygen saturation,

flow, or

chest movement

Body position Documented or

objectively measured

Possible Possible No

Leg movement EMG or motion sensor

desirable but optional

Optional Optional No

Personnel

in attendance

Yes No No No

Interventions during

the study

Possible No No No

Abbreviations: EEG, electroencephalography; EOG, electrooculography; EMG, electromyography; ECG, electrocardiography.

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295284

and report each value at a frequency up to 10 Hz [11].

A sampling rate of every 12 seconds has been shown

in one study to give falsely low rates of oxygen

desaturations per hour [12].

Methods of automated analysis of the oxygen satu-

ration signal are also variable: most rely on detection

of a drop in oxygen saturation from 2% to 5%, some

detect resaturation [13], and others are designed to use

both criteria [14]. Some automated analyses calculate

baseline oxygen saturation [14], but most do not.

Some oximetry-based monitors do not score dis-

crete events but instead identify sleep apnea from an

overall pattern or distribution of oxygen saturations.

The CT90 is the cumulative percentage of time that

oxygen saturations are below 90%; a CT90 that

exceeds 1% has been used as a criterion for diagnos-

ing sleep apnea [15]. The delta index is a measure of

variability in oxygen saturation over constant time

intervals; the higher the delta index, the higher the

likelihood of sleep apnea [16].

Some type 4 monitors are oximetry based but also

record snoring [17] and heart rate variability [18].

Using a nasal pressure cannula, one type 4 monitor

detects reduction in nasal airflow as the primary

criterion for breathing-disordered events [19–21].

There is no consensus about the best method for

interpreting data from home monitors. Some methods

identify and count events automatically, but these

may fail to identify poor quality recordings and can

give misleading results. Others depend on manual

review by a sleep technician or physician, which

raises the issue of interobserver and intraobserver

variability. Still others score events automatically

but produce printouts of raw data that can be

reviewed manually to detect problems, such as

artifact or poor quality data. So far, researchers using

manual scoring or manual review have not published

data on the reliability of their scoring methods.

Existing guidelines and reviews

In 1994, the American Sleep Disorders Asso-

ciation practice parameters recommended that poly-

somnography remain the standard for the diagnosis,

determination of severity, and treatment of sleep

apnea [22]. Unattended portable recording was

viewed as an acceptable alternative only under the

following circumstances: (1) when initiation of treat-

ment was urgent and polysomnography unavailable,

(2) when patients could not undergo polysomnogra-

phy because of mobility issues, or (3) as a follow-up

to treatment study. The use of type 4 studies was not

considered acceptable at that time. Since 1994, many

validation studies for various types of portable mon-

itoring and three reviews have been published; an

update of the existing guidelines is warranted.

In 1994, the American Sleep Disorders Asso-

ciation reviewed 23 studies [7], and in 1997 it pub-

lished a review [23] and practice parameters [24] for

polysomnography and related procedures that in-

cluded a section on type 3 and 4 monitors. These

practice parameters suggested that attended type 3

monitors might be appropriate in patients with a high

pretest probability (eg, >70%) of sleep apnea and that

negative type 3 monitor studies in symptomatic

patients should be followed up with a full polysomno-

gram [24]. Also in 1997, the Agency for Healthcare

Research and Quality (formerly the Agency for Health

Care Policy and Research) commissioned a systematic

review of the research on the diagnosis of sleep apnea.

The section of that review devoted to portable mon-

itors reviewed 25 studies of multi-channel devices,

including 12 studies on oximetry alone [25]. The

quality of each reviewed study was rated using a scale

developed by the authors. The system for assigning

‘‘quality ratings’’ to the articles was somewhat differ-

ent from published methods for rating evidence on

diagnostic studies, however [26]. Higher rated (ac-

cording to the system of Sackett et al [26]), quality

research studies that compared portable monitoring to

polysomnography will be the focus of this article. A

complete, updated systematic review of the literature

on portable monitoring for sleep apnea is required but

is beyond the scope of this article.

Evidence

Rating the evidence

To avoid bias in assessing a diagnostic test such as a

portable monitor, several key factors must be consid-

ered. Selection biasmay be introduced if consecutively

referred patients are not used. Verification bias may be

introduced if the decision to perform a reference

standard (in this case, polysomnography) is influenced

by the results of the test being evaluated (a portable

monitor). Table 2 depicts how the system by Sackett

et al [26] for rating evidence would apply to studies

evaluating a portable monitor.

Comparing results of portable monitoring

to polysomnography

Several methods are in use for evaluating agree-

ment between the results of two diagnostic tests, such

as the AHI from polysomnography and the RDI from a

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295 285

portable monitor. The product-moment (Pearson) cor-

relation coefficient is most commonly used but is not

recommended [27]. Although it describes how closely

the two test results are associated (ie, how closely they

cluster along a straight line when one result is plotted

against the other), it does not indicate whether one

result is the same as the other. For example, a monitor

that always gives an RDI exactly half the AHI would

have the same high correlation coefficient as one that

gives an RDI equal to theAHI. Correlation coefficients

also are strongly influenced by the range of values of

the test results. The coefficient might be high in cases

with a high AHI and RDI, but there may be a great deal

of scatter at the lower end of the AHI range (impor-

tantly, near the diagnostic cut-off), which makes the

portable monitor useless at identifying anything other

than severe disease.

The Bland Altman approach is to calculate the

difference between each pair of results (AHI and

corresponding RDI) and plot that against the mean of

the two numbers [27]. The ‘‘limits of agreement’’ (ie,

the mean F 2 standard deviations of the differences)

can be misleading, however, because they are often

strongly influenced by data in the range of high AHI,

where it is irrelevant. The limits of agreement in the

important low range of AHI, near the diagnostic cut-

off, may be better than the statistic calculated for the

whole group.

Because ultimately a clinician’s main concern is

whether a test correctly classifies patients as having

or not having sleep apnea, sensitivity, specificity, and

likelihood ratios for the RDI as a predictor of the AHI

seem more appealing. This approach dictates that a

patient be classified with or without sleep apnea

based on an arbitrary cut-off, such as an AHI of 10;

by dichotomizing results into simply positive or

negative, a good deal of information is lost. Most

research studies on portable monitoring report sen-

sitivity and specificity; some studies also list mean

differences and limits of agreement. The only way to

compare the performance of most portable monitors

is to use their reported sensitivity and specificity and

their calculated likelihood ratios.

Sensitivity is the proportion of patients with

disease who have a positive test result, or the ‘‘true-

positive’’ rate, whereas specificity is the proportion of

patients without disease who have a negative result,

or the ‘‘true-negative’’ rate. These numbers indicate

the probability that the test result will be positive if

the patient has the disease and the probability that the

test result will be negative if the patient does not have

the disease, respectively. These numbers by them-

selves are not sufficient to guide a clinician’s deci-

sion-making process, however, because clinicians do

not know whether a patient has the disease. What a

physician must know is the probability that the

patient has the disease if the test result is positive

or negative (positive and negative predictive values

of the test, respectively).

Sensitivity and specificity can be determined by

analyzing columns in a 2� 2 table (Table 3), whereas

the positive and negative predictive values are

obtained by analyzing rows. By convention, the

reference standard is at the top; for sleep apnea this

is usually based on the AHI (the most common cut-

offs used are 10 or 15). The new diagnostic test to

which it is being compared is on the side; for sleep

apnea these are the results of the portable monitor

or RDI.

Changing the threshold of what constitutes a

normal or abnormal diagnostic test changes the

sensitivity and specificity. Lowering the threshold

increases sensitivity but lowers specificity, which

causes more true-positive results (and fewer false-

negative results) but also more false-positive results.

The converse—increasing the threshold—has the

opposite effect (it lowers sensitivity and increases

specificity). Because positive and negative predictive

values depend on the combination of sensitivity and

specificity, using either of these statistics in isolation

to infer the usefulness of a diagnostic test for ruling in

or ruling out a disorder can be misleading.

Table 2

Levels of evidence for studies of portable monitors for the

diagnosis of sleep apnea

Level

of evidence Criteria

1 Independent, blind comparison

between the PM and PSG

Appropriate spectrum of

consecutive patients

PM and PSG performed on

all patients

2 Independent, blind comparison

between the PM and PSG

Narrow spectrum of individuals or

nonconsecutive patients

PM and PSG performed on

all patients

3 Independent, blind comparison

between the PM and PSG

Appropriate spectrum of

consecutive patients

PSG not performed on all patients

4 Comparison between the PM and

PSG was not independent or blind

Abbreviations: PM, portable monitor; PSG, polysomno-

graphy.

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295286

The utility of a test is best captured in a single

number, the likelihood ratio. The likelihood ratio for

a positive test result is the ratio of the proportion of

patients with disease who have a positive test (true-

positive rate or sensitivity) to the proportion of people

without disease who have a positive test (false-

positive rate). Similarly, the likelihood ratio for a

negative test result is the ratio of the proportion of

patients with disease who have a negative test (false-

negative rate) to the proportion of people without

disease who have a negative test (true-negative rate or

specificity). Using the example of the 2� 2 table

(see Table 3), the likelihood ratio for a positive result

is 0.9/0.112, which is 8; the likelihood ratio for a

negative result is 0.1/0.882, which is 0.11. Mathemat-

ically, when using likelihood ratios to convert pretest

to posttest probabilities, the pretest probability esti-

mate (ie, the estimated prevalence) is first converted

to an odds expression (pretest odds = pretest prob-

ability/1� pretest probability) and then multiplied by

the likelihood ratio to obtain the posttest odds, which

are then converted back to a probability statement

(posttest probability = posttest odds/posttest odds + 1).

This process can be simplified greatly with the use of

a nomogram (Fig. 1) [28]. The nomogram also high-

lights the interaction between pretest probability and

likelihood ratio on posttest probability.

A guide to the interpretation of likelihood

ratios follows.

Likelihood ratio influence on disease probability

< 0.05 Very large reduction

0.05–0.1 Large reduction

0.1–0.2 Modest reduction

0.21–5 Little change

5.1–10 Modest increase

10.1–20 Large increase

>20 Very large increase

In this article the authors have included the best

reported sensitivity and specificity for the portable

monitors evaluated. In some studies, the best sen-

sitivity and best specificity are obtained at different

RDI cut-offs. If this is the case, then some patients in

the study population will have a ‘‘negative’’ result (an

RDI below the cut-off for best sensitivity) and others

will have a ‘‘positive’’ result (an RDI above the cut-

off for best specificity), but a certain percentage of

patients will have an RDI between these cut-offs and

will have neither a negative nor positive result. If this

percentage of ‘‘unclassified’’ patients is high, then the

portable monitoring test may have little clinical use

despite having a high sensitivity and specificity. This

potential problem is circumvented if the best sen-

sitivity and specificity occur at the same RDI cut-off,

in which case all patients can be classified as either

negative or positive.

Type 2 monitors

A potential advantage of type 2 monitors is that

they provide information about non–sleep apnea dis-

orders, such as periodic limb movements. Patients

usually must come to the laboratory to have electrodes

applied by a technician before the home study, how-

ever. Data loss rates of 20% have been reported in the

unattended setting [29], and patients may sleep poorly

Table 3

Calculating sensitivity, specificity, positive and negative predictive values, and the effect of prevalence

(prevalence = 150/1000 or 15%) (prevalence = 600/1000 or 60%)

RS + ve RS� ve RS + ve RS� ve

DT+ ve 135TP 100FP 235 DT+ ve 540 47 587

DT� ve 15FN 750TN 765 DT� ve 60 353 413

150 850 1000 600 400 1000

In this hypothetical example (left side), 150 patients have sleep apnea (prevalence = 15%) and 135 of these patients have a

positive diagnostic test result (sensitivity = 135/150 = 90%). Of the 850 patients who do not have the disease, 750 have a

negative test result (specificity = 750/850 = 88.2%). The positive predictive value is 135/235 (57.4%). The negative predictive

value is 750/765 (98%). In the example on the right side, the prevalence has increased to 60% with no change in sensitivity or

specificity; however, the positive predictive value has increased substantially to 92%, and the negative predictive value has

dropped to 85.5%. The formulas are as follows:

Sensitivity: TP/TP + FN

Specificity: TN/TN+FP

Positive predictive value: TP/TP + FP

Negative predictive value: TN/TN+FN

Abbreviations: TP, true positives; FP, false positives; TN, true negatives; FN, false negatives; RS, reference standard

(polysomnography); DT, diagnostic test (portable monitor).

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295 287

because of concerns about safety or equipment fail-

ures. Two studies found that patients preferred labora-

tory polysomnography to a type 2 portable monitor

[29,30]. The best quality study (level 2 evidence) had

only 20 patients, from which only gross estimates of

sensitivity (80%) and specificity (90%) can be drawn.

The calculated likelihood ratio for a positive test result

was 8, and the likelihood ratio for a negative test result

was 0.22 [31]. Currently, it is not proven that type 2

monitors are reliable or offer any advantage over

laboratory polysomnography.

Type 3 monitors

With fewer channels, type 3 monitors are easier

for patients to sleep with, and technicians are not

required for the initial set up. One study reported an

at-home failure rate of 10% [32]. Three studies that

were Level 1 evidence compared type 3 monitors to

simultaneous laboratory polysomnography. Sensitiv-

ity rates ranged from 92% to 100%, and specificity

rates ranged from 96% to 100% [9,33,34]. Calculated

likelihood ratios were more than 20 for a positive test

result and less than 0.10 for a negative test result. One

study noted that the sensitivity rate dropped to 55% at

an AHI cut-off of 40 when events were scored

automatically, but visual editing of the raw data

improved the sensitivity rate to 91% [33].

To date, no level 1 studies have compared un-

attended type 3 monitors to laboratory polysomnog-

raphy. Two level 2 studies reported best sensitivity

rates of 91% to 95% and best specificity rates of 83%

to 93%, with likelihood ratios of 5.1 to 9 for a

positive test result and 0.13 to 0.15 for a negative

test result; however, it should be noted that 22% to

37% of patients would have been ‘‘unclassified’’ in

these studies [8,32].

Overall, type 3 monitors have been shown in

level 1 attended laboratory studies to have like-

lihood ratios that can alter substantially the posttest

probability of sleep apnea. Manual scoring or

review of raw data with editing seems to improve

the specificity of some of these devices at higher

AHI cut-offs; however, it is not clear that this would

have an impact on clinical decision making. In the

home setting, level 2 evidence has shown low

likelihood ratios for negative tests, and these mon-

itors could be used to ‘‘rule out’’ sleep apnea. In

these studies, portable monitoring and polysomnog-

raphy were performed on different nights, and night-

to-night variation in a patient’s disease may have

played a role in explaining the modest likelihood

ratios for a positive result. Some authors also have

postulated that patients may have slept more in the

Fig. 1. A nomogram for converting pretest to posttest

probability (probabilities listed as percentages), using

likelihood ratios. To use the nomogram, anchor a straight

edge at the pretest probability and direct it through the

appropriate likelihood ratio. The intersection of the straight

edge with the third (right) line produces the probability

result. (From Fagan TJ. Nomogram for Bayes’ theorem.

N Engl J Med 1975;293:257; with permission.)

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295288

home environment and experienced more breathing-

disordered events, which lead to ‘‘false-positive’’

results on portable monitoring but were in fact

true-positive results with a falsely negative poly-

somnogram result [8].

All type 3 monitors evaluated in the literature have

used thermistors as a measurement of flow. Limited

data are available on their accuracy, but laboratory

models that compared thermistors to pneumotacho-

graphs suggest that thermal signals are nonlinearly

related to actual airflow and tend to overestimate

ventilation [35]. Nasal pressure has been shown to

have excellent agreement with a pneumotachograph

[36], and although theoretically some false-positive

events may result from mouth breathing, this tech-

nology seems superior to thermistors for detecting

apneas and hypopneas. Type 3 monitors might be

improved by substituting nasal pressure for thermis-

tors, but to date no unattended study using nasal

pressure-based monitors has been reported.

Type 4 monitors

Oximetry alone

In 1993, Series et al published a level 1 study that

compared nocturnal home oximetry to subsequent

polysomnography in 240 patients with suspected

sleep apnea [37]. Oximetry, with a sampling rate of

0.5 Hz using a finger probe, was classified as ‘‘nor-

mal’’ or ‘‘abnormal’’ according to the absence or

presence of repetitive episodes of transient desatura-

tion; no minimum decrease in saturation levels or

threshold saturation was used. Repeat oximetry was

required in 8% of patients. The authors reported a high

sensitivity (98%) but a low specificity (48%), which

corresponded to calculated likelihood ratios of 1.88

for a positive test result and 0.037 for a negative test

result. The low likelihood ratio for a negative test

result indicates that this approach was useful for

‘‘ruling out’’ sleep apnea; however, a positive test

result would have required additional testing.

Most other studies of oximetry alone have used a

desaturation threshold to identify and quantify breath-

ing-disordered events, including two level 1 evidence

studies. In a home setting, Gyulay et al reported a best

sensitivity rate of 93% and a best specificity rate of

98% (calculated likelihood ratio of 20 for a positive

test result and 0.14 for a negative test result), although

49% of patients would have been ‘‘unclassified’’ using

these RDI cut-offs [15]. A study by Chiner et al in a

laboratory setting reported a best sensitivity rate of

82% and best specificity rate of 93% (calculated

likelihood ratio of 8.9 for a positive test result and

0.24 for a negative test result); 19% of these patients

would have been ‘‘unclassified’’ [38].

Snoring and oximetry

Issa et al reported on a monitor that measured

snoring via a laryngeal microphone and 1 Hz oxime-

try [17]. In their level 2 evidence study, they reported

a best sensitivity rate of 89% and a best specificity

rate of 98% (calculated likelihood ratio of 45 for a

positive test result and 0.12 for a negative test result),

although 22% of patients would have been ‘‘unclas-

sified.’’ A subsequent version of the device modified

the automated oximetry analysis algorithm and elim-

inated snoring from the definition of a breathing-

disordered event. A level 1 validation study of the

newer device compared with simultaneous polysom-

nography reported a best sensitivity rate of 97% and a

best specificity rate of 88% (calculated likelihood

ratio of 8.2 for a positive test result and 0.04 for a

negative test result), with 11% of tests ‘‘unclassified’’

[14]. The increased specificity of this device com-

pared with other oximeters is likely a result of the

unique analysis software, which uses a moving base-

line and desaturation and resaturation criteria for

defining an event. Both of these studies took place

in a laboratory setting, and their results must be

confirmed in a home study.

Nasal pressure

Several published studies that are level 2 evidence

have been conducted using a monitor that measures

nasal flow via a pressure transducer [19–21]. In

these studies, an RDI was defined by a reduction

in nasal flow of 50% or more. Oximetry also was

measured but was not one of the criteria for defining

an event. On comparison with simultaneous labora-

tory polysomnography, best sensitivity rate ranged

from 97% to 100%, and best specificity rate ranged

from 77% to 93% (calculated likelihood ratio of

4.2–12.5 for a positive test result and 0–0.06 for a

negative test result). In one study, 48% of patients

would have been ‘‘unclassified’’ [20]; in the other

two studies the best sensitivity and specificity rates

were obtained at the same RDI cut-off. These prom-

ising likelihood ratios must be confirmed by level 1

studies, and the devices should be tested unattended

in the home.

Oximetry, snoring, and heart rate variability

A 1992 level 2 evidence study reported a device

that measured oximetry, snoring, heart rate, and body

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295 289

position but used only oximetry in the algorithm for

defining a breathing-disordered event [39]. The oxy-

gen saturation sampling frequency was not reported.

When compared with simultaneous laboratory poly-

somnography, the best sensitivity rate was 97% and the

best specificity rate was 92% (calculated likelihood

ratio of 12.1 for a positive test result and 0.03 for a

negative test result). A subsequent study (level 2

evidence) used snoring pauses in the scoring algorithm

and had a higher calculated likelihood ratio for a

positive result (34.5) and a similar calculated like-

lihood ratio for a negative result. However, 26% of the

patients would have been ‘‘unclassified’’ [18]. The

most recent version of this monitor used an algorithm

that combined oximetry with heart rate. The best

reported sensitivity and specificity rates were lower,

which resulted in less useful likelihood ratios, but it

has only been evaluated in a level 4 evidence, unat-

tended home study [40].

In summary, oximetry has demonstrated use for

‘‘ruling out’’ sleep apnea in the attended and unat-

tended settings. The sampling rate and interpretation

algorithm of the particular device must be considered

carefully. In devices that measure other channels,

such as snoring and heart rate variability, the best

likelihood ratios for a positive test result were

obtained using algorithms that used oximetry alone

to define breathing-disordered events. Limited data

are available on oximetry-based portable monitors in

the unattended setting for ‘‘ruling in’’ sleep apnea.

Further research is required to determine if the find-

ings in the attended setting remain valid in the

unattended setting. Nasal pressure-based monitors

also have useful negative likelihood ratios and rea-

sonably helpful positive likelihood ratios. There is a

pressing need for further validation studies of type 4

monitors in the unattended setting.

Further research directions

Additional research on portable monitors is

required to address several issues. Most monitors

have been studied by only a single group of inves-

tigators. All studies have taken place on patients

referred to a sleep center. It is yet to be proven what

the effect of changing the studied clinical population

would have on the diagnostic performance of these

monitors. Primary care populations, women, non-

whites, and patients with comorbid illness have not

been studied adequately; therefore, the published

results on portable monitoring cannot necessarily

be generalized to these groups. Changing the popu-

lation of patients could have two effects: (1) it could

alter prevalence (pretest probability) of the condition

and impact positive and negative predictive values

and (2) it could affect the operating characteristics

(sensitivity, specificity, likelihood ratios) of the

monitor. Studies published in the future should plan

to address key methodologic issues such as selection

bias, verification bias, and blinded data interpre-

tation. Investigators are encouraged to provide

detailed information on their study population,

methods used for acquiring and analyzing portable

monitoring data, and polysomnography data.

Using laboratory polysomnography as a reference

standard is often criticized because many patients do

not sleep well in a laboratory and it is difficult to

account for night-to-night variability. Although a dif-

ference of 5 between the AHI and RDI may not be

clinically significant, it can result in a portable moni-

toring study being labeled ‘‘falsely’’ positive or nega-

tive. A more rigorous validation study would address

important clinical outcomes, such as improvement in

quality of life (including symptoms such as daytime

sleepiness) and compliance with treatment.

Portable monitors in a clinical decision algorithm

Like any diagnostic test, the results of testing

with portable monitors are most useful when applied

to the appropriate clinical context. The results of a

negative portable monitoring study would have

different implications for a mildly symptomatic

patient with a low pretest probability compared with

a symptomatic patient with a high pretest probabil-

ity. The probability that a patient has sleep apnea

based on clinical factors alone can be estimated

using one of several clinical prediction rules [41].

The sensitivity rate of a risk stratification algorithm

that combined a clinical prediction rule and oxime-

try has been reported by Gurubhagavatula et al [42]

to be 95% for detection of sleep apnea (AHI� 5)

and 85% for severe sleep apnea (AHI� 30). Cor-

responding specificity rates were 68% and 97%,

respectively. Although this is a well-validated

method, the complexity of it may limit its clinical

application. A simplified approach to assigning

clinical probability and incorporating it into a strat-

egy for managing patients with suspected sleep

apnea recently was published (Fig. 2) [1]. It is

derived from a sleep apnea clinical prediction rule

that was developed using multiple linear regression

[43]. The ‘‘adjusted neck circumference’’ in centi-

meters is calculated by adding 4 cm if the patient

has hypertension, 3 cm if the patient is a habitual

snorer, and 3 cm if the patient is reported to choke

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295290

or gasp most nights [1]. Table 4 illustrates how the

adjusted neck circumference corresponds with a

patient’s clinical probability of having a positive

test result for sleep apnea.

The following scenarios are examples of how

this clinical decision algorithm might be applied

to patients.

Example 1

Mrs. A is a healthy 53-year-old schoolteacher

whose husband has complained of her heavy snor-

ing. They have started to sleep in different rooms

because of this. Her neck circumference is 40 cm,

she does not have systemic hypertension, and she

Fig. 2. A suggested clinical decision algorithm for evaluating patients with suspected sleep apnea. (From Flemons WW.

Obstructive sleep apnea. N Engl J Med 2002;347:498–504; with permission.)

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295 291

has not been reported to choke or gasp while

sleeping. She does not complain of significant

daytime somnolence.

Mrs. A’s adjusted neck circumference is 43 cm

(40 cm + 3 cm for snoring), and she has an inter-

mediate clinical probability of sleep apnea. Testing at

home with a portable monitor similar to the one

studied by Vazquez et al [14] produced results illus-

trated in Fig. 3. Although this monitor has more

channels than the one reported by Vazquez et al,

the automated scoring algorithm based on the oxygen

saturation signal is identical. The updated monitor

records and reports airflow (using nasal pressure) and

heart rate in addition to the standard signals of

oxygen saturation, snoring, and body position. The

tracing in Fig. 3 demonstrates snoring but normal

flow and oxygen saturation. The patient’s RDI was

4.1, which, combined with her low pretest probabil-

ity, is sufficient to ‘‘rule out’’ clinically important

sleep apnea. Because she is asymptomatic, further

investigations are not indicated, and a discussion of

her treatment options for primary snoring can ensue.

Example 2

Mr. B is a 37-year-old executive who presents

because of excessive daytime somnolence. He has

been falling asleep in meetings and in front of his

computer at work. On several occasions, he has

dozed off while driving home and swerved off the

road. He has systemic hypertension and a neck

circumference of 41 cm. His wife describes him as

an occasional snorer, but she does not report choking

or gasping during sleep.

The patient has an adjusted neck circumference of

45 cm (41 cm + 4 cm for hypertension), and thus an

intermediate clinical probability of sleep apnea. Fig. 4

is taken from his portable monitor study, which

demonstrates no snoring, normal flow, and a normal

oxygen profile. The RDI was 1.8.

Mr. B’s portable monitor study makes sleep apnea

an unlikely cause of his daytime symptoms. Because

of the severity of his somnolence, further investi-

gations such as polysomnography and a multiple

sleep latency test are indicated.

Example 3

Mr. C is a 49-year-old carpenter who is referred

for assessment of excessive daytime somnolence. He

can fall asleep in any situation if he is not physically

active or mentally stimulated. His wife claims that he

is a ‘‘heroic’’ snorer who frequently chokes, gasps,

snorts, and stops breathing when he is asleep. He has

mild type II diabetes mellitus and a blood pressure of

160/90. His neck circumference is 47 cm.

Mr. C has a high clinical probability of sleep

apnea, with an adjusted neck circumference of

57 cm (47 cm + 4 cm for hypertension + 3 cm for

snoring + 3 cm for choking/gasping). His portable

monitor study, shown in Fig. 5, demonstrates fre-

quent, cyclic oxygen desaturations associated with

Table 4

Adjusted neck circumference and corresponding clinical

probability

Adjusted neck circumference (cm) Clinical probability

< 43 Low

43–48 Intermediate

> 48 High

Fig. 3. Heavy snoring; no evidence of sleep apnea. The oxygen saturation profile (red tracing), air flow (blue tracing), and

heart rate (green tracing) are all normal. The patient is lying in a nonsupine position as indicated by the lack of a horizontal

line adjacent to the ‘‘Supine’’ label (compare with Fig. 5). The vertical black lines at the bottom indicate heavy snoring. (Time

frame = 10 minutes.)

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295292

intermittent snoring and reductions in flow. He has an

RDI of 65 and clinically important daytime som-

nolence, so a trial of CPAP is indicated. It is impor-

tant for him to have a follow-up test to ensure that

these abnormalities normalize on CPAP.

Summary

Many different portable monitors have been used

to assess patients with suspected sleep apnea. There is

limited evidence for the use of type 2 monitors,

especially in the unattended setting in which there

may be high rates of data loss. Type 3 monitors have

low likelihood ratios for negative tests and can be

used to ‘‘rule out’’ sleep apnea. The ability of type 3

monitors to ‘‘rule in’’ sleep apnea is less convincing,

but this may improve with the use of improved

technology, such as nasal pressure transducers. Type

4 monitors usually use oximetry and can be used to

‘‘rule out’’ sleep apnea. Higher sampling rates and

improved analysis algorithms can improve the spe-

cificity of these monitors; hence, likelihood ratios for

a positive test result can be high enough with some

monitors to ‘‘rule in’’ sleep apnea as well. Not all

monitors record and analyze signals in the same way;

it is not possible to generalize results from one

monitor across all monitors of a particular type.

Limited evidence is available for many portable

monitors in the unattended setting, and further

research is required in this area.

Clinicians should identify how they plan to use a

portable monitor: as a mechanism to exclude disease

in asymptomatic snorers, to confirm disease in

Fig. 4. No evidence of sleep apnea. The oximetry recording (red tracing), air flow (blue line), and heart rate (green line) are all

normal. The patient is lying in a nonsupine position as indicated by the lack of a horizontal line adjacent to the ‘‘Supine’’ label

(compare with Fig. 5). There are no vertical bars adjacent to the ‘‘Snore’’ label, which indicates that the patient was not snoring.

(Time frame = 10 minutes.)

Fig. 5. Severe sleep apnea. Cyclic oxygen desaturations are present, to as low as 74% (red tracing). The black vertical bars at the

nadir of the oxygen saturation indicate that the monitor scored this as a respiratory disturbance. There is intermittent cessation of

air flow (blue tracing) and tachycardia with termination of most apneas (green tracing). The patient is in the supine position

(magenta line), and intermittent snoring is also present (black vertical lines). (Time frame = 10 minutes.)

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295 293

patients with a high clinical probability of disease, or

to risk stratify patients so that proper priority for

polysomnography can be determined. This deter-

mination allows them to select a portable monitor

with signals most appropriate to their needs. The

quality of the validation studies for each portable

monitor also should be evaluated carefully before

implementation in clinical practice. The ability for a

clinician to review raw data manually and consider

artifact is a necessary feature. Measurement of oxy-

gen saturation also is important to identify patients

with previously unsuspected serious desaturation that

would indicate the need for more urgent treatment.

In centers in which polysomnography is not

readily available, a clinical decision algorithm that

incorporates a clinical prediction rule with the use of

portable monitors can guide clinicians toward insti-

tution of therapy or further investigations. Intuitively,

this approach could reduce waiting times for poly-

somnography and delays in diagnosis, but additional

evidence for the validity and cost effectiveness of this

approach is required.

References

[1] Flemons WW. Obstructive sleep apnea. N Engl J Med

2002;347:498–504.

[2] Young T, Palta M, Dempsey J, Skatrud J, Weber S,

Badr S. The occurrence of sleep-disordered breathing

among middle-aged adults. N Engl J Med 1993;

328:1230–5.

[3] Peppard PE, Young T, Palta M, Skatrud J. Prospective

study of the association between sleep-disordered

breathing and hypertension. N Engl J Med 2000;

342:1378–84.

[4] McTigue KM, Garrett JM, Popkin BM. The natural

history of the development of obesity in a cohort of

young US adults between 1981 and 1998. Ann Intern

Med 2002;136:857–64.

[5] AASM Task Force. Sleep-related breathing disorders

in adults: recommendations for syndrome definition

and measurement techniques in clinical research. Sleep

1999;22:667–89.

[6] Douglas NJ, Thomas S, Jan MA. Clinical value of

polysomnography. Lancet 1992;339:347–50.

[7] Ferber R, Millman R, Coppola M, Fleetham J, Mur-

ray CF, Iber C, et al. Portable recording in the as-

sessment of obstructive sleep apnea. Sleep 1994;17:

378–92.

[8] White DP, Gib TJ, Wall JM, Westbrook PR. Assess-

ment of accuracy and analysis time of a novel device to

monitor sleep and breathing in the home. Sleep 1995;

18:115–26.

[9] Emsellem HA, Corson WA, Rappaport BA, Hackett S,

Smith LG, Hausfeld JN. Verification of sleep apnea

using a portable sleep apnea screening device. South

Med J 1990;83:748–52.

[10] Farre R, Montserrat JM, Ballester E, Hernandez L,

Rotger M, Navajas D. Importance of the pulse oxim-

eter averaging time when measuring oxygen desatura-

tion in sleep apnea. Sleep 1998;21:386–90.

[11] Baltzan MA, Verschelden P, Al-Jahdali H, Olha AE,

Kimoff RJ. Accuracy of oximetry with thermistor (oxi-

flow) for diagnosis of obstructive sleep apnea and hy-

popnea. Sleep 2000;25:61–9.

[12] Wiltshire N, Kendrick AH, Catterall JR. Home oxime-

try studies for diagnosis of sleep apnea/hypopnea syn-

drome: limitation of memory storage capabilities. Chest

2001;120:384–9.

[13] Rauscher H, Popp W, Zwick H. Model for investigat-

ing snorers with suspected sleep apnea. Thorax 1993;

48:275–9.

[14] Vazquez J, Tsai WH, Flemons WW, Masuda A, Brant

R, Hajduk E, et al. Automated analysis of digital

oximetry in the diagnosis of obstructive sleep apnoea.

Thorax 2000;55:302–7.

[15] Gyulay S, Olson LG, Hensley MJ, King MT, Murree

Allen K, Saunders NA. A comparison of clinical as-

sessment and home oximetry in the diagnosis of ob-

structive sleep apnea. Am Rev Respir Dis 1993;147:

50–3.

[16] Pepin JL, Levy P, Lepaulle B, Brambilla C, Guillemi-

nault C. Does oximetry contribute to the detection of

apneic events? Mathematical processing of the SaO2

signal. Chest 1991;99:1151–7.

[17] Issa F, Morrison D, Hadjuk E, Iyer A, Feroah T,

Remmers JE. Digital monitoring of sleep-disordered

breathing using snoring sound and arterial oxygen

saturation. Am Rev Respir Dis 1993;148:1023–9.

[18] Esnaola S, Duran J, Infante-Rivard C, Rubio R, Fer-

nandez A. Diagnostic accuracy of a portable recording

device (MESAM IV) in suspected obstructive sleep

apnoea. Eur Respir J 1996;9:2597–605.

[19] Kiely JL, Delahunty C, Matthews S, McNicholas WT.

Comparison of a limited computerized diagnostic sys-

tem (ResCare Autoset) with polysomnography in the

diagnosis of obstructive sleep apnoea syndrome. Eur

Respir J 1996;9:2360–4.

[20] Mayer P, Meurice J-C, Philip-Joet F, Cornette A, Rako-

tonanahary D, Meslier N, et al. Simultaneous labora-

tory-based comparison of ResMed Autoset with

polysomnography in the diagnosis of sleep apnoea/hy-

popnoea syndrome. Eur Respir J 1998;12:770–5.

[21] Gugger M. Comparison of ResMed Autoset (version

3.03) with polysomnography in the diagnosis of the

sleep apnoea/hypopnoea syndrome. Eur Respir J

1997;10:587–91.

[22] Standards of Practice Committee of the American

Sleep Disorders Association. Practice parameters for

the use of portable recording in the assessment of ob-

structive sleep apnea. Sleep 1994;17:372–7.

[23] Chesson Jr AL, Ferber RA, Fry JM, Grigg-Damberger

M, Hartse KM, Hurwitz TD, et al. The indications for

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295294

polysomnography and related procedures. Sleep 1997;

20:423–87.

[24] Polysomnography Task Force. American Sleep Disor-

ders Association Standards of Practice Committee:

practice parameters for the indications for polysomnog-

raphy and related procedures. Sleep 1997;20:406–22.

[25] Ross SD, Sheinhait IA, Harrison KJ, Kvasz M, Con-

nelly JE, Shea SA, et al. Systematic review and meta-

analysis of the literature regarding the diagnosis of

sleep apnea. Sleep 2000;23:519–32.

[26] Sacket DL, Straus SE, Richardson WS, Rosenberg W,

Haynes RB. Guidelines. In: Evidence-based medicine:

how to practice and teach EBM. Edinburgh: Harcourt;

2000. p. 169–82.

[27] Bland JM, Altman DG. Statistical methods for assess-

ing agreement between two methods of clinical mea-

surement. Lancet 1986;1:307–10.

[28] Fagan TJ. Nomogram for Bayes theorem [letter].

N Engl J Med 1975;293:257.

[29] Portier F, Portmann A, Czernichow P, Vascaut L, De-

vin E, Benhamou D, et al. Evaluation of home versus

laboratory polysomnography in the diagnosis of sleep

apnea syndrome. Am J Respir Crit Care Med 2000;

162:814–8.

[30] Fry JM, DiPhillipo MA, Curran K, Goldberg R, Baran

AS. Full polysomnography in the home. Sleep 1998;

21:635–42.

[31] Mykytyn IJ, Sajkov D, Neill AM, McEvoy RD. Por-

table computerized polysomnography in attended and

unattended settings. Chest 1999;115:114–22.

[32] Parra O, Garcia-Esclasans N, Monteserrat JM, Garcia

Eroles L, Ruiz J, Lopez JA, et al. Should patients with

sleep apnoea/hypopnoea syndrome be diagnosed and

managed on the basis of home sleep studies? Eur

Respir J 1997;10:1720–4.

[33] Zucconi M, Ferini-Strambi L, Castronovo V, Oldanni

A, Smirne S. An unattended device for sleep-related

breathing disorders: validation study in suspected ob-

structive sleep apnoea syndrome. Eur Respir J 1996;9:

1251–6.

[34] Verse T, Pirsig W, Junge-Hulsing BJ, Kroker B. Vali-

dation of the POLY-MESAM seven-channel ambula-

tory recording unit. Chest 2000;117:1613–8.

[35] Farre R, Montserrat JM, Rotger M, Ballester E, Na-

vajas D. Accuracy of thermistors and thermocouples

as flow-measuring devices for detecting hypopneas.

Eur Respir J 1998;11:179–82.

[36] Heitman SJ, Atkar RS, Hajduk EA, Wanner RA, Fle-

mons WW. Validation of nasal pressure for the identi-

fication of apneas/hypopneas during sleep. Am J

Respir Crit Care Med 2002;166:386–91.

[37] Series F, Marc I, Cormier Y, La Forge J. Utility of

nocturnal home oximetry for case finding in patients

with suspected sleep apnea hypopnea syndrome. Ann

Intern Med 1993;119:449–53.

[38] Chiner E, Jaime S, Arriero JM, Marco J, Fuentes I,

Sergado A. Nocturnal oximetry for the diagnosis of

the sleep apnoea hypopnoea syndrome: a method to

reduce the number of polysomnographies? Thorax

1999;54:968–71.

[39] Stoohs R, Guilleminault C. MESAM 4: an ambulatory

device for the detection of patients at risk for obstruc-

tive sleep apnea syndrome (OSAS). Chest 1992;101:

1221–7.

[40] Schafer H, Ewig S, Hasper E, Luderitz B. Predictive

diagnostic value of clinical assessment and nonlabo-

ratory monitoring system recordings in patients with

symptoms suggestive of obstructive sleep apnea syn-

drome. Respiration 1997;64:194–9.

[41] Flemons WW, McNicholas WT. Clinical prediction of

the sleep apnea syndrome. Sleep Med Reviews 1997;

1:19–32.

[42] Gurubhagavatula I, Maislin G, Pack AI. An algorithm

to stratify sleep apnea risk in a sleep disorders clinic

population. Am J Respir Crit Care Med 2001;164:

1904–9.

[43] Flemons WW, Whitelaw WA, Brant R, Remmers JE.

Likelihood ratios for a sleep apnea clinical prediction

rule. Am J Respir Crit Care Med 1994;150:1279–85.

C.K. Li, W.W. Flemons / Clin Chest Med 24 (2003) 283–295 295

Monitoring respiration during sleep

Teofilo L. Lee-Chiong Jr, MD

Division of Pulmonary and Critical Care Medicine, University of Arkansas for Medical Sciences,

4301 West Markham, Slot 555, Little Rock, AR 72205, USA

Sleep-related breathing disorders

The sleep-related breathing disorders have been

categorized in various ways. The most basic sche-

ma divides them into obstructive or central apneic

events. An American Academy of Sleep Medicine

(AASM) Task Force Report published in 1999 defined

four separate syndromes associated with abnormal

respiratory events during sleep among adults,

namely, obstructive sleep apnea-hypopnea syndrome

(OSAHS), central sleep apnea-hypopnea syndrome,

Cheyne-Stokes breathing syndrome, and sleep hypo-

ventilation syndrome [1]. In this classification, the

upper airway resistance syndrome was not regarded

as a distinct syndrome; instead, respiratory event-

related arousals (RERAs) were considered part of the

syndrome of OSAHS.

Obstructive sleep apnea-hypopnea syndrome

OSAHS is characterized by repetitive reduction

or cessation of airflow during sleep caused by partial

or complete upper airway occlusion in the presence

of respiratory efforts. Mixed apnea, in which an

initial period of apnea caused by an absence of res-

piratory efforts precedes upper airway obstruction, is

included in this syndrome. These events are typically

accompanied by oxygen desaturation, arousals, and

sleep disruption.

Apnea is characterized by the cessation of airflow

for 10 seconds or longer. Although there is almost

universal consensus regarding the definition of apnea

in adults, the presence of hypopnea continues to be

identified using various criteria, including (1) a 50%

reduction in airflow accompanied by a 4% fall in

oxygen saturation (SaO2) or an arousal, (2) a 50%

reduction in airflow accompanied by any fall in SaO2,

or (3) any reduction in airflow with or without

oxygen desaturation or arousal [2].

The criteria used for scoring hypopneas influence

the diagnosis of OSAHS and the rating of its

severity. Different scoring criteria for hypopneas

may result in varying apnea-hypopnea indices [3].

Interpretation of polysomnographic records ideally

should include a description of the scoring method

used to derive hypopneas.

The sum of apneas and hypopneas divided by the

total sleep time is commonly referred to as the apnea-

hypopnea index. The respiratory disturbance index

(RDI) is the sum of apneas, hypopneas, and RERAs

divided by the total sleep time.

Estimates of the severity of sleep-disordered

breathing depend on the approach to measuring

RDI. Redline et al examined the relationships among

RDIs defined by different definitions of apneas and

hypopneas in 5046 participants in the Sleep Heart

Health Study who underwent overnight unattended

12-channel polysomnography. The correlation bet-

ween RDIs based on various definitions ranged from

0.99 to 0.68, and the magnitude of the median RDI

varied from 29.3 when it was based on events iden-

tified on the basis of flow or volume amplitude criteria

alone to 2 for an RDI that required a 5% oxygen

desaturation with events [4].

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00021-2

Portions of the text have appeared previously in Lee-

Chiong TL. Monitoring respiration during sleep. In: Lee-

Chiong TL, Sateia MJ, Carskadon MA, editors. Sleep

medicine. Philadelphia: Hanley and Belfus, Inc.; 2002.

E-mail address: leechiongteofilol@uams.edu

Clin Chest Med 24 (2003) 297–306

It is generally not necessary to distinguish apneas

from hypopneas in routine clinical care, and often the

two respiratory events are scored and reported togeth-

er. The diagnostic criteria for apneas and hypopneas

recommended by the AASM Task Force include a

reduction ( > 50%) in the amplitude of breathing from

baseline during sleep or a reduction ( < 50%) in the

amplitude of breathing from baseline during sleep

associated with either an oxygen desaturation (>3%)

or an arousal plus an event duration of at least

10 seconds [1]. RERAs, which do not fulfill the criteria

for either apnea or hypopnea, consist of increasing

respiratory efforts that last 10 seconds or longer and

culminate in an arousal or a progressively more

negative esophageal pressure preceding a change in

esophageal pressure to a less negative level.

The reference standard for measuring an obstruc-

tive apnea-hypopnea is a reduction in total oronasal

airflow detected by a pneumotachometer placed in a

well-fitted facemask [1]. Other methods used to

identify obstructive apnea-hypopneas include mea-

surement of nasal pressure, respiratory inductance

plethysmography (RIP), piezo sensors, strain gauges,

thoracic impedance, thermal sensors, and expired

carbon dioxide (CO2). Whereas measurement tech-

niques that identify apneas also are able to detect

hypopneas, methods that measure hypopneas may

not necessarily be adequate in identifying apneic

events. The reference standard for identifying a RERA

is themeasurement of esophageal pressure [1]. RERAs

also can be detected using measurements of nasal

pressure and surface diaphragmatic electromyography.

The demonstration of five or more obstructive

apneas-hypopneas or RERAs per hour of sleep dur-

ing an overnight study, plus excessive daytime sleepi-

ness (that is not caused by other factors) or two or

more of the following manifestations, including

choking or gasping during sleep, recurrent awaken-

ings from sleep, unrefreshing sleep, daytime fatigue,

or impaired concentration, establishes the diagnosis

of OSAHS [1].

Central sleep apnea-hypopnea syndrome

This syndrome is characterized by repetitive epi-

sodes of sleep-related apnea unaccompanied by upper

airway obstruction. Each respiratory event consists of

reduced airflow, 10 seconds or longer in duration,

associated with a reduction in esophageal pressure

excursions from baseline levels and often with oxy-

gen desaturation and arousals.

The diagnostic criteria for central sleep apnea-

hypopnea syndrome consist of (1) excessive day-

time sleepiness or frequent arousals/awakenings,

and (2) at least five central apnea-hypopneas per hour

of sleep during an overnight study, and (3) awake

arterial carbon dioxide tension (PaCO2) of less than

45 mm Hg [1].

Esophageal pressure monitoring is the reference

standard measurement of central apnea-hypopneas

[1]. Other methods, such as RIP, surface diaphrag-

matic electromyography, thermal sensors, expired

CO2, piezo sensors and strain gauges, are relatively

insensitive in identifying these events.

Cheyne-Stokes breathing syndrome

In this syndrome, cyclical waxing and waning of

respiration develops, with central apnea or hypopnea

alternating with hyperpnea. Transient arousals that

occur at the crest of hyperpnea may lead to sleep

fragmentation and excessive somnolence.

The reference standards of measuring airflow and

respiratory effort are pneumotachometry and esopha-

geal pressure monitoring, respectively [1]. Other tech-

niques for detecting Cheyne-Stokes breathing include

RIP, surface diaphragmatic electromyography, oro-

nasal airflow monitoring, and oximetry. Cheyne-

Stokes breathing syndrome is diagnosed based on

the following criteria: (1) presence of congestive heart

failure or cerebral neurologic disorders, (2) three or

more consecutive cycles of respiratory irregularity

characterized by crescendo-decrescendo amplitude of

breathing lasting at least 10 consecutive minutes, and

(3) five or more central apnea-hypopneas per hour of

sleep [1].

Sleep hypoventilation syndrome

Persons with sleep hypoventilation syndrome may

have oxygen desaturation and hypercapnia during

sleep unrelated to distinct periods of apnea-hypopnea.

Periods of hypoventilation are more frequent and

severe during rapid eye movement sleep than in

non–rapid eye movement sleep. PaCO2 monitoring

is the reference standard measurement for identifying

sleep hypoventilation [1]. Continuous oximetry

(demonstrating a decline in SaO2 without accom-

panying respiratory events), transcutaneous carbon

dioxide (PtcCO2) monitoring, calibrated RIP (show-

ing reduced tidal volume and minute ventilation), and

end-tidal carbon dioxide (PetCO2) measurements also

have been used to monitor sleep hypoventilation. The

diagnosis of sleep hypoventilation syndrome is based

on the presence of cor pulmonale, pulmonary hyper-

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306298

tension, excessive somnolence not secondary to other

factors, erythrocytosis or awake PaCO2 of more than

45 mm Hg, and an increase in PaCO2 during sleep by

more than 10 mm Hg compared with levels during

wakefulness or sleep-related oxygen desaturation not

caused by apnea-hypopnea [1].

Monitoring respiration during sleep

Accurate monitoring of respiration during sleep,

including measurements of airflow, respiratory effort,

oxygenation, and ventilation, is indispensable in

identifying sleep-disordered breathing.

Measurement of respiratory effort

Measurement of respiratory effort using either

esophageal pressure monitoring or surface diaphrag-

matic electromyography is vital in distinguishing

central from obstructive apneas.

Esophageal pressure

Changes in pleural pressure accompany respi-

ratory effort. Esophageal pressure monitoring during

polysomnography, using either esophageal balloons

or newer catheter transducers, is considered the

reference standard for detecting respiratory effort

during sleep and is a direct measure of respiratory

load [1]. This method requires a transnasal insertion

of an esophageal catheter with a pressure transducer

placed on its tip after topical anesthesia of the nares

and pharynx. During episodes of RERAs in patients

with upper airway resistance syndrome, esophageal

pressures become increasingly more negative imme-

diately preceding an arousal, followed by a rapid

return to baseline levels [1]. Virkkula et al reported

that esophageal pressure monitoring improved the

diagnostic value of limited polygraphic recording of

oxygen saturation, respiratory and leg movements,

airflow, body position, and snoring in detecting sleep-

disordered breathing [5].

Transnasal insertion of esophageal catheters in

sleep studies may increase ipsilateral nasal resistance,

as measured by anterior rhinomanometry, but does not

affect combined nasal resistance [6]. Changes in nasal

pressure and airflow during esophageal pressure moni-

toring may be particularly relevant in persons with

already compromised nasal airflow. The amount of

apneas and arousals has been shown to increase with

nasal airflow obstruction. The use of nasoesophageal

catheters is generally associated with only minimal

changes in sleep architecture [5]. Patient compliance

with esophageal catheter is generally good [5].

Surface diaphragmatic electromyography

Although the presence of respiratory efforts may

be inferred by analysis of signal tracings derived from

electrodes placed on the chest wall during poly-

somnography, surface diaphragmatic electromyogra-

phy by itself is seldom helpful in detecting RERAs or

central apnea-hypopneas [1].

Measurement of airflow

Airflow during sleep can be measured either

directly or indirectly. The only method that measures

airflow directly is pneumotachography. Thermal sen-

sors and PetCO2 monitors detect changes in the

thermal and chemical characteristics of inspired

ambient air and expired air originating from the

airways; both methods provide only an indirect

estimate of airflow [7].

Although indirect methods of measuring airflow

can detect episodes of apnea reliably, they are less

consistent in identifying hypopneas. Simultaneous

measurement of lung volume or effort and thermal

or PetCO2 sensors is required to distinguish among

central apneas, obstructive apneas, and a prolonged

inspiration [7].

Pneumotachometer

A pneumotachometer, attached to a well-fitted

facemask, can measure total oronasal airflow by

detecting changes in pressure between inspiration

and expiration and is the reference standard for

measuring airflow [1]. Patient discomfort from a

tightly fitting facemask may disturb sleep and limit

its use in clinical sleep studies.

Nasal pressure

Nasal airflow can be measured quantitatively and

directly with a pneumotachograph that detects

changes in nasal pressure during respiration. Nasal

airway pressure decreases during inspiration and

increases during expiration. The fluctuations pro-

duced on the transducer signals are proportional to

flow [8]. The device consists of a standard oxygen

nasal cannula connected to a pressure transducer and

placed in the nares.

The shape and amplitude of signals obtained from

a nasal cannula are comparable to those from a

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306 299

facemask pneumotachograph [9]. A plateau on the

inspiratory flow signal is associated with increased

upper airway resistance and airflow limitation. In one

study, airway resistance was increased for breaths

with flattened or intermediate inspiratory flow signal

contours compared with breaths with normal flow

contours [8].

Measurement of pressure by nasal prongs is

superior to the use of thermistors in detecting respi-

ratory events during sleep studies [9]. Nasal cannula/

pressure sensors may recognize additional events

characterized by flow limitation that are missed by

thermistors [10]. Nasal pressure monitoring is not

recommended for persons who are predominantly

mouth breathers or who have nasal obstruction

[7,11]. In persons with narrow nares or a deviated

septum, nasal prongs used to assess nasal flow during

sleep can increase nasal airflow resistance—as esti-

mated by posterior rhinomanometry—and possibly

alter the diagnosis of OSAHS and its severity [12].

Nasal prongs that partly occlude the nasal passages

can cause sleep breathing disorders associated with

brief arousals. Thurnheer et al observed that com-

pared with facemask pneumotachography, nasal can-

nula pressure recordings provided accurate clinical

assessment of ventilation during sleep even in pa-

tients who reported nasal obstruction [13].

Thermal sensors

Thermal sensors (thermistors or thermocouples)

afford an indirect and semiquantitative measurement

of airflow. These devices are placed over the nose and

mouth and infer airflow by sensing differences in the

temperature of the warmer expired air and the cooler

inhaled ambient air. The flow signal generated is

related directly to the sensor temperature and indi-

rectly to airflow. Unfortunately, temperature changes

of respiratory air often bear little correlation to air-

flow. The flow signal also is influenced by the pattern

of airflow and the placement of the sensor in relation

to the nostril. Even minor displacements of the

thermal sensors or alternations in the proportion of

nasal and oral breathing relative to the sensor position

can lead to large changes in signal amplitude [14].

Although temperature-sensing receptors can de-

tect apneas reliably, they are less accurate in iden-

tifying hypopnea [10]. Farre et al noted that thermal

sensors were imprecise in monitoring airflow and,

when a reduction in thermal sensor signal is used

to quantify hypopneas, they tend to underestimate

hypopneic events [15]. Thermistors do not allow the

detection of inspiratory flow limitation, which is

suggestive of upper airway narrowing.

Oronasal thermistors are typically located at the

upper lip; in this location, thermistors may be unable

to differentiate between high and low rates of airflow

and detect hypopneas. Akre et al introduced the use

of internal thermistors to measure airflow in the

pharynx. They reported that this method was more

sensitive than external thermistors in detecting minor

changes in air flow and hypopneas [16,17]. In awake,

normal subjects, the reliability of internal thermistors

in diagnosing hypopneas is comparable to that of

pneumotachography [18].

In summary, signals obtained from thermocouples

and thermistors provide only qualitative data regard-

ing airflow, and as a rule, thermal sensors are unable

to identify reliably the presence of hypopnea and

cannot distinguish central from obstructive apnea-

hypopneas [1].

Expired carbon dioxide

Ambient air contains negligible amounts of CO2

compared with expired air from the lungs, which has

a higher concentration of CO2. A qualitative measure

of airflow can be obtained using infrared analyzers of

expired CO2 placed in front of the nose and mouth.

An advantage of PetCO2 monitoring over thermal

sensing techniques is its ability to infer the occur-

rence of hypoventilation by a rising PetCO2 level.

Minute fluctuations in lung volume that accom-

pany each heart beat also may be transmitted to the

sensor via a patent upper airway during central

apneas [7]. These fluctuations may appear as cardiac

oscillations in the CO2 tracings, further corroborating

the diagnosis of central apneas.

Tracheal sound recording

Tracheal sound recordings, made by using a

stethoscope head taped over the manubrium sternum

and air-coupled to a microphone, have been proposed

as a method of detecting and monitoring airflow. This

method is limited by interference from environmental

noise [19].

Strain gauges

Rib cage and abdominal excursions can be mea-

sured by placing length-sensitive strain gauges below

the axilla and at the level of the umbilicus, respec-

tively [20]. Respiratory movements can be detected

by a single uncalibrated abdominal or chest gauge.

Calibration of the rib cage and abdominal gauges

against another volume-measuring device is required

to measure volume changes quantitatively. The

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306300

summed rib cage-abdominal volume signals do not

distinguish central events (no net volume change

caused by absence of respiratory effort) from ob-

structive sleep apnea (no net volume change caused

by rib cage-abdominal paradox). Loss of tone of the

diaphragm or the accessory respiratory muscles also

can lead to paradoxical motion of the rib cage and

abdomen [7]. Esophageal pressure monitoring may

be needed to verify respiratory efforts whenever most

apneas detected by strain gauges appear central in

origin [20].

Displacement of the strain gauges during the

monitoring period because of changes in sleep posi-

tion or body movements influences signal quality

[20]. Accuracy of measurements is affected by over-

stretching or understretching of the gauges and altera-

tions in muscle tone during sleep [7].

Respiratory inductance plethysmography

Respiratory inductance plethysmography (RIP)

can be used to measure changes semi-quantitatively

in chest and abdominal volume during respiration.

Transducers are placed around the chest and abdomen

to monitor changes in the cross-sectional area of

the respective body compartments as reflected by

changes in inductance (resistance to change in flow

of current) of the transducers [7]. RIP is based on the

principle of a two-compartment model of thoraco-

abdominal wall movement during respiration [21].

With a closed glottis, the sum of chest and abdominal

volume is fixed, and any increase or loss of volume

of the rib cage is accompanied by a simultaneous,

equal but opposite change in volume of the abdomen

[22]. The sum of the signals from calibrated chest and

abdominal sensors can estimate tidal volume and

respiratory pattern during sleep but cannot provide

data regarding airflow [11].

Thoracoabdominal asynchrony during breathing is

currently most commonly identified by visual analy-

sis of records. Brown et al described a novel auto-

mated analysis approach using a recursive linear

regression to identify synchrony or asynchrony

between ribcage and abdominal movements during

breathing in 15 infants [23]. Paradoxical ribcage

motion also can be assessed by measuring thoracic

delay based on the degree to which peaks in ribcage

and abdominal signals are synchronized in time [23].

Hypopneas could be scored reproducibly using

RIP to monitor thoracoabdominal movement with or

without a simultaneous flow sensor signal [24].

Hypopnea is scored if there is a at least a 50%

reduction of RIP sum from baseline of either cali-

brated or uncalibrated signals; at least a 50% reduc-

tion from baseline in chest and abdominal signals

(dual channel) in the absence of an RIP sum; or more

than a 50% reduction from baseline or less than a

50% reduction from baseline accompanied by either

an arousal or an oxygen desaturation (� 3%) in either

chest or abdominal signal (single channel) [1].

The accuracy of RIP in monitoring the volume

and duration of respiration depends on its initial

calibration and the constancy of calibration with body

movements and changes in lung volumes [25]. Vari-

ous procedures, such as the simultaneous equation

method, isovolume maneuver method, and least

squares regression method, can be used to calibrate

RIP [25,26]. Displacements of the transducer bands

or alterations in posture during sleep can lead to

inaccuracies in measurements. Bands should be taped

firmly to the skin to avoid slippage during overnight

monitoring. Sleep-related thoracoabdominal distor-

tion or movement asynchrony also can affect accu-

racy of RIP measurements during sleep [26,27].

Thoracic impedance

Thoracic impedance can be used to measure

airflow qualitatively. Impedance varies with the

relative amount of conductive materials (body fluids

and tissue) and nonconductive air between a pair of

electrodes placed at opposite sides of the thoracic

cage. It decreases as the volume of conductive

material increases in proportion to air and vice

versa. The volume of air contained within the

thoracic cage during the different phases of respira-

tion can be estimated based on changes in recorded

impedance [7].

Measurement of snoring intensity

Another method that has been used to measure

airflow is measurement of snoring intensity. One

study demonstrated a linear correlation, albeit weak,

between snoring intensity and respiratory effort and

flow limitation during sleep [28].

Piezo sensors

Piezo sensors can monitor changes in airflow

qualitatively but cannot distinguish central apnea-

hypopneas from obstructive respiratory events [1].

Magnetometers

Respiratory magnetometer recordings of chest and

abdominal motion have been shown to be able to

distinguish between obstructive and central apneic

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306 301

events by differences in patterns of motion (ie,

paradoxical motion of the rib cage and abdomen with

obstructive events) [29]. The recordings also can be

used to monitor changes in body position during the

sleep study.

Canopy with a neck seal

The use of a canopy ventilation monitor to

measure ventilation quantitatively during sleep has

been described [30]. The device directly measures

gas flow using a pneumotachograph and consists of a

rigid canopy fitted over the head. It is sealed at

the neck, which creates an airtight enclosure through

which a continuous flow of air or oxygen is pro-

vided. Inflow of gas is kept equal to outflow. Airflow

is measured as respiration alters the flow in and out

of the canopy. Canopy ventilation monitoring has a

reported accuracy of approximately 92% in mea-

suring tidal volume [30].

Flow-volume loop analysis

The presence of airway obstruction during wake-

fulness and sleep can be inferred by analyzing

abnormalities of the flow-volume loop. Flow limita-

tion and an elevated upper airway resistance are

suggested by the presence of a plateau (normally

rounded) on the contour of the inspiratory flow

tracing obtained during continuous positive pressure

(CPAP) therapy for OSAHS. In one study, breath-by-

breath analysis of the flow-volume curve of a tidal

breath was accurate in identifying inspiratory flow

limitation during sleep in persons with OSAHS on

CPAP therapy [31]. Inspiratory flow limitation was

defined by the presence of an inspiratory plateau or

reduction in inspiratory flow independent of any

increase in inspiratory efforts.

Cardiac oscillometry

Small oscillations at cardiac frequency may be

appreciated in the airflow signal tracing during epi-

sodes of central apnea. These cardiogenic oscillations

are believed to be related to persistence of airway

patency possibly coupled with relaxation of the

thoracic muscles during central apneas [32].

Air mattress

Chow et al described the use of an air mattress

system that consists of multiple air compartments to

monitor noninvasively thoracic and abdominal move-

ments separately. The sensitivity and accuracy rates

of the air mattress for detecting hypopnoeas were

above 90% compared with respiratory inductive

phlethysmography [33].

Measurement of oxygenation and ventilation

Oxygenation and ventilation change rapidly dur-

ing sleep in patients with sleep-disordered breathing.

To be accurate and reliable, methods to assess oxy-

genation and ventilation must be capable of rapid and

repetitive measurements. Direct measurements of

arterial oxygen tension (PaO2), arterial carbon diox-

ide tension (PaCO2), and SaO2 via arterial blood

sampling are more accurate than estimates derived

from noninvasive methods such as pulse oximetry,

transcutaneous oxygen tension (PtcO2) measurement,

transcutaneous carbon dioxide tension (PtcCO2)

measurement, or airway CO2 (PetCO2) monitoring.

Arterial blood gas sampling provides only a static

measure of oxygenation and ventilation rather than a

continuous monitoring, however. Repetitive sampling

of arterial blood during sleep studies is painful, time

consuming, inconvenient, expensive, and intrusive of

sleep and is associated with more complications than

noninvasive assessments.

Pulse oximetry

With pulse oximetry, a pulsating vascular bed

(eg, earlobe or fingertip) is placed between a two-

wavelength light source and a sensor. This arrange-

ment is designed to eliminate any artifact that might

originate from absorption of light by venous blood or

tissue [34].

Pulse oximeters are used routinely during over-

night polysomnography to monitor SaO2. They are

easy to use, portable, relatively inexpensive, readily

available, noninvasive, respond rapidly to changes in

SaO2, and allow continuous monitoring of SaO2 [7].

Several factors influence the accuracy and reli-

ability of pulse oximetry. Pulse oximetry response

time can be affected by changes in heart rate and

circulation time. Altering the pulse oximeter response

time influences the accuracy of pulse oximeters in

measuring changes in SaO2. For instance, SaO2

recordings may be inaccurate if the oximeter response

time approximates the duration of oxygen desatura-

tion events. In one study that involved subjects with

severe OSAHS, increasing the pulse oximeter aver-

aging time from 3 seconds to 12 and 21 seconds

resulted in significant differences in the measured

SaO2, with underestimation of oxygen desaturation

by up to 60% [35].

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306302

SaO2 measurement and response characteristics

using pulse oximetry also vary with sensor location

(eg, earlobe or fingertip) and type [36]. Finally,

sensitivity of pulse oximetry is greater with shorter

sampling intervals, and the least filtering to achieve

the most rapid response is recommended [7,37].

Several factors limit the use of oximetry in the

evaluation of persons with sleep-disordered breath-

ing. Oximetry alone is inadequate in persons without

oxygen desaturation [37]. The presence of dyshemo-

globin species, such as carboxyhemoglobin or met-

hemoglobin, produces errors in measurement because

of its reliance on only two light wavelengths [7].

Reduced skin perfusion caused by hypothermia,

hypotension, or vasoconstriction and by poor sensor

attachment may alter signal amplitude [14]. Finally,

oximetry readings may overestimate low oxygen

saturation values [38].

As a screening test for OSAHS, nocturnal pulse

oximetry has a reported sensitivity rate of 69% and a

specificity rate of 97%. Accuracy was decreased in

persons who had higher awake baseline SaO2, were

less overweight, and had milder disease [39]. Yama-

shiro and Kryger noted that nocturnal oximetry may

not be able to detect breathing disorders during sleep

with sufficient sensitivity and specificity and is inef-

fective in identifying other disorders of sleep [40]. In

another study that compared clinical assessment,

unsupervised home oximetry, and formal poly-

somnography in the diagnosis of OSAHS, clinical

assessment was superior to home oximetry analyzed

by counting the number of recorded arterial oxygen

desaturations [41].

Epstein et al compared polysomnography to two

patterns of oxyhemoglobin desaturation used as a

method of screening for OSAHS: (1) a ‘‘deep’’

pattern that consisted of more than 4% fall in SaO2

to less than or equal to 90% and (2) a ‘‘fluctuating’’

pattern that consisted of repetitive, brief drops in

SaO2 [42]. As screening tools for sleep-disordered

breathing, the ‘‘deep’’ pattern had greater specificity

and positive predictive value and the ‘‘fluctuating’’

pattern had a greater sensitivity and negative predic-

tive value. For mild disease, screening nocturnal

oximetry using the ‘‘fluctuating’’ pattern is less

sensitive compared with polysomnography, with

61% of patients with abnormal polysomnographic

studies having normal oximetry results [42].

Transcutaneous oxygen monitoring

Oxygen tension at the skin surface (PtcO2), which

is measured using a modified Clark electrode, is

influenced by cutaneous perfusion, temperature, and

metabolism. The application of PtcO2 monitoring

during adult polysomnography is limited by the

variable relationship between PaO2 and PtcO2 and

its slow response time that fails to mirror rapid

changes in PaO2. It requires meticulous skin prepara-

tion. Blood flow to the skin can be increased by local

application of heat, with periodic site changes every

4 to 6 hours to prevent cutaneous thermal injury [37].

A delay in recording in the warm-up period after site

changes is expected [37].

Transcutaneous carbon dioxide

Transcutaneous carbon dioxide (PtcCO2) refers to

the CO2 tension at the epidermal surface. It can be

monitored noninvasively and continuously during

sleep using a silver chloride electrode or an infrared

capnometer. PtcCO2 monitoring may provide useful

information during pediatric polysomnography

because pediatric OSAHS is associated with partial

airway obstruction, alveolar hypoventilation, and

hypercarbia. PtcCO2 monitoring is most commonly

used in neonates. It requires meticulous skin pre-

paration and arterial blood gas sampling for cali-

bration [37].

Among adults, PtcCO2 often differs significantly

from a simultaneously obtained PaCO2 [1,43]. Rou-

tine PtcCO2 monitoring has minimal clinical use

during adult polysomnography. Its slow response time

makes it unsuitable for monitoring blood gas tensions

during sleep, in which rapid and short-lasting changes

can occur [7]. PtcCO2 monitoring may be of some

use in adults with waking hypercapnia or suspected

sleep-related alveolar hypoventilation.

Expired end tidal carbon dioxide

Airway carbon dioxide (PetCO2) measured at the

end of a complete expiration is related to PaCO2.

PetCO2 can be monitored continuously during poly-

somnography using infrared spectrophotometers or

respiratory mass spectrometers. PetCO2 measure-

ments are affected by conditions that alter the rela-

tionships among ventilation, perfusion, and PaCO2

[38]. PetCO2 may underestimate PaCO2 when dead

space to tidal volume ratio is increased during sleep

because of a reduction in tidal volume. PetCO2

measurements using facemasks or nasal cannula or

during nasal CPAP ventilation may not reflect PaCO2

reliably because of gas dilution with room air or

continuous gas leakage via the CPAP mask, respec-

tively. Hypoventilation, mouth breathing, or concom-

itant use of supplemental oxygen therapy also can

give rise to inaccuracies in measurement [37,43].

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306 303

In one study, neither PetCO2 nor PtcCO2 accurately

or consistently reflected simultaneously recorded

PaCO2 values during polysomnography in persons

who were spontaneously breathing room air, re-

ceiving supplemental oxygen given via nasal cannula,

or receiving nocturnal positive pressure ventilatory

assistance [43].

Newer approaches

Pulse transit time analysis

Blood pressure fluctuates during sleep in persons

with OSA. Blood pressure transiently increases dur-

ing arousals from sleep and falls during inspiration.

Davies et al reported that the degree of inspiratory fall

in blood pressure progressively increased from nor-

mal sleep, through snoring, to obstructive respiratory

events. The frequency of arousal-related increases in

blood pressure also rose during obstructive apnea and

during snoring accompanied by arousals [44].

Pulse transit time (PTT) is the transmission time

for the arterial pulse pressure wave to travel from the

aortic valve to the periphery. It is measured using

electrocardiography as the interval between the

R-wave and the subsequent pulse shock wave

detected at the finger. PTT is typically approximately

250 milliseconds. The speed of the shock wave is

affected by the stiffness of the arterial walls and blood

pressure. PTT is inversely related to blood pressure: as

blood pressure rises, PTT falls because of increases in

arterial wall stiffness and pulse wave speed. PTT

increases during inspiratory falls in blood pressure

and decreases during arousal-induced increases in

blood pressure [45].

With esophageal pressure as a reference, PTT has

been reported to have high sensitivity and specificity

rates in distinguishing between central and obstruc-

tive apnea-hypopnea [46]. Among persons with

OSAHS, PTT studies also have been demonstrated

to differentiate reliably between persons who require

nasal CPAP and persons who do not [47].

Forced oscillation technique

Forced oscillation technique has been proposed as

a method for detecting upper airway obstruction

during sleep and titrating CPAP therapy [48–51].

This technique is a noninvasive measure of input

impedance of the respiratory system that uses high-

frequency pressure oscillation to the upper airway

[49]. Forced oscillation techniques are able to par-

tition reliably the airway component of respiratory

impedance from that of lung tissue [50]. This tech-

nique does not require patient cooperation and may

prove useful for assessing uncooperative patients.

Contrary to earlier concerns, Badia et al observed

that the use of forced oscillation technique does not

alter upper airway muscle tone or affect electroence-

phalographic variables [49]. This novel approach

requires further standardization before it can be used

in clinical sleep studies [50].

Steltner et al evaluated the performance of a new

algorithm for automated detection and classification

of apneas and hypopneas based on time series analy-

sis of nasal mask pressure and a forced oscillation

signal related to respiratory input impedance [52].

They noted no significant difference in the variability

and discrepancy between automated analysis and

visual analysis of standard polysomnographic signals.

Acknowledgment

The author wishes to thank Grace Zamudio for her

assistance in the preparation of the manuscript.

References

[1] American Academy of Sleep Medicine Task Force.

Sleep-related breathing disorders in adults: recommen-

dations for syndrome definition and measurement tech-

niques in clinical research. Sleep 1999;22:667–89.

[2] Phillips BA, Anstead MI, Gottlieb DJ. Monitoring

sleep and breathing: methodology. Part I: Monitoring

breathing. Clin Chest Med 1998;1:203–12.

[3] Manser RL, Rochford P, Pierce RJ, Byrnes GB, Camp-

bell DA. Impact of different criteria for defining hypo-

pneas in the apnea-hypopnea index. Chest 2001;120:

909–14.

[4] Redline S, Kapur VK, Sanders MH, Quan SF, Gottlieb

DJ, Rapoport DM, et al. Effects of varying approaches

for identifying respiratory disturbances on sleep apnea

assessment. Am J Respir Crit Care Med 2000;161:

369–74.

[5] Virkkula P, Silvola J, Maasilta P, Malmberg H, Salmi

T. Esophageal pressure monitoring in detection of

sleep-disordered breathing. Laryngoscope 2002;112:

264–70.

[6] Virkkula P, Silvola J, Lehtonen H, Salmi T, Malmberg

H. The effect of esophageal pressure monitoring on

nasal airway resistance. Otolaryngol Head Neck Surg

2001;125:261–4.

[7] Kryger MH. Monitoring respiratory and cardiac func-

tion. In: Kryger MH, Roth T, Dement WC, editors.

Principles and practice of sleep medicine. 3rd edition.

Philadelphia: WB Saunders Co.; 2000. p. 1217–30.

[8] Hosselet JJ, Norman RG, Ayappa I, Rapoport DM.

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306304

Detection of flow limitation with a nasal cannula/pres-

sure transducer system. Am J Respir Crit Care Med

1998;157:1461–7.

[9] Ruhle KH, Fahrner A, Randerath W. Evaluating oro-

nasal flow with temperature (thermistor) and obstruc-

tive pressure (prongs). Pneumologie 2001;55:4–6.

[10] Norman RG, Ahmed MM, Walsleben JA, Rapoport

DM. Detection of respiratory events during NPSG:

nasal cannula/pressure sensor versus thermistor. Sleep

1997;20:1175–84.

[11] Series F, Marc I. Nasal pressure recording in the diag-

nosis of sleep apnoea hypopnoea syndrome. Thorax

1999;54:506–10.

[12] Lorino AM, Lorino H, Dahan E, d’Ortho MP, Coste A,

Harf A, et al. Effects of nasal prongs on nasal airflow

resistance. Chest 2000;118:366–71.

[13] Thurnheer R, Xie X, Bloch KE. Accuracy of nasal

cannula pressure recordings for assessment of ventila-

tion during sleep. Am J Respir Crit Care Med 2001;

164:1914–9.

[14] Berg S, Haight JSJ, Yap V, Hoffstein V, Cole P. Com-

parison of direct and indirect measurements of respira-

tory airflow: implications for hypopneas. Sleep 1997;

20:60–4.

[15] Farre R, Montserrat JM, Rotger M, Ballester E, Nava-

jas D. Accuracy of thermistors and thermocouples as

flow-measuring devices for detecting hypopneas. Eur

Respir J 1998;11:179–82.

[16] Akre H, Skatvedt O. Advantages of measuring air flow

in the pharynx with internal thermistors. Eur Arch Oto-

rhinolaryngol 2000;257:251–5.

[17] Akre H, Skatvedt O, Borgersen AK. Diagnosing res-

piratory events and tracing air flow by internal thermis-

tors. Acta Otolaryngol 2000;120:414–9.

[18] Akre H, Borgersen AK, Mair IW, Skatvedt O. Tracing

air flow and diagnosing hypopnoeas in normal sub-

jects. Physiol Meas 2000;21:221–7.

[19] Cummiskey J, Williams TC, Krumpe PE, Guillemi-

nault C. The detection and quantification of sleep ap-

nea by tracheal sound recordings. Am Rev Respir Dis

1982;126:221–4.

[20] BoudewynsA,WillemanM,WagemansM,DeCockW,

Van de Heyning P, De Backer W. Assessment of respi-

ratory effort by means of strain gauges and esophageal

pressure swings: a comparative study. Sleep 1997;20:

168–70.

[21] Konn K, Mead J. Measurement of the separate volume

changes of rib cage and abdomen during breathing.

J Appl Physiol 1967;22:407–22.

[22] Staats BA, Bonekat HW, Harris CD, Offord KP. Chest

wall motion in sleep apnea. Am Rev Respir Dis 1984;

130:59–63.

[23] Brown KA, Platt R, Bates JHT. Automated analysis of

paradoxical ribcage motion during sleep in infants. Pe-

diatr Pulmonol 2002;33:38–46.

[24] Whyte KF, Allen MB, Fitzpatrick MF, Douglas NJ.

Accuracy and significance of scoring hypopneas. Sleep

1992;15:257–60.

[25] Chadha TS, Watson H, Birch S, Jenouri GA, Schneider

AW, Cohn MA, et al. Validation of respiratory induc-

tive plethysmography using different calibration proce-

dures. Am Rev Respir Dis 1982;125:644–9.

[26] Cantineau JP, Escourrou P, Sartene R, Gaultier C,

Goldman M. Accuracy of respiratory inductive pleth-

ysmography during wakefulness and sleep in pa-

tients with obstructive sleep apnea. Chest 1992;102:

1145–51.

[27] Whyte KF, GuggerM, Gould GA,Molloy J,Wraith PK,

Douglas NJ. Accuracy of respiratory inductive plethys-

mograph in measuring tidal volume during sleep. J Appl

Physiol 1991;71:1866–71.

[28] Stoohs R, Skrobal A, Harter R, Guilleminault C. Snor-

ing sound level as a predictor of flow limitation. Am

Rev Respir Dis 1992;145:A724.

[29] Sharp JT, Druz WS, Foster JR, Wicks MS, Chokrov-

erty S. Use of the respiratory magnetometer in diag-

nosis and classification of sleep apnea. Chest 1980;77:

350–3.

[30] Sorkin B, Rapoport DM, Falk DB, Goldring RM. Can-

opy ventilation monitor for quantitative measurement

of ventilation during sleep. J Appl Physiol 1980;48:

724–30.

[31] Series F, Marc I. Accuracy of breath-by-breath analysis

of flow-volume loop in identifying sleep-induced flow-

limited breathing cycles in sleep apnoea-hypopnoea

syndrome. Clin Sci 1995;88:707–12.

[32] Ayappa I, Norman RG, Rapoport DM. Cardiogenic

oscillations on the airflow signal during continuous

positive airway pressure as a marker of central apnea.

Chest 1999;116:660–6.

[33] Chow P, Nagendra G, Abisheganaden J, Wang YT.

Respiratory monitoring using an air-mattress system.

Physiol Meas 2000;21:345–54.

[34] Yelderman M, New W. Evaluation of pulse oximetry.

Anesthesiology 1983;59:349–52.

[35] Farre R, Montserrat JM, Ballester E, Hernandez L,

Rotger M, Navajas D. Importance of the pulse oxim-

eter averaging time when measuring oxygen desatura-

tion in sleep apnea. Sleep 1998;21:386–90.

[36] West P, George CF, Kryger MH. Dynamic in vivo

response characteristics of three oximeters: Hewlett-

Packard 47201A, Biox III, and Nellcor N-100. Sleep

1987;10:263–71.

[37] Clark JS, Votteri B, Ariagno RL, Cheung P, Eichhorn

JH, Fallat RJ, et al. Noninvasive assessment of blood

gases. Am Rev Respir Dis 1992;145:220–32.

[38] Kesten S, Chapman KR, Rebuck AS. Response char-

acteristics of a dual transcutaneous oxygen/carbon di-

oxide monitoring system. Chest 1991;99:1211–5.

[39] Collard P, Aubert G, Rodenstein DO. Value of noctur-

nal pulse oximetry as a screening tool for sleep ap-

nea-hypopnea syndrome. Am Rev Respir Dis 1992;

145:A724.

[40] Yamashiro Y, Kryger MH. Nocturnal oximetry: is it

a screening tool for sleep disorders? Sleep 1995;18:

167–71.

[41] Gyulay S, Olson LG, Hensley MJ, King MT, Allen

KM, Saunders NA. A comparison of clinical assess-

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306 305

ment and home oximetry in the diagnosis of obstruc-

tive sleep apnea. Am Rev Respir Dis 1993;147:50–3.

[42] Epstein LJ, Dorlac GR. Cost-effectiveness analysis of

nocturnal oximetry as a method of screening for sleep

apnea-hypopnea syndrome. Chest 1998;113:97–103.

[43] Sanders MH, Kern NB, Costantino JP, Stiller RA,

Strollo PJ, Studniki KA, et al. Accuracy of end-tidal

and transcutaneous PCO2 monitoring during sleep.

Chest 1994;106:472–83.

[44] Morielli A, Desjardins D, Brouillette RT. Transcutane-

ous and end-tidal carbon dioxide pressures should be

measured during pediatric polysomnography. Am Rev

Respir Dis 1993;148:1599–604.

[45] Davies RJO, Vardi-Visy K, Clarke M, Stradling JR.

Identification of sleep disruption and sleep disordered

breathing from systolic blood pressure profile. Thorax

1993;48:1242–7.

[46] Argod J, Pepin JL, Levy P. Differentiating obstructive

and central sleep respiratory events through pulse

transit time. Am J Respir Crit Care Med 1998;158:

1778–83.

[47] Pitson DJ, Sandell A, van der Hout R, Stradling JR.

Use of pulse transit time as a measure of inspiratory

effort in patients with obstructive sleep apnoea. Eur

Respir J 1995;8:1669–74.

[48] Navajas D, Farre R. Forced oscillation technique: from

theory to clinical applications. Monaldi Arch Chest Dis

2001;56:555–62.

[49] Badia JR, Farre R, Rigau J, Uribe ME, Navajas D,

Montserrat JM. Forced oscillation measurements do

not affect upper airway muscle tone or sleep in clinical

studies. Eur Respir J 2001;18:335–9.

[50] MacLeod D, Birch M. Respiratory input impedance

measurement: forced oscillation methods. Med Biol

Eng Comput 2001;39:505–16.

[51] Reisch S, Daniuk J, Steltner H, Ruhle KH, Timmer J,

Guttmann J. Detection of sleep apnea with the forced

oscillation technique compared to three standard poly-

somnographic signals. Respiration 2000;67:518–25.

[52] Steltner H, Staats R, Timmer J, Vogel M, Guttmann J,

Matthys H, et al. Diagnosis of sleep apnea by auto-

matic analysis of nasal pressure and forced oscillation

impedance. Am J Respir Crit Care Med 2002;165:

940–4.

T.L. Lee-Chiong Jr / Clin Chest Med 24 (2003) 297–306306

Indications for treatment of obstructive sleep apnea in adults

Patrick J. Strollo, Jr, MD, FCCP

Division of Pulmonary, Allergy, and Critical Care Medicine, University of Pittsburgh School of Medicine,

University of Pittsburgh Medical Center, Montefiore Hospital, Suite 628 West, 3459 Fifth Avenue, Pittsburgh,

PA 15213-2582, USA

Why treat obstructive sleep apnea (OSA)? OSA

is associated with significant daytime sleepiness,

reduced quality of life, insulin resistance, motor

vehicle crashes, and vascular morbidity and mortality

[1–3]. Current evidence supports the belief that all

these parameters can be impacted favorably by treat-

ment. Medical therapy with positive pressure elimi-

nates snoring and favorably affects daytime sleepiness,

driving risk, vascular function, vascular risk, and

quality of life [4–8]. The conundrum for the clinician

is that patients are variably affected by OSA of similar

severity (Fig. 1). Treatment may be difficult to accept

or adhere to, and some treatment options are not

uniformly effective. The long-term impact of treatment

is uncertain.

The current convention is to grade the severity of

OSA by the apnea-hypopnea index (AHI). The Amer-

ican Academy of Sleep Medicine recommends grad-

ing sleep apnea as mild (AHI 5–15), moderate (AHI

15–30), and severe (AHI > 30) [9]. This metric sta-

tistically correlates the presence of sleepiness, neuro-

cognitive impairment, and vascular risk [10–12]. It is

relatively easy to treat patients with severe, symp-

tomatic OSA. The difficulty with regard to treatment

frequently occurs when patients with severe OSA are

not symptomatic or when patients are profoundly

symptomatic with a low AHI.

Treatment of the minimally symptomatic patient

with severe OSA can be challenging. The medical

therapy of choice—positive pressure via a mask—is

unique and not discrete [13]. The treatment is

administered in one of the most intimate settings,

the bedroom.

In the absence of definitive long-term outcome

data, there is uncertainty regarding how hard to push

therapy in patients with mild to moderate OSA with

minimal symptoms [14]. Patients who are profoundly

symptomatic with relatively mild OSA may not

accept positive pressure therapy. The long-term effect

of alternative treatments to positive pressure is un-

known but may be of value in select circumstances.

Patient assessment

Successful treatment cannot be accomplished with-

out proper patient assessment. It is helpful to under-

stand what a patient hopes to gain from the evaluation.

This expectation is best handled by seeing the patient

before polysomnography. The clinician can under-

stand what is driving the evaluation: the complaint of

snoring, the complaint of fatigue or daytime sleepi-

ness, or the concern of vascular risk. It is also helpful to

understand up front whether the patient, spouse, or

referring physician is most concerned about OSA.

If the patient is most concerned with the pos-

sibility of OSA and he or she is subjectively sleepy,

there is a good chance that medical therapy with

positive pressure will be accepted. These patients are

good candidates for split-night polysomnography

[15–17]. If the patient does not complain of daytime

fatigue or sleepiness or does not regard snoring as a

significant problem, acceptance and adherence to

positive pressure therapy may be difficult to estab-

lish, and split-night polysomnography may not be

the best approach [18,19]. In this circumstance, it is

generally best to obtain a full night of diagnostic

polysomnography data and review the findings

before a trial of positive pressure.

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00025-X

E-mail address: strollopj@msx.upmc.edu

Clin Chest Med 24 (2003) 307–313

The clinician must know if insufficient sleep or

depression contribute to the complaint of daytime

sleepiness or fatigue [20]. Does shift work or a

possible sleep phase shift contribute to daytime

impairment? Could concomitant narcolepsy without

cataplexy or idiopathic hypersomnolence be present?

Does the patient have difficulty with sleep mainte-

nance unrelated to OSA? If so, adequate therapy may

involve treatment of insomnia or restless leg syn-

drome. Can non–sleep-related pathology, such as

chronic pain, contribute to alterations in sleep archi-

tecture and continuity?

Before positive pressure therapy is attempted,

several issues that are likely to impact on acceptance

or adherence of positive pressure should be consid-

ered. Is the patient familiar with positive pressure

therapy? If not, an educational intervention is neces-

sary before the introduction of therapy [21,22]. Is

nasal obstruction present? If so, medical and possibly

mechanical treatment of the nose may be necessary for

effective treatment [23–25]. Is the patient claustro-

phobic? If this is the case, an attempt at desensitization

may be beneficial before instituting therapy [26].

Tailoring the treatment to a given patient

Once the decision has been made that a patient

potentially would benefit from a trial of therapy, the

first intervention in conjunction with lifestyle recom-

mendations (ie, avoiding alcohol and sedatives,

ensuring proper sleep hygiene, beginning smoking

cessation, and maintaining fitness) should be a trial of

positive pressure via a mask [13,27]. The trial is best

accomplished in the laboratory with a technician in

attendance. Attended positive pressure titrations

allow for further patient education and reassurance

by the technical staff and proper mask fit, optimal

modality (ie, continuous positive airway pressure

[CPAP] or bi-level pressure) and an accurate pressure

prescription [21]. Whether this is accomplished in the

context of a split- or full-night study depends on the

previously discussed considerations.

In-line heated humidification may be particularly

useful in elderly patients and patients with nasal

congestion or mouth leaks [28,29]. It should be

prescribed for patients who are treated with sys-

temic anticoagulation [30]. Chin straps and oronasal

masks may be tried for mouth leaks but are poorly

tolerated compared with nasal interfaces with

heated humidification.

Second-line therapy: alternatives to

positive pressure

Despite adequate preparation and an effective

attended titration, several patients with an elevated

AHI or daytime symptoms will not accept or adhere

to positive pressure therapy. This possibility high-

lights the need for follow-up with objective measure-

ment of adherence to positive pressure therapy. In

these patients, it is important to revisit the primary

complaint that drove the evaluation in the context of

the severity of OSA and underlying vascular risk

(Fig. 2).

Primary concern: snoring

In patients with mild OSA (AHI 5–15), minimal

symptoms of fatigue or daytime sleepiness, and the

primary complaint of snoring, trial of an oral appli-

ance or a palatal procedure is a reasonable option

[27]. Patients may prefer an oral appliance to positive

pressure [31]. The response to treatment is not

complete, which mandates follow-up [7]. Despite

expert adjustment, treatment with oral appliance

therapy may be limited by tooth movement and bite

discomfort [32,33]. The long-term outcomes with

oral appliance therapy are not well characterized.

Fig. 1. The variable effect of OSA on physiologic outcomes.

P.J. Strollo, Jr / Clin Chest Med 24 (2003) 307–313308

Palatal procedures include conventional scalpel

technique uvulopalatopharyngoplasty, laser assisted

uvulopalatoplasty, and radiofrequency treatment of

the palate (somnoplasty) [34,35]. The pros and cons

of the palatal procedures are discussed in detail

elsewhere in this issue. Overall, palatal procedures

alone can be effective treatments of snoring. If a

tonsillectomy is included, mild OSA can be impacted

favorably, although as in the case of oral appliances,

the response to treatment may not be complete and

follow-up is mandatory [36].

Optimal treatment of nasal pathology can modify

snoring favorably and may be an important contri-

bution to the treatment plan. This treatment may

require medical interventions (ie, antihistamines,

nasal steroids, or leukotriene antagonists) [23,25].

Mechanical treatment of nasal obstruction may pro-

vide additional added value. Radiofrequency treat-

ment of the nasal turbinates can be effective and may

avoid an operating room procedure [24].

Primary concern: vascular risk

Patients with OSA are at risk for vascular

morbidity or mortality [37]. If vascular comorbidity

is present in the absence of significant daytime

impairment, treatment with positive pressure may

not be accepted [19]. Similar difficulty may be

encountered with oral appliance therapy. No defin-

itive data support surgery–other than tracheos-

tomy–as an effective treatment option to impact

vascular comorbidities related to OSA [38,39].

Burgeoning evidence supports the concept that

intermittent hypoxia may be the primary determi-

nant of vascular risk related to OSA [40]. This may

be mediated, in part, by reactive oxygen species

that are precipitated by an ischemia-reperfusion

insult related to the intermittent cell hypoxia [41].

In animal experiments, intermittent hypoxia has

been shown to upregulate sympathetic tone, which

results in catecholamine release and elevated blood

pressure [42].

Nocturnal oxygen may be accepted in patients

who do not tolerate positive pressure therapy [43].

Although definitive evidence is lacking, it is bio-

logically plausible that nocturnal oxygen would affect

vascular risk favorably. One current limitation to this

treatment option is the inconvenience of transporting

oxygen concentrators that are bulky and weigh on

average between 20 and 50 lbs [44].

Primary concern: daytime symptoms

It is always helpful to determine the response of

impaired daytime function (ie, fatigue and sleepi-

ness) to positive pressure therapy. It is a considerable

problem to sort out this effect when patients are

unwilling to accept treatment with positive pressure.

Chronic sleep deprivation (the most common cause

of daytime impairment) and depression as confound-

ers should be excluded [20]. An objective assess-

ment of daytime sleepiness, such as the multiple

sleep latency test, can be helpful in determining the

degree of daytime impairment and providing insight

into the possibility of a concomitant diagnosis of

narcolepsy without cataplexy or idiopathic hyper-

somnolence [45].

Fig. 2. Focusing the treatment on the primary patient complaint.

P.J. Strollo, Jr / Clin Chest Med 24 (2003) 307–313 309

A judicious trial of a daytime stimulant may

improve quality of life. This trial is best accom-

plished in conjunction with treatment with positive

pressure therapy. Certain patients may have contin-

ued daytime sleepiness despite treatment with CPAP

or bi-level pressure. Pack et al reported success with

modafinil as adjunctive therapy for daytime sleepi-

ness in OSA [46]. In their 4-week double blind

treatment trial (n = 157), inclusion criteria required

that patients adhere to CPAP (7.1 + 2.9 hours placebo

versus 7 + 1.2 modafinil). Modafinil at a dose of

400 mg/day resulted in a significant improvement in

subjective daytime sleepiness and objective daytime

sleepiness measured by the multiple sleep latency test.

There was no difference between the two treatment

groups in the percentage who normalized their mul-

tiple sleep latency test scores to more than 10 minutes

(25% placebo versus 29% modafinil, P = 0.613) [46].

Nonamphetamine daytime stimulants seem to be

reasonably safe as an adjunct to treatment with

positive pressure for daytime sleepiness [47]. Cur-

rently, stimulant therapy alone cannot be recom-

mended for patients with sleep apnea (AHI >5)

[46,48]. If the patient does not accept positive pressure

therapy, second-line therapy for OSA should be pur-

sued, whether medical, surgical, or dental, before

contemplating adjunctive stimulant treatment. It is

imperative that the potential impact on vascular risk

be examined carefully. Follow-up monitoring of blood

pressure is necessary.

Special circumstances

Upper airway resistance syndrome

There is uncertainty regarding the use of stimulant

therapy alone in patients with the upper airway

resistance syndrome [49–51]. Ideally, a trial of treat-

ment with positive pressure is advisable. Unfortu-

nately, a significant percentage of these patients may

not accept treatment with positive pressure. This

approach is frequently hampered by the fact that third

party payers will not reimburse homecare companies

for a positive pressure treatment trial of upper airway

resistance syndrome, and the patient may be unwill-

ing to bear the cost.

Down syndrome

Patients with Down syndrome have upper airway

abnormalities that place them at risk for sleep-disor-

dered breathing [52]. In the adult patient with Down

syndrome, the challenge is therapeutic, not diagnostic

[53]. Many of these patients have difficulty accepting

positive pressure therapy. Oxygen may be easier to

tolerate and worth trying if CPAP or bi-level pressure

is not an option [43]. It is essential that the caregiver

responsible for the patient be trained to help the

patient with the prescribed therapy.

Hospitalized patients

Obstructive sleep apnea can be found in medical

patients hospitalized with another primary diagnosis.

Clinical experience dictates that the prevalence is

increased compared with healthy outpatients. This

rate undoubtedly reflects the high incidence of obe-

sity, cardiovascular disease, cerebrovascular disease,

and diabetes in this patient population. These patients

present a challenge to diagnose and treat while

acutely hospitalized. The need for monitoring and

intravenous medications poses problems for the sleep

laboratory in which nursing personnel may not be

available to provide additional care. The patient may

be reluctant to pursue treatment with positive pres-

sure during the hospitalization. Sleep deprivation, the

use of sedatives and narcotics, and suboptimal volume

status also may tend to worsen the severity of the

underlying OSA. It may be important to identify OSA

acutely, but definitive treatment with CPAP or bi-level

pressure may be best reserved when the patient is

stabilized as an outpatient. Head of bed elevation and

supplemental oxygen may be better tolerated acutely

[43,54–56].

Elderly patients

Elderly patients (particularly older than 80 years),

much like hospitalized patients, are challenging to

treat. Major abnormalities of the sleep schedule are

frequently present. Concomitant insomnia and ad-

vanced phase disorders make it problematic to assess

a response to positive pressure if OSA is present [57].

Many of these patients have significant vascular risk,

and treatment makes good clinical sense. Second-line

therapy with oxygen or head of bed elevation is

frequently the best fit in these patients and may

provide significant benefit [43,54–56].

Hypoventilation syndromes

Hypercapnia is common in OSA but frequently

overlooked. One recent series found that 17% of

patients referred for polysomnography had evidence

of daytime hypercapnia [58]. There is uncertainty

whether CPAP is contraindicated. If a patient com-

plains of frequent morning headaches or has evidence

P.J. Strollo, Jr / Clin Chest Med 24 (2003) 307–313310

of persistent right heart failure or hypercapnia—or

both—at the time of follow-up, a bi-level pressure

titration should be considered [59].

Summary

The primary treatment modality for OSA remains

positive pressure therapy. Differential susceptibility

to daytime sleepiness and vascular risk exists. In

patients who do not accept positive pressure therapy

despite careful attempts to optimize the treatment,

second-line therapy should be explored.

A careful assessment of the primary treatment con-

cern shouldguide further intervention(s).Althoughpal-

atal surgery can treat snoring effectively, the effect on

the AHI and daytime sleepiness is less robust. Oral ap-

pliances may help some patients [31]. Recent data

suggest that the durability of the treatment over time is

uncertain and subject to frequent dental complications

[32,33].

Treatment with oxygen should be considered in

patients who do not accept positive pressure therapy

and are believed to be at increased risk for vascular

complications [43]. Current generation oxygen con-

centrators are difficult to transport and limit the use of

this treatment option in highly mobile patients [44].

Special populations, including patients with Down

syndrome, hospitalized patients, and elderly persons,

may be more accepting of treatment with oxygen via

nasal cannula alone. Although this approach makes

biologic sense, definitive outcome evidence is lacking.

Future expectations

Cumulative epidemiology data provide a con-

vincing argument that patients with OSA are at risk for

impaired daytime performance (sleepiness or fatigue),

insulin resistance, automobile crashes, and vascular

complications. It also has become evident that whereas

a dose-response relationship exists with regard to the

AHI and risk for the group as a whole, differential

susceptibility may exist for a given patient [10–12].

The challenge for the future is to define the risk in

a given patient. Physiologic tests that provide added

value to the current evaluation are welcome. Quan-

tifying daytime impairment with vigilance testing and

better assessing vascular risk with new technology

may prove to be useful [60–62]. On the horizon,

insights gained from functional genomics, proteo-

mics, and possibly metabonomics undoubtedly will

provide powerful data for future clinical decision

making in OSA [63].

References

[1] Ip MS, Lam B, Ng MM, Lam WK, Tsang KW, Lam

KS. Obstructive sleep apnea is independently associ-

ated with insulin resistance [comment]. Am J Respir

Crit Care Med 2002;165:670–6.

[2] Punjabi NM, Sorkin JD, Katzel LI, Goldberg AP,

Schwartz AR, Smith PL. Sleep-disordered breathing

and insulin resistance in middle-aged and overweight

men [comment]. Am J Respir Crit Care Med 2002;

165:677–82.

[3] Strollo Jr PJ, Rogers RM. Obstructive sleep apnea.

N Engl J Med 1996;334:99–104.

[4] George CF. Reduction in motor vehicle collisions fol-

lowing treatment of sleep apnoea with nasal CPAP

[comment]. Thorax 2001;56:508–12.

[5] Pepperell JC, Ramdassingh-Dow S, Crosthwaite N,

et al. Ambulatory blood pressure after therapeutic

and subtherapeutic nasal continuous positive airway

pressure for obstructive sleep apnoea: a randomised

parallel trial [comment]. Lancet 2002;359:204–10.

[6] Jenkinson C, Davies RJ, Mullins R, Stradling JR.

Comparison of therapeutic and subtherapeutic nasal

continuous positive airway pressure for obstructive

sleep apnoea: a randomised prospective parallel trial

[comment]. Lancet 1999;353:2100–5.

[7] White J, Cates C, Wright J. Continuous positive air-

ways pressure for obstructive sleep apnoea [update].

Cochrane Database of Systematic Reviews 2002;

CD001106.

[8] Sin DD, Mayers I, Man GC, Pawluk L. Long-term

compliance rates to continuous positive airway pres-

sure in obstructive sleep apnea: a population-based

study. Chest 2002;121:430–5.

[9] American Academy of Sleep Medicine Task Force.

Sleep-related breathing disorders in adults: recommen-

dations for syndrome definition and measurement tech-

niques in clinical research [comment]. Sleep 1999;

22:667–89.

[10] Gottlieb DJ, Whitney CW, Bonekat WH, et al. Relation

of sleepiness to respiratory disturbance index: the

Sleep Heart Health Study. Am J Respir Crit Care

Med 1999;159:502–7.

[11] Shahar E, Whitney CW, Redline S, et al. Sleep-disor-

dered breathing and cardiovascular disease: cross-

sectional results of the Sleep Heart Health Study

[comment]. Am J Respir Crit Care Med 2001;163:

19–25.

[12] Peppard PE, Young T, Palta M, Skatrud J. Prospective

study of the association between sleep-disordered

breathing and hypertension [comment]. N Engl J

Med 2000;342:1378–84.

[13] Strollo Jr PJ, Sanders MH, Atwood CW. Positive pres-

sure therapy. Clin Chest Med 1998;19:55–68.

[14] Barbe F, Mayoralas LR, Duran J, et al. Treatment with

continuous positive airway pressure is not effective in

patients with sleep apnea but no daytime sleepiness:

a randomized, controlled trial [comment]. Ann Intern

Med 2001;134:1015–23.

P.J. Strollo, Jr / Clin Chest Med 24 (2003) 307–313 311

[15] Strollo Jr PJ, Sanders MH, Costantino JP, Walsh SK,

Stiller RA, Atwood Jr CW. Split-night studies for the

diagnosis and treatment of sleep-disordered breathing.

Sleep 1996;19:S255–9.

[16] McArdle N, Grove A, Devereux G, Mackay-Brown L,

Mackay T, Douglas NJ. Split-night versus full-night

studies for sleep apnoea/hypopnoea syndrome. Eur Res-

pir J 2000;15:670–5.

[17] Sanders MH, Constantino JP, Strollo PJ, Studnicki K,

Atwood CW. The impact of split-night polysomnogra-

phy for diagnosis and positive pressure therapy titra-

tion on treatment acceptance and adherence in sleep

apnea/hypopnea. Sleep 2000;23:17–24.

[18] Meslier N, Lebrun T, Grillier-Lanoir V, et al. A French

survey of 3,225 patients treated with CPAP for obstruc-

tive sleep apnoea: benefits, tolerance, compliance and

quality of life. Eur Respir J 1998;12:185–92.

[19] McArdle N, Devereux G, Heidarnejad H, Engleman

HM, Mackay TW, Douglas NJ. Long-term use of

CPAP therapy for sleep apnea/hypopnea syndrome.

Am J Respir Crit Care Med 1999;159:1108–14.

[20] Douglas NJ. Why am I sleepy? Sorting the somnolent.

Am J Respir Crit Care Med 2001;163:1310–3.

[21] Hoy CJ, Vennelle M, Kingshott RN, Engleman HM,

Douglas NJ. Can intensive support improve continuous

positive airway pressure use in patients with the sleep

apnea/hypopnea syndrome? Am J Respir Crit Care

Med 1999;159:1096–100.

[22] Chervin RD, Theut S, Bassetti C, Aldrich MS. Com-

pliance with nasal CPAP can be improved by simple

interventions. Sleep 1997;20:284–9.

[23] Berger WE. Treatment update: allergic rhinitis. Allergy

Asthma Proc 2001;22:191–8.

[24] Powell NB, Zonato AI, Weaver EM, et al. Radiofre-

quency treatment of turbinate hypertrophy in subjects

using continuous positive airway pressure: a random-

ized, double-blind, placebo-controlled clinical pilot tri-

al. Laryngoscope 2001;111:1783–90.

[25] Zozula R, Rosen R. Compliance with continuous pos-

itive airway pressure therapy: assessing and improving

treatment outcomes. Curr Opin Pulm Med 2001;7:

391–8.

[26] Edinger JD, Radtke RA. Use of in vivo desensitization

to treat a patient’s claustrophobic response to nasal

CPAP. Sleep 1993;16:678–80.

[27] Henderson II JH, Strollo Jr PJ. Medical management of

obstructive sleep apnea. Prog Cardiovasc Dis 1999;

41:377–86.

[28] Martins De Araujo MT, Vieira SB, Vasquez EC, Fleury

B. Heated humidification or face mask to prevent upper

airway dryness during continuous positive airway pres-

sure therapy [comment]. Chest 2000;117:142–7.

[29] Massie CA, Hart RW, Peralez K, Richards GN. Effects

of humidification on nasal symptoms and compliance

in sleep apnea patients using continuous positive air-

way pressure [comment]. Chest 1999;116:403–8.

[30] Strumpf DA, Harrop P, Dobbin J, Millman RP. Mas-

sive epistaxis from nasal CPAP therapy. Chest 1989;

95:1141.

[31] Ferguson KA, Ono T, Lowe AA, Keenan SP, Fleetham

JA. A randomized crossover study of an oral appliance

vs nasal-continuous positive airway pressure in the

treatment of mild-moderate obstructive sleep apnea

[comment]. Chest 1996;109:1269–75.

[32] Petit FX, Pepin JL, Bettega G, Sadek H, Raphael B,

Levy P. Mandibular advancement devices: rate of con-

traindications in 100 consecutive obstructive sleep ap-

nea patients. Am J Respir Crit Care Med 2002;166:

274–8.

[33] Rose EC, Staats R, Virchow Jr C, Jonas IE. Occlusal

and skeletal effects of an oral appliance in the treatment

of obstructive sleep apnea. Chest 2002;122:871–7.

[34] Li KK, Powell NB, Riley RW, Troell RJ, Guilleminault

C. Radiofrequency volumetric reduction of the palate:

an extended follow-up study. Otolaryngol Head Neck

Surg 2000;122:410–4.

[35] Troell RJ, Powell NB, Riley RW, Li KK, Guilleminault

C. Comparison of postoperative pain between laser-

assisted uvulopalatoplasty, uvulopalatopharyngoplasty,

and radiofrequency volumetric tissue reduction of the

palate. Otolaryngol Head Neck Surg 2000;122:402–9.

[36] McGuirt Jr WF, Johnson JT, Sanders MH. Previous

tonsillectomy as prognostic indicator for success of

uvulopalatopharyngoplasty. Laryngoscope 1995;105:

1253–5.

[37] Partinen M, Guilleminault C. Daytime sleepiness and

vascular morbidity at seven-year follow-up in obstruc-

tive sleep apnea patients. Chest 1990;97:27–32.

[38] He J, Kryger MH, Zorick FJ, Conway W, Roth T.

Mortality and apnea index in obstructive sleep apnea:

experience in 385 male patients. Chest 1988;94:9–14.

[39] Bridgman SA, Dunn KM. Surgery for obstructive sleep

apnoea. Cochrane Database of Systematic Reviews

2000;CD001004.

[40] Neubauer JA. Invited review: physiological and patho-

physiological responses to intermittent hypoxia. J Appl

Physiol 2001;90:1593–9.

[41] Prabhakar NR. Sleep apneas: an oxidative stress?

[comment]. Am J Respir Crit Care Med 2002;165:

859–60.

[42] Sica AL, Greenberg HE, Ruggiero DA, Scharf SM.

Chronic-intermittent hypoxia: a model of sympathetic

activation in the rat. Respir Physiol 2000;121:173–84.

[43] Landsberg R, Friedman M, Ascher-Landsberg J. Treat-

ment of hypoxemia in obstructive sleep apnea. Am J

Rhinol 2001;15:311–3. [Erratum appears in Am J Rhi-

nol 2002;16:67.]

[44] Kacmarek RM. Delivery systems for long-term oxygen

therapy. Respiratory Care 2000;45:84–92 discussion

92–4.

[45] Carskadon MA, Dement WC, Mitler MM, Roth T,

Westbrook PR, Keenan S. Guidelines for the multiple

sleep latency test (MSLT): a standard measure of sleep-

iness. Sleep 1986;9:519–24.

[46] Pack AI, Black JE, Schwartz JR, Matheson JK. Moda-

finil as adjunct therapy for daytime sleepiness in ob-

structive sleep apnea. Am J Respir Crit Care Med 2001;

164:1675–81.

P.J. Strollo, Jr / Clin Chest Med 24 (2003) 307–313312

[47] Arnulf I, Homeyer P, Garma L, WhitelawWA, Derenne

JP. Modafinil in obstructive sleep apnea-hypopnea syn-

drome: a pilot study in 6 patients. Respiration 1997;64:

159–61.

[48] Kingshott RN, Vennelle M, Coleman EL, Engleman

HM, Mackay TW, Douglas NJ. Randomized, double-

blind, placebo-controlled crossover trial of modafinil in

the treatment of residual excessive daytime sleepiness in

the sleep apnea/hypopnea syndrome. Am J Respir Crit

Care Med 2001;163:918–23.

[49] Douglas NJ. Upper airway resistance syndrome is not a

distinct syndrome [comment]. Am J Respir Crit Care

Med 2000;161:1413–6.

[50] Guilleminault C, Chowdhuri S. Upper airway resist-

ance syndrome is a distinct syndrome [comment].

Am J Respir Crit Care Med 2000;161:1412–3.

[51] Rees K, Kingshott RN, Wraith PK, Douglas NJ. Fre-

quency and significance of increased upper airway re-

sistance during sleep. Am J Respir Crit Care Med

2000;162:1210–4.

[52] Marcus CL, Keens TG, Bautista DB, von Pechmann

WS, Ward SL. Obstructive sleep apnea in children

with Down syndrome. Pediatrics 1991;88:132–9.

[53] Smith DS. Health care management of adults with

Down syndrome. Am Fam Physician 2001;64:1031–8.

[54] McEvoy RD, Sharp DJ, Thornton AT. The effects of

posture on obstructive sleep apnea. Am Rev Respir Dis

1986;133:662–6.

[55] Neill AM, Angus SM, Sajkov D, McEvoy RD. Effects

of sleep posture on upper airway stability in patients

with obstructive sleep apnea. Am J Respir Crit Care

Med 1997;155:199–204.

[56] Hakala K, Maasilta P, Sovijarvi AR. Upright body

position and weight loss improve respiratory mechan-

ics and daytime oxygenation in obese patients with

obstructive sleep apnoea. Clin Physiol 2000;20:50–5.

[57] Avidan AY. Sleep changes and disorders in the elderly

patient. Current Neurology and Neuroscience Reports

2002;2:178–85.

[58] Resta O, Foschino Barbaro MP, Bonfitto P, et al. Hy-

percapnia in obstructive sleep apnoea syndrome. Neth

J Med 2000;56:215–22.

[59] Resta O, Guido P, Picca V, et al. Prescription of nCPAP

and nBIPAP in obstructive sleep apnoea syndrome:

Italian experience in 105 subjects. A prospective two

centre study. Respir Med 1998;92:820–7.

[60] Dinges DF, Pack F, Williams K, et al. Cumulative

sleepiness, mood disturbance, and psychomotor vigi-

lance performance decrements during a week of sleep

restricted to 4–5 hours per night. Sleep 1997;20:

267–77.

[61] O’Donnell CP, Allan L, Atkinson P, Schwartz AR. The

effect of upper airway obstruction and arousal on pe-

ripheral arterial tonometry in obstructive sleep apnea.

Am J Respir Crit Care Med 2002;166:965–71.

[62] Pitson DJ, Sandell A, van den Hout R, Stradling JR.

Use of pulse transit time as a measure of inspiratory

effort in patients with obstructive sleep apnoea. Eur

Respir J 1995;8:1669–74.

[63] Brindle JT, Antti H, Holmes E, et al. Rapid and non-

invasive diagnosis of the presence and severity of

coronary heart disease using (1)H-NMR-based me-

tabonomics. Nat Med 2002;8:1439–45.

P.J. Strollo, Jr / Clin Chest Med 24 (2003) 307–313 313

Continuous positive airway pressure: new generations

Francoise J. Roux, MD, PhDa,b,*, Janet Hilbert, MDa,c

aSection of Pulmonary and Critical Care Medicine, Yale University School of Medicine, 333 Cedar Street,

Post Office Box 208057, New Haven, CT 06520-8057, USAbWest Haven VA Medical Center Sleep Laboratory, 950 Campbell Avenue, West Haven, CT 06516, USA

cYale Center for Sleep Medicine, 333 Cedar Street, Post Office Box 208057, New Haven, CT 06520-8057, USA

Continuous positive airway pressure (CPAP) ther-

apy for obstructive sleep apnea (OSA) was first

described in 1981 [1]. Since that time, CPAP has

become the mainstay of therapy for OSA [2]. CPAP

effectively prevents repetitive upper airway obstruc-

tion, most likely by acting as a pneumatic splint [3],

and is associated with improved respiratory [1,4] and

sleep parameters [5] and clinical outcomes [5–15].

CPAP therapy continues to evolve, and, since the last

review of positive airway pressure therapy in this

Clinics issue [16], further advancements have been

made in newer generations of CPAP. Automatic (also

known as automated, autotitrating, or autoadjusting)

positive airway pressure (APAP) devices detect and

respond to changes in upper airway resistance by

variably increasing or decreasing the pressure gener-

ated. As such, APAP potentially may be able to (1)

assist with the initial diagnosis of OSA, (2) act

therapeutically in patients with OSA instead of con-

ventional CPAP, and (3) assist with CPAP titration to

determine an effective conventional CPAP pressure in

patients with confirmed OSA. In this article, the

authors present an updated review of the technical

aspects of APAP and the diagnostic, therapeutic, and

titrating capabilities. They also discuss the current

clinical recommendations for use of these devices.

Technical aspects of automatic positive

airway pressure

Background

The original CPAP device described by Sullivan

et al in 1981 [1] (Fig. 1) consisted of a vacuum

cleaner blower motor with variable speed control

installed in a box lined with acoustic material. This

was connected to a wide bore tube, into which were

inserted soft plastic tubes to fit into the patient’s nares

and which then distally narrowed with mechanical

resistance. A range of pressures could be generated.

Although effective in maintaining the patency of the

upper airway in patients with OSA, the original

CPAP machines in the 1980s were heavy (approx-

imately 15–20 lbs), loud, and fairly simple, with

limited capabilities.

Over the past 20 years, machines have become

lighter (typically ranging from 3.5–6 lbs), quieter,

and more sophisticated. In some CPAP machines,

microprocessors allow compliance data (most cur-

rently using mask-on-time rather than the earlier

machine-on-time) to be stored for variable amounts

of time and downloaded. Various options, including

specialized filters, ramps, automatic altitude adjust-

ment, automatic leak compensation, internal power

adaptors, and internal humidification, have been

incorporated into various machines. Options for

CPAP accessories, including nasal and oronasal

masks, headgear, and humidification, also have

increased, seemingly exponentially.

Another level of sophistication has been added

with the development of APAP devices. APAP devices

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00017-0

* Corresponding author. Section of Pulmonary and

Critical Care Medicine, Yale University School of Medi-

cine, 333 Cedar Street, PO Box 208057, New Haven, CT

06520-8057.

E-mail address: francoise.roux@yale.edu (F.J. Roux).

Clin Chest Med 24 (2003) 315–342

use noninvasive methods to detect evidence of upper

airway obstruction, including snoring, apneas, hypo-

pneas, or airflow limitation. As shown in Fig. 2, a

diagram of prototype device that detects and responds

to pharyngeal wall vibration [17], APAP devices

incorporate one or more sensors to detect a signal (in

this case, a pressure transducer to detect pharyngeal

wall vibration) and a central processing unit to inter-

pret the signal(s) (according to specific diagnostic

algorithms) and determine the resultant voltage for

the APAP blower in response to the signal(s) (accord-

ing to specific therapeutic algorithms). Additional

band filters and rectifiers are needed to process the

signal, and analog-digital and digital-analog convert-

ers also are required downstream and upstream of the

central processing unit. Technology does not come

without a price; the cost of currently available APAP

devices can be 1.5 to 3 times that of conventional

CPAP machines, depending on incorporated features.

APAP devices can function exclusively in a diagnostic

mode and recognize and record abnormal respiratory

events without correcting them. APAP devices also

function in a therapeutic mode, responding to events

(or lack of them) by adjusting the positive airway

pressure accordingly.

Detection of upper airway obstruction by

automatic positive airway pressure

As shown in Table 1, current APAP devices detect

multiple abnormalities, such as snoring, apneas, hypo-

pneas, or flow limitation, which are surrogates of

upper airway obstruction. Clinical studies have been

published in the peer-reviewed literature to date on

versions of the Autoset (ResMed, Sydney, Australia)

[18–32], Goodknight 418A (Puritan Benett/Malinck-

rodt, Les Ulis, France) [33] and its precursor, REM +

auto (SEFAM/Nellcor Puritan Benett, Nancy, France)

[34–37], Horizon AutoAdjust (DeVilbiss/Sunrise

Medical, Somerset, PA) [38–42], Morphee Plus/

Cloudnine (Pierre Medical/Nellcor Puritan Benett,

Verrieres-Le-Buisson, France, and Minneapolis, MN)

[41,43–46], REM + with MC + (SEFAM/Nellcor

Puritan Benett) [47], Somnosmart (Weimann, Ham-

Fig. 2. Diagram of prototype APAP system’s major components. LCD, liquid crystal display; CPU, central processing unit;

ROM, read only memory; RAM, random access memory. (From Behbehani K, Yen FC, Burk JR, Lucas EA, Axe JR. Automatic

control of airway pressure for treatment of obstructive sleep apnea. IEEE Trans Biomed Eng 1995;42:1007; with permission.)

Fig. 1. Diagram of apparatus used to provide CPAP from the

nares. In the experimental system, pressure (Pa) wasmeasured

via a catheter in one nasal tube, and airway CO2 (CO2) was

sampled via a catheter in the other nasal tube. (From Sullivan

CE, Issa FG, Berthon-Jones M, Eves L. Reversal of

obstructive sleep apnoea by continuous positive airway

pressure applied through the nares. Lancet 1981;1:862;

with permission.)

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342316

burg, Germany) [48–53], and Virtuoso (Respironics,

Murrysville, PA) [33,54–56]. Sensors used to detect

evidence of upper airway obstruction and diagnostic

algorithms vary among devices.

Most, but not all, APAP devices have the capabil-

ity of detecting snoring. Snoring is typically detected

by a high frequency response pressure transducer in

the presence of mask pressure vibration [17,25,

34,37,47,54,57]. The signal is then band-pass filtered

using high pass and low pass filters specific to the

device (eg, 30 Hz and 280 Hz in the REM + auto

[34,37] and the REM + with MC plus [47] devices

and 20 Hz and 120 Hz in another prototype [17]), and

the amplitude is then analyzed to detect amplitude

variations, such as snoring.

Many APAP devices also can identify apneas, as

indicated by absence of flow or pressure, and hypo-

pneas, as indicated by decrements of flow or pressure.

Early versions of the Autoset in diagnostic mode

detected apneas (and later, apneas and hypopneas) by

analyzing the pressure tracing from nasal prongs [19],

whereas later therapeutic models detected changes in

flow with a built-in pneumotachograph [30]. The

Horizon AutoAdjust [40] also uses a pneumotacho-

graph to detect apneas and hypopneas. The Morphee

Plus determines patency of the upper airway by

monitoring the breath-by-breath difference between

maximal inspiratory and expiratory flow based on

machine compressor speed [43]. The default defini-

tions of apneas or hypopneas used by the detection

software vary with the specific device and software

version, as does the ability for the clinician or

investigator to change the detection algorithm. For

example, on the Horizon Autoadjust, the criteria for

Table 1

Comparison of parameters detected by automatic positive airway devices

Parameters detected

Device Manufacturer Sn A/H (flow) A/H (FOT) FL

Autoset

Autoset Clinical

Autoset Portable

Autoset T

Autoset Spirit

Eclipse Auto

Goodknight

TM418A

Goodknight

TM418P

Horizon

AutoAdjust

Morphee Plus

Cloudnine

REM + with MC +

REM + Auto

REMstar Auto

Somnosmart

Tranquility Auto

Virtuoso LX

ResCare/ResMed,

Sydney, Australia,

San Diego, CA,

Saint Priest,

France

Taema, Antony, France

Puritan Benett, Pleasanton,

CA/Mallinckrodt, Les Ulis, France

Puritan Benett, Pleasanton,

CA/Mallincrodt, Les Ulis, France

DeVilbiss Healthcare,Inc./Sunrise Medical,

Somerset, PA, Parcay Meslay, France

Pierre Medical, Verrieres-Le-Buisson,

France/Nellcor-Puritan Benett, Minneapolis, MN

SEFAM/Nellcor-Puritan Benett, Nancy, France

Respironics Inc., Murrysville, PA

Weiman, Hamburg,Germany

Respironics Inc., Murrysville, PA

Respironics Inc., Murrysville,PA

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

+

Abbreviations: Sn, snoring (detected by mask pressure vibration); A, apnea; H, hypopnea; flow, airflow detected by pneu-

motachograph, nasal pressure, or changes in compressor speed; FOT, forced oscillation technique; FL, flow limitation (detected

by flow versus time profile).

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 317

hypopnea (eg, percent flow decrement and duration)

can be programmed [39].

Forced oscillation technique (FOT), initially

described by DuBois et al in 1956 [58], is a unique

noninvasive method that detects changes in airway

resistance or impedance [59]. Since the initial

description, this technique has been validated in a

research model of airway obstruction [60] and in

patients with OSA [61,62]. A pump, connected to

the CPAP system, generates a sinusoidal pressure

signal at a constant frequency during spontaneous

breathing; respiratory impedance (Z) or respiratory

system resistance (Rrs) is derived from the oscillatory

pressure and flow signal at the nasal mask. In patients

with OSA, apneas are associated with sustained

increases in impedance throughout the respiratory

cycle, whereas hypopneas are associated with inter-

mittent increases in impedance (Fig. 3). With CPAP

treatment, as CPAP is progressively increased to the

effective range (Fig. 4), breathing flow normalizes,

esophageal pressure swings become less negative,

and respiratory resistance or impedance decreases to

normal. This technique has since been used to control

the pressure algorithm in several devices, including

the Somnosmart [48 –53] and other prototypes

[60,63].

As shown in Table 1, few of the currently avail-

able devices can detect flow limitation. The char-

acteristic inspiratory airflow flattening seen with

increased upper airway resistance [64] may be the

most sensitive indicator of upper airway obstruction

[65,66]. In the Autoset, the flow-time profile (with

flow measured by pressure transducer or pneu-

motachograph) is expressed as a curvature index

(Fig. 5) [25]. A low curvature index suggests inspir-

atory airflow limitation, whereas a higher curvature

index suggests more normal airflow.

Adjustment of positive airway pressure by

automatic positive airway pressure

Once a respiratory event is detected, APAP devices

increase the pressure automatically in a progressive

fashion until an effective therapeutic pressure is

reached. Conversely, in the absence of respiratory

events, the pressure level decreases until evidence of

upper airway obstruction recurs. APAP devices are

inherently unstable [67]. The algorithms for pressure

adjustment vary among specific APAP devices. Even

within a specific device, the amount of pressure

increase and the time course of pressure change vary

Fig. 3. Representative compressed polysomnographic recording (6 epochs, 3 minute). Note that impedance (Z) increases during

apnea and is low during arousal. The pattern of Z shows a cyclic increase before apnea and during hypopnea (the last event in

this figure). EOG, electrooculogram; C4-A1 and C4-A2, electroencephalogram channels; EMG-GC, chin electromyogram; Flow,

flow by pneumotachograph; Effort, effort by thoraco-abdominal bands; Sum, thoraco-abdominal sum; SaO2, arterial oxygen

saturation; Gen DC Z, respiratory impedance. (From Badia JR, Farre R, Montserrat JM, Ballester E, Hernandez L, Rotger M,

et al. Forced oscillation technique for the evaluation of severe sleep apnoea/hypopnoea syndrome: a pilot study. Eur Respir J

1998;11:1128; with permission.)

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342318

with the specific respiratory event detected, with

larger changes for apneas and smaller changes for

more subtle upper airway obstruction, such as snoring

or flow limitation. For example, in the Autoset

device, positive pressure increases in response to

snoring, apneas, and flow limitation [25]: snoring

(depending on the loudness of the snoring) results in

an increase in pressure by 1 cm H2O/breath; flow

limitation (depending on the curvature index) results

in an increase in pressure by 1.5 cm H2O/minute;

apneas (depending on calculated airway conductance)

result in an increase in pressure by 1 cm H2O/

Fig. 5. Schematic of Autoset software response to various flow versus time curves. The curvature index is a measure of the

deviation from unit scaled flow over the middle 50% of inspiratory time (indicated by shading). (Left) A severely flattened

curve with a low curvature index typical of inadequate CPAP pressure. The software responds by increasing CPAP pressure.

(Center) A breath showing slight flattening. The CPAP pressure remains unchanged. (Right) Rounded curve with high curvature

index; the software assumes that this breath represents hyperadequate CPAP pressure. The software reduces the CPAP pressure.

(From Teschler H, Berthon-Jones M, Thompson AB, Henkel A, Henry J, Konietzko N. Automated continuous positive airway

pressure titration for obstructive sleep apnea syndrome. Am J Respir Crit Care Med 1996;154:734; with permission.)

Fig. 4. Breathing flow (V), esophageal pressure (Pes), and respiratory resistance (Rrs) in a patient at different levels of CPAP.

Note that at low subtherapeutic CPAP levels (CPAP = 4 cm H20), obstructive apnea—characterized by minimal V, wide swings

in Pes, and high Rrs—occurs. With increasing, but still suboptimal CPAP (CPAP = 8 cm H20), hypopnea occurs, characterized by

variable Rrs. With therapeutic CPAP (CPAP = 12 cm H20) and return of normal V and Pes, Rrs decreases to normal. (From

Navajas D, Farre R, Rotger M, Badia R, Puig-de-Morales M, Montserrat JM. Assessment of airflow obstruction during CPAP

by means of forced oscillation in patients with sleep apnea. Am J Respir Crit Care Med 1998;157:1526; with permission.)

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 319

15 second of apnea. If no further abnormalities are

detected, pressure decreases with a time constant of

20 minutes for snoring and flow limitation and

40 minutes for apnea. In contrast, in the Morphee

Plus, if upper airway obstruction is detected (based

on changes in compressor speed), the pressure

increases between 1 and 3 cm H2O at a rate of 1 cm

H2O/second depending on the severity of the respir-

atory abnormality. In the absence of a respiratory

abnormality, pressure decreases at a rate of 1 cm

H2O / 30 seconds [43]. In some devices, the clinician

can change the default algorithms [39].

Most APAP devices operate in the range of 3 to 4

to 18 to 20 cm H2O pressure, and the upper and lower

limits of acceptable pressure usually can be set by the

clinician. Typically, the pressure starts low and auto

adjusts depending on the therapeutic algorithm. Some

devices, such as the Morphee Plus [43] and the REM

+ auto [35], operate around an acceptable range (such

as + 2 cm and � 4 cm H2O) of a reference pressure

(with the reference pressure set by the clinician using

either a previous CPAP titration or a formula); the

clinician also can set the acceptable range or use the

default settings for range.

Special issues: detection of leaks, central apneas,

and hypoventilation

Mask and mouth leaks and mouth breathing are

not uncommon in conventional CPAP titrations, with

continuous leaks occurring in 7 of 14 patients in one

study [68] and lasting from 4% to 70% of total sleep

time. Mouth leaks limited APAP therapy in 2 of 15

patients with OSA using a device based on snoring

detection (REM + Control with MC +, with apnea/

hypopnea detection disabled) [47]. With the Autoset,

leaks in excess of 0.4 L/second occur for an average

of 10% of a supervised night and 15% of an unsu-

pervised night [27]. Leaks tend to be interpreted by

many APAP systems as apneas or hypopneas and

result in increases in pressure, which, in turn, increase

leak [69]. Examining the raw flow signal can allow

for the detection of leaks [42]. Some devices have

algorithms that limit pressure increases when mask

pressure goes to zero, which indicates that the mask is

off or when there are excessive leaks as detected by

mean mask flow [25]. Leak alarms have been incor-

porated into some units. Some devices, such as the

Autoset and the Goodknight 418P, can record leaks

during the night for later use in interpretation. In

devices that use FOT, mouth leaks can induce false

low impedance values and lead to an underestimate of

upper airway obstruction, whereas mouth expiration

or change in the route of breathing can lead to a false

increase in impedance and an overestimate of upper

airway obstruction. Incorporating a pneumotacho-

graph along with FOT can be helpful in avoiding

misinterpretation of the FOT signal [68].

Another problem for APAP devices is the distinc-

tion between central and obstructive apneas. The

Autoset software classifies apneas as having closed

and open airways through the calculation of airway

conductance by modulating the mask pressure during

apnea and measuring the resultant induced airflow.

The pressure increases for obstructive apneas and for

central apneas in which the airway is closed [25].

Using FOT, whereas obstructive apneas are associ-

ated with sustained increases in impedance, central

apneas can be associated with high or low impedance

values, which suggests that different mechanisms

may be involved [61]. Cardiac oscillations may be

visible in the FOT signal to help distinguish central

from obstructive apneas [70]. Many clinical trials of

APAP devices have excluded patients at risk for

central apnea, such as persons with congestive heart

failure [69].

Current APAP devices do not incorporate oxime-

try and cannot detect sustained oxygen desaturation

in the absence of upper airway obstruction. Many

clinical trials of APAP have excluded patients at risk

for hypoventilation, including persons with chronic

obstructive pulmonary disease or other respiratory

disease/respiratory failure [69].

Comparison of commercially available automatic

positive airway pressure devices

Few studies have been published that directly

compare the available APAP devices listed in

Table 1. Farre et al [71] used a bench model with a

waveform generator to simulate normal breathing,

apneas, hypopneas, and flow limitation, with or with-

out snoring, and tested the response of five different

APAP devices (AutoAdjust LT, Autoset Portable II

Plus, Autoset T, Virtuoso LX, and Goodknight 418P)

to different combinations of signals. All these devices

responded to snoring; however, they responded differ-

ently to various degrees of hypopneas, and some did

not even respond to repetitive apneas. The time course

of pressure adjustment after normalization of breathing

also varied with devices, as did the behavior after a

simulated leak. Lofaso et al [33] also used a bench

model to study the performance of six commercially

available devices (Horizon AutoAdjust, Goodknight

418A, Goodknight 418P, Autoset T, Virtuoso LX, and

Eclipse Auto), this time in response to simulated

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342320

snoring at varying frequencies and at different CPAP

pressures. Threshold pressure-amplitude differences

of up to threefold were found across devices, and the

sensitivity of snoring detection decreased as CPAP

increased in all devices.

A follow-up study was performed in six patients

with OSA to test the clinical significance of these

findings. The percentage of snoring events followed

by a pressure increase was higher in the device that

was more sensitive in the bench study (Goodknight

418P) than in the less sensitive device (Virtuoso LX).

The only other comparative study in patients reported

to date [41] compared the response of two different

devices to nasal obstruction induced by local instilla-

tion of histamine. One device was regulated by

analysis of changes in compressor speed (Morphee

Plus), and the other device was regulated by analysis

of changes in flow determined with a pneumotacho-

graph (Horizon); neither device analyzed flow versus

time curves. The authors found that increases in nasal

resistance occurred with histamine and that flow

limitation and arousals generally accompanied the

increase in nasal resistance. The behavior of the

two devices differed and was sometimes paradoxical,

however. Mask pressure initially decreased as nasal

resistance increased in the Morphee Plus and some-

times subsequently increased, whereas mask pressure

did not change with the Horizon. Whether and to

what extent these differences between devices impact

long-term clinical outcomes is not known.

Role of automatic positive airway pressure in

diagnosis of obstructive sleep apnea

Rationale for use of automatic positive airway

pressure in diagnosis of obstructive sleep apnea

The diagnosis of OSA is conventionally made by

level 1 polysomnography (PSG) performed by a

trained technologist who attends the patient (usually

in a sleep laboratory), with recording or documenta-

tion of sleep and respiratory variables including

electroencephalogram, electrooculogram, chin elec-

tromyogram, airflow, respiratory effort, arterial oxy-

gen saturation (SaO2), and body position, with or

without leg movement recording [72,73]. In general,

the number of respiratory disturbances per hour of

sleep (RDI) or the sum of apneas and hypopneas per

hour of sleep (AHI) is used as the summary statistic

to diagnose OSA and determine the severity of sleep-

disordered breathing. It has become increasingly clear

that the RDI can vary as much as tenfold [74]

depending on the technique used to measure airflow

[75] and the definitions used to define specific

respiratory events (eg, percent flow change, degree

of oxygen desaturation, presence/absence of arousal

on encephalogram). The best metric to define OSA

and the cutoff between normal and abnormal is

controversial [76,77]. The RDI can have internight

and intranight variability, and it is possible for OSA

to be missed on one night of monitoring [78,79].

Despite these limitations, in the United States, level I

PSG is the gold standard for diagnosing OSA and

guiding treatment options, and it is against this

standard that other options must be compared.

In light of the growing clinical recognition of OSA

[80], there have been attempts in the United States and

elsewhere [81–83] to perform unattended home mon-

itoring for OSA to reduce the diagnostic waiting time.

Portable and generally unattended studies may range

from comprehensive portable PSG (level II) to modi-

fied portable sleep apnea testing (level III) to continu-

ous single or dual bioparameter recording (level IV).

The benefits and limitations of portable monitoring

have been reviewed elsewhere in this issue and are not

discussed herein. APAP devices have inherent di-

agnostic capability that is consistent with a level IV

recording (or even a level III recording if additional

monitoring such as respiratory effort, oximetry, and

electrocardiogram or heart rate, with or without body

position is added to assessment of airway patency by

APAP), and APAP theoretically could be helpful in

diagnosing OSA. The rationale for the use of APAP in

the diagnosis of OSA is that it might provide an

accurate enough diagnosis in some patient groups, it

might reduce diagnostic waiting time, and it might

reduce health care costs.

Studies that evaluate automatic positive airway

pressure for the diagnosis of obstructive

sleep apnea

A limited number of published studies have eval-

uated whether APAP is reliable enough in the diag-

nostic mode to recognize sleep-disordered breathing

(Table 2). All [19–23,28,29,70] except one [42] of

these studies were performed in a supervised envi-

ronment, with concurrent PSG, with airflow during

PSG usually [19,21–23,28,29,70] but not always

[20] monitored with thermistor. Sometimes these

studies were conducted with unselected consecutive

patients [19,21,23], but often they were conducted

in patients who were suspected of having OSA

[20,22,28,29,42,70]. Exclusions were not always

stated, but several groups excluded technically unsat-

isfactory recordings [20,22] or technically unsatisfac-

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 321

Table 2

Summary of diagnostic studies of automatic positive airway pressure devices

Author Device n

AHI Mean

(SE or SD)

Correlation APAP:PSG

(AI:AI) AHI:AHI

Limits of agreement APAP-PSG

(AI-AIdiff ) (95% CI) AHI-AHIdiff (95% CI)

Sensitivity at threshold

AHI(AI) >5

AHI(AI) >10

AHI(AI) >15

AHI(AI) >20

Specificity at threshold

AHI(AI) >5

AHI(AI) >10

AHI(AI) >15

AHI(AI) >20

Gugger et al [19] Autoset — (r = 0.85) — — —

‘‘Autumn’’ — — — —

27 — —

82% 90%

Bradley et al [21] Autoset 25 — — — —

Version 2.0 (SE = 4) r = 0.85 + 3.1 (+ 8.4, � 1.6) — —

37 100% 92%

— —

Kiely et al [22] Autoset 19.4 (r = 0.85) � (� 15.5, + 10.5) — —

Version 3.03 (SD = 24.7) r = 0.92 � (� 15.5, + 13) 85% 87%

36 100% 92%

88% 93%

Fleury et al [20] Autoset — (r = 0.98) + 2.6 (� 11.6, + 6.4) (100%) 76%

— — — (100%) 87%

44 (100%) —

(100%) 88%

Mayer et al [29] Autoset 43.3 (SD = 33.4) — — 97% 50%

— r = 0.87 � 9.6 (� 2.2, + 23.7) 92% 79%

95 86% 86%

79% 93%

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

322

Gugger [23] Autoset 26.2 (SE = 2.9) (r = 0.95) (+ 2.5) (+ 15.6, � 10.6) — —

Version 3.03 r = 0.95 + 4.2 (+ 18.7,10.3) — —

67 — —

97% 100%

Rees et al [28] Autoset 39 (SD = 26) — — — —

Version 3.03 r = 0.9a � 3.1a (+ 11.2, � 17.4)a — —

27 — —

— —

Fletcher et al [42] Horizon — — — — —

63 r = 0.85b — — —

— —

— —

Steltner et al [70] Prototype

(FOT)

19

AHI1= 34.2

(SD = 17.4)

AHI2= 25.4

(SD = 19.6)

—

—

Kw1 = 0.45

Kw2 = 0.5

—

—

—

—

—

—

—

—

Abbreviations: AHI, apneas plus hypopneas per hour (AHI determined by APAP per recording time and determined by PSG per hour of sleep except where noted); AHI-AHIdiff, mean

difference between AHI determined by APAP and AHI determined by PSG; AHI1 and AHI2, apneas plus hypopneas per hour of sleep for scorer 1 and scorer 2, respectively; AI, apneas per

hour; AI-AIdiff, mean difference between AI determined by APAP and AHI determined by PSG; APAP, automatic positive airway pressure; CI, confidence interval; FOT, forced oscillation

technique; Kw1 and Kw

2, weighted kappa for apneas plus hypopneas per hour of sleep computed on second-by-second basis to evaluate agreement between APAP and scorer 1 and scorer 2,

respectively; PSG, polysomnography; r, correlation coefficient; SD, standard deviation; SE, standard error.a This study compared AHI as detected by APAP to AHI per time in bed on PSG.b This study compared AHI as detected by APAP to AHI determined from visual analysis of flow from APAP study.

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

323

tory portions of recordings [28] from analysis. Chronic

obstructive pulmonary disease [22], respiratory failure

[70], awake hypoxemia [22], complicating medical

illnesses [42], severe arrhythmia [70], or suspicion for

complicating sleep disorders [42] were exclusions in

several reports.

Most of the diagnostic studies of APAP have been

conducted with the Autoset device, with early studies

using devices that detected only apneas [19,20] and

later studies using devices that detected apneas and

hypopneas, with [20,22,28,29] or without [21] differ-

entiating different types of respiratory events. In

general, as noted in Table 2, the correlation between

respiratory events detected by APAP and those

detected by PSG was good, with correlation coeffi-

cients of 0.85 to 0.98. The limits of agreement

between APAP and PSG were wide, however, with

mean differences in AHI detected by APAP compared

with PSG ranging from � 9.6 to + 4.2, with 95% of

the true values for the difference ranging from � 15.5

to + 23.7. Often, APAP overscored respiratory events

compared with PSG [19–23], but two groups found

that APAP underscored events [28,29]. In part, this

may be related to different denominators for the AHI

determined from APAP (ie, time in bed) and AHI

determined from PSG (ie, total sleep time in usual

practice and in most of the noted comparative studies).

APAP devices cannot detect sleep. Irregular breathing

in wakefulness may be detected as respiratory distur-

bances and may increase artificially the AHI deter-

mined by APAP, whereas long periods of wakefulness

with regular breathing or time off the device artificially

may decrease the AHI.

The use of thermistors to monitor airflow in PSG

also may lead to a lower AHI than the nasal pressure

used by this APAP device [23]. Sensitivity and spec-

ificity varied with the AHI threshold used to define

disease, with increased specificity with increasing

severity of OSA. In one multicenter study [29], a high

pretest probability added—not unexpectedly—to the

diagnostic accuracy of APAP. In two studies, APAP

was found to be superior to oximetry alone in diagnos-

ing OSA [21,23]. In a study that used a different

prototype APAP device incorporating FOT to monitor

upper airway obstruction [70], Steltner et al found that

APAP yielded diagnostic results similar to visual

analysis of standard PSG performed by two scorers.

In the only unattended (home) study reported to

date [42], Fletcher et al studied 63 patients (screened

for the presence of symptoms of OSA and the

absence of complicating illnesses) using a device that

monitors airflow with a built in pneumotachograph

(Horizon). No gold standard PSG was performed,

although analog flow tracings were examined in an

unblinded fashion by the investigators. The RDIs

determined by APAP correlated with the RDIs deter-

mined by visual analysis (r = 0.85), although the

former were systematically lower than the latter, in

part because the APAP software did not correct for

time off the device. Nine patients could not complete

the study because of failure to tolerate the mask,

inability to hook up the equipment properly, or failure

to return for follow-up. Of the 53 remaining patients

with successful studies, 45 were diagnosed with OSA

by APAP, 35 of whom ultimately returned for APAP

titration studies. The authors reported an average of

1.4 diagnostic studies and 2.4 titration studies to

establish the diagnosis and reach satisfactory treat-

ment pressures. Cost analysis showed that in this

group of patients, the estimated cost for all in-home

APAP studies was less than one fourth the estimated

cost for in-laboratory PSG. There are no published

data on subsequent compliance with CPAP therapy

when the diagnosis of OSA is made using APAP in

an unattended setting. Kreiger et al reported a lower

subsequent objective compliance with CPAP in a

group of patients diagnosed in the ambulatory setting

with a MESAM IV ambulatory monitoring device,

however, as compared with patients diagnosed by

PSG [84].

Taken together, these studies suggest that APAP in

diagnostic mode, after examination of the raw data to

exclude technical problems [19,23], can diagnose

severe OSA effectively, particularly in the presence

of high clinical suspicion and in the absence of

complicating factors. Given the wide limits of agree-

ment between APAP and conventional PSG, the use

of APAP to diagnose less severe OSA is problematic

[21,22]. In the presence of a high clinical suspicion

for OSA and a negative result on APAP, conventional

PSG still plays a role [29]. Although APAP studies

are less expensive than conventional PSG studies

[21,42] and could potentially save costs in some

patients, the overall health care cost-benefit of this

approach on a large scale remains to be clarified.

Role of automatic positive airway pressure in

therapy for obstructive sleep apnea

Rationale for the use of automatic positive airway

pressure in therapy for obstructive sleep apnea

Conventional CPAP, used at an effective pressure

in patients with OSA, has been shown to reduce

nocturnal respiratory disturbances and improve noc-

turnal oxygenation [1,4] and sleep architecture [5].

Regarding longer term clinical outcomes, CPAP

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342324

improves neurocognitive function, including daytime

sleepiness [5–8], cognitive performance [9], driving

performance [10,11], and perceived health status

[7,8,12]. Treatment with CPAP also may be associ-

ated with improvement in mortality [13] and cardio-

vascular endpoints, such as systemic hypertension

[14,15], cardiac arrhythmias [85], nocturnal ischemia

[86], and left ventricular function [87]. Finally,

patients with OSA who are treated with CPAP have

decreased health care use compared with untreated

patients [88].

Despite these benefits, conventional CPAP is not

accepted by all patients [89,90]. An early study using

covert monitoring demonstrated that 35 patients who

were followed over 3 months attempted to use CPAP

only 66% of the monitored days, with a median use

of 4.9 hours [89]. ‘‘Inconvenience’’ and ‘‘stuffy

nose’’ were frequently cited problems, but only a

complaint of ‘‘claustrophobia’’ distinguished regular

users from irregular users; CPAP pressure was similar

in the two groups. In a larger long-term study of

1211 patients in Edinburgh followed for a median of

22 months [90], 68% used CPAP at least 2 hours/

night, with a median use of 5.7 hours. Reasons for

discontinuing CPAP included lack of benefit and

discomfort (including noise and feeling of claustro-

phobia). In multivariate analysis, CPAP pressure was

not found to be a determinant of long-term CPAP use.

In another study of 193 patients with moderate to

severe OSA who were followed an average of

19 months [91], 88% used CPAP every night, with

a mean use of 6.5 hours, despite side effects related to

the mask in 50%, dry nose or mouth in 65%, sneezing

or nasal drip in 35%, and nasal congestion in 25%.

There was no correlation between side effects and

level of pressure. Only 1% reported lack of benefit

from CPAP. It is generally believed, although not

formally studied, that having an inadequate pressure,

(whether too high, perhaps causing increased side

effects, or too low, resulting decreased benefit), can

be associated with decreased CPAP compliance [18].

Intensive CPAP education and support by staff

has been shown to improve compliance [92,93],

whereas lack of technician interaction may decrease

compliance [84].

In conventional CPAP therapy, after the diagnosis

of OSA has been established, patients typically

undergo an attended level I PSG for CPAP titration

[72]. After appropriate patient education and mask

fitting, while sleep and respiratory parameters are

being monitored, CPAP is titrated up manually in a

progressive fashion in the laboratory until an effective

pressure (Peff) is reached. This single Peff then deter-

mines the level of fixed CPAP for long-term home

use. The goals of titration and definition of Peff have

not been formally established and vary in different

studies, however. It has been suggested that the

endpoint of titration should be the abolition of

apneas, hypopneas, snoring, and airflow limitation

[64] with a concomitant decrease in the number of

arousals. Higher pressures are needed to eliminate

respiratory effort related arousals with airflow limita-

tion [65]. CPAP requirements or the RDI within the

same patient may have intranight and internight

variability [94], depending on sleep stage [32,95],

body position [95–97], consumption of alcohol or

other sedatives [98–100], nasal resistance [41,101],

inspiratory airflow [102,103], airway humidification

[103], and body weight [104].

Peff may, although not universally [27], decrease

during the first 8 months of CPAP therapy [105],

perhaps secondary to resolution of upper airway

edema on therapy [106]. Peff cannot be assumed

to be constant. The proportion of patients with

OSA with variable CPAP requirements and the mag-

nitude of that variability have not been well studied.

Rather than the goal of eliminating all respiratory

disturbances and respiratory arousals under all con-

ditions, other clinicians or investigators have used as

an endpoint of CPAP titration elimination of most of

these abnormalities. Varying targets for an acceptable

RDI at Peff (eg, < 5/hour or 10/hour or 15/hour) have

been used. Methods to monitor airflow impact the

measured RDI and the resultant titration results. The

art of titration thus encompasses a fine line between

efficacy (however defined) and side effects. Titration

is time consuming, labor intensive, and expensive

and requires highly qualified technologists. Conven-

tional CPAP therapy, delivered at a fixed Peff deter-

mined by in-laboratory PSG, is the gold standard to

which APAP therapy must be compared.

The rationale for APAP in treatment of OSA is

that the variable pressure delivered by APAP in

response to dynamic changes in airway resistance

might result in improved clinical outcomes compared

with conventional CPAP at fixed Peff, with perhaps

more favorable airway pressures, fewer side effects,

and better compliance. Despite the increased cost of

APAP devices, if fewer therapeutic PSGs need to be

performed or if overall health outcomes improve,

health care costs might be reduced. Alternatively, if

there are adverse outcomes (such as if APAP results

in increased leaks and arousals with higher pressures,

if oxygenation remains suboptimal, or if lack of

technician interaction limits compliance) or if the

group of patients who are candidates for APAP is

small, overall long-term benefits of APAP would

be limited.

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 325

Table 3

Summary of therapeutic studies of automatic positive airway pressure devices

Author

Device n

study design

AHIdiag Mean

(SD or SE)

AHIAPAP

No. < 5, 10, or 15

Mean

Oxygenation:

%TST < 90APAP

mean Sa02APAP

nadir Sa02APAP

Sleep:

Arousal-IAPAP

DeltaAPAP

REMAPAP

Symptoms:

ESSAPAP

Other testAPAP

Pressure:

Mean pAPAP

Peak pAPAP

%TSTAPAP < pConv

ComplAPAP

Pref APAP

Side effects

Berthon-Jones [18] Prototype — <5 in 19/20 — — — 6/20zc, 14/20 #c —

(A,Sn,FL) — — — — — —

20 — — — —

Clinical series

Behbehani

et al [17]

Prototype

(Sn)

5

Clinical series

49.9 < 5 in 3/5,

< 10 in 5/5

4.76

—

—

—

—

—

—

—

—

6.8 #c12.2 X c—

—

—

—

Llorberes Autoset 59 — — 12F 7 Xm — — —

et al [24] 20 (subgroup 9) (SD = 21) 5.6F 6 — 36F 14 Xm — — —

RCT-CO (subgroup) — 27F 13 Xm — —

(Clinical series)

Lofaso et al [33] REM+ with MC+ 51 < 10 in 12/15 39F 101m #d 13F 20 #d — 7.5F 2.5 in 12 —

(only Sn enabled)

15

Clinical series

(SD = 30) 12F 21 #d —

89F 3 zd102F 149m zd55F 31m X d

— 9.9F 2.8 in 12

—

—

—

Meurice

et al [43]

Morphee Plus

16 (8/8)

RCT-parallel

43.6

(SD = 19.8)

—

1.7F 1.2 X c—

96.0F 0.3

—

10.1F 2.5 X c� dz, X c� dz

5.6F 3.7 #d, X cMWT: zd, X cTMT-A: #d, X cTMT-B: X d, X c

—

—

49.3F 14.9

6.5F 1.0

(3wk)zc—

—

Scharf et al [38] Horizon 57.3 — — 9.9F 9.5 X c — — —

Autoadjust (SD = 30.8) 4.4F 2.2 X c 82.6F 3.4 zd, X c 8.6F 7.5 zc — — —

12 — 23.5F 6.0 X c 63.1F 34.2 —

RCT-CO

Sharma

et al [54]

Prototype

(Sn)

20

RCT-CO

50.8

(SD = 28.8)

—

6.1F 5.3 #d, Xm13.9F 25.6 #d, Xm—

79.9F 9.7 zd,#m

11.3F 0.3 #d, Xm17.1F 9.3zd, Xm25.3F 7.4zd, Xm

—

—

—

10.1F 3.8 #m—

—

11/18

(61%) Xm—

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

326

Teschler et al [25] Autoset 60.3 Success 19/20 — 8.9F 0.6 #m, X c — — —

20 (SE = 5.7) 2.8F 0.9 #m, X c — 29.8F 3.1 Xm, X c — — —

RCT-CO 90.4F 0.8 #m, X c 21F1.3 Xm, X c — —

Series [44] Morphee Plus

36 (12/12/12)

RCT-parallel

46.8a

(SD = 22.3)

< 15 in 12/12

�#d, X c6.5F 6.8 #d, X c—

—

�#d, X c�zd, X c� X c

7F 3.7 #dMWT: zd, X c

�#c—

51F 7.9%

6.5F 0.9

(3wk) X c—

—

61.5b

(SD = 27.7)

< 15 in 11/12

�#d, X c11.8F 12.5 #d, X c

—

—

�#d, X c�zd, X c� X c

7.9F 4.0 #dMWT: zd, X c

�#c—

51F 7.9%

6.4F 1.1

(3wk) X c

—

—

Behbehani Prototype 55.2 — — — — 8.4F SE 3.3 #c —

et al [57] (Sn) (SD = 33.7) 5.4F 5.4 #d, X c — �zd, X c — 12.8F SE 4.3 X c —

31 — �zd, X c — —

RCT-CO

Ficker et al [35] REM+Auto 54.1 < 10 in 14/16 — 7.4F 4.1 X c 5.3F 3.9 X c 8.1F 2.1 mbar zc —

16

RCT-CO

(SD = 24) 4.2F 5.1 #d, X c —

—

20.7F 11.9 X c18.1F 5.5 X c

VT: X c —

—

6/16

(37.5%) X c� X c

Konermann

et al [40]

Horizon

50 (48 completed)

(23:25)

RCT-parallel

35.5

(SD = 9.6)

—

2.4F SE 1.6 #d, X c0.1 #d, X c94.7F 1.4 zd, X c90.3F 3.6 zd, X c

2.3F 7.4 #d, #c27.2F 16.5 zd, zc

20.1F10 zd, X c

—

—

6.5F 1.7 #c—

—

5.9F 1.6

(3–6mo) X c—

—

Boudewyns

et al [36]

REM+Auto (1.6)

15

non-RCT

Median: 65.8

(CI:48.6–80.3)

—

2.1 (0.9,3.2) #d, X c—

X c—

8.4 (5.4,12.8) #d, X c�zd, X c�zd, X c

5 (3,11) X c

—

5.2 (4.9,6.8) X c—

—

6.1(5.2,6.8)

(2 mo)

—

� X cGagnadoux Autoset 69.6 < 10 in 21/24 0.2F 1 #d — — — —

et al [31] 24 (SD = 29.8) 5.7F 4.6 — 39.5F 15.9 zd — — —

Clinical series — 18.8F 8.8 zd —

Miyazaki Virtuoso 68.3 — — — — — —

et al [56] 11 (SD = 20.2) 9.6F 14.5 Xm — — — 9.4F 2.0 —

RCT 89F 3.7 #m — — —

Randerath Somnosmart 31.6 — — 21.2F 13.1 #d — 5.4F 1.0 mbar #c —

et al [48] 11 (SD = 26.6) 3.4F 4.5c #d 94.4F 2.4 X d 12.3F 8.8 X d — 12.3F 3.2 mbar —

RCT-COc,d 85.6F 7.4 zd 22F 7.7 zd 91.7F 9.3 —

(continued on next page)

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

327

Table 3 (continued )

Author

Device n

study design

AHIdiag Mean

(SD or SE)

AHIAPAP

No. < 5, 10, or 15

Mean

Oxygenation:

%TST < 90APAP

mean Sa02APAP

nadir Sa02APAP

Sleep:

Arousal-IAPAP

DeltaAPAP

REMAPAP

Symptoms:

ESSAPAP

Other testAPAP

Pressure:

Mean pAPAP

Peak pAPAP

%TSTAPAP < pConv

ComplAPAP

PrefAPAP

Side effects

— — 24.5F 10.2 #d — 5.1F 0.7 mbar #c —

5F 7.2d #d 93.7F 2.9 X d 12.3F 7.6 X d — 11.8F 2.1 mbar —

86.7F 6.9 zd 23F 7.9 zd 90.4F 6.3 —

d’Ortho et al [37] REM+Auto(2.1)

25

RCT-CO

57.8

(SD = 5.8)

< 10 in 16/25

10.6F 9.3 #d, X c8.8F 20.5 m #d, X c95.6F 1.6 zd, X c85.2F 9.0 zd, X c

15.5F 8.9 #d, X c87F 40 mzd, X c21F 8 m zd, X c

9.3F 4.8 #d, X cSQ: 32F 11 X c

8.8F 1.8 #c—

—

4.1F1.8

(2 mo) X c15/25 (60%)

� X cFicker et al [49] Somnosmart 48.0 < 10 in 17/18 — 6.6F 2.1 X c 5.6F 1.8 X c 0.84F 0.26 kPa #c —

18

RCT-CO

(28.1) 3.4F 3.4 X c —

—

19.3F 6.6 X c21.7F 4.9 X c

— —

—

8/18

(44%) X c� X c

Fletcher et al [42] Horizon 34.1 — — — 10.5F 0.9 #d 9.4F 0.6 —

30 (SD = 4) 8.6F 0.8 #d — — MSLT: 5.7F 0.8zd 12.9F 0.6 —

Clinical series — — — —

Hudgel [55] Virtuoso

60(39 completed)

RCT-CO

30

(SE = 4)

—

—

—

—

—

—

—

—

9F 1 #d, X c

—

6.4F 0.4 #c—

—

6F 0.3

(12 wk) zc—

—

Randerath Somnosmart 18.2 — — 22.2F 9.7 X d — 5.6 + /12.1 mbar #c —

et al [50] 10 (SD = 13.3) 2.5F 1.9e #d — 20.2F 10.4 zd — 13.9F 3.2 mbar —

RCT-COe,f — 19.6F 2.3 X d 73.6F 31.4 —

— — 22.9F 8.1 X d — 7.3F 1.6 mbar #c —

1.8F 0.7f #d — 22.3F 9.3 X d — 13.4F 3.5 —

— 18.3F 6.4 X d 48.6F 45.1 —

Teschler

et al [30]

Autoset

10

RCT-CO

52.9

(SD = 8.1)

—

3.5F 1.7 #d—

—

—

7.7F 2.4 #d24.6F 2.8 zd25.9F 1.4 zd

—

—

Median 7.6F 0.4 #c�zc

6.3F 0.4

(2 mo) X c—

—

Randerath Somnosmart 32.2 — — 16.5F 9.4 #d, X c — 5.7F 2.1 mbar #c —

et al [52] 25 (SD = 18.1) 5.5F 3.8 #d, X c — 21.6F 10.9 zd, X c SQ:6.8F 2.6 zc 12.6F 4.6 —

RCT-CO 87.0F 4.2 zd 20.3F 7.3 zd, X c � X c

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

328

Randerath

et al [51]

Somnosmart

52 (47 completed)

RCT-CO

35.1

(SD = 26)

—

5.3F 5.6 #d—

—

92F 5 X d

10.3F 6.4 #d18F 12 X d

57F 19 zd

� 7.8F 4.7 #d—

6.6F 2.4 c#14.3F 4zc81.5F 21

5.3F 1.6

(6 wk) X c35/47

(75%) zc�X c

Fuchs et al [53] Somnosmart 47.7 — — 14.5F 6.6 — — —

30 (SD = 21.9) 4.7F 4.7 — — — — —

Clinical series — — — —

Marrone et al [32] Autoset Clinical 64.8 — — 11.5F 6.5 — — —

15 (SD = 25.4) 1.8F 1.5 — 7F 8.4%TIB — — —

Clinical series 91.6F 3.5 14.2F 8%TIB —

Data are presented as means F standard deviation or standard error, as appropriate to each study, unless noted as median and 95% confidence interval.

Abbreviations: A, apnea; AHIAPAP, number of apneas plus hypopneas per hour of sleep on APAP; AHIdiag, number of apneas plus hypopneas per hour of sleep on baseline diagnostic night;

APAP, automatic positive airway pressure; Arousal-IAPAP, number of arousals per hour of sleep on APAP; CI, 95% confidence interval; CO, cross-over; ComplAPAP, compliance with APAP

in h/day of use over defined follow-up period; CPAP, continuous positive airway pressure; DeltaAPAP, amount of delta (slow-wave, stage 3 + 4) sleep on APAP [in % total sleep time unless

stated as %time in bed (%TIB) or minutes(m)]; ESSAPAP, Epworth Sleepiness Scale on APAP; FL, flow limitation; mean Sa02APAP, mean nocturnal arterial oxygen saturation on APAP (%);

Mean pAPAP, mean positive airway pressure level on APAP (cm/H20 unless noted as mbar or kPa); MSLT, mean sleep latency test (minutes); MWT, maintenance of wakefulness test

(minutes); nadir Sa02APAP, nadir arterial oxygen saturation on APAP (%); Other test APAP, semiquantitative or objective test of symptoms (sleepiness or performance) on APAP; Peak pAPAP,

peak positive airway pressure level on APAP (cm/H20 unless noted as mbar or kPa); Pref APAP, proportion (percentage) preferring APAP to CPAP; RCT, randomized controlled trial; SD,

standard deviation; SE, standard error; Sn, snoring; SQ, sleep questionnaire (specific to study and not standardized); %TST < 90APAP, percent total sleep time with arterial oxygen saturation

less than 90% on APAP (in %, unless stated as minutes (m)); %TSTAPAP < pConv, percent total sleep time on APAP at positive pressure less than conventional fixed CPAP as determined by

manual titration; TMT-A, trail-making test A; TMT-B, trail-making test B; VT, vigilance test.

X d, #d, zd: no change from, lower than ( P < 0.05), or higher than ( P < 0.05) diagnostic night, respectively.

Xm, #m, zm: no change from, lower than ( P < 0.05), or higher than ( P < 0.05) manual CPAP-titration night, respectively.

X c, #c, zc: no change compared with, lower than ( P < 0.05), or higher than ( P < 0.05) with conventional fixed CPAP as determined by manual titration.a APAP device reference pressure set at effective pressure determined by manual titration.b APAP device reference pressure set at effective pressure estimated by a formula.c APAP device pressure range set at widest possible range (4–15.5 mbar).d APAP device pressure range set with maximum acceptable pressure calculated from formula based on effective pressure determined by manual titration and lower limit set at 4 mbar.e APAP device pressure range set at widest possible range (4–15.5 mbar).f APAP device pressure range set with minimum acceptable pressure calculated from formula based on effective pressure determined by manual titration and upper limit set at

15.5 mbar.

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

329

Studies that evaluate automatic positive airway

pressure for the therapy for obstructive sleep apnea

Multiple studies that evaluated APAP devices for

the therapy of OSA have been published in the

English literature (Table 3). Most have been single-

night studies, although some have been performed for

3 weeks to 6 months to assess certain outcomes

[30,37,40,43,44,51,55]. APAP usually has been

studied in an attended setting, in which a technician

could assess for leaks or other problems and intervene

as necessary. Study designs, APAP devices, defini-

tions, monitoring techniques, and outcome measures

have varied across studies. Most of the patients

studied have had moderately severe or severe OSA,

as indicated by the mean baseline AHI for each study

typically being in the 30/hour to 50/hour range, with

a large standard deviation (see Table 3). Only one

study included a small group of four patients with

upper airway resistance syndrome [55]. All patients

were diagnosed by PSG, except in one study in which

patients were diagnosed at home with APAP and

subsequently treated at home with APAP [42]. Many

patients were exposed to CPAP before treatment

with APAP. Most studies also listed exclusions

[18,25,30,35,37,38,40,42,44,47,49,52,54,55,57].

Patients with nonobstructive sleep-related breathing

disorders (eg, hypoventilation syndromes, Cheyne-

Stokes respiration, central sleep apnea) or compli-

cating medical illnesses (eg, congestive heart failure,

chronic obstructive pulmonary disease, respiratory

failure, cerebrovascular disease) were frequently

excluded. Patients with other sleep disorders (eg,

narcolepsy, periodic limb movement disorder, restless

legs syndrome), previous velopharyngeal surgery, or

need for increased CPAP level on CPAP titration

night (eg, >14–15 cm H20) also were excluded in

some studies.

Indexes of upper airway obstruction

As noted in Table 3, the mean AHI was signifi-

cantly reduced with APAP as compared with the

baseline diagnostic night in every study to date. The

AHI was not always reduced to normal in all

patients, however. In some patients, therapy with

APAP was not possible or was problematic because

of inability of the device to detect evidence of upper

airway obstruction or because of significant leaks.

Lofaso et al, using a device that exclusively detects

snoring, reported that APAP was ineffective in 3 of

15 patients, one with non-heavy snoring and two

with mouth breathing/leak [47]. The device in-

creased its pressure in response to snoring in only

84%Ff

6% of snoring events. In that study and in

another report by Miyazaki et al using a similar

device [56], esophageal pressure swings were

reduced with APAP, which indicated improved upper

airway obstruction. Miyazaki et al found that the

improvement in esophageal pressure was less than

that with manually adjusted CPAP, however. Tesch-

ler et al, using a different device that detects apnea,

snoring, and flow limitation, also noted that high

leak precluded single-night APAP use in 1 of 21 pa-

tients [25]. There is also one case report of a patient

with moderately severe OSA (AHI: 35.3/hour) who

was stable on CPAP 8 cm H20 who subsequently

developed central apneas and arousals when treated

with APAP [107]. In general, in the studies that

compared APAP to conventional fixed CPAP, the

improvement in AHI was similar in the two groups,

with no advantage of one mode of therapy over

the other.

Nocturnal oxygenation

In all of the studies in which the effect of APAP

on nocturnal oxygenation was examined, some or all

measures of oxygenation (eg, time with Sa02 < 90%,

mean Sa02, and mean nadir Sa02) improved com-

pared with the baseline diagnostic night (see Table 3).

Oxygenation did not necessarily normalize in all

patients, however. The improvement in oxygenation

with APAP was generally similar to the improvement

with manually titrated or conventional fixed CPAP. In

three studies that used different devices, however, the

mean nadir Sa02 was less with APAP than with

manually titrated CPAP [25,54,56].

Sleep architecture

The varying pressure supplied by APAP might be

expected to result in disturbed sleep; however, most

studies have shown improvement in sleep architec-

ture with APAP compared with the baseline dia-

gnostic night (see Table 3). The amount of sleep

fragmentation improved with APAP, as indicated by a

decrease in the number of arousals per hour of sleep

(arousal index) in almost all studies. Similarly, most

studies showed an increase in delta or slow wave

(stage 3 + 4 non-rapid eye movement[REM]) sleep

on APAP, and some studies showed an increase in

REM sleep. Sleep architecture was usually similar

with APAP and with manually titrated or conven-

tional fixed CPAP. Teschler et al [25], however,

reported a lower arousal index with APAP compared

with manually titrated CPAP but not with conven-

tional fixed CPAP. Scharf et al [38] and Konermann

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342330

et al [40], each using a device that monitors snoring

and respiratory events, reported more delta sleep with

APAP compared with conventional fixed CPAP.

Improvements in sleep architecture, along with

improvements in AHI and oxygenation, have been

shown to be maintained over 6 weeks [51] and up to

6 months [40]. A more detailed analysis of whether

changes in pressure with APAP could induce arousal

was performed by Fuchs et al [53]. Thirty patients

with OSA were studied with PSG during APAP

therapy using a device controlled by impedance. As

in other studies, the overall number of arousals during

sleep improved with APAP compared with baseline.

During periods of sleep time in which there was a

pressure variation by more than 0.5 mbar, however,

there was an increase in arousals as compared with pe-

riods of time in which pressure was constant. There al-

so was considerable interindividual variability among

patients, with some having relatively few pressure-

dependent arousals and others having up to 61% of

arousals being classified as pressure dependent.

Obstructive sleep apnea symptoms

Changes in nocturnal and daytime symptoms of

OSA have been assessed with subjective and objec-

tive tests (see Table 3). Randerath et al [52] used a

sleep questionnaire to evaluate the quality of sleep in

25 patients randomly assigned to conventional fixed

CPAP or impedance-controlled APAP in a single-

blind cross-over comparison. The 16 patients who

completed the questionnaire rated the quality of

their sleep higher with APAP than with CPAP. In con-

trast, d’Ortho et al [37], using a randomized cross-

over design with 2-month treatment periods to study a

different APAP device that senses snoring and respi-

ratory events, found similar OSA symptom scores on

their sleep questionnaire in the APAP and conven-

tional fixed CPAP groups. Daytime sleepiness has

been assessed most commonly with the Epworth

Sleepiness Scale (ESS) [108], which ranges from 0

(least sleepy) to 24 (most sleepy), depending on the

patient’s perception of level of sleepiness in eight

situations. In general, ESS has decreased (improved)

with APAP therapy, similar to conventional fixed

CPAP. Hudgel and Fung reported similar improve-

ments in ESS in a subgroup of four patients with upper

airway resistance syndrome treated with APAP or

conventional fixed CPAP over a 12-week period

[55]. In the only study that did not note improvement

in ESS with APAP, the baseline ESS was already in

the normal range [36]. Multiple Sleep Latency Test

[109,110] and Maintenance of Wakefulness Test [111]

results have confirmed objective improvements in the

latency to sleep in the daytime in patients treated with

APAP (two studies used devices that sense respiratory

events and one used a device that senses snoring and

respiratory events), again similar to conventional fixed

CPAP [42–44]. Meurice et al [43] used two trailmak-

ing tests (TMT-A and TMT-B) to assess alertness and

concentration after 3 weeks of APAP or 3 weeks of

conventional fixed CPAP. TMT-A score improved to a

similar degree with APAP and CPAP, whereas TMT-B

did not change. Finally, Ficker et al [35], using a device

that senses snoring and respiratory events, reported

that a standardized vigilance test normalized in all

patients after a single night of treatment with APAP or

conventional fixed CPAP.

Cardiovascular outcomes

There are no published reports to date on the acute

or chronic effects of APAP on blood pressure or other

cardiovascular outcomes.

Positive airway pressure levels

Almost all studies show a decrease in mean

treatment pressure with APAP compared with con-

ventional fixed CPAP (see Table 3). In some studies,

more than 50% of total sleep time on APAP was

spent at a pressure level less than Peff determined by

conventional manual CPAP titration (Peffconv) [45,48,

51]. The average mean APAP pressure was lower

than Peffconv by 0.9 cm H20 [37] to 3.1 cm H20 [57];

however, this was not true for all patients in each

study. The mean peak pressure delivered by APAP

was often higher than Peffconv, however (see Table 3).

In the original report of APAP therapy using a device

that responded to apnea, snoring, and flow limitation

[18], expiratory leak through the lips confused the

auto-setting algorithm in 6 of 20 patients, which led

to increased pressure in these patients. In another

report, Teschler et al noted that high leak caused

unnecessary increases in pressure in 3 of 21 patients

[25]. Randerath found that whereas the average of

mean APAP pressure was lower than Peffconv by 2.1 cm

H20, the range of differences was 6 cm H20 lower to

4 cm H20 higher than Peffconv [51]. The magnitude of

the difference between APAP mean pressure and

Peffconv has been shown to depend at least partially

on the algorithm used to select Peffconv and the

algorithm controlling the APAP device. Sleep stage

and body position are also important in some patients.

Mean positive airway pressure levels with APAP have

been shown to decrease during delta sleep compared

with stage I-II non-REM sleep and REM sleep [32,35,

43,44] and in the lateral position compared with the

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 331

supine position [44]. Many patients with body posi-

tion–dependent OSA also may have sleep stage–

dependent OSA [46]. In a randomized parallel group

study [46], the night-to-night variability of pressure

levels with APAP was higher in body position– and

sleep stage–dependent patients than independent

patients. Six patients with body position and sleep

stage dependence treated with 3 weeks of APAP had

less subjective and objective sleepiness than a similar

group of six patients treated with CPAP, which sug-

gests that there may be a treatment advantage for

APAP over CPAP in this group.

Side effects

Side effects with APAP as compared with con-

ventional fixed CPAP have been examined, either in

the form of questionnaires [35–37,49,51,52]or sur-

rogates, such as use of humidifiers to correct nasal

symptoms [44] (see Table 3). Patients were more

aware of pressure variations when treated with

APAP than CPAP in one report [51], felt the

pressure was lower on APAP than CPAP in another

report [35], and had a tendency to report breathing

against the machine more with APAP than CPAP in

yet another report [49], but no other differences have

been noted.

Compliance

Compliance with APAP therapy has been exam-

ined over 3-week to 3-to- 6-month time periods (see

Table 3). Some studies [40,43,55], but not all

[30,37,44,51], have found that some measures of

compliance improved with APAP compared with

conventional fixed CPAP. Meurice et al reported

7.1 Ff

1 hour nightly use in the APAP group com-

pared with 5.1 Ff

1.1 hours in a parallel CPAP group

[43]. In another randomized parallel group study,

Konermann et al reported similar compliance be-

tween APAP and CPAP in terms of hours per night

but increased compliance with APAP compared

with CPAP in terms of nights per week (6.5 Ff

0.4

and 5.7 Ff

0.7, respectively) [40]. Hudgel and Fung,

in a randomized cross-over study [55], found in-

creased nightly use of APAP compared with CPAP

(6Ff

0.3 hours and 5.5Ff

0.3 hours, respectively) but

no difference in nights of use, cumulative hours of

use, or patterns of use. In a subgroup of four patients

with upper airway resistance syndrome, cumulative

hours of use and nights of use were higher with

APAP than CPAP.

Preference

Several single night and longer-term studies have

evaluated whether patients prefer APAP or conven-

tional fixed CPAP (see Table 3). In general, patient

preferences were not different between the two mo-

dalities, with the exception of one single blind study

in which 35 of 47 patients (75%) preferred APAP.

Health care costs

In a strategy that used APAP for in-home diagnosis

and therapy [42], Fletcher et al reported cost savings

with APAP compared with conventional therapy. No

other systematic comparisons have been published.

In summary, these studies suggest that APAP can

be an effective therapy for OSA in patients without

complicating sleep or medical diagnoses. APAP ther-

apy can result in a reduced AHI, although devices

that predominantly detect snoring as a measure of

upper airway obstruction may be less effective. Not

all patients can achieve equivalent results. Sleep and

oxygenation parameters improve, although there may

be a somewhat lower Sa02 nadir with APAP than

CPAP. OSA symptoms also improve. Mean airway

pressures tend to be lower with APAP, without

significant change in side effect profile. Compliance

and preference tend to be similar or somewhat better

with APAP. Patients with sleep-stage and body posi-

tion-dependent OSA may gain the most from APAP

therapy, but further work is needed to define the most

appropriate patients for this modality. The effects of

APAP on cardiovascular outcomes and health cares

costs and the differences between devices also require

further study.

Role of automatic positive airway pressure in the

titration of continuous positive airway pressure

for obstructive sleep apnea

Rationale for use of automatic positive airway

pressure to determine an effective continuous positive

airway pressure in patients with obstructive

sleep apnea

Traditionally, in patients with OSA who are

treated with conventional fixed CPAP, a full-night

attended PSG for manual CPAP titration to determine

Peff (as described earlier) follows the initial diagnostic

night, which requires two separate studies for diag-

nosis and therapy. ‘‘Split-night’’ PSG, with the first

half of the night to establish the diagnosis and the

second half of the night to titrate CPAP, is an

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342332

accepted alternative used by many centers in patients

who meet certain criteria [72,73]. Although this

technique has demonstrated efficacy and saves the

cost of one PSG, not all patients ultimately found to

have OSA meet the diagnostic criteria early enough

in the night to initiate a CPAP titration on the same

night. Even in patients who do undergo a CPAP trial

in a split-night format, not all patients are titrated

successfully by the end of the night. A second study

may be required to titrate CPAP.

Rather than using full-night or split-night in-

laboratory PSG, some investigators have used pre-

diction formulae [112], patient and bedpartner

reports [113], and limited unattended or attended

respiratory F sleep portable monitoring [81] to help

determine Peff. Home titration to determine Peff in

patients with established OSA using a four channel

portable device in the home (EdenTech, Eden

Prairie, MN) guided by patient or bedpartner inter-

view [113] was found to be feasible and lead to

improvement in AHI on follow-up PSG at machine-

derived Peff (Peffdevice). In a study of 17 patients with

OSA attended by a registered nurse or polysomno-

graphic technician in the home using the same

device for titration [114], AHI was reduced on the

titration night at a lower cost than conventional PSG.

Compliance on CPAP determined by Peffdevice was

similar to historical controls at 18 and 13 months

follow-up in both studies, respectively. Using a

different device that records cardiorespiratory data,

airway pressure, and sleep (VITPAP, Vitalog HMS-

5000, Vitalog Monitoring Inc., Redwood City, CA),

unattended machine-controlled titration was per-

formed in 21 unselected patients with OSA [115].

In the 19 patients who completed the machine

titration, the Peff was determined after the recording

was reviewed visually and scored by the investiga-

tors. This Peffdevice was highly correlated (r = 0.90)

with Peff determined by conventional manual CPAP

titration (Peffconv), with a mean difference of 0.21 F

1.08 cm H2O. Cardiorespiratory complications

occurred in six patients, however, including mild

discomfort that required a resetting of CPAP pres-

sure, central apnea with oxygen desaturation of more

than 85%, and ventricular ectopy, with termination

of the procedure required in two patients.

Subsequently, unattended home CPAP titration

was studied in 30 patients with OSA using a portable

respiratory and sleep monitoring system with modem

technology that allowed transfer of data from home to

the laboratory (NightWatch, Healthdyne) [116] and

compared with in-laboratory titration in a parallel

group of 30 patients. In this study, patients were

excluded if they had severe cardiopulmonary or renal

disease or important arrhythmia or if they required

oxygen or nocturnal ventilation, and all patients in

both groups met with a respiratory therapist for a

pretitration education session. AHI and sleep stage

distribution at follow-up PSG at 6 to 8 weeks and

objective compliance were similar in the group using

fixed CPAP at Peffdevice and the group using fixed

CPAP at Peffconv.

Instead of using APAP with the goal of long-term

treatment, there have been attempts to use APAP in

the short term (one or several nights), similar to other

portable monitoring systems, to determine Peff after

an initial diagnostic PSG. This PeffAPAP then could be

administered long term at a fixed level at home using

a conventional CPAP device. The Peffconv is the gold

standard to which the PeffAPAP must be compared.

Titration with APAP can be done during an attended

study, in which the advantage over traditional CPAP

titration might be freeing up technician time. It also

could be done in the unattended home setting over 1

or more days as a way of determining a more

effective level of CPAP for the long term, given that

sleep might be expected to be more normal at home

than in the laboratory. Turnaround time between

diagnosis and therapy potentially could be improved.

By eliminating the need for a second study, cost

savings also could be realized, especially compared

with a traditional 2-night approach to diagnosis and

therapy. Because patient-technician interaction is lim-

ited with the use of home titration with APAP,

however, if the patient does not have a successful

autotitration, long-term adherence and compliance

might be adversely affected.

Studies that evaluated automatic positive airway

pressure for determining an effective continuous

positive airway pressure in patients with

obstructive sleep apnea

Studies that evaluated APAP in determining Peffare summarized in Table 4. Patients included in

these studies were previously diagnosed with OSA

by laboratory-based or portable PSG and gener-

ally had a baseline mean AHI in the severe range

(Table 4). Usually patients were not previously

treated with CPAP [24,26,30,31,34,45,68], but this

was not always stated [25,39]. As in studies that

evaluated the role of APAP for diagnosis and

therapy, patients with complicating medical or sleep

disorders were often [25,26,30,34], but not always

[24,31,39,45,68], excluded. Studies often were per-

formed in an attended setting so that the technician

(or another health professional in the case of par-

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 333

Table 4

Summary of titration studies using automatic positive airway pressure devices to determine an effective continuous positive airway pressure level for treatment of obstructive sleep apnea

Author

Device

n

Study design

Setting of APAP titration

AHIdiag

Mean (SD or SE)

Peffconv

MeanF SD or SE

PeffAPAP (a)

MeanF SD or SE

AHI-PeffAPAP

Mean F SD or SE

Llorberes et al [24] Autoset 53.3 10.1F1.8 10.3F 1.5 (A1) Xm —

20 (SD = 19) 11.5F 2.9 (B4) XmRCT-CO 10.7F 2.7 (B4) XmPartially attended—hospital

Teschler et al [25] Autoset 60.3 8.6F 0.4 9.9F 0.4 (A4) zm 2.5F 0.7

20 (SE = 5.7)

RCT-CO

Attended

Stradling et al [39] Horizon (H definition changed) — 8.7F 2.5 8.2F 2.1(A3) Xm —

112 (52/61)

RCT-Parallel

Unattended—laboratory

Teschler et al [26] Autoset

20

RCT-CO

Attended (at 3 mo and 8 mo

follow-up of prior study

group from Teschler [25]

60.3

(SE = 5.7)

11.0F 0.5 (at 3 mo)

10.4F 0.4 (at 8 mo)

10.6F 0.4 (A4) Xm (at 3 mo)

9.7F 0.5 (A4) Xm (at 8 mo)

4.3F 0.6 # (at 3 mo)

3.6F 0.5 # (at 8 mo)

Berkani et al [34] REM + auto (Only Sn enabled) 55 — 10.5F 2.2 (B1) 7F 5 #

10 (SD = 16)

Clinical series

Unattended—hospital

Badia et al [68] Prototype (FOT)

28

Simultaneous recording

Attended—laboratory

63.7F 3.1

(nap study, n = 14)

67.3F 2.89

(overnight study, n = 14)

10.6F 0.6

9.9F 0.7

11.1F 0.6 (A2) Xm9.9F 0.6 (A2) Xm

—

—

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

334

Gagnadoux et al [31] Autoset 69.6 — 11.2F 1.6 (A4) 4.1F 3.2 # at 3 mo

24 (SD = 29.8) < 10 in 17/18

Clinical series

Attended—laboratory

Series [45] MorpheePlus 43.6 — 10F 1.7 (B6) 4.8F 6 at 2 wk

42 (SD = 19.8) (1 wk) < 10 in 38/40

Clinical series 9.7F 1.1 (B6)

Unattended—home

(1- or 2-wk titration)

(2 wk)

Teschler et al [30] Autoset 52.9 9.4F 0.6 10.3F 0.4 (A3)zm —

10 (SE = 8.1) (attended)

RCT-CO 10.1F 0.5 (A3)zmAttended—laboratory

(3 d titration at d 0,60,120)

(unattended)

Unattended—home

(12 d titration over 2 mo)

Data are presented as means F standard deviation or standard error, as appropriate for each study.

Abbreviations: AHI-PeffAPAP, apneas plus hyponeas per hour of sleep on fixed CPAP at effective pressure as determined by APAP titration (AHI determined by conventional PSG unless #

to indicate portable home study); APAP, automatic positive airway pressure; FOT, forced oscillation technique; H, hypopnea; PeffAPAP, effective pressure (cm H20) as determined by APAP

titration; Peffconv, effective pressure (cm H20) as determined by conventional manual titration; SD, standard deviation; SE, standard error; Sn, snoring.

Xm, #m, zm: no change from, lower than (P < 0.05), or higher than (P < 0.05) Peff from manual CPAP-titration night, respectively.a Method of determination of Peff

APAP: A, review of raw data to exclude periods of leak or poor recording prior to determining Peff; B, no mention of review of raw data prior to

determining Peff; 1, highest pressure; 2, pressure that eliminates upper airway obstruction events; 3, pressure that eliminates most upper airway obstruction events; 4, P95: pressure that is

exceeded only 5% of the time: 5, P90: pressure that is exceeded only 10% of the time; 6, pressure determined by percentage of time spent below reference pressure (in turn determined by a

formula incorporating body mass index, neck circumference, and AHI), constrained by a range + 3cm H20/� 4cm H20.

F.J.

Roux,

J.Hilb

ert/Clin

Chest

Med

24(2003)315–342

335

tially attended studies) could monitor for leaks or

other technical problems (see Table 4). Only one

study, a clinical series with historical controls [45],

was performed in an unattended home setting in

CPAP-naıve patients. The amount of time for APAP

titration varied from a single night in most studies,

to 1 or 2 weeks [45], to an average of 12 nights

over 2 months [30].

Studies differed in device used, study design,

setting, and outcomes assessed (see Table 4). The

primary outcome was usually PeffAPAP, which was

either compared directly with Peffconv or assessed for

effectiveness by follow-up conventional fixed CPAP

therapy at PeffAPAP. Methods of determining Peff

APAP

varied across studies (see Table 4). The raw data

usually were excluded to eliminate periods of high

leak or poor recording. In some reports, however,

data review was not specifically mentioned or was

not performed. Subsequently, PeffAPAP was deter-

mined by analysis of pressure during the APAP

titration night(s) and was variably defined as the

highest pressure of the recording, the pressure that

eliminated all or most upper airway obstruction

events, the pressure that was exceeded only 5% or

10% of the time (P95 or P90, respectively), or the

percent of time spent below a reference pressure as

determined by a formula (see Table 4). Methods to

determine Peffconv varied, and the goals of the

conventional manual titration were not necessarily

the same as the goals of the APAP titration [26].

As outlined in Table 4, most patients had a

successful APAP titration, and PeffAPAP and Peff

conv

were similar in most studies. PeffAPAP also has been

shown to be stable over 8 months of follow-up [26].

Teschler et al, using a device that detects snoring,

apnea, hypopnea, and flow limitation, initially noted

that PeffAPAP was higher than Peff

conv by an average

of 1.3 F 0.3 cm H20 [25], despite excluding periods

of leak. This difference subsequently decreased in a

follow-up study of this same group of patients after

changing the goals of the manual titration to be more

similar to those used by the device [26].

Not all patients were able to have PeffAPAP deter-

mined with APAP titration. Teschler reported that

high leak prevented autotitration in one patient and

caused unnecessary increases in pressure in 3 of 21

patients, although for most of the night, leak was low

(< 0.4 L/second) [25]. During the APAP titration

night, the technician reseated the mask an average

of 1.9 F 0.4 occasions per patient per night, a

frequency similar to the manual CPAP titration night.

Gagnadoux et al also reported leaks of more than 0.4

L/second in 3.1% F 4.8% of titration time (range 0%

to 15%), with the technician repositioning the mask

an average of 0.93 F 0.46 times per night [31].

Periods of continuous leak occurred in 7 of 14 pa-

tients and ranged from 4% to 70% of total sleep time

during a manual titration when impedance was simul-

taneously monitored with FOT and subsequently

interpreted in a blinded fashion by the investigators

[68]. As in the previous studies, these periods of leak

were excluded before determining PeffAPAP.

Llorbes et al found that PeffAPAP determined by

review of raw data to exclude mask leaks and atypical

pressure changes followed by visual inspection to

determine the highest level of pressure was similar to

PeffAPAP determined by P90 or P95. Berkani et al

reported that APAP titration, using a device adjusted

to detect only snoring, was unsuccessful in two of ten

patients, one of whom had a laryngectomy for laryn-

geal cancer and one of whom underwent uvulopalato-

pharyngoplasty [34]. These two patients ultimately

had successful titration when the APAP pressure

range was less constrained. Increased mouth leak,

even at low CPAP pressures, previously has been

reported in patients who have undergone uvulopalato-

pharyngoplasty [117]. Series reported that 2 of 42

patients were not successful with home APAP titra-

tion using a device that detects apneas and hypopneas

and operates within a set range of a reference pres-

sure, 1 because of central apnea and 1 because of

machine malfunction [45].

Gagnadoux et al, using a device that detects

snoring, apnea, hypopnea, and flow limitation, found

that APAP titration was unsuccessful (defined as an

AHI > 10/hours on subsequent PSG at fixed PeffAPAP)

in 3 of 24 patients, perhaps because of severity of

OSA [31]. All 3 patients had a high AHI at baseline

that ranged from 95/hour to 123/hour. Finally, the

variability of APAP pressure levels was studied in

relation to sleep architecture in 15 patients on home

therapy with APAP [32]. The highest pressures gen-

erally occurred during periods of drowsiness or

fragmented non-REM sleep, which suggested that

if APAP had been used for titration at home in

patients with poor sleep quality, PeffAPAP could have

been overestimated.

After successful determination of PeffAPAP, pa-

tients have been treated with CPAP at fixed PeffAPAP

and other outcomes have been assessed. Mean AHI

on repeat PSG or limited home monitoring with

CPAP at fixed PeffAPAP improved compared with

the baseline diagnostic night (see Table 4). Improve-

ments in sleep architecture [34,45], including a

decrease in arousal index and an increase in delta

sleep and REM sleep and improvements in nocturnal

oxygenation [34,45], also have been reported. ESS

also improved compared with baseline [31,45] and

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342336

was similar for patients treated with CPAP at fixed

Peffconv [39].

Compliance with CPAP at fixed PeffAPAP has been

evaluated in several studies. Subjective compliance

was assessed 6 weeks after APAP titration in 52

patients and compared with a parallel group of 61

patients after manual CPAP titration. The proportion

of successful patients who were established on CPAP

in the APAP group was similar to the manual group

(73% and 64%, respectively; 25% and 23%, respec-

tively, were undecided about CPAP therapy). Fewer

patients in the APAP group (2%) had decided defi-

nitely against CPAP therapy compared with the

manual group, however (13%). In two other studies,

objective 3-month compliance by hour meter was

4.9 F 0.3 hours/night in 20 patients [25] and 5.25 F1.82 hours/night in 18 of 22 patients [31], both

similar to historical controls. Finally, in a study in

which PeffAPAP was determined by home titration, 86%

of patients initially accepted CPAP for home therapy,

and objective compliance was 6.1 F 1.7 hours/night

in 36 of the 40 patients who were successful with

APAP titration [45].

Minimal data are available on the impact of

APAP titration on health care resources. In using

APAP for titration in the attended setting, Teschler et

al and Gagnadoux et al each reported an average of

one to two technician interactions per patient per

night—fewer than would be expected on a manual

titration night, thus potentially reducing technician

workload [25,31]. Berkani et al, using APAP in an

unattended setting, estimated that the cost of the 12

ambulatory studies required to determine PeffAPAP in

10 patients was less than that of conventional manual

titration [34].

Overall, these studies suggested that APAP can be

a useful modality in uncomplicated patients to deter-

mine Peff for long-term conventional CPAP therapy.

The best device and best method for determining Peffare not known. Regardless of device, final Peff

APAP is

generally similar to Peffconv, but some patients do not

have an effective titration. Patients who do not snore

may not have an adequate APAP titration using a

device based on snoring detection. Supervised APAP

titration may be required because leaks and the need

for intervention occur. Unsupervised titration can be

Table 5

American Academy of Sleep Medicine practice parameters (2002) for the use of autotitrating positive airway pressure devices in

adult patients with obstructive sleep apnea

Recommendation Level of recommendation

1 A diagnosis of OSA must be established by an acceptable method. Standard

2 Patients with the following conditions are not candidates for APAP titration

and APAP treatment:

Standard

congestive heart failure

significant lung disease (eg, chronic obstructive pulmonary disease),

daytime hypoxemia, or respiratory failure

prominent nocturnal desaturation other than that from OSA

(eg, obesity-hypoventilation syndrome).

Patients who do not snore should not be titrated with an APAP device that

relies on vibration or sound in the device’s algorithm.

3 APAP devices are not currently recommended for split-night studies. Standard

4 Certain APAP devices may be used during attended titration to identify, by

polysomnography, a single pressure for use with standard CPAP for treatment of OSA.

Guideline

5 Once an initial successful attended CPAP or APAP titration has been determined by

polysomnography, certain APAP devices may be used in the self-adjusting mode for

unattended treatment of OSA.

Guideline

6 Use of unattended APAP to either initially determine pressures for fixed CPAP or provide

for self-adjusting APAP treatment in CPAP naıve patients is not currently established.

Option

7 Patients being treated with fixed CPAP on the basis of APAP titration or being treated

with APAP must be followed to determine treatment efficacy and safety.

Standard

8 A reevaluation and, if necessary, a standard attended CPAP titration should be

performed if symptoms do not resolve or the CPAP or APAP treatment seems

to lack efficacy.

Standard

Modified from Berry RB, Parish JM, Hartse KM. The use of auto-titrating continuous positive airway pressure for treatment of

adult obstructive sleep apnea: an American Academy of Sleep Medicine review. Sleep 2002;25:148; with permission.

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 337

successful in some patients, however. Use of APAP

for titration does not seem to impact compliance

negatively. Data on the impact on health care resources

of a patient-care strategy that incorporates APAP

devices for titration are preliminary.

Recommendations for the clinical use of automatic

positive airway pressure

Several authors have presented algorithms for the

use of APAP in the unattended setting for diagnosis

and therapy [23,42]. In general, in the straightforward

patient with classic signs and symptoms of OSA

[118,119] without complicating disorders, such as

respiratory insufficiency or congestive heart failure

or reasons for mask/mouth leak, APAP could be used

for diagnosis. If the study is positive and of satisfac-

tory quality, APAP then could be used for therapy,

with close patient follow-up for resolution of symp-

toms and compliance. If the study is negative or of

poor quality or if the patient’s symptoms are persist-

ent, conventional in-laboratory attended PSG would

be recommended. Careful patient selection, patient

education and support, and close follow-up must be

incorporated into the algorithm to ensure the success

of such a strategy [27]. Although appealing in many

respects, the effect of this strategy on long-term

outcomes of OSA has yet to be tested formally in a

large series of patients.

Practice parameters for the use of APAP devices

for titrating pressures and treating patients with OSA

have been published recently by the American Acad-

emy of Sleep Medicine [120]. Available data on the

therapeutic and titrating but not diagnostic roles of

APAP were reviewed by the Standards of Practice

Committee of the American Academy of Sleep Medi-

cine [69], and studies were graded according to levels

of evidence [121]. Based on this review, the commit-

tee made recommendations for the clinical use of

APAP, which were approved by the Board of Direc-

tors of the American Academy of Sleep Medicine. As

noted in Table 5, practice parameters were divided

into standards (a generally accepted patient-care

strategy, which reflects a high degree of clinical

certainty), guidelines (a patient-care strategy which

reflects a moderate degree of clinical certainty), and

options (uncertain patient-care strategy) [122].

Summary

Automatic positive airway pressure devices are the

most technologically advanced positive airway pres-

sure devices available for use in OSA. Although

heterogeneous, they have in common the ability to

detect and respond to changes in upper airway resist-

ance. Data cannot necessarily be extrapolated from

one device to another, and the field is rapidly advan-

cing. Most studies of APAP have been performed in a

supervised setting, or patients have been carefully

selected to have a high likelihood of OSA uncompli-

cated by disorders such as alveolar hypoventilation or

central apnea or technical problems such as mask

leaks. Studies of APAP for the diagnosis of OSA have

shown that APAP can diagnose severe OSA effec-

tively, but the diagnosis of mild-moderate OSA is less

reliable. APAP devices also can be effective therapy

for selected patients with OSA, with overall similar

results to conventional fixed CPAP in terms of respi-

ratory disturbances, sleep quality, nocturnal oxygena-

tion, and daytime sleepiness and performance, with

less known or other long-term outcomes. In most

studies, mean treatment pressures are lower, without

change in side effect profile. Compliance and pref-

erence with APAP are similar to or somewhat better

than CPAP in most studies. APAP also can be used in

an attended setting to titrate an effective pressure for

use in long-term conventional CPAP therapy, also

with similar results to CPAP in many patients. APAP

devices are more expensive than CPAP devices, but

the cost may be outweighed if a group of patients who

can be diagnosed, treated, or titrated safely in the

unattended setting can be identified. Although diag-

nostic and therapeutic algorithms for APAP have been

proposed, the best candidates for this modality must

be defined better.

References

[1] Sullivan CE, Issa FG, Berthon-Jones M, Eves L. Re-

versal of obstructive sleep apnoea by continuous pos-

itive airway pressure applied through the nares. Lancet

1981;1:862.

[2] Loube DI, Gay PC, Strohl KP, Pack AI, White DP,

Collop NA. Indications for positive airway pressure

treatment of adult obstructive sleep apnea patients: a

consensus statement. Chest 1999;115:863.

[3] Rapoport DM. Methods to stabilize the upper airway

using positive pressure. Sleep 1996;19:S123.

[4] Iber C, O’Brien C, Schluter J, Davies S, Leatherman

J, Mahowald M. Single night studies in obstructive

sleep apnea. Sleep 1991;14:383.

[5] McArdle N, Douglas NJ. Effect of continuous posi-

tive airway pressure on sleep architecture in the sleep

apnea-hypopnea syndrome: a randomized controlled

trial. Am J Respir Crit Care Med 2001;164:1459.

[6] Lamphere J, Roehrs T, Wittig R, Zorick F, Conway

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342338

WA, Roth T. Recovery of alertness after CPAP in

apnea. Chest 1989;96:1364.

[7] Jenkinson C, Davies RJ, Mullins R, Stradling JR.

Comparison of therapeutic and subtherapeutic nasal

continuous positive airway pressure for obstructive

sleep apnoea: a randomised prospective parallel trial.

Lancet 1999;353:2100.

[8] Montserrat JM, Ferrer M, Hernandez L, Farre R,

Vilagut G, Navajas D, et al. Effectiveness of CPAP

treatment in daytime function in sleep apnea syn-

drome: a randomized controlled study with an opti-

mized placebo. Am J Respir Crit Care Med 2001;

164:608.

[9] Bardwell WA, Ancoli-Israel S, Berry CC, Dimsdale

JE. Neuropsychological effects of one-week continu-

ous positive airway pressure treatment in patients

with obstructive sleep apnea: a placebo-controlled

study. Psychosom Med 2001;63:579.

[10] Findley LJ, Fabrizio MJ, Knight H, Norcross BB,

LaForte AJ, Suratt PM. Driving simulator perform-

ance in patients with sleep apnea. Am Rev Respir Dis

1989;140:529.

[11] Findley L, Smith C, Hooper J, Dineen M, Suratt PM.

Treatment with nasal CPAP decreases automobile ac-

cidents in patients with sleep apnea. Am J Respir Crit

Care Med 2000;161:857.

[12] D’Ambrosio C, Bowman T, Mohsenin V. Quality of

life in patients with obstructive sleep apnea: effect of

nasal continuous positive airway pressure. A prospec-

tive study. Chest 1999;115:123.

[13] He J, Kryger MH, Zorick FJ, Conway W, Roth T.

Mortality and apnea index in obstructive sleep apnea:

experience in 385 male patients. Chest 1988;94:9.

[14] Worsnop CJ, Pierce RJ, Naughton M. Systemic hy-

pertension and obstructive sleep apnea. Sleep 1993;

16:S148.

[15] Faccenda JF, Mackay TW, Boon NA, Douglas NJ.

Randomized placebo-controlled trial of continuous

positive airway pressure on blood pressure in the

sleep apnea-hypopnea syndrome. Am J Respir Crit

Care Med 2001;163:344.

[16] Strollo Jr PJ, Sanders MH, Atwood CW. Positive

pressure therapy. Clin Chest Med 1998;19:55.

[17] Behbehani K, Yen FC, Burk JR, Lucas EA, Axe JR.

Automatic control of airway pressure for treatment

of obstructive sleep apnea. IEEE Trans Biomed Eng

1995;42:1007.

[18] Berthon-Jones M. Feasibility of a self-setting CPAP

machine. Sleep 1993;16:S120.

[19] Gugger M, Mathis J, Bassetti C. Accuracy of an intel-

ligent CPAP machine with in-built diagnostic abilities

in detecting apnoeas: a comparison with polysomno-

graphy. Thorax 1995;50:1199.

[20] Fleury B, Rakotonanahary D, Hausser-Hauw C, Leb-

eau B, Guilleminault C. A laboratory validation study

of the diagnostic mode of the Autoset system for sleep-

related respiratory disorders. Sleep 1996;19:502.

[21] Bradley PA, Mortimore IL, Douglas NJ. Comparison

of polysomnography with ResCare Autoset in the

diagnosis of the sleep apnoea/hypopnoea syndrome.

Thorax 1995;50:1201.

[22] Kiely JL, Delahunty C, Matthews S, McNicholas WT.

Comparison of a limited computerized diagnostic sys-

tem (ResCare Autoset) with polysomnography in the

diagnosis of obstructive sleep apnoea syndrome. Eur

Respir J 1996;9:2360.

[23] Gugger M. Comparison of ResMed AutoSet (version

3.03) with polysomnography in the diagnosis of the

sleep apnoea/hypopnoea syndrome. Eur Respir J

1997;10:587.

[24] Lloberes P, Ballester E, Montserrat JM, Botifoll E,

Ramirez A, Reolid A, et al. Comparison of manual

and automatic CPAP titration in patients with sleep

apnea/hypopnea syndrome. Am J Respir Crit Care

Med 1996;154:1755.

[25] Teschler H, Berthon-Jones M, Thompson AB, Henkel

A, Henry J, Konietzko N. Automated continuous pos-

itive airway pressure titration for obstructive sleep ap-

nea syndrome. Am J Respir Crit Care Med 1996;

154:734.

[26] Teschler H, Farhat AA, Exner V, Konietzko N, Ber-

thon-Jones M. AutoSet nasal CPAP titration: con-

stancy of pressure, compliance and effectiveness at

8 month follow-up. Eur Respir J 1997;10:2073.

[27] Teschler H, Berthon-Jones M. Intelligent CPAP

systems: clinical experience. Thorax 1998;53

(Suppl 3):S49.

[28] Rees K, Wraith PK, Berthon-Jones M, Douglas NJ.

Detection of apnoeas, hypopnoeas and arousals by the

AutoSet in the sleep apnoea/hypopnoea syndrome.

Eur Respir J 1998;12:764.

[29] ?Mayer P, Meurice JC, Philip-Joet F, Cornette A,

Rakotonanahary D, Meslier N, et al. Simultaneous

laboratory-based comparison of ResMed Autoset

with polysomnography in the diagnosis of sleep

apnoea/hypopnoea syndrome. Eur Respir J 1998;

12:770.

[30] Teschler H, Wessendorf TE, Farhat AA, Konietzko N,

Berthon-Jones M. Two months auto-adjusting versus

conventional nCPAP for obstructive sleep apnoea

syndrome. Eur Respir J 2000;15:990.

[31] Gagnadoux F, Rakotonanahary D, Martins de Araujo

MT, Barros-Vieira S, Fleury B. Long-term efficacy of

fixed CPAP recommended by Autoset for OSAS.

Sleep 1999;22:1095.

[32] Marrone O, Insalaco G, Bonsignore MR, Romano S,

Salvaggio A, Bonsignore G. Sleep structure correlates

of continuous positive airway pressure variations dur-

ing application of an autotitrating continuous positive

airway pressure machine in patients with obstructive

sleep apnea syndrome. Chest 2002;121:759.

[33] Lofaso F, Lerou K, Quera S, Mroue G, Dortho MP,

Isabey D, et al. Snoring detection during auto-nasal

continuous positive airway pressure. Eur Respir J

2002;19:108.

[34] Berkani M, Lofaso F, Chouaid C, Pia d’Ortho M,

Theret D, Grillier-Lanoir V, et al. CPAP titration by

an auto-CPAP device based on snoring detection: a

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 339

clinical trial and economic considerations. Eur Respir

J 1998;12:759.

[35] Ficker JH, Wiest GH, Lehnert G, Wiest B, Hahn EG.

Evaluation of an auto-CPAP device for treatment of

obstructive sleep apnoea. Thorax 1998;53:643.

[36] Boudewyns A, Grillier-Lanoir V, Willemen MJ, De

Cock WA, Van de Heyning PH, De Backer WA.

Two months follow up of auto-CPAP treatment in

patients with obstructive sleep apnoea. Thorax 1999;

54:147.

[37] d’Ortho MP, Grillier-Lanoir V, Levy P, Goldenberg F,

Corriger E, Harf A, et al. Constant vs. automatic con-

tinuous positive airway pressure therapy: home eval-

uation. Chest 2000;118:1010.

[38] Scharf MB, Brannen DE, McDannold MD, Berko-

witz DV. Computerized adjustable versus fixed

nCPAP treatment of obstructive sleep apnea. Sleep

1996;19:491.

[39] Stradling JR, Barbour C, Pitson DJ, Davies RJ. Au-

tomatic nasal continuous positive airway pressure ti-

tration in the laboratory: patient outcomes. Thorax

1997;52:72.

[40] Konermann M, Sanner BM, Vyleta M, Laschewski F,

Groetz J, Sturm A, et al. Use of conventional and self-

adjusting nasal continuous positive airway pressure

for treatment of severe obstructive sleep apnea syn-

drome: a comparative study. Chest 1998;113:714.

[41] Lafond C, Series F. Influence of nasal obstruction on

auto-CPAP behaviour during sleep in sleep apnoea/

hypopnoea syndrome. Thorax 1998;53:780.

[42] Fletcher EC, Stich J, Yang KL. Unattended home di-

agnosis and treatment of obstructive sleep apnea with-

out polysomnography. Arch Fam Med 2000;9:168.

[43] Meurice JC, Marc I, Series F. Efficacy of auto-CPAP

in the treatment of obstructive sleep apnea/hypopnea

syndrome. Am J Respir Crit Care Med 1996;153:794.

[44] Series F, Marc I. Efficacy of automatic continuous

positive airway pressure therapy that uses an esti-

mated required pressure in the treatment of the

obstructive sleep apnea syndrome. Ann Intern Med

1997;127:588.

[45] Series F. Accuracy of an unattended home CPAP ti-

tration in the treatment of obstructive sleep apnea.

Am J Respir Crit Care Med 2000;162:94.

[46] Series F, Marc I. Importance of sleep stage- and body

position-dependence of sleep apnoea in determining

benefits to auto-CPAP therapy. Eur Respir J 2001;

18:170.

[47] Lofaso F, Lorino AM, Duizabo D, Najafi Zadeh H,

Theret D, Goldenberg F, et al. Evaluation of an auto-

nCPAP device based on snoring detection. Eur Respir

J 1996;9:1795.

[48] Randerath WJ, Parys K, Feldmeyer F, Sanner B,

Ruhle KH. Self-adjusting nasal continuous positive

airway pressure therapy based on measurement of

impedance: a comparison of two different maximum

pressure levels. Chest 1999;116:991.

[49] Ficker JH, Fuchs FS, Wiest GH, Asshoff G, Schmelz-

er AH, Hahn EG. An auto-continuous positive airway

pressure device controlled exclusively by the forced

oscillation technique. Eur Respir J 2000;16:914.

[50] Randerath W, Parys K, Lehmann D, Sanner B, Feld-

meyer F, Ruhle KH. Self-adjusting continuous positive

airway pressure therapy based on the measurement of

impedance: a comparison of free pressure variation and

individually fixed higher minimum pressure. Respira-

tion 2000;67:272.

[51] Randerath WJ, Schraeder O, Galetke W, Feldmeyer F,

Ruhle KH. Autoadjusting CPAP therapy based on

impedance efficacy, compliance and acceptance. Am

J Respir Crit Care Med 2001;163:652.

[52] Randerath WJ, Galetke W, David M, Siebrecht H,

Sanner B, Ruhle K. Prospective randomized compar-

ison of impedance-controlled auto-continuous posi-

tive airway pressure (APAP(FOT)) with constant

CPAP. Sleep Med 2001;2:115.

[53] Fuchs FS, Wiest GH, Frank M, Harsch IA, Schahin

SP, Hahn EG, et al. Auto-CPAP therapy for obstruc-

tive sleep apnea: induction of microarousals by au-

tomatic variations of CPAP pressure? Sleep 2002;

25:514.

[54] Sharma S, Wali S, Pouliot Z, Peters M, Neufeld H,

Kryger M. Treatment of obstructive sleep apnea with

a self-titrating continuous positive airway pressure

(CPAP) system. Sleep 1996;19:497.

[55] Hudgel DW, Fung C. A long-term randomized, cross-

over comparison of auto-titrating and standard nasal

continuous airway pressure. Sleep 2000;23:645.

[56] Miyazaki S, Itasaka Y, Ishikawa K, Togawa K. Eval-

uation of the auto continuous positive airway pressure

efficacy by upper airway pressure measurement. Psy-

chiatry Clin Neurosci 1999;53:327.

[57] Behbehani K, Yen FC, Lucas EA, Burk JR. A sleep

laboratory evaluation of an automatic positive airway

pressure system for treatment of obstructive sleep ap-

nea. Sleep 1998;21:485.

[58] DuBois AB, Brody AW, Lewis DH, Burgess BF. Os-

cillation mechanics of lungs and chest in man. J Appl

Physiol 1956;8:587.

[59] Montserrat JM, Farre R, Navajas D. New technolo-

gies to detect static and dynamic upper airway ob-

struction during sleep. Sleep Breath 2001;5:193.

[60] Farre R, Peslin R, Rotger M, Navajas D. Inspiratory

dynamic obstruction detected by forced oscillation

during CPAP: a model study. Am J Respir Crit Care

Med 1997;155:952.

[61] Badia JR, Farre R, Montserrat JM, Ballester E, Her-

nandez L, Rotger M, et al. Forced oscillation tech-

nique for the evaluation of severe sleep apnoea/

hypopnoea syndrome: a pilot study. Eur Respir J

1998;11:1128.

[62] Navajas D, Farre R, Rotger M, Badia R, Puig-de-

Morales M, Montserrat JM. Assessment of airflow

obstruction during CPAP by means of forced oscilla-

tion in patients with sleep apnea. Am J Respir Crit

Care Med 1998;157:1526.

[63] Ficker JH, Wiest GH, Asshoff G, Fuchs FS, Schmelz-

er AH, Harsch IA, et al. Sensitivity of a simplified

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342340

forced oscillation technique for detection of upper

airway obstruction. Respir Physiol 2001;124:243.

[64] Condos R, Norman RG, Krishnasamy I, Peduzzi N,

Goldring RM, Rapoport DM. Flow limitation as a

noninvasive assessment of residual upper-airway re-

sistance during continuous positive airway pressure

therapy of obstructive sleep apnea. Am J Respir Crit

Care Med 1994;150:475.

[65] Montserrat J, Ballester E, Olivi H, Reolid A, Llo-

beres P, Morello A, et al. Time-course of stepwise

CPAP titration: behavior of respiratory and neuro-

logical variables. Am J Respir Crit Care Med 1995;

152:1854.

[66] Ayappa I, Norman RG, Hosselet JJ, Gruenke RA,

Walsleben JA, Rapoport DM. Relative occurrence

of flow limitation and snoring during continuous pos-

itive airway pressure titration. Chest 1998;114:685.

[67] Krieger J. Therapeutic use of auto-CPAP. Sleep Med

Rev 1999;3:159.

[68] Badia JR, Farre RO, John Kimoff R, Ballester E,

Hernandez L, Rotger M, et al. Clinical application

of the forced oscillation technique for CPAP titration

in the sleep apnea/hypopnea syndrome. Am J Respir

Crit Care Med 1999;160:1550.

[69] Berry RB, Parish JM, Hartse KM. The use of auto-

titrating continuous positive airway pressure for treat-

ment of adult obstructive sleep apnea: an American

Academy of Sleep Medicine review. Sleep 2002;

25:148.

[70] Steltner H, Staats R, Timmer J, Vogel M, Guttmann J,

Matthys H, et al. Diagnosis of sleep apnea by auto-

matic analysis of nasal pressure and forced oscillation

impedance. Am J Respir Crit Care Med 2002;165:940.

[71] Farre R, Montserrat JM, Rigau J, Trepat X, Pinto P,

Navajas D. Response of automatic continuous posi-

tive airway pressure devices to different sleep breath-

ing patterns: a bench study. Am J Respir Crit Care

Med 2002;166:469.

[72] American Sleep Disorders Association. Practice pa-

rameters for the indications for polysomnography and

related procedures. Sleep 1997;20:406.

[73] Chesson Jr AL, Ferber RA, Fry JM, Grigg-Damberg-

er M, Hartse KM, Hurwitz TD, et al. The indications

for polysomnography and related procedures. Sleep

1997;20:423.

[74] Redline S, Kapur VK, Sanders MH, Quan SF, Got-

tlieb DJ, Rapoport DM, et al. Effects of varying ap-

proaches for identifying respiratory disturbances on

sleep apnea assessment. Am J Respir Crit Care Med

2000;161:369.

[75] American Academy of Sleep Medicine Task Force.

Sleep-related breathing disorders in adults: recom-

mendations for syndrome definition and measurement

techniques in clinical research. Sleep 1999;22:667.

[76] Thomas RJ. Definitions of respiratory events in sleep-

disordered breathing. Sleep Med 2002;3:89.

[77] Hosselet J, Ayappa I, Norma RG, Krieger AC, Rapo-

port DM. Classification of sleep-disordered breathing.

Am J Respir Crit Care Med 2001;163:398.

[78] Meyer TJ, Eveloff SE, Kline LR, Millman RP. One

negative polysomnogram does not exclude obstruc-

tive sleep apnea. Chest 1993;103:756.

[79] Le Bon O, Hoffmann G, Tecco J, Staner L, Noseda A,

Pelc I, et al. Mild to moderate sleep respiratory

events: one negative night may not be enough. Chest

2000;118:353.

[80] Namen AM, Dunagan DP, Fleischer A, Tillett J, Bar-

nett M, McCall WV, et al. Increased physician-

reported sleep apnea(*): the national ambulatory

medical care survey. Chest 2002;121:1741.

[81] Ferber R, Millman R, Coppola M, Fleetham J, Mur-

ray CF, Iber C, et al. Portable recording in the assess-

ment of obstructive sleep apnea: ASDA standards of

practice. Sleep 1994;17:378.

[82] Thorpy M, Chesson A, Ferber R, Kader G, Millman

R, Reite M, et al. Practice parameters for the use of

portable recording in the assessment of obstructive

sleep apnea: an American Sleep Disorders Associa-

tion report. Sleep 1994;17:372.

[83] Indications for the clinical use of unattended portable

recording for the diagnosis of sleep-related breathing

disorders. ASDA News 1999;6:19.

[84] Krieger J, Sforza E, Petiau C, Weiss T. Simplified

diagnostic procedure for obstructive sleep apnoea

syndrome: lower subsequent compliance with CPAP.

Eur Respir J 1998;12:776.

[85] Javaheri S. Effects of continuous positive airway pres-

sure on sleep apnea and ventricular irritability in pa-

tients with heart failure. Circulation 2000;101:392.

[86] Peled N, Abinader EG, Pillar G, Sharif D, Lavie P.

Nocturnal ischemic events in patients with obstructive

sleep apnea syndrome and ischemic heart disease:

effects of continuous positive air pressure treatment.

J Am Coll Cardiol 1999;34:1744.

[87] Parker JD, Brooks D, Kozar LF, Render-Teixeira CL,

Horner RL, Douglas Bradley T, et al. Acute and

chronic effects of airway obstruction on canine left

ventricular performance. Am J Respir Crit Care Med

1999;160:1888.

[88] Bahammam A, Delaive K, Ronald J, Manfreda J,

Roos L, Kryger MH. Health care utilization in males

with obstructive sleep apnea syndrome two years

after diagnosis and treatment. Sleep 1999;22:740.

[89] Kribbs NB, Pack AI, Kline LR, Smith PL, Schwartz

AR, Schubert NM, et al. Objective measurement of

patterns of nasal CPAP use by patients with obstruc-

tive sleep apnea. Am Rev Respir Dis 1993;147:887.

[90] McArdle N, Devereux G, Heidarnejad H, Engleman

HM, Mackay TW, Douglas NJ. Long-term use of

CPAP therapy for sleep apnea/hypopnea syndrome.

Am J Respir Crit Care Med 1999;159:1108.

[91] Pepin JL, Leger P, Veale D, Langevin B, Robert D,

Levy P. Side effects of nasal continuous positive air-

way pressure in sleep apnea syndrome: study of 193

patients in two French sleep centers. Chest 1995;

107:375.

[92] Hoy CJ, Vennelle M, Kingshott RN, Engleman HM,

Douglas NJ. Can intensive support improve continu-

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342 341

ous positive airway pressure use in patients with the

sleep apnea/hypopnea syndrome? Am J Respir Crit

Care Med 1999;159:1096.

[93] Chervin RD, Theut S, Bassetti C, Aldrich MS. Com-

pliance with nasal CPAP can be improved by simple

interventions. Sleep 1997;20:284.

[94] Wittig RM, Romaker A, Zorick FJ, Roehrs TA, Con-

way WA, Roth T. Night-to-night consistency of ap-

neas during sleep. Am Rev Respir Dis 1984;129:244.

[95] Oksenberg A, Silverberg DS, Arons E, Radwan H.

The sleep supine position has a major effect on opti-

mal nasal continuous positive airway pressure: rela-

tionship with rapid eye movements and non-rapid eye

movements sleep, body mass index, respiratory dis-

turbance index, and age. Chest 1999;116:1000.

[96] Cartwright RD. Effect of sleep position on sleep ap-

nea severity. Sleep 1984;7:110.

[97] Pevernagie DA, Shepard Jr JW. Relations between

sleep stage, posture and effective nasal CPAP levels

in OSA. Sleep 1992;15:162.

[98] Issa FG, Sulllivan CE. Alcohol, snoring, and sleep

apnea. J Neurol Neurosurg Psychiatry 1982;45:353.

[99] Teschler H, Berthon-Jones M, Wessendorf T, Meyer

HJ, Konietzko N. Influence of moderate alcohol con-

sumption on obstructive sleep apnoea with and with-

out AutoSet nasal CPAP therapy. Eur Respir J 1996;

9:2371.

[100] Dolly FR, Block AJ. Effect of florazepam on sleep-

disordered breathing and nocturnal oxygen desatura-

tion in asymptomatic subjects. Am J Med 1982;

73:239.

[101] Richards GN, Cistulli PA, Ungar RG, Berthon-

Jones M, Sullivan CE. Mouth leak with nasal con-

tinuous positive airway pressure increases nasal

airway resistance. Am J Respir Crit Care Med

1996;154:182.

[102] Farre R, Peslin R, Montserrat JM, Rotger M, Navajas

D. Flow-dependent positive airway pressure to main-

tain airway patency in sleep apnea-hypopnea syn-

drome. Am J Respir Crit Care Med 1998;157:1855.

[103] Bacon JP, Farney RJ, Jensen RL, Walker JM, Clo-

ward TV. Nasal continuous positive airway pressure

devices do not maintain the set pressure dynamically

when tested under simulated clinical conditions.

Chest 2000;118:1441.

[104] Browman CP, Sampson MG, Gujavarty KS, Weiler

SJ, Walseben JA, Hahn PM, et al. Obstructive sleep

apnea and body weight. Chest 1984;85:435.

[105] Series F, Marc I, Cormier Y, La Forge J. Required

levels of nasal continuous positive airway pressure

during treatment of obstructive sleep apnoea. Eur Res-

pir J 1994;7:1776.

[106] Ryan CF, Lowe AA, Li D, Fleetham JA. Magnetic

resonance imaging of the upper airway in obstructive

sleep apnea before and after chronic nasal continuous

positive airway pressure therapy. Am Rev Respir Dis

1991;144:939.

[107] Boudewyns A, Van de Heyning P, De Backer W.

Appearance of central apnoea in a patient treated by

auto-CPAP for obstructive sleep apnoea. Respir Med

1998;92:891.

[108] Johns M. A new method of measuring daytime sleep-

iness: the Epworth Sleepiness Scale. Sleep 1991;

14:540.

[109] Richardson GS, Carskadon MA, Flagg W, Vanden-

Hoed J, Dement WC, Mitler MM. Excessive daytime

sleepiness in man: multiple sleep latency test mea-

surement in narcoleptics and control subjects. Electro-

encephalogr Clin Neurophysiol 1978;45:621.

[110] American Sleep Disorders Association. The clinical

use of the multiple sleep latency test. Sleep 1992;

15:268.

[111] Mitler MM, Gujavarty KS, Broman CP. Maintenance

of wakefulness test: a polysomnographic technique

for evaluating treatment efficacy in patients with ex-

cessive somnolence. Electroencephalogr Clin Neuro-

physiol 1982;53:658.

[112] Miljeteig H, Hoffstein V. Determinants of continuous

positive airway pressure level for treatment of ob-

structive sleep apnea. Am Rev Respir Dis 1993;

147:1526.

[113] Coppola MP, Lawee M. Management of obstructive

sleep apnea syndrome in the home: the role of port-

able recording. Chest 1993;104:19.

[114] Waldhorn RE, Wood K. Attended home titration of

nasal continuous positive pressure therapy for ob-

structive sleep apnea. Chest 1993;104:1707.

[115] Juhasz J, Schillen J, Urbigkeit A, Ploch T, Penzel T,

Peter JH. Unattended continuous positive airway

pressure titration: clinical relevance and cardiorespi-

ratory hazards of the method. Am J Respir Crit Care

Med 1996;154:359.

[116] White DP, Gibb TJ. Evaluation of the Healthdyne

NightWatch system to titrate CPAP in the home.

Sleep 1998;21:198.

[117] Mortimore IL, Bradley PA, Murray JA, Douglas NJ.

Uvulopalatopharyngoplasty may compromise nasal

CPAP therapy in sleep apnea syndrome. Am J Respir

Crit Care Med 1996;154:1759.

[118] Strohl KP, Redline S. Recognition of obstructive

sleep apnea. Am J Respir Crit Care Med 1996;

154:279.

[119] Flemons WW, McNicholas WT. Clinical prediction

of the sleep apnea syndrome. Sleep Med Rev 1997;

1:19.

[120] Littner M, Hirshkowitz M, Davila D, Anderson WM,

Kushida CA, Woodson BT, et al. Practice parameters

for the use of auto-titrating continuous positive air-

way pressure devices for titrating pressures and treat-

ing adult patients with obstructive sleep apnea

syndrome: an American Academy of Sleep Medicine

report. Sleep 2002;25:143.

[121] Sackett D. Rules of evidence and the clinical recom-

mendations for the management of patients. Can J

Cardiol 1993;9:487.

[122] Eddy DM. A manual for assessing health practices

and designing practice policies: the explicit approach.

Philadelphia: American College of Physicians; 1992.

F.J. Roux, J. Hilbert / Clin Chest Med 24 (2003) 315–342342

Behavioral and pharmacologic therapy of

obstructive sleep apnea

Ulysses J. Magalang, MDa,b,*, M. Jeffery Mador, MDa,c

aDivision of Pulmonary, Critical Care, and Sleep Medicine, University at Buffalo,

State University of New York, 3435 Main Street, Buffalo, NY 14214, USAbAssociated Sleep Center, 1400 Sweet Home Road, Amherst, NY 14228, USA

cBuffalo Veterans Affairs Medical Center Sleep Disorders Center, 3495 Bailey Avenue, Buffalo, NY 14215, USA

Behavioral therapy of obstructive sleep apnea

In this section, the authors discuss the role of

weight loss and modification of sleep posture in the

treatment of obstructive sleep apnea (OSA).

Weight loss

Obesity is strongly correlated with OSA in clinic

populations and population-based epidemiologic

studies [1,2]. In the Wisconsin Sleep Cohort Study,

a group of state employees were prospectively

studied [2]. In this study, 4% of men and 2% of

women had an apnea-hypopnea index (AHI) of more

than 5/hour and symptoms of daytime hypersomno-

lence, and 24% of men and 9% of women had an

AHI of more than 5/hour with or without symptoms.

In this study, an increase in body mass index (body

weight in kilograms divided by height2 in meters) of

one standard deviation was associated with a fourfold

increase in the risk of having an AHI of more than

5/hour. All measurements of body habitus, including

weight, significantly influenced the AHI.

Obesity can promote OSA by various mecha-

nisms. A detailed discussion of potential mechanisms

has been provided elsewhere [3]. It is believed that

obesity can reduce the size or change the shape of the

upper airway, which promotes airway occlusion.

Some CT scan studies of the upper airway have

shown a smaller and differently shaped retropalatal

airway in patients with OSA than control subjects

[4,5]. MRI studies, which are better at identifying fat,

also have shown increased fat deposits in the upper

airway in patients with OSA compared with weight-

matched controls [6]. All of these studies were

conducted while patients were awake. Obesity also

seems to alter upper airway function. Various indirect

measurements have suggested that the upper airway

is more collapsible in patients with sleep apnea [7,8].

Weight loss in overweight patients with sleep apnea

reduced the pharyngeal critical closing pressure dur-

ing sleep, which indicated a reduction in upper air-

way collapsibility [9].

One study has examined the effects of changes in

weight on the AHI in a longitudinal population study

[10]. In this study, a group of healthy volunteers

underwent repeat sleep studies 4 years after their

initial polysomnogram. The changes in AHI were

correlated to changes in weight after potential co-

variates were taken into account. For each percentage

change in weight, there was approximately a 3%

change in the AHI. For example, a 10% reduction

in weight was associated with a 26% reduction in the

AHI. For subjects with normal or mildly increased

AHI at baseline (AHI < 15/hour), a 10% increase in

weight was associated with a sixfold increase in the

chance of developing moderate to severe sleep-dis-

ordered breathing (AHI > 15/hour).

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00022-4

* Corresponding author. Suite 162, Erie County Med-

ical Center, 462 Grider Street, Buffalo, NY 14215.

E-mail address: magalang@buffalo.edu

(U.J. Magalang).

Clin Chest Med 24 (2003) 343–353

Short-term effects of weight loss

Several small clinical studies have evaluated the

short-term effects of varying degrees of weight loss in

patients with OSA (Table 1) [9,11–18]. Most of these

studies were uncontrolled, and the severity of obesity

at baseline varied widely. Virtually all of these studies

showed that weight loss improved sleep apnea to

some extent, however, at least in some patients. In

one controlled study, 23 mild to moderately obese

patients were randomized to receive dietary counsel-

ing (15 patients) or no intervention (control group of

8 subjects) [11]. The sleep study was repeated when

they had lost at least 5% of their initial body weight

(intervention group) or when their weight had

remained stable (control group). The mean fall in

body weight was 9%. This modest reduction in

weight was associated with a significant reduction

in the apnea index, an improvement in nocturnal

oxygenation, improvements in sleep architecture,

and a borderline improvement in the multiple sleep

latency test. No changes were observed in the control

group. 4 patients had a reduction in the apnea index

to normal.

In a follow-up study, 23 additional patients re-

ceived dietary advice and follow-up [9]. 13 patients

lost at least 5% of their initial body weight (ie,

the therapy worked and the patients were restudied

and compared with 13 matched controls). The dietary

intervention group lost approximately 17% of their

body weight. The AHI decreased from 83.3/hourF31/hour to 32.5/hour F 35.9/hour. In 7 of the patients

the AHI decreased to below 20/hour (close to 0 in

5 patients). The pharyngeal critical closing pressure

was reduced significantly after weight loss, which

indicated a reduction in upper airway collapsibility.

When the pharyngeal critical closing pressure was

below � 4 cm H2O (ie, more negative), sleep apnea

was virtually abolished. These results provide an

attractive potential mechanism by which weight loss

influences the AHI. The extent to which the AHI is

improved byweight loss depends on howmuchweight

loss improves upper airway collapsibility.

Several case studies have examined the effects of

dramatic weight loss on sleep apnea in morbidly obese

persons. Weight loss has been achieved by surgical

procedures (Table 2) [19–24] or very low calorie diets

[14] (see Table 1). When significant weight loss has

been achieved, improvements in sleep apnea have

been observed, with total resolution of sleep apnea

in some patients. Similar to the small studies per-

formed in moderately obese sleep apnea patients, the

amount of weight loss achieved did not always

correlate with the extent of improvement, possibly

because a given degree of weight loss affects upper

Table 1

Dietary weight loss: effect on sleep apnea

n

Length of

follow-up Method of weight loss Weight change kg (%) AHI pre AHI post

Smith et al [11] 15 5.3 mo Dietary advice/follow-up � 9.6 (� 9) 55 29.2

Schwartz et al [9] 13 17 mo Dietary advice/follow-up � 11.8 (� 17.4) 83.3 32.5

Rubinstein et al [12] 12 8–18 mo Diet/gastroplasty � 24 (� 20.5) 57 14

Kiselak et al [13] 19 18–20 wk Diet/exercise/behavioral therapy � 27.2 (� 23.9) 17.6 ?

Suratt et al [14] 8 24 mo Very low calorie diet � 21 (� 14) 90 62

Pasquali et al [18] 23 ? Diet or very low calorie diet/follow-up � 18.5 (� 17.5) 66.5 33

Rajala et al [15] 8 ? Diet ? (� 13) 39.5 31.6

Lojander et al [16] 24 1 y Very low calorie diet/diet/follow-up � 11 (� 10) ? ?

Kansanen et al [17] 15 3 mo Very low calorie diet � 9 (� 7.9) 31 19

Table 2

Surgical weight loss: effect on sleep apnea

n

Length of

follow-up Surgical procedure Weight change kg (%) AHI pre AHI post

Harman et al [21] 4 24 mo Jejuno-ileal bypass � 108 (� 47) 78 1.4

Peiser et al [19] 15 2–4 mo Gastric bypass � 35.1 (� 25) 81.9 15

Scheuller et al [23] 15 1–12 y Gastric bypass/gastroplasty � 54.7 (� 34) 96.9 11.3

Pillar et al [24] 14 4.5 mo Gastric bypass/gastroplasty � 35.6 (� 27) 40 11

Pillar et al [24] 14 5–10 y Gastric bypass/gastroplasty � 29.9 (� 23) 40 24

Sugerman et al [20] 40 ? Gastric bypass/gastroplasty � 57 (� 32) 64 26

Charuzi et al [22] 13 6 mo Gastric bypass ?(� 72.5) 88.8 8

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353344

airway collapsibility to different degrees in individual

patients. In contrast, in the large population study

described previously, changes in weight affected the

AHI in the expected dose-response manner (ie, the

more the weight loss, the greater the improvement in

the AHI). In this study, only changes in weight of less

than 20% of initial body weight were examined. In a

recent review, the average weight loss and average

reduction in AHI from the various clinical studies

were plotted [25]. A clear relationship between the

extent of weight loss and AHI could be appreciated.

In most published studies, many or all of the

patients studied had severe OSA defined arbitrarily

as an AHI of more than 30/hour [26]. In the clinical

arena, weight loss is often considered in patients with

mild to moderate disease who are reluctant to try or

are noncompliant with more definitive therapies for

sleep apnea, such as continuous positive airway

pressure (CPAP). Studies that particularly address

this patient population are largely lacking. The popu-

lation study [10] suggests that weight loss might be

efficacious, however, at least in the short term in this

patient group.

In a recent systematic review, the effects of weight

loss on sleep apnea were specifically assessed [27].

No study to date (last reviewed July 2, 2002) met the

entry criteria pointing out the limitations of the

existing database. Only the study by Smith et al

was randomized and included a control group [11].

The reason why this study was excluded was not

specifically reported but may have been because the

investigators were not blinded to treatment allocation.

Long-term effects of weight loss

Long-term data on the effects of weight loss are

sparse. Studies that involve obese patients without

sleep apnea indicate that whereas achieving weight

loss is difficult, maintaining weight loss is even

harder [28,29]. Unfortunately, most patients who lose

weight ultimately regain it. One recent study exam-

ined the long-term effects of weight loss in sleep

apnea patients [30]. Two hundred sixteen mildly

obese patients with sleep apnea were treated with a

weight reduction program that consisted of a hypo-

caloric diet, encouragement to increase physical

activity, and periodic appointments for reinforcement.

One hundred four patients lost at least 10% of their

initial weight. One hundred one patients underwent a

follow-up sleep study. Thirty-four patients had a

follow up AHI of less than 10 /hour with resolution

of daytime hypersomnolence and were considered

cured. Four patients also stopped excessive alcohol

or sedative usage. Six patients were lost to follow-up.

Twenty-four patients were followed for 5 to 11 years.

Over this time period, 11 of the patients regained a

significant amount of weight defined as at least 50%

of the initial weight loss. Not surprisingly, sleep

apnea recurred in 8 of these patients. Most impor-

tantly, in the 13 patients who maintained their weight

loss, sleep apnea recurred in 7.

Similarly, Pillar et al followed a group of morbidly

obese patients after bariatric surgery [24]. After

surgery, there was an impressive weight loss associ-

ated with a dramatic reduction in the apnea index

from 40/hourF 29/hour to 11/hourF 16/hour. Forty-

eight percent of patients had complete resolution of

apneas. 7.5 years later, the apnea index had increased

to 24/hourF 23/hour in these patients despite only a

modest increase in weight from their postoperative

minimum. 5 patients had an increase in their apnea

index despite absolutely no gain in weight. These

studies showed clearly that sleep apnea can recur in

overweight patients in the absence of weight gain.

Not all patients with obesity have sleep apnea.

Additional factors must be present—such as upper

airway size and function—that predispose some

obese patients to sleep apnea [18]. Presumably, these

factors can progress over time sufficiently to induce

sleep apnea at the reduced body weight.

Summary

Although data that address weight loss in patients

with sleep apnea are somewhat limited, the data

available suggest that weight loss can be a highly

effective treatment of sleep apnea in the short term.

Although long-term data are sparse, recurrence of

sleep apnea seems to be common either because of

failure to maintain weight loss or recurrence of sleep

apnea despite maintenance of weight loss. Because of

these factors, clinicians remain appropriately skepti-

cal of the overall efficacy of weight loss in patients

with OSA. Further study of weight loss in less

severely affected patients (AHI < 30/hour) in whom

acceptance of standard therapies for sleep apnea may

be difficult is warranted.

Positional therapy

In patients with OSA, the frequency of apnea and

hypopneas is influenced by body position in 50% to

60% of patients [31,32]. The AHI increases in the su-

pine position and is lower in the lateral position or with

the head of the bed elevated to 30� to 60� [31–33].

Even in patients in whom the AHI is not influenced by

body position, the duration of apnea/hypopnea and the

degree of associated desaturation are worse in the

supine position [34].

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353 345

This observation led early investigators to explore

methods to avoid sleep in the supine position. Ini-

tially, investigators considered patients eligible for

this therapy if the AHI in the supine position was at

least twice that in the lateral position [35,36]. If the

AHI is 80/hour in the supine position and 30/hour in

the lateral position, however, even if the therapy is

totally effective in eliminating supine sleep, the

patient still has an AHI likely to cause continued

symptoms. A better definition for eligible patients

would be an elevated AHI in the supine position and

an AHI in the lateral position less than a predefined

threshold value. In prior studies, an AHI of less than

15/hour has been used [37]. Depending on the

clinician’s threshold for distinguishing what is an

elevated AHI, a different threshold value of 5/hour

or 10/hour could be used. The prevalence of posi-

tional sleep apnea when this alternative definition is

used has not been determined.

In the United States, so-called split-night studies

(diagnostic and CPAP titration performed on the

same night) are becoming increasingly popular

because of pressures from commercial payors.

Whether positional sleep apnea can be diagnosed

accurately during a split-night study must be deter-

mined. Given the limited amount of time typically

available for the diagnostic portion of the study, it

seems doubtful that positional sleep apnea could be

assessed accurately during a split-night study. For-

tunately, positional sleep apnea seems to be more

common in patients with milder disease [31], whereas

split-night studies are generally reserved for patients

who display sleep study findings of severe disease.

Mechanisms for the effect of posture on sleep apnea

In awake, normal subjects [38] and patients with

sleep apnea, upper airway size increases in the seated

position compared with the supine position [7,39]. In

contrast, upper airway size does not seem to increase

when patients with sleep apnea move from the supine

to the lateral position [39]. Upper airway collapsibil-

ity is reduced in the seated position compared with

the supine position [40,41]. Conflicting results

between studies have been obtained in the lateral

position, but at least some measures in some studies

have shown a reduction in upper airway collapsibility

in the lateral position compared with the supine

position, which provides a potential explanation for

the improvement in the AHI [40,41].

Methods for avoiding supine sleep position

In the original studies of positional therapy, Cart-

wright et al used a posture alarm [35,36]. The patient

wore a positional monitor that triggered an alarm if

the patient remained in the supine position for more

than 15 seconds. The posture alarm was highly

effective in preventing supine sleep posture. In a

study of 15 patients, 1 slept in the supine position

for 35.5 minutes, 4 slept in the supine position for

less than 10 minutes, and supine sleep was com-

pletely eliminated in 10 patients [35]. Interestingly,

after 8 weeks of therapy, 8 of the patients slept

minimally in the supine posture during one night of

monitoring without the posture alarm [35]. In another

study, patients wore a backpack with a softball inside

positioned to prevent them from sleeping in the

supine position [42]. This modality also was highly

effective at preventing supine sleep posture. In a

study of 13 patients, 3 slept in the supine position

for 18 to 32 minutes, 1 slept in the supine position for

less than 10 minutes, and in 9 patients supine sleep

was totally prevented. Other methods to prevent

supine sleep position include pinning a tennis ball

to the patient’s pajama top or placing a wedge pillow

lengthwise in the bed.

Effectiveness of positional therapy

Surprisingly few studies have evaluated positional

therapy formally. In one study, 13 patients who were

studied during a single overnight sleep study spent

half the night in the supine position and half the

night in the semi-seated position with the bed inclined

at a 60� angle [33]. The AHI decreased significantly

from 68/hourF 12/hour in the supine position to

47/hourF 30/hour in the semi-seated position. Two

patients had an AHI of less than 10/hour in the semi-

seated position. This study showed that positional

therapy is not effective in unselected patients with

severe sleep apnea. Further studies are required to

evaluate the semi-seated position in patients with

positional sleep apnea identified on their initial sleep

study and in patients with milder disease.

In another study, 15 patients with an AHI in the

supine posture more than twice that in the lateral

posture were evaluated with the posture alarm [35].

The AHI was reduced from 33/hourF 21/hour to

21/hourF 29/hour with positional therapy. The

AHI was reduced to less than 10/hour in 10 of the

15 patients. Interestingly, equivalent results were

obtained in this study when subjects were just told to

learn to sleep on their side, lose weight, moderately

exercise, and avoid alcohol after 6 PM. In 15 patients

given these instructions, the AHI was reduced from

27/hourF 13/hour to 8/hourF 10/hour. The AHI was

reduced to less than 10/hour in 11 of the 15 patients.

Positional therapy has been compared with nasal

CPAP in a randomized cross-over study in 13 patients

who had an AHI in the supine posture more than twice

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353346

that in the lateral posture with an AHI of less than

15/hour in the lateral posture [42]. Each therapy was

delivered for 2 weeks, and the patient then switched to

the other therapy with no washout period between

therapies. Various outcome variables were evaluated.

The patients in this study had relatively mild disease,

with an AHI of 18/hourF 5/hour on the initial baseline

study. The AHI was reduced to 9.5/hourF 1.9/hour

with positional therapy and 3.4/hourF 0.5/hour with

nasal CPAP therapy. This difference was statistically

significant. The AHI during positional therapy corre-

lated with the severity of sleep apnea during the

baseline study (ie, the milder the disease the more

likely positional therapy was to be successful). The

subjective sensation of sleepiness as assessed by the

Epworth Sleepiness Scale improved significantly with

both therapies, and the degree of improvement was not

significantly different between therapies. Objective

alertness as assessed by the maintenance of wakeful-

ness test was not significantly different after the two

treatments. Tests of cognitive function improved

equally with both therapies. Four patients preferred

positional therapy, 7 preferred CPAP, and 2 had no

preference. In this group of patients with mild disease,

positional therapy seemed to be almost as effective as

nasal CPAP therapy. Although nasal CPAP improved

the AHI to a greater extent than positional therapy, it

did not result in greater improvements in subjective

and objective sleepiness or cognitive performance.

Positional therapy seems to be a reasonable alternative

to nasal CPAP in patients with mild disease with a

positional component. The long-term effects of posi-

tional therapy have not been evaluated.

A recent systematic review evaluated the effects

of positional therapy on sleep apnea [27]. No study

met the entry criteria. The study by Jokic et al [42]

came closest but was rejected because it compared

positional therapy to nasal CPAP rather than placebo.

This may not be fair because comparison to a therapy

that is known to be effective for sleep apnea (nasal

CPAP) is not an unreasonable approach and has been

used successfully to evaluate dental appliances. This

study only included 13 patients, which clearly poin-

ted out the need for additional studies to evaluate this

treatment modality.

Summary

Positional therapy can be considered in patients

with sleep apnea who have at least twice the number

of respiratory events in the supine position than in the

lateral position and have an AHI of less than 15/hour

and preferably less than 10/hour in the lateral posi-

tion. The number of such patients seen in a typical

sleep laboratory has not been determined adequately.

If a patient’s overall AHI is more than 15/hour (ie, the

patient’s sleep apnea is at least moderate [26]), a

follow-up sleep study that documents that the posi-

tional therapy chosen is effective at reducing the AHI

should be performed. In the authors’ sleep center, a

tennis ball attached to the pajamas or in a backpack or

wedge pillows are used to train patients to sleep in the

lateral position because these methods are much

simpler and less expensive than the posture alarm.

Pharmacologic therapy for obstructive

sleep apnea

An effective pharmacologic therapy for OSA is

desirable because all current forms of treatment have

significant limitations. Over the past several years,

much has been discovered about the pathogenesis of

OSA. Although ventilation may be normal during

wakefulness in patients with OSA, a sleep-induced

reduction in upper airway dilator muscle activity

results in collapse of an anatomically narrowed

upper airway [43]. Augmenting the activity of upper

airway dilator muscles during sleep by excitation of

motoneurons that innervate them is an attractive

approach in the development of an effective pharma-

cologic agent. Other approaches that have been used

include modifying sleep architecture (eg, reducing

rapid eye movement [REM] sleep because OSA

tends to be worse during this sleep stage) and using

respiratory stimulants. Several agents have been

tried, but none has been found to be consistently

efficacious to be recommended as standard therapy.

A detailed review of trials of medications in OSA

has been published [44].

Protriptyline

Two randomized, double-blind, placebo-con-

trolled, cross-over trials of protriptyline, a non-

sedating tricyclic antidepressant and REM sleep

suppressant, have been performed involving only a

total of 15 patients with OSA, with conflicting results.

Brownell et al [45] did not find a significant change in

the overall apnea index after 2 weeks of protriptyline

(20 mg/day) compared with placebo in 5 male patients

with OSA with relatively severe disease. The apnea

index during REM sleep (but not during non-REM

sleep) was reduced in association with a decrease in

REM apnea time, which is expressed as a proportion

of total sleep time and improvement in nocturnal

oxygenation. Subjective daytime sleepiness was

improved in 4 patients. The reduction in REM sleep

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353 347

seen during treatment accounted for the decrease in

REM apnea time. In 3 patients, follow-up sleep

studies after 6 months of protriptyline did show a

statistically significant reduction in the overall apnea

index, but the changes seen were modest (56/hourF8.1/hour compared with 70.9/hourF 12.2/hour at

baseline). The REM apnea index was decreased from

15.7/hourF 4.2/hour during placebo to 3.7/hourF0.7/hour during protriptyline, although this change

did not attain statistical significance.

Whyte et al [46], using a similar study design,

found that protriptyline (20 mg/day) for 2 weeks did

not have significant effects on symptoms, frequency

of apneas and hypopneas, oxyhemoglobin desatura-

tion, and arousals in 10 patients with OSA who also

had relatively severe disease. Surprisingly, protripty-

line did not reduce significantly the amount of REM

sleep in this study. In an unblinded, uncontrolled

study, Hanzel et al [47] reported that protriptyline

(10 mg/day) for 4 weeks reduced the AHI from

57/hourF 9/hour to 33/hourF 8/hour. The AHI dur-

ing REM sleep did not change significantly, but this

was difficult to interpret given the significant reduc-

tion of REM sleep with therapy. Protriptyline sig-

nificantly reduced the AHI during non-REM sleep,

however. Two other unblinded, uncontrolled studies

showed improvements in the AHI and nocturnal

oxygenation [48] and daytime hypersomnolence

[49] with protriptyline.

Summary

Protriptyline may reduce modestly (but not abol-

ish) the AHI in some patients with OSA that may be

associated with improvement in daytime sleepiness.

Aside from reducing REM sleep, other mechanisms,

such as stimulation of hypoglossal motoneurons, may

be responsible for the effects on sleep-disordered

breathing [50]. Given the small number of patients

involved in these trials, the occurrence of anticholi-

nergic side effects, including dry mouth, constipation,

and urinary retention, in a significant number of

patients, and modest reduction in the AHI in only

one controlled study, protriptyline cannot be recom-

mended currently as an effective pharmacologic agent

in the treatment of OSA. Further studies are required

to determine its efficacy in persons with mild to

moderate disease (AHI < 30/hour) or in patients with

only REM-related OSA.

Progesterone

Progesterone, a ventilatory stimulant, has been

tried in the treatment of OSA. An uncontrolled study

reported a possible role of medroxyprogesterone

acetate (MPA) in the treatment of OSA, especially

in hypercapnic patients [51]. Other uncontrolled

studies did not show any significant effects of

MPA, however [52,53]. Progesterone hormone

replacement in postmenopausal women with OSA

also has not been found to be effective [54]. Most

importantly, a randomized, double-blind, placebo-

controlled cross-over trial that involved ten male

patients with OSA also failed to show any effect of

MPA on sleep-disordered breathing [55]. This study

included four patients with daytime hypercapnea

(PaCO2>45 mm Hg).

Medroxyprogesterone acetate also has been tried

in the treatment of patients with obesity-hypoventila-

tion syndrome (in whom OSA is frequently present).

In an uncontrolled study of ten patients with the

obesity-hypoventilation syndrome (Pickwickian syn-

drome), MPA (20 mg every 8 hours) significantly

reduced the daytime pCO2 by 13F 2.6 mm Hg (SEM)

and increased daytime pO2 by 12.6F 2.7 mmHg after

4 to 9 months of treatment [56]. There was no

significant change in body weight during treatment.

Withdrawal of MPA for 1 month in seven patients

resulted in deterioration to pretreatment levels, and

reinstitution of MPA resulted in improvement of

arterial blood gas values. Randomized, controlled

trials in a larger sample of patients are lacking,

however, and the role of progesterone in association

with nocturnal positive airway pressure therapy in

obesity-hypoventilation syndrome is unclear.

Currently, there is no good evidence that proges-

terone is a useful agent in the treatment of OSA. Its

role in the treatment of patients who develop obesity-

hypoventilation syndrome is also unclear, because no

long-term, controlled studies have been conducted

regarding its efficacy and safety in this condition.

Thyroid hormone replacement

Hypothyroidism has been associated with OSA.

In small case series, the presence of OSA was

reported in 25% to 82% of diagnosed hypothyroid

patients [57–60]. Not all of the patients in these

reports were obese, and other mechanisms aside from

obesity have been implicated, including hypotonia of

upper airway dilator muscles caused by myopathy

[58], narrowing of the upper airway by deposition of

mucopolysaccharides and protein extravasation into

the tissues of the oropharynx [61], and impaired

ventilatory control [62].

In a group of 200 patients referred for polysom-

nography for suspected OSA and screened for hypo-

thyroidism, Skjodt et al [63] reported on 3 patients

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353348

who were confirmed to have OSA and undiagnosed

hypothyroidism. These 3 patients were treated with

thyroxine alone without the use of CPAP or a dental

device. Subjective symptoms, oxyhemoglobin desa-

turations, and the AHI all improved with thyroid

replacement therapy. The AHI decreased from

30/hour, 14/hour, and 24/hour, respectively, before

treatment to 1.7/hour, 1/hour, 16 /hour, respectively,

after therapy. There was no significant change in the

body mass index. In an earlier study, nine patients

with hypothyroidism with OSA were treated with

thyroxine for 3 to 12 months [59]. The apnea index

decreased from 71.8/hourF 18/hour to 12.7/hourF6.1/hour after treatment without any significant

change in body weight. The reason why thyroid

replacement improves OSA is unclear, but mecha-

nisms other than weight loss also seem to be important.

Not all patients have responded to thyroid replace-

ment therapy alone. In six of eight hypothyroid

patients with relatively severe OSA, Grunstein et al

reported that normalization of thyroid status with

thyroxine therapy did not improve the apnea index

[58]. The apnea index was 51/hourF 6/hour before

treatment and 45/hourF 8/hour after correction of the

hypothyroid state. CPAP therapy was required in

these patients.

It would be reasonable to start CPAP therapy in

patients with hypothyroidism with severe OSA and in

patients with an urgent reason to treat the sleep apnea,

in combination with thyroid replacement therapy. An

evaluation of whether CPAP therapy is still required

can be performed after euthyroid status has been

achieved. In persons with less severe sleep apnea,

treatment with thyroid replacement alone can be tried

and a follow-up study performed after achievement of

euthyroid state to ensure that OSA has been elimi-

nated. Whether it is cost effective to screen all

patients diagnosed with OSA for hypothyroidism is

controversial [63], but it seems to be unwarranted

[64,65] unless clinical symptoms suggest the pres-

ence of hypothyroidism.

Serotonergic agents

Obstructive sleep apnea is characterized by repet-

itive episodes of upper airway obstruction during

sleep. Airway obstructions are associated with a

decrease in the activity of upper airway dilator

muscles, such as the genioglossus (which controls

tongue movements) [43]. If upper airway dilator

muscle activity can be maintained or augmented dur-

ing sleep, then pharyngeal collapse may be prevented.

Several animal studies have suggested that serotonin is

important in the maintenance of upper airway patency.

Serotonergic neurons exert an excitatory effect on

upper airway dilator motoneurons [66,67]. In the

English bulldog, a natural animal model of OSA, the

systemic administration of serotonin antagonists

resulted in suppression of upper airway dilator muscle

activity, which led to a reduction in upper airway cross-

sectional area and oxyhemoglobin desaturations [68].

On the other hand, administration of the serotonergic

agents, trazodone and L-tryptophan, was effective in

treating sleep-disordered breathing in the English

bulldog, and the effectiveness of this therapy was

related to increased upper airway dilator muscle activi-

ty during sleep [69].

In humans, administration of a selective serotonin

reuptake inhibitor (SSRI) increased activity of upper

airway dilator muscle muscles during wakefulness in

normal subjects [70] and during non-REM sleep in

patients with OSA [71], which suggested that these

agents may be effective in treating OSA. Adminis-

tration of the serotonin precursor, L-tryptophan, was

reported to be effective in decreasing obstructive

apneas in non-REM sleep in an uncontrolled study

of 12 patients with OSA [72].

At least three published studies have used SSRI as

treatment for OSA. In an unblinded, uncontrolled

study, Hanzel et al [47] found that fluoxetine

(20 mg/day) reduced the AHI from 57/hourF 9/hour

to 34/hourF 6/hour after 4 weeks of treatment. The

AHI and the number of desaturation events per hour

of sleep were reduced by at least 50% in 4 of

12 patients. The reduction in AHI was seen only

during non-REM sleep and not during REM sleep.

Berry et al [71] studied the effects of a single 40-mg

dose of paroxetine in a group of eight adult men with

severe OSA in a double-blind cross-over manner.

Paroxetine did not decrease the AHI, although it

did increase genioglossus muscle activity. It would

be hard to assess the efficacy of a medication after a

single dose, however. Kraiczi et al conducted a

double-blind, randomized, placebo-controlled trial

[73] and determined the effects of a relatively low

dose of paroxetine (20 mg/day) for 6 weeks in

patients with OSAwithout known psychiatric disease.

The AHI was 36.3/hourF 24.7/hour (F standard

deviation) during placebo and was 30.2/hourF18.5/hour during treatment. The reduction was statis-

tically significant, albeit small, and was not attrib-

uted to changes in total sleep time or sleep

architecture. The mild reduction in AHI was mainly

caused by a decrease in the frequency of obstructive

apneas rather than hypopneas, and again this occurred

only during non-REM sleep. The number of apneas

and hypopneas during REM sleep was unchanged.

Overall, there was no change in psychopathologic

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353 349

symptoms (assessed by the Comprehensive Psycho-

pathological Rating Scale) and OSA-related daytime

complaints, including sleepiness, morning headache,

difficulties in concentration, memory complaints, and

low mood. Some patients did report improvements in

their well-being during paroxetine therapy compared

with placebo.

Summary

There is a growing body of evidence that sero-

tonin is important in the maintenance of upper airway

patency. SSRI therapy evaluated in a single random-

ized, controlled trial for several weeks resulted only

in a small reduction in the number of obstructive

apneas during non-REM sleep that was not accom-

panied by improvement in daytime symptoms.

Whether higher doses of SSRI will be more effective

is unknown. SSRI currently cannot be recommended

as an effective treatment for OSA. Further studies that

examine the effects of SSRI in persons with milder

disease are needed.

Fourteen different serotonin receptor subtypes

have been identified so far [74,75]. The specific type

of serotonin receptor that mediates the excitatory

effects of serotonin in upper airway motoneurons is

unclear and must be determined. Of interest is that in

trials of SSRIs in OSA, no effect on the AHI during

REM sleep has been found. The effect of SSRI

depends on remaining serotonin release [76]. In ani-

mal studies, activity of nerve cells that contain sero-

tonin that innervates upper airway motoneurons is

profoundly suppressed during REM sleep [77,78].

One can speculate that the absence of an effect of

SSRI during REM sleep may be caused partially by

the lack of available extracellular serotonin, and

reuptake inhibition cannot prevent the suppression

of upper airway motoneuron activity. To be effective

for OSA, it appears that a drug also should have direct

serotonin receptor agonist activity aside from inhibi-

ting serotonin reuptake. Although serotonin seems to

be important in maintaining upper airway patency

during sleep, other neurotransmitters also may play

a role in modulating the activity of motoneurons,

which innervate the muscles of the upper airway [79].

Other agents

Acetazolamide was found to decrease the AHI

from 50/hour to 26/hour in ten patients with OSA in a

randomized, double-blind, cross-over trial [46]. The

decrease in AHI was not accompanied by improve-

ment in symptoms, however, and paresthesias were

common. Theophylline [80,81] and transdermal nic-

otine [82] also do not seem to be helpful in OSA and

frequently cause sleep disruption. It is doubtful that

future studies in the treatment of OSA will involve

these medications.

Adjunctive therapy

Some patients with OSA continue to have residual

daytime sleepiness despite good compliance with

nasal CPAP. Two randomized, double-blind, pla-

cebo-controlled trials have been performed involving

this type of patients with OSA using modafinil, a

nonamphetamine wake-promoting medication with

unknown mechanism of action [83,84]. Modafinil

initially was investigated in the treatment of excessive

daytime sleepiness in narcolepsy. It has a favorable

side-effect profile [85] and lacks abuse potential [86].

In view of its efficacy in vigilance promotion with

minor side effects, it was believed to have a potential

role in the management of patients with OSA with

residual daytime sleepiness despite regular use of

CPAP. In a multicenter trial, Pack et al studied

157 patients with OSA (80 treated with placebo and

77 treated with modafinil) who were compliant with

CPAP therapy [83]. Treatment with CPAP and mod-

afinil (400 mg daily) significantly improved both

subjective (Epworth Sleepiness Scale) and objective

(multiple sleep latency test) measures of daytime

sleepiness compared with CPAP and placebo at 4

weeks. The percentage of patients with normalized

daytime sleepiness, defined as an Epworth Sleepiness

Scale score of less than 10, was significantly higher

with modafinil (51%) compared with placebo (27%).

The AHI andmean duration of CPAP usage (6.2 hours/

night) were the same in both groups.

In another study that involved 30 patients with

OSA, Kingshott et al found significant improvements

in alertness as measured by the maintenance of wake-

fulness test after 2 weeks of CPAP and modafinil but

found no effects on subjective and objective measure-

ments of daytime sleepiness. Based on the results of

these two well-designed trials, modafinil may be

considered as an adjunctive therapy in patients with

OSA who complain of persistent daytime sleepiness

and in whom good compliance with optimal levels of

CPAP has been checked objectively [87]. Before

committing to long-term treatment with modafinil,

one is advised first to embark on a thorough investiga-

tion of the cause of persistent daytime sleepiness that

can be specifically addressed, such as inappropriate

CPAP pressure, insufficient sleep, presence of another

sleep disorder (eg, narcolepsy), or drug effects.

Modafinil does not seem to affect sleep-disor-

dered breathing. In studies that involved untreated

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353350

patients with OSA [88,89] modafinil did not change

the AHI but improved objective measures of sleepi-

ness compared with placebo. One major concern for

using modafinil to treat the symptom of daytime

sleepiness in OSA patients who are not on definitive

or are intolerant of treatment (CPAP or oral appli-

ance) is that it does not prevent the cardiovascular

consequences associated with OSA because it does

not eliminate upper airway obstruction during sleep

[90]. Currently, there have been no published con-

trolled, long-term studies of modafinil in patients who

are not on definitive treatment for their sleep apnea,

and its use cannot be recommended for these patients.

Summary

Previous attempts at using pharmacologic agents

in the treatment of OSA have been disappointing.

Medroxyprogesterone has not been found to be useful

in the treatment of OSA. Use of protriptyline is

limited by frequent side effects, but its role in mild

and REM-related OSA must be clarified. SSRIs seem

to be ineffective in treatment of severe OSA. Further

studies are needed to determine their effect in persons

with mild disease. This is important because patients

with mild OSA (AHI < 15 hours) are most likely to

be noncompliant with CPAP therapy [91].

A recent systematic review of drug treatments for

OSA concluded that the current data do not support

the use of any drug as an alternative to CPAP [92]. Of

56 studies identified, only 9 studies met methodo-

logic criteria. Clearly, basic research and adequately

powered clinical trials are needed to identify an

effective medication for OSA.

References

[1] Grunstein R, Wilcox I, Yang TS, et al. Snoring and

sleep apnoea in men: association with central obesity

and hypertension. Int J Obes Relat Metab Disord 1993;

22:39–47.

[2] Young T, Palta M, Dempsey J, et al. The occurrence of

sleep-disordered breathing among middle aged adults.

N Engl J Med 1993;328:1230–5.

[3] Strobel RJ, Rosen RC. Obesity and weight loss in

obstructive sleep apnea: a critical review. Sleep 1996;

19:104–15.

[4] Haponik E, Smith P, Bohlman M, et al. Computerized

tomography in obstructive sleep apnea: correlation of

airway size with physiology during sleep and wakeful-

ness. Am Rev Respir Dis 1983;127:221–6.

[5] Schwab R, Gefter W, Hoffman E, et al. Dynamic upper

airway imaging during awake respiration in normal

subjects and patients with sleep disordered breathing.

Am Rev Respir Dis 1993;148:1385–400.

[6] Horner R, Mohiaddin R, Lowell D, et al. Sites and sizes

of fat deposits around the pharynx in obese patients with

obstructive sleep apnoea and weight matched controls.

Eur Respir J 1989;2:613–22.

[7] Brown I, McClean P, Boucher R, et al. Changes in pha-

ryngeal cross sectional area with posture and application

of continuous positive airway pressure in patients with

obstructive sleep apnea. Am Rev Respir Dis 1987;136:

628–32.

[8] Suratt P, McTier R, Findley L, et al. Collapsibility of

the nasopharyngeal airway in obstructive sleep apnea.

Am Rev Respir Dis 1985;132:967–71.

[9] Schwartz AR, Gold AR, Schubert N, et al. Effect of

weight loss on upper airway collapsibility in obstructive

sleep apnea. Am Rev Respir Dis 1991;144:494–8.

[10] Peppard PE, Young T, Palta M, et al. Longitudinal

study of moderate weight change and sleep disordered

breathing. JAMA 2000;284:3015–21.

[11] Smith PL, Gold AR, Meyers DA, et al. Weight loss in

mildly to moderately obese patients with obstructive

sleep apnea. Ann Intern Med 1985;103:850–5.

[12] Rubinstein I, Colapinto N, Rotstein L, et al. Improve-

ment in upper airway function after weight loss in

patients with obstructive sleep apnea. Am Rev Respir

Dis 1988;138:1192–5.

[13] Kiselak J, Clark M, Pera V, et al. The association be-

tween hypertension and sleep apnea in obese patients.

Chest 1993;104:775–80.

[14] Suratt P, McTier R, Findley L, et al. Effect of very-low

calorie diets with weight loss on obstructive sleep ap-

nea. Am J Clin Nutr 1992;56:182S–4S.

[15] Rajala R, Partinen M, Sane T, et al. Obstructive sleep

apnea in morbidly obese patients. J Intern Med 1991;

230:125–9.

[16] Lojander J, Mustajoki P, Ronka S, et al. A nurse-man-

aged weight reduction programme for obstructive sleep

apnoea syndrome. J Intern Med 1998;244:251–5.

[17] KansanenM, Vanninen E, TuunainenA, et al. The effect

of very low-calorie diet-induced weight loss on the se-

verity of obstructive sleep apnoea and autonomic ner-

vous function in obese patients with obstructive sleep

apnoea syndrome. Clin Physiol 1998;18:377–85.

[18] Pasquali R, Colella P, Cirignotta F, et al. Treatment

of obese patients with obstructive sleep apnea syn-

drome (OSAS): effect of weight loss and interference

of otorhinolaryngoiatric pathology. Int J Obes 1990;

14:207–17.

[19] Peiser J, Lavie P, Ovnat A, et al. Sleep apnea syndrome

in the morbidly obese as an indication for weight re-

duction surgery. Ann Surg 1984;199:112–5.

[20] Sugerman HJ, Fairman RP, Sood RK, et al. Long-term

effects of gastric surgery for treating respiratory insuffi-

ciency of obesity. Am J Clin Nutr 1992;55:597S–601S.

[21] Harman EM,Wynne JW, Block AJ. The effect of weight

loss on sleep disordered breathing and oxygen desatura-

tion in morbidly obese men. Chest 1982;82:291–4.

[22] Charuzi I, Ovnat A, Peiser J, et al. The effect of sur-

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353 351

gical weight reduction on sleep quality in obesity-re-

lated sleep apnea syndrome. Surgery 1985;97:535–8.

[23] Scheuller M, Weider D. Bariatric surgery for treatment

of sleep apnea syndrome in 15 morbidly obese pa-

tients: long-term results. Otolaryngol Head Neck Surg

2001;125:299–302.

[24] Pillar G, Peled R, Lavie P. Recurrence of sleep apnea

without concomitant weight increase 7.5 years after

weight reduction surgery. Chest 1994;106:1702–4.

[25] Young T, Peppard PE, Gottlieb DJ. Epidemiology of

obstructive sleep apnea: a population health perspec-

tive. Am J Respir Crit Care Med 2002;165:1217–39.

[26] Sleep-related breathing disorders in adults: recommen-

dations for syndrome definition and measurement tech-

niques in clinical research. The Report of an American

Academy of Sleep Medicine Task Force. Sleep 1999;

22:667–89.

[27] Shneerson J, Wright J. Lifestyle modification for ob-

structive sleep apnoea. Cochrane Database of System-

atic Reviews 2002;4.

[28] NIH Technology Assessment Conference Panel. Meth-

ods for voluntary weight loss and control. Ann Intern

Med 1993;119:764–70.

[29] Garner D, Wooley S. Confronting the failure of behav-

ioural and dietary treatment for obesity. Clin Psychol

Rev 1991;11:729–80.

[30] Sampol G, Munoz X, Sagales MT, et al. Long-term

efficacy of dietary weight loss in sleep apnoea/hypo-

pnoea syndrome. Eur Respir J 1998;12:1156–9.

[31] Oksenberg A, Silverberg DS, Arons E, et al. Positional

vs nonpositional obstructive sleep apnea patients: an-

thropomorphic, nocturnal polysomnographic, and mul-

tiple sleep latency test data. Chest 1997;112:629–39.

[32] Cartwright RD. Effect of sleep position on sleep apnea

severity. Sleep 1984;7:110–4.

[33] McEvoy RD, Sharp DJ, Thornton AT. The effects of

posture on obstructive sleep apnea. Am Rev Respir Dis

1986;133:662–6.

[34] Oksenberg A, Khamaysi I, Silverberg DS, et al. Asso-

ciation of body position with severity of apneic events

in patients with severe nonpositional obstructive sleep

apnea. Chest 2000;118:1018–24.

[35] Cartwright R, Ristanovic R, Diaz F, et al. A comparative

study of treatments for positional sleep apnea. Sleep

1991;14:546–52.

[36] Cartwright R, Lloyd S, Lilie J, et al. Sleep position

training as treatment for sleep apnea syndrome: a pre-

liminary study. Sleep 1985;8:87–94.

[37] Jokic R, Klimaszewski A, Croosley M, et al. Positional

treatment vs continuous positive airway pressure in

patients with positional obstructive sleep apnea syn-

drome. Chest 1999;115:771–81.

[38] Jan MA, Marshall I, Douglas NJ. Effect of posture on

upper airway dimensions in normal human. Am J Res-

pir Crit Care Med 1994;149:145–8.

[39] Martin SE, Marshall I, Douglas NJ. The effect of

posture on airway calibre with the sleep-apnea/hypo-

pnea syndrome. Am J Respir Crit Care Med 1995;

152:721–4.

[40] Neill AM, Angus SM, Sajkov D, et al. Effects of sleep

posture on upper airway stability in patients with ob-

structive sleep apnea. Am J Respir Crit Care Med 1997;

155:199–204.

[41] Penzel T, Moller M, Becker HF, et al. Effect of sleep

position and sleep stage on the collapsibility of the up-

per airways in patients with sleep apnea. Sleep 2001;

24:909–95.

[42] Jokic R, Klimaszewski A, Croosley M, et al. Positional

treatment vs continuous positive airway pressure in

patients with positional obstructive sleep apnea syn-

drome. Chest 1999;115:771–81.

[43] Remmers JE, Degroot WJ, Sauerland EK, et al. Patho-

genesis of upper airway occlusion during sleep. J Appl

Physiol 1978;44:931–8.

[44] Hudgel DW, Thanakitcharu S. Pharmacologic treat-

ment of sleep-disordered breathing. Am J Respir Crit

Care Med 1998;158:691–9.

[45] Brownell LG, West P, Sweatman P, et al. Protriptyline

in obstructive sleep apnea. N Engl J Med 1982;307:

1037–42.

[46] Whyte KF, Gould GA, Airlie MAA, et al. Role of

protriptyline and acetazolamide in the sleep apnea/hy-

popnea syndrome. Sleep 1988;11:463–72.

[47] Hanzel DA, Proia NG, Hudgel DW. Response of ob-

structive sleep apnea to fluoxetine and protriptyline.

Chest 1991;100:416–21.

[48] Smith PL, Haponik EF, Allen RP, et al. The effects of

protriptyline in sleep-disordered breathing. Am Rev

Respir Dis 1983;127:8–13.

[49] Conway WA, Zorck F, Piccione P, et al. Protriptyline in

the treatment of sleep apnoea. Thorax 1982;37:49–53.

[50] Bonora M, St. John WM, Beldsoe TA. Differential

elevation by protriptyline and depression by diazepam

of upper airway respiratory muscle activity. Am Rev

Respir Dis 1985;131:41–5.

[51] Strohl KP, Hensley NJ, Saunders NA, et al. Progester-

one administration and progressive sleep apnea. JAMA

1981;245:1230–2.

[52] Orr WC, Imes NK, Martin RJ. Progesterone therapy in

obese patients with sleep apnea. Arch Intern Med 1979;

139:109–11.

[53] Rajagopal KR, Abbrecht PH, Jabbari B. Effects of me-

droxyprogesterone acetate in obstructive sleep apnea.

Chest 1986;90:815–21.

[54] Cistulli PA, Barnes DJ, Grunstein RR, et al. Effect of

short term hormone replacement in the treatment of

obstructive sleep apnoea in postmenopausal women.

Thorax 1994;49:699–702.

[55] Cook WR, Benich JJ, Wooten SA. Indices of severity

of obstructive sleep apnea syndrome do not change

during medroxyprogesterone acetate therapy. Chest

1989;96:262–6.

[56] Sutton FD, Zwillich CW, Creagh CE, et al. Progester-

one for outpatient treatment of Pickwickian syndrome.

Ann Intern Med 1975;83:476–9.

[57] Lin CC, Tsan KW, Chen PJ. The relationship between

sleep apnea syndrome and hypothyroidism. Chest

1994;105:1296–7.

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353352

[58] Grunstein RR, Sullivan CE. Sleep apnea and hypothy-

roidism: mechanisms and management. Am J Med

1988;85:775–9.

[59] Rajagopal KR, Abbrecht PH, Derderian SS, et al. Ob-

structive sleep apnea in hypothyroidism. Ann Intern

Med 1984;101:491–4.

[60] Pelttari L, Rauhala E, Polo O, et al. Upper airway ob-

struction in hypothyroidism. J Intern Med 1994;236:

177–81.

[61] Orr WC, Males JL, Imes NK. Myxedema and obstruc-

tive sleep apnea. Am J Med 1981;70:1061–6.

[62] Skatrud J, Iber C, Ewart R, et al. Disordered breathing

during sleep in hypothyroidism. Am Rev Respir Dis

1981;124:325–9.

[63] Skjodt NM, Atkar R, Easton PA. Screening for hypo-

thyroidism in sleep apnea. Am J Respir Crit Care Med

1999;160:732–5.

[64] Winkelman JW, Goldman H, Piscatelli N, et al. Are

thyroid function tests necessary in patients with sus-

pected sleep apnea? Sleep 1996;19:790–3.

[65] Kapur VK, Koepsell TD, deMaine J, et al. Association

of hypothyroidism and obstructive sleep apnea. Am J

Respir Crit Care Med 1998;158:1379–83.

[66] Berger AJ, Bayliss DA, Viana F. Modulation of neo-

natal rat hypoglossal motoneuron excitability by sero-

tonin. Neurosci Lett 1992;143:164–8.

[67] Kubin L, Tojima H, Davies RO, et al. Serotonergic

excitatory drive to hypoglossal motoneurons in the

decerebrate cat. Neurosci Lett 1992;139:243–8.

[68] Veasey SC, Panckeri KA, Hoffman EA, et al. The

effects of serotonin antagonists in an animal model

of sleep-disordered breathing. Am J Respir Crit Care

Med 1996;153:776–86.

[69] Veasey SC, Fenik P, Panckeri KA, et al. The effects of

trazodone with L-tryptophan on sleep-disordered

breathing in the English bulldog. Am J Respir Crit

Care Med 1999;160:1659–67.

[70] Sunderram J, Parisi RA, Stobel RJ. Serotonergic stim-

ulation of the genioglossus and the response to nasal

continuous positive airway pressure. Am J Respir Crit

Care Med 2000;162:925–9.

[71] Berry RB, Yamura EM, Gill K, et al. Acute effects of

paroxetine on genioglossus activity in obstructive sleep

apnea. Sleep 1999;22:1087–92.

[72] Schmidt HS. L-tryptophan in the treatment of impaired

respiration in sleep. Bull Eur Physiopathol Respir 1983;

19:625–9.

[73] Kraiczi H, Hedner J, Dahlof P, et al. Effect of serotonin

uptake inhibition on breathing during sleep and day-

time symptoms in obstructive sleep apnea. Sleep 1999;

22:61–7.

[74] Hoyer D, Martin G. 5HT receptor classification and

nomenclature: towards a harmonization with the hu-

man genome. Neuropharmacology 1997;36:419–28.

[75] Hoyer D, Martin GR. Classification and nomenclature

of 5 HT receptors: a comment on current issues. Behav

Brain Res 1996;73:263–8.

[76] Fuller RW. Serotonin uptake inhibitors: uses in clinical

therapy and in laboratory research. Prog Drug Res

1995;45:167–204.

[77] Jacobs BL, Azmitia EC. Structure and function of the

brain serotonin system. Physiol Rev 1992;72:165–229.

[78] Woch G, Davies RO, Pack AI, et al. Behaviour of

raphe cells projecting to the dorsomedial medulla dur-

ing carbachol-induced atonia in the cat. J Physiol

1996;490:745–58.

[79] Yamuy J, Fung SJ, Xi M, et al. Hypoglossal motoneu-

rons are postsynaptically inhibited during carbachol-

induced rapid eye movement sleep. Neuroscience 1999;

94:11–5.

[80] Guilleminault C, Hayes B. Naloxone, theophylline, bro-

mocriptine, and obstructive sleep apnea: negative re-

sults. Bull Eur Physiopathol Respir 1983;19:632–4.

[81] Mulloy E,McNicholasWT. Theophylline in obstructive

sleep apnea: a double blind evaluation. Chest 1992;101:

753–7.

[82] Davila DG, Hunt RD, Offord KP, et al. Acute effects of

transdermal nicotine on sleep architecture, snoring, and

sleep-disordered breathing in nonsmokers. Am J Re-

spir Crit Care Med 1994;150:469–74.

[83] Pack AI, Black JE, Schwartz JR, et al. Modafinil as

adjunct therapy for daytime sleepiness in obstructive

sleep apnea. Am J Respir Crit Care Med 2001;164:

1675–81.

[84] Kingshott RN, Vennelle M, Coleman EL, et al. Ran-

domized, double-blind, placebo-controlled crossover

trial of modafinil in the treatment of residual exces-

sive daytime sleepiness in the sleep apnea/hypopnea

syndrome. Am J Respir Crit Care Med 2001;163:

918–23.

[85] Besset A, Cherit M, Carlander B, et al. Use of mod-

afinil in the treatment of narcolepsy: a long term fol-

low-up study. Neurophysiol Clin 1996;26:60–6.

[86] Jasinski DR. An evaluation of the abuse potential of

modafinil using methyphenidate as a reference. J Psy-

chopharmacol 2000;14:53–60.

[87] Black J. Pro: Modafinil has a role in management of

sleep apnea. Am J Respir Crit Care Med 2003;167:

105–6.

[88] Arnulf I, Homeyer P, Garma L, et al. Modafinil in

obstructive sleep apnea-hypopnea syndrome: a pilot

study in 6 patients. Respiration 1997;64:159–61.

[89] Heitmann J, Cassel W, Grote L, et al. Does short-term

treatment with modafinil affect blood pressure in pa-

tients with obstructive sleep apnea? Clin Pharmacol

Ther 1999;65:328–35.

[90] Pollak CP. Con: Modafinil has no role in management

of sleep apnea. Am J Respir Crit Care Med 2003;167:

106–7.

[91] McArdle N, Devereux G, Heidarnejad H, et al. Long-

term use of CPAP therapy for sleep apnea/hypopnea

syndrome. Am J Respir Crit Care Med 1999;159:

1108–14.

[92] Smith I, Lasserson T, Wright J. Drug treatments for

obstructive sleep apnea. Cochrane Database of System-

atic Reviews 2002;4.

U.J. Magalang, M.J. Mador / Clin Chest Med 24 (2003) 343–353 353

The role of oral appliance therapy in the treatment of

obstructive sleep apnea

Kathleen A. Ferguson, MD, FRCPC, FCCP

Division of Respirology, London Health Sciences Centre, University of Western Ontario, 375 South Street, London,

Ontario, N6A 4G5, Canada

Oral appliances are an established treatment op-

tion for simple snoring and obstructive sleep apnea

(OSA). Early evidence led to the recommendation

that they be used for the treatment of mild OSA or

simple snoring [1,2]. Recently published, randomized

controlled clinical trials have shown them to be an

effective treatment option in many patients, and some

studies have suggested a reasonable level of effec-

tiveness in more severe OSA. Oral appliances are

appealing because they are simple to use, reversible,

and portable and generally have a low complication

rate. This article reviews available types of oral ap-

pliances, their mechanism of action, and the evidence

for using oral appliance therapy. The role of the phy-

sician and dentist is discussed. The article also re-

views the side effects and complications of appliance

therapy and the evidence about predictors of outcome

of treatment.

Appliance types and mechanisms of action

There are two main appliance groups in common

clinical use: tongue repositioning devices and man-

dibular repositioning appliances (MRAs) (Figs. 1–3).

An infrequently used design is a palatal lifting device,

which contacts the soft palate directly. Because of the

limited effectiveness of this device in the treatment of

snoring [3] and obstructive sleep apnea (OSA) [4], it

is not discussed in this article.

Effects of mandibular and tongue advancement on

upper airway patency

The effects of oral appliances on upper airway size

are variable and depend on the method of imaging the

airway, when the studies are performed (ie, wakeful-

ness versus sleep), the subject’s body position (ie,

supine versus upright), the type of appliance, and the

amount of mandibular protrusion. Oral appliancesmay

improve upper airway patency by enlarging the upper

airway or by decreasing upper airway collapsibility

(eg, improving upper airway muscle tone). Simple

active anterior movement of the tongue or mandible

can increase cross-sectional airway size in subjects

with and without OSA [5]. Passive mandibular ad-

vancement during general anesthesia stabilized the

upper airway by increasing airway size in the retro-

palatal and retroglossal area and reducing closing

pressure [6]. The effect of passive pharyngeal advance-

ment during anesthesia in the retropalatal area is great-

er in nonobese subjects [7].

Several studies have evaluated the effects of MRAs

on upper airway size using upright lateral cephalom-

etry (during wakefulness) (Fig. 4). These results are

sometimes conflicting. In two studies, an MRA in-

creased the posterior airway space in most subjects

[8,9]. In another study in which the amount of protru-

sion was individualized in each patient, there was no

change in the size of the posterior airway space with

the appliance on a cephalogram [10]. Other studies that

used upright lateral cephalometry have shown that

MRAs lower the tongue position, reduce the mandib-

ular-plane-to-hyoid distance, advance the mandible,

and widen the upper oropharynx (retropalatal and ret-

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00015-7

E-mail address: kafergus@uwo.ca

Clin Chest Med 24 (2003) 355–364

roglossal) in some subjects [9,11–13]. Similar reduc-

tions in mandibular-plane-to-hyoid distance [14],

increases in oropharyngeal airway size [14,15], and

velopharyngeal airway size [16] have been seen using

supine cephalograms.

Other imaging modalities (eg, CT, MRI) also have

demonstrated increases in pharyngeal airway size

[9,17] and volume [18]. Direct imaging of the awake

supine airway with videoendoscopy confirms that an

MRA increases the cross-sectional area of the airway,

particularly in the velopharynx [19].

The presence of an intraoral device affects upper

airway muscle tone. Tongue retaining devices (TRDs)

affect genioglossus muscle activity in patients with

OSA (awake or asleep), but effects of a TRD on other

upper airwaymuscles have not been evaluated [20,21].

A TRD worn during sleep reduced the AHI and

decreased genioglossus electromyogram (EMG) activ-

ity [21]. The modified TRD (no bulb) also reduced the

apnea-hypopnea index (AHI) and increased the peak

genioglossus activity measured just before airway

reopening. The presence of the device without tongue

advancement did have an impact on genioglossus

activity and on apnea severity. The mechanism for this

effect is not certain. A study using an MRA found that

upper airway muscle tone increased with an MRA

except in the postapnea period in the genioglossus,

where tone was lower [22]. This study suggested that

activation of the upper airway muscles may contribute

to upper airway patency during sleep. In a more recent

placebo-controlled trial, the simple presence of an

intraoral appliance had no impact on the AHI or

oxygen saturation [23]. The study suggested that

mandibular advancement is required for the appliance

to improve OSA because the presence of an intraoral

device without advancement showed no clinical effect.

Effectiveness of oral appliance therapy

Mandibular repositioning appliances

Several studies have evaluated the efficacy of

mandibular advancers. A detailed review of oral

Fig. 1. The Klearway adjustable oral appliance. (Courtesy of

Great Lakes Orthodontics, Ltd., Tonawanda, NY.)

Fig. 2. An adjustable Herbst appliance. (Courtesy of Great

Lakes Orthodontics, Ltd., Tonawanda, NY.)

Fig. 3. A Monobloc appliance. (Courtesy of Dr. Konrad

Bloch, University of Zurich.)

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364356

appliance therapy was published in 1995 [1]. The

literature at that time consisted of case reports and

retrospective and prospective case series (before and

after design), and most, but not all, were peer

reviewed. The authors pooled the results for the

TRD and MRA of different designs. Seventy percent

of the 304 subjects had a reduction in AHI by 50%,

51% had a posttreatment AHI of less than 10 per

hour, but as many as 40% had a posttreatment AHI of

more than 20 per hour. Snoring was reported to be

improved in most patients. Since 1995, several pro-

spective studies have been published, including

randomized and controlled trials. In the more recent

prospective case series of oral appliance therapy, 54%

to 81% of patients had a reduction in AHI by 50%

[24–27], and 51% to 64% of patients had a posttreat-

ment AHI of less than 10 per hour [24–26,28,29].

Ten prospective controlled clinical studies have

been published: five cross-over studies that compared

oral appliances to continuous positive airway pressure

(CPAP) (four randomized [30–33] and one non-

randomized [34]), three randomized studies that com-

pared two different appliance designs [35–37], and

two randomized, placebo-controlled trials [23,38].

Three of the cross-over studies that compared oral ap-

pliances to CPAP have been described in detail else-

where [39].

Clark et al published a cross-over study of the

Herbst appliance compared with CPAP therapy in

23 men with OSA [34]. The choice of initial therapy

was not randomized, and most patients used CPAP

before they used the MRA. Although not reported

directly, from the figure provided it seems that 4 pa-

tients (19%) had an AHI of less than 10 per hour with

the MRA set at roughly two thirds of maximal pro-

trusion. The mean decrease in AHI was 39%. Sleep

quality was improved more by CPAP than by the

MRA, and CPAP was more effective at reducing

the AHI. Symptoms of excessive daytime sleepiness

were equally improved by the two treatments. The

first cross-over study by Ferguson et al assessed a

fixed position, boil and bite MRA [30], and the sec-

ond study assessed a partly adjustable custom ap-

pliance [31]. Patients were randomly assigned to

4 months of treatment first with the MRA or with

CPAP and then they crossed over to the other treat-

Fig. 4. Diagrammatic representation of the anatomic points and planes used to identify craniofacial and soft tissue parameters and

areas on lateral cephalometric radiographs. S, center of the sella turcica; N, nasion; PNS, posterior nasal spine tip; ANS, anterior

nasal spine tip; Gn, gnathion; RGN, retrognathion; Me, menton; Go, gonion; H, anterior superior tip of hyoid bone; TT, tongue

tip; Eb, base of epiglottis; P, inferior tip of palate; SN-MP angle, angle between the cranial base (line between S and N) and the

mandibular plane. Linear measurements: 1. TGH, tongue height; 2. PNSP, soft palate length; 3. boundary between velopharynx

and nasopharynx; 4. RPAS, retropalatal airway space; 5. superior margin oropharyngeal airway space; 6. PAS, posterior airway

space; 7. inferior margin oropharyngeal airway space (upper boundary hypopharynx); 8. OB, overbite; 9. OJ, overjet; 10. MP,

mandibular plane, line between Me and Go; 11. MPH, mandibular plane to hyoid. Areas: Tongue area, area outlined by the dorsal

configuration of the tongue surface and lines which connect TT, RGN, H and Eb; Soft Palate Area, area confined by the outline

of the soft palate which starts and ends at PNS through P.

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364 357

ment for 4 months. Treatment success for the fixed

position MRA was 48% (reduction in AHI to

V10 per hour with relief of symptoms). The appliance

was well tolerated and had fewer side effects than

CPAP, but some patients (24%) were unable or un-

willing to use the fixed position MRA because of

poor overnight retention or discomfort. The MRAwas

effective in reducing snoring in most patients and re-

ducing excessive daytime sleepiness. The partly ad-

justable custom MRAwas successful in treating 55%

of patients (AHI V10 per hour and relief of symp-

toms). In these three published prospective cross-over

studies of MRA therapy versus CPAP in the treatment

of mild to moderate OSA, CPAP was more effective

in reducing snoring, improving oxygenation, and de-

creasing the AHI. In two of the three studies they

were equally effective in relieving excessive daytime

sleepiness. The MRA had a lower side effect rate (in

one study) and was the form of therapy preferred by

patients in all three studies.

Randerath et al conducted a randomized cross-

over study of an intraoral sleep apnea device (ISAD)

versus CPAP in patients with mild to moderate OSA

(AHI between 5 and 30 per hour) [32]. The appliance

was arbitrarily set at two thirds of maximum man-

dibular protrusion and was not further adjusted during

the study. CPAP was titrated to an effective pressure

in the laboratory. After 6 weeks of therapy, CPAP was

more effective at improving snoring, AHI, and oxy-

genation. The ISAD was not particularly effective at

reducing the AHI (baseline AHI 17.5F 7.7 to 13.8Fhsp sp="0.17">11.1 at 6 weeks; P = NS), although

patients reported greater ease of use and higher

compliance with the ISAD. Overall only 30% of

patients (6/20) had an AHI of less than 10 per hour

with the ISAD. The relatively low level of efficacy of

the ISAD may be related to the lack of titration of the

appliance during the 6-week period of therapy.

Engleman et al published a randomized cross-over

study of CPAP and an oral appliance in patients with a

range of severity of OSA (AHI 11–43 per hour) and at

least two symptoms of OSA [33]. The patients were

selected for the presence of reported sleepiness. In

addition to the usual outcomes, the study included a

maintenance of wakefulness test, the functional out-

comes of sleep questionnaire, the Short Form 36 health

survey (SF-36), and an assessment of cognitive per-

formance. The appliance was set at roughly 80% of

maximum mandibular protrusion. CPAP was more

effective than the oral device for improving AHI and

subjective ratings of daytime function, even in the

patients with milder OSA (AHI between 5 and 15).

There were no differences between the treatments in

the effect on objective measures of sleepiness or

cognition or patient preference. Preference for CPAP

therapy over the oral appliance was related to a higher

body mass index and greater daytime impairment. The

authors concluded that CPAP would be the preferred

first-line therapy in patients with OSA who have

significant functional impairment and sleepiness over

an oral appliance, even in patients with mild OSA

(defined by a lower AHI).

Three studies have compared different oral appli-

ances or designs. Hans et al evaluated a fixed position

appliance (SnoreGuard) and a modified device in

24 patients with mild OSA [35]. The device that pro-

truded the mandible (Device A) was more effective in

reducing the AHI than the device that minimally

opened the vertical dimension but did not protrude

the mandible (Device B). Three out of 10 patients with

Device A (30%) had an AHI of less than 10 per hour

with the appliance. Four of the 7 subjects who

switched to Device A after failing on Device B had

an improvement in AHI. Some patients had an increase

in AHI using Device A or Device B. Bloch et al con-

ducted a randomized, 21 February 2003controlled,

cross-over study of the Herbst (Fig. 2) and Monobloc

(Fig. 3) appliances, both of which set approximately

75% of maximum protrusion [36]. The AHI was less

than 10 in 75% of patients with the Monobloc ap-

pliance and in 67% of patients with the Herbst ap-

pliance. Both devices reduced sleepiness and snoring,

but patients felt that the Monobloc device was more

effective in reducing symptoms and preferred it for

long-term therapy.

A recent randomized, cross-over study evaluated

the effect of vertical dimension opening on the effica-

cy of an oral appliance [37]. The splint was construct-

ed with 4 mm of interincisal opening (MAS-1) or

14 mm of opening (MAS-2). Twenty-three patients

wore each appliance for 2 weeks in a random order.

Both appliances had similar efficacy in reducing the

AHI (complete and partial response 74% with MAS-1

and 61% with MAS-2). Both appliances improved

snoring and sleepiness, but there was a trend to more

jaw discomfort with MAS-2. Overall, the patients pre-

ferred the MAS-1 for long-term therapy. In this short-

term study, increasing the vertical opening did not

have an impact on appliance efficacy, but there is con-

cern that with long-term use this could have an impact

on side effects and complications.

Mehta et al published the first prospective, ran-

domized, placebo-controlled cross-over trial of an

MRA for the treatment of OSA [23]. Twenty-eight

patients had an acclimatization period during which

the mandible was incrementally advanced until symp-

toms resolved or maximum tolerated protrusion was

obtained. Patients were then randomly assigned to

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364358

treatment with the placebo followed by the active

device or treatment with the active device followed by

the placebo (lower plate of the appliance only). A

partial response was defined as symptomatic improve-

ment with an AHI reduced by 50% or more, but more

than 5 per hour and a complete response was defined

as a resolution of symptoms along with an AHI of less

than 5 per hour. The active appliance resulted in a

partial or complete response in 15 patients or 62.5%

(complete response in 9%–37.5%). Seventy-one per-

cent of patients had an AHI of less than 10 per hour

with the active appliance. The placebo device had no

impact on the AHI or oxygen saturation. The active

appliance improved snoring, sleep structure, oxygena-

tion, and daytime symptoms. There were few impor-

tant side effects and no complications.

A recent study has evaluated the effect of oral

appliance therapy on symptoms of OSA in a random-

ized, cross-over design that compared a mandibular

advancement splint to a placebo device [38]. In con-

trast to most other trials, the study included a multiple

sleep latency test to assess the impact of oral appliance

therapy on an objective measure of sleepiness. Most

of the patients (62 of 73; 85%) had moderate to severe

OSA (AHI� 15 per hour). 38 patients (52%) reported

subjective sleepiness (defined as an Epworth Sleepi-

ness Scale score > 10). On average the appliance was

set at 80% of maximum mandibular protrusion. The

active splint improved symptoms such as snoring and

reduced the AHI by 52% overall, with 63% of patients

having a complete or partial response. The active

splint reduced the Epworth Sleepiness Scale score

and increased the mean sleep latency significantly

when compared with the placebo device.

In summary, MRAs are an effective treatment

option for many patients with OSA, including some

patients with more severe OSA (higher AHI). They

improve snoring and daytime symptoms and reduce

the AHI and improve oxygenation during sleep. They

are not as effective as CPAP in reducing the AHI or

snoring. In some studies they were not as effective in

reducing symptoms of sleepiness as CPAP but in

other studies they were. Overall, CPAP is a more

effective treatment than an MRA and should be

considered first-line therapy in patients with more

severe symptoms and perhaps in patients with more

severe OSA, particularly if there is significant impair-

ment of oxygenation.

Tongue repositioners

Tongue repositioning devices include the TRD,

which is the best studied of these devices. The TRD is

a custom-made soft acrylic appliance that covers the

upper and lower teeth and has an anterior plastic bulb.

It uses negative suction pressure to hold the tongue in

a forward position inside the bulb. In 1982, Cartwright

and Samelson reported their initial experience with the

TRD in 20 patients [40]. Fourteen of the 20 patients

had undergone polysomnography before and with the

TRD. There was a reduction in AHI of approximately

50%, although patients only wore the TRD half the

night. Cartwright reported a second uncontrolled

study of the TRD in 16 patients [41]. Treatment

success in this study was defined as a reduction in

apnea index to the normal range (0–6 per hour) or a

50% reduction in apnea index. 69% were successfully

treated by the TRD by these criteria. In another case

series that evaluated the TRD in 15 patients, the

success rate was reported as 73% for the reduction

of the AHI to less than 10 per hour [42].

Side effects and complications

In a review published in 1995, the authors found

nine studies that reported side effects and complica-

tions [1]. Excessive salivation and temporary discom-

fort after awakening were commonly reported. In one

long-term study, 3 out of 20 patients stopped the

device because of temporomandibular joint pain, but

the pain ceased when they stopped treatment [43]. In

another study, 3 out of 14 patients reported a sense of

altered occlusion, but it was not systematically

studied [44]. In most short-term studies of oral

appliance therapy published since the 1995 review

article, side effects were common but generally minor

and no serious complications were generally observ-

ed. Several long-term studies have been published

that systematically have evaluated side effects and

complications from oral appliance therapy.

Pantin et al assessed 132 of 191 (69%) patients

consecutively treated with a mandibular advance-

ment splint over a 5-year period and performed a

dental examination on 106 of them [45]. Ten patients

had discontinued using the appliance because of

minor dental side effects. They documented occlusal

changes in 14% cases, and in two cases the changes

were great enough to recommend that the patient

stop treatment. Marklund et al investigated orthodon-

tic side effects of a soft and a hard acrylic MRA in

75 patients who reported using the device more than

50% of nights for approximately 2.5 years [46].

Overbite and overjet decreased, and 3 patients re-

ported a permanent change in occlusion. Hard acrylic

appliances and larger amounts of protrusion were

associated with more occlusal changes.

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364 359

Fritsch et al evaluated 22 patients who had used

either a Monobloc or a Herbst oral appliance for the

treatment of OSA [47]. Common side effects included

mucosal dryness (86%), tooth discomfort (59%), ex-

cessive salivation (55%), jaw pain (41%), and occlu-

sal changes (32%), but they were described as minor

side effects. Long-term appliance use was associated

with small orthodontic changes: decreased overjet and

overbite, retroclined maxillary incisors, and slight

anterior movement of the first mandibular molars.

Patients reported that symptoms caused by these

changes generally resolved after a few minutes in

the morning. A detailed study of skeletal and dental

changes with mandibular advancers in 100 patients

found similar results [48]. At 6 months of follow-up,

a vertical change in condylar position was noted,

the total anterior and posterior facial height was in-

creased, and overbite and overjet were decreased.

After 24 months of treatment, similar changes were

noted but the decrease in overbite and overjet was

more marked related to proclination of the mandibular

incisors. By 30 months of reported regular MRA use,

the proclination of the mandibular incisors was more

pronounced. The author did not comment on whether

these changes led to any clinical problems for the

patients who used the oral device. Overall, there is a

degree of occlusal change in patients with long-term

MRA use, and these changes must be monitored and

dealt with when they arise. Patients must be informed

of the potential for occlusal change when they embark

on oral appliance therapy.

Worsening of sleep apnea

Occasionally, an oral appliance can worsen apnea

severity [8,27,31,35,41]. In one of the more recent tri-

als, 4 of 28 subjects (14%) had an increase in AHI with

the appliance. The reason for this increase could not be

determined from a review of the patient data [27].

Treatment compliance

Some studies in the 1995 review reported the

long-term compliance of patients using an oral appli-

ance. Reported regular appliance use was in the range

of 75% to 100% for most of the studies, with one

study having a low compliance rate of only 50%.

More recent studies have had 76% to 90% of patients

reporting regular use [14,26]. In two of the cross-over

studies that compared oral appliances to CPAP, com-

pliance was measured by patient reports [30,31].

There was no difference in reported nightly use of

approximately 60% for all treatment arms. Until

objective compliance monitors are available, the

actual long-term compliance rates will be uncertain

given the unreliability of patient self-report for treat-

ment usage.

Titration of oral appliance therapy

Relative medical contraindications to first-line

therapy with an oral appliance include severe OSA,

severe excessive daytime sleepiness, and marked

arterial oxygen desaturations during sleep (eg, obe-

sity-hypoventilation). It may take time to optimize the

anterior position of the appliance and optimize treat-

ment success. Two studies have assessed overnight

titration of an oral appliance to determine the effective

therapeutic position [49,50]. This is a promising ap-

proach that may allow better identification of patients

in whom an oral device might be effective. CPAP ther-

apy can be titrated to the optimal pressure in a single

night and overall is more effective than oral appliance

therapy at reducing the AHI and correcting abnormal-

ities of oxygenation [30,31,34]. If an appliance could

be titrated more rapidly, then patients with more se-

vere OSA could be treated without delay.

Predictors of treatment outcome

Clinical predictors

Many studies have evaluated variables that may

be associated with treatment outcome (Box 1). Most

studies have been underpowered to find a signifi-

cant relationship between treatment outcome and

these variables. A younger age [32,51], lower body

mass index [41,51], lower neck size [23], position-

al OSA [41,52,53], and lower AHI [8,23,34,51,54]

and further amounts of mandibular protrusion [55]

have been related to improved treatment response.

Some studies, however, have demonstrated reasonably

good success rates in patients with more severe OSA

[4,23,25,27,50,56].

Craniofacial and dental predictors

Published studies have used various imaging

techniques to assess the upper airway and the factors

associated with treatment response. Several features

from cephalometry, including a smaller or narrow

oropharynx [11,51], smaller overjet [51], normal

mandible length [57], shorter mandibular plane to

hyoid distance [10], shorter soft palate length [10],

smaller upper to lower facial height ratios [58],

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364360

normal or reduced lower facial height [57], small soft

palate and tongue [57], increased retropalatal airway

space [23], and larger angle between the anterior

cranial base and mandibular plane [23] are associated

with improved outcome. Some authors have sug-

gested that a more micrognathic or retrognathic

mandible is associated with improved treatment

response [59]. Finally, hypopharyngeal closure that

causes OSA may be associated with improved treat-

ment outcome, but many patients with velopha-

ryngeal closure still get a good result [27].

Indications for oral appliance therapy

The American Academy of Sleep Medicine has

published guidelines about the use of oral appliance

therapy in the treatment of OSA [2]. These guidelines

stated that oral appliances are indicated as first-line

therapy in patients with simple snoring and mild OSA

and as second-line therapy for patients with moderate

to severe OSA when other therapies have failed. At

the time the guidelines were published the available

studies of oral appliance therapy were only uncon-

trolled, largely retrospective case series. Since then,

many prospective studies have been published,

including controlled clinical trials with comparisons

to CPAP, other appliances, and placebo [23,30,31,

34–36]. With evidence of effectiveness from ran-

domized controlled trials it is reasonable to expand

the indications for first-line therapy with an oral appli-

ance to the treatment of patients with moderate OSA.

The guidelines defined the roles of the physician

and dentist in the provision of oral appliance therapy

[2]. Physicians, preferably trained in sleep disorders,

perform the initial assessment and determine whether

the patient is ‘‘medically’’ suitable for oral appliance

therapy. A dentist skilled in this type of treatment

determines the patients’ ‘‘dental’’ suitability for oral

appliance treatment from a full assessment of oral and

dental health.

Treatment must be individualized to each patient,

with the dentist choosing the most appropriate oral

appliance. Tongue repositioning devices, such as the

TRD, are used particularly in patients with large

tongues or inadequate healthy teeth to use an MRA.

In general, MRAs require an adequate number of

healthy teeth for good retention. Severe temporoman-

dibular joint problems, inadequate protrusive ability,

and advanced periodontal disease are relative contra-

indications to the use of an MRA. In a study of 100

patients consecutively assessed by oral and maxillo-

facial surgeons, 34% of patients had primary contra-

indications to therapy and 16% had dental problems

or concerns about temporomandibular joint function

that would require careful dental follow-up [60].

Although many patients may be medically suitable

for oral appliance therapy, they require a careful

assessment by a qualified dental practitioner to deter-

mine if dental contraindications are present.

Long-term dental follow-up includes optimizing

the appliance, monitoring retention, and assessing

effectiveness. Periodic adjustments and repairs may

be required. Monitoring dental health, side effects,

and complications of therapy is also important. Med-

ical follow-up is necessary to evaluate treatment

response and assess for recurrence of OSA. It is

recommended that follow-up sleep studies be per-

formed to verify the improvement in apnea, oxygena-

tion, and sleep fragmentation by the oral appliance

[2]. This recommendation is supported by the evi-

dence that some patients have an increase in AHI with

oral appliance treatment [8,27,31,35,41].

Future directions

Future randomized controlled trials are needed to

compare the effectiveness of different types of appli-

Box 1. Predictors of oral appliance efficacy

Clinical predictors

Younger ageLower body mass indexLower neck circumferencePositional OSA (worse supine)Lower AHI (not a consistent predictor)Increased protrusion of appliance

Dental and craniofacial variables

Smaller and/or narrow oropharynxSmaller overjetNormal mandible lengthShorter mandibular plane to hypoid

distanceShorter soft palate lengthSmaller upper to lower facial height

ratiosNormal or reduced lower facial heightSmall soft palate and tongueIncreased retropalatal airway spaceLarger angle cranial base to mandibu-

lar plane

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364 361

ances and different design features (eg, the amount of

vertical opening). The effect of oral appliances on

excessive daytime sleepiness and performance must

be determined with objective and validated tools. The

precise indications, complication rates, and reasons

for treatment failure must be determined for each oral

appliance if it is going to be used in clinical practice.

Ongoing refinements of appliance design eventually

may lead to improved treatment outcomes. Only

when the mechanisms of action of oral appliance

therapy are fully understood can more effective

appliances be developed. On the horizon for the field

of oral appliance therapy is the introduction of a

compliance monitor that will allow an objective

determination of appliance usage. Several investiga-

tors also are developing systems that would allow

overnight titration of oral appliances in the sleep

laboratory. This might ultimately shorten the time

from initiation of oral appliance therapy to optimiza-

tion of the appliance.

Summary

The development of oral appliance treatment for

OSA represents an important step in the management

of this disease. Randomized, controlled clinical trials

have shown them to be an effective treatment option

for snoring and OSA in some patients, particularly

patients with less severe OSA or simple snoring and

patients who have failed other treatment modalities.

Although oral appliances are not as effective as CPAP

therapy, they work in most patients to relieve symp-

toms and apnea and are well tolerated by patients.

Most patients report improvements in sleep quality

and excessive daytime sleepiness. Short-term side

effects are generally minor and are related to excessive

salivation, jaw and tooth discomfort, and occasional

joint discomfort. These symptoms may lead to dis-

continuation of appliance therapy but usually improve

in most patients over time. Serious complications are

not common, but occlusal changes are more common

than previously believed.

References

[1] Schmidt-Nowara W, Lowe A, Wiegand L, Cartwright

R, Perez-Guerra F, Menn S. Oral appliances for the

treatment of snoring and obstructive sleep apnea: a

review. Sleep 1995;18:501–10.

[2] American Sleep Disorders Association. Practice param-

eters for the treatment of snoring and obstructive sleep

apnea with oral appliances. Sleep 1995;18:511–3.

[3] Marklund M, Franklin KA. Dental appliances in the

treatment of snoring: a comparison between an activa-

tor, a soft-palate lifter, and a mouth-shield. Swed Dent

J 1996;20:183–8.

[4] Barthlen GM, Brown LK, Wiland MR, Sadeh JS, Pat-

wari J, Zimmerman M. Comparison of three oral ap-

pliances for treatment of severe obstructive sleep apnea

syndrome. Sleep Medicine 2000;1:299–305.

[5] Ferguson KA, Love LL, Ryan CF. Effect of mandibu-

lar and tongue protrusion on upper airway size. Am J

Respir Crit Care Med 1997;155:1748–54.

[6] Isono S, Tanaka A, Sho Y, Konno A, Nishino T. Ad-

vancement of the mandible improves velopharyngeal

airway patency. J Appl Physiol 1995;79:2132–8.

[7] Isono S, Tanaka A, Tagaito Y, Sho Y, Nishino T. Pha-

ryngeal patency in response to advancement of the man-

dible in obese anesthetized persons. Anesthesiology

1997;87:1055–62.

[8] Schmidt-Nowara WW, Meade TE, Hays MB. Treat-

ment of snoring and obstructive sleep apnea with a

dental orthosis. Chest 1991;99:1378–85.

[9] Ishida M, Inoue Y, Suto Y, Okamoto K, Ryoke K, Hi-

gami S, et al. Mechanism of action and therapeutic in-

dication of prosthetic mandibular advancement in

obstructive sleep apnea syndrome. Psychiatry Clin Neu-

rosci 1998;52:227–9.

[10] Eveloff SE, Rosenberg CL, Carlisle CC, Millman RP.

Efficacy of a Herbst mandibular advancement device in

obstructive sleep apnea. Am J Respir Crit Care Med

1994;149:905–9.

[11] Mayer G, Meier-Ewert K. Cephalometric predictors for

orthopaedic mandibular advancement in obstructive

sleep apnoea. Eur J Orthod 1995;17:35–43.

[12] Battagel JM, L’Estrange PR, Nolan P, Harkness B. The

role of lateral cephalometric radiography and fluoros-

copy in assessing mandibular advancement in sleep-

related disorders. Eur J Orthod 1998;20:121–32.

[13] Liu Y, Lowe AA. Factors related to the efficacy of an

adjustable oral appliance for the treatment of obstruc-

tive sleep apnea. Chin J Dent Res 2000;3:15–23.

[14] Johal A, Battagel JM. An investigation into the

changes in airway dimension and the efficacy of man-

dibular advancement appliances in subjects with ob-

structive sleep apnoea. Br J Orthod 1999;26:205–10.

[15] Battagel JM, Johal A, L’Estrange PR, Croft CB, Kote-

cha B. Changes in airway and hyoid position in response

to mandibular protrusion in subjects with obstructive

sleep apnoea (OSA). Eur J Orthod 1999;21:363–76.

[16] Tsuiki S, Hiyama S, Ono T, Imamura N, Ishiwata Y,

Kuroda T, et al. Effects of a titratable oral appliance on

supine airway size in awake non-apneic individuals.

Sleep 2001;24:554–60.

[17] Gale DJ, Sawyer RH, Woodcock A, Stone P, Thomp-

son R, O’Brien K. Do oral appliances enlarge the air-

way in patients with obstructive sleep apnoea? A

prospective computerized tomographic study. Eur J

Orthod 2000;22:159–68.

[18] Cobo J, Canut JA, Carlos F, Vijande M, Llamas JM.

Changes in the upper airway of patients who wear a

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364362

modified functional appliance to treat obstructive sleep

apnea. Int J Adult Orthodon Orthognath Surg 1995;10:

53–7.

[19] Ryan CF, Love LL, Peat D, Fleetham JA, Lowe AA.

Mandibular advancement oral appliance therapy for

obstructive sleep apnoea: effect on awake calibre of

the velopharynx. Thorax 1999;54:972–7.

[20] Ono T, Lowe AA, Ferguson KA, Pae EK, Fleetham JA.

The effect of the tongue retaining device on awake ge-

nioglossus muscle activity in patients with obstructive

sleep apnea. Am J Orthod Dentofacial Orthop 1996;

110:28–35.

[21] Ono T, Lowe AA, Ferguson KA, Fleetham JA. A

tongue retaining device and sleep-state genioglossus

muscle activity in patients with obstructive sleep ap-

nea. Angle Orthod 1996;66:273–80.

[22] Yoshida K. Effect of a prosthetic appliance for treatment

of sleep apnea syndrome on masticatory and tongue

muscle activity. J Prosthet Dent 1998;79:537–44.

[23] Mehta A, Qian J, Petocz P, Darendeliler MA, Cistulli

PA. A randomized, controlled study of a mandibular

advancement splint for obstructive sleep apnea. Am J

Respir Crit Care Med 2001;163:1457–61.

[24] Millman RP, Rosenberg CL, Carlisle CC, Kramer NR,

Kahn DM, Bonitati AE. The efficacy of oral appliances

in the treatment of persistent sleep apnea after uvulo-

palatopharyngoplasty. Chest 1998;113:992–6.

[25] Pancer J, Al-Faifi S, Al-Faifi M, Hoffstein V. Evalua-

tion of variable mandibular advancement appliance for

treatment of snoring and sleep apnea [see comments].

Chest 1999;116:1511–8.

[26] Yoshida K. Effects of a mandibular advancement device

for the treatment of sleep apnea syndrome and snoring

on respiratory function and sleep quality. Cranio 2000;

18:98–105.

[27] Henke KG, Frantz DE, Kuna ST. An oral elastic man-

dibular advancement device for obstructive sleep ap-

nea. Am J Respir Crit Care Med 2000;161:420–5.

[28] Menn SJ, Loube DI, Morgan TD, Mitler MM, Berger

JS, Erman MK. The mandibular repositioning device:

role in the treatment of obstructive sleep apnea. Sleep

1996;19:794–800.

[29] Marklund M, Sahlin C, Stenlund H, Persson M, Frank-

lin KA. Mandibular advancement device in patients

with obstructive sleep apnea: long-term effects on ap-

nea and sleep. Chest 2001;120:162–9.

[30] Ferguson KA, Ono T, Lowe AA, Keenan SP, Fleetham

JA. A randomized crossover study of an oral appliance

vs nasal-continuous positive airway pressure in the

treatment of mild-moderate obstructive sleep apnea.

Chest 1996;109:1269–75.

[31] Ferguson KA, Ono T, Lowe AA, Al-Majed S, Love LL,

Fleetham JA. A short term controlled trial of an ad-

justable oral appliance for the treatment of mild to mod-

erate obstructive sleep apnoea. Thorax 1997;52:362–8.

[32] Randerath WJ, Heise M, Hinz R, Ruehle KH. An in-

dividually adjustable oral appliance vs continuous pos-

itive airway pressure in mild-to-moderate obstructive

sleep apnea syndrome. Chest 2002;122:569–75.

[33] Engleman HM, McDonald JP, Graham D, Lello GE,

Kingshott RN, Coleman EL, et al. Randomized cross-

over trial of two treatments for sleep apnea/hypopnea

syndrome: continuous positive airway pressure and

mandibular repositioning splint. Am J Respir Crit Care

Med 2002;166:855–9.

[34] Clark GT, Blumenfeld I, Yoffe N, Peled E, Lavie P. A

crossover study comparing the efficacy of continuous

positive airway pressure with anterior mandibular po-

sitioning devices on patients with obstructive sleep

apnea. Chest 1996;109:1477–83.

[35] Hans MG, Nelson S, Luks VG, Lorkovich P, Baek SJ.

Comparison of two dental devices for treatment of

obstructive sleep apnea syndrome (OSAS). Am J Or-

thod Dentofacial Orthop 1997;111:562–70.

[36] Bloch KE, Iseli A, Zhang JN, Xie X, Kaplan V,

Stoeckli PW, et al. A randomized, controlled crossover

trial of two oral appliances for sleep apnea treatment.

Am J Respir Crit Care Med 2000;162:246–51.

[37] Pitsis AJ, Darendeliler MA, Gotsopoulos H, Petocz P,

Cistulli PA. Effect of vertical dimension on efficacy of

oral appliance therapy in obstructive sleep apnea. Am J

Respir Crit Care Med 2002;166:860–4.

[38] Gotsopoulos H, Chen C, Qian J, Cistulli PA. Oral ap-

pliance therapy improves symptoms in obstructive

sleep apnea: a randomized, controlled trial. Am J Res-

pir Crit Care Med 2002;166:743–8.

[39] Ferguson KA. Oral appliance therapy for the treatment

of obstructive sleep apnea. Semin Respir Crit Care Med

1998;19:157–64.

[40] Cartwright RD, Samelson CF. The effects of a nonsur-

gical treatment for obstructive sleep apnea. JAMA

1982;248:705–9.

[41] Cartwright RD. Predicting response to the tongue retain-

ing device for sleep apnea syndrome. Arch Otolaryngol

1985;111:385–8.

[42] Cartwright R, Ristanovic R, Diaz F, Caldarelli D, Al-

der G. A comparative study of treatments for positional

sleep apnea. Sleep 1991;14:546–52.

[43] Clark GT, Arand D, Chung E, Tong D. Effect of ante-

rior mandibular positioning on obstructive sleep apnea.

Am Rev Respir Dis 1993;147:624–9.

[44] Nakazawa Y, Sakamoto T, Yasutake R, Yamaga K, Ko-

torii T, Miyahara Y, et al. Treatment of sleep apnea with

prosthetic mandibular advancement (PMA). Sleep

1992;15:499–504.

[45] Pantin CC, Hillman DR, Tennant M. Dental side ef-

fects of an oral device to treat snoring and obstructive

sleep apnea. Sleep 1999;22:237–40.

[46] Marklund M, Franklin KA, Persson M. Orthodontic

side-effects of mandibular advancement devices during

treatment of snoring and sleep apnoea. Eur J Orthod

2001;23:135–44.

[47] Fritsch KM, Iseli A, Russi EW, Bloch KE. Side effects

of mandibular advancement devices for sleep apnea

treatment. Am JRespir Crit CareMed 2001;164:813–8.

[48] Robertson CJ. Dental and skeletal changes associated

with long-term mandibular advancement. Sleep 2001;

24:531–7.

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364 363

[49] Raphaelson MA, Alpher EJ, Bakker KW, Perlstrom

JR. Oral appliance therapy for obstructive sleep apnea

syndrome: progressive mandibular advancement dur-

ing polysomnography. Cranio 1998;16:44–50.

[50] Petelle B, Vincent G, Gagnadoux F, Rakotonanahary D,

Meyer B, Fleury B. One-night mandibular advance-

ment titration for obstructive sleep apnea syndrome:

a pilot study. Am J Respir Crit Care Med 2002;165:

1150–3.

[51] Liu Y, Lowe AA, Fleetham JA, Park YC. Cephalomet-

ric and physiologic predictors of the efficacy of an ad-

justable oral appliance for treating obstructive sleep

apnea. Am J Orthod Dentofacial Orthop 2001;120:

639–47.

[52] Marklund M, Persson M, Franklin KA. Treatment suc-

cess with a mandibular advancement device is related

to supine-dependent sleep apnea. Chest 1998;114:

1630–5.

[53] Yoshida K. Influence of sleep posture on response to

oral appliance therapy for sleep apnea syndrome. Sleep

2001;24:538–44.

[54] O’Sullivan RA, Hillman DR, Mateljan R, Pantin C,

Finucane KE. Mandibular advancement splint: an ap-

pliance to treat snoring and obstructive sleep apnea.

Am J Respir Crit Care Med 1995;151:194–8.

[55] de Almeida FR, Bittencourt LR, de Almeida CI, Tsuiki

S, Lowe AA, Tufik S. Effects of mandibular posture

on obstructive sleep apnea severity and the temporo-

mandibular joint in patients fitted with an oral appli-

ance. Sleep 2002;25:507–13.

[56] Lowe AA, Sjoholm TT, Ryan CF, Fleetham JA, Fergu-

son KA, Remmers JE. Treatment, airway and compli-

ance effects of a titratable oral appliance. Sleep 2000;

23:S172–8.

[57] L’Estrange PR, Battagel JM, Harkness B, Spratley

MH, Nolan PJ, Jorgensen GI. A method of studying

adaptive changes of the oropharynx to variation in

mandibular position in patients with obstructive sleep

apnoea. J Oral Rehabil 1996;23:699–711.

[58] Liu Y, Zeng X, Fu M, Huang X, Lowe AA. Effects of a

mandibular repositioner on obstructive sleep apnea.

Am J Orthod Dentofacial Orthop 2000;118:248–56.

[59] Yoshida K. Prosthetic therapy for sleep apnea syn-

drome. J Prosthet Dent 1994;72:296–302.

[60] Petit FX, Pepin JL, Bettega G, Sadek H, Raphael B,

Levy P. Mandibular advancement devices: rate of con-

traindications in 100 consecutive obstructive sleep ap-

nea patients. Am J Respir Crit Care Med 2002;166:

274–8.

K.A. Ferguson / Clin Chest Med 24 (2003) 355–364364

Surgical management of obstructive sleep apnea

Kasey K. Li, MD, DDS*

Stanford University Sleep Disorders and Research Center, 401 Quarry Road, Stanford, CA 94305, USA

Despite the effectiveness of nasal continuous pos-

itive airway pressure (CPAP) in the management of

obstructive sleep apnea (OSA), patient acceptance and

tolerance remain a significant problem. Consequently,

surgery remains a highly desirable option for many

patients and should be considered. Several major

surgical advances have improved significantly the

understanding and treatment of OSA since the first

tracheotomy performed by Kuhlo [1] for the treatment

of upper airway obstruction in ‘‘Pickwickian’’ subject.

Uvulopalatopharyngoplasty (UPPP) was initially

described by Ikematsu [2] and later popularized by

Fujita [3]. UPPP improves oropharyngeal obstruction

and is the most commonly performed procedure for the

treatment of OSA. With the increased recognition of

hypopharyngeal airway obstruction as a major con-

tributing factor of OSA, genioglossus and hyoid

advancement were later developed [4,5] to improve

surgical treatment outcomes. In the early 1980s,

numerous investigators reported that surgical advance-

ment of the mandible can improve OSA [6–8]. To

maximize the extent of mandibular advancement,

concurrent maxillary advancement was subsequently

advocated [9]. Maxillomandibular advancement also

has been noted to widen the retropalatal airway, which

further improves the outcomes. Currently, UPPP, ge-

nioglossus and hyoid advancement, and maxilloman-

dibular advancement (MMA) are used widely to

improve upper airway obstruction in OSA. Of the

available surgical interventions, MMA has been

shown to have the highest success rate [9–11]. Several

years ago, radiofrequency (RF) energy was investi-

gated as a potential treatment of OSA by ablation of the

excessive upper airway tissues [12–15]. Based on the

initial animal study and subsequent human clinical

trials, RF has been shown to improve OSA [15,16].

Clinical evaluation

Before embarking on any surgical procedure, a

thorough head and neck evaluation combined with

fiberoptic pharyngolaryngoscopy is performed to iso-

late and direct treatment at the region or regions of

obstruction. A lateral cephalometric radiograph also is

used to assist in treatment planning. Although ceph-

alometric radiography is only a static two-dimensional

method of evaluating a dynamic three-dimensional

area, it does provide useful information on the pos-

terior airway space. The posterior airway space mea-

surement on lateral cephalometric radiography has

been shown to correlate with the volume of hypo-

pharyngeal airway on three-dimensional CT scans

[17]. It also is a valuable study for assessing the

relation of the maxillofacial skeleton and the hyoid

bone to the airway. Based on the evaluations, the sites

of airway obstruction are identified and a surgical plan

is formulated based on the severity of the anatomic

obstruction, the severity of sleep apnea, and—more

importantly—the patient’s desire and health status.

Oropharyngeal surgery

Uvulopalatopharyngoplasty is an effective sur-

gical procedure to improve airway obstruction in

the oropharynx. UPPP consists of the removal of a

portion of the soft palate and uvula and a limited

amount of the lateral pharyngeal wall and tonsillar

0272-5231/03/$ – see front matter D 2003, Elsevier Inc. All rights reserved.

doi:10.1016/S0272-5231(03)00016-9

* 750 Welch Road, Suite 317, Palo Alto, CA 94304.

E-mail address: kaseyli@hotmail.com

Clin Chest Med 24 (2003) 365–370

tissues (if present). The most crucial aspect of the

operation lies not in the amount of the tissue removal

but rather in the way the wound is sutured to widen

the airway aperture. The temptation to remove an

excessive amount of the tissues should be resisted

because the risk of complications dramatically

increases. At the authors’ center, the uvulopalatal flap

[18] is the preferred procedure as opposed to the

conventional UPPP. The uvulopalatal flap achieves

similar result as UPPP but removes less tissue, which

potentially reduces the risk of complications. In

general, the success rate of UPPP as the sole treat-

ment of OSA is only approximately 40% [19]

because many patients have hypopharyngeal and

oropharyngeal obstruction. Improvement of the oro-

pharyngeal airway alone is thus insufficient.

Hypopharyngeal surgery

The hypopharyngeal airway is intimately related

to the mandible, tongue, and hyoid complex [20,21].

It has been shown that advancing the genioglossus

muscle improves the tension of the genioglossus

muscle and decreases its collapsibility during sleep,

which alleviates airway obstruction. Obstruction at

the hypopharyngeal level can be improved further by

anterior movement of the hyoid bone, and numerous

reports have supported the concept that surgical

intervention at the hyoid level improves the hypo-

pharyngeal airway [22,23].

Initially, advancement of the genioglossus muscle

and the hyoid bone was performed simultaneously to

improve the hypopharyngeal airway [4]. The tech-

nique has evolved over the years to improve outcome

and minimize morbidity. The current technique of

advancement of genioglossus muscle involves a lim-

ited osteotomy intraorally to isolate and advance the

genial tubercle [24]. We have not performed both

operations simultaneously on a routine basis, how-

ever, because most patients with OSA have diffused

airway obstruction, and genioglossus muscle

advancement is generally combined with UPPP. The

added insult to the infrahyoid region by combining the

genioglossus muscle and hyoid bone advancement

results in increased edema and was believed to be

inappropriate in some patients. The authors also have

Fig. 1. Lateral cephalometric radiograph before maxillo-

mandibular advancement.

Fig. 2. Lateral cephalometric radiograph after maxilloman-

dibular advancement.

K.K. Li / Clin Chest Med 24 (2003) 365–370366

found that the hypopharyngeal airway obstruction is

resolved with only genioglossus muscle advancement

in some patients; thus hyoid bone may not always be

necessary. In some elderly patients ( > 60 years old),

airway edema after simultaneous genioglossus muscle

and hyoid bone can result in prolonged dysphagia that

may require days to recover. For these reasons the

authors perform hyoid bone advancement only in

some patients as a separate surgical step.

Maxillomandibular advancement has been shown

to be the most effective surgical option in the treat-

ment of OSA [9–11]. MMA achieves enlargement of

the pharyngeal and hypopharyngeal airway by phys-

ically expanding the skeletal framework. The forward

movement of the maxillomandibular complex also

improves the tension and collapsibility of the supra-

hyoid and velopharyngeal musculature. When MMA

is performed in patients with persistent OSA after

UPPP with genioglossus muscle and hyoid bone

advancement, MMA creates further tension and phys-

ical room in the upper airway, which relieves residual

obstructions. To maximize airway expansion, a major

advancement of the maxillomandibular complex is

required to facilitate a successful result (Figs. 1–6). It

is important, however, to achieve maximal advance-

ment while maintaining a stable dental occlusion and

a balanced esthetic appearance. Over the past

17 years, patients with and without ‘‘disproportion-

ate’’ craniomaxillofacial features have undergone

MMA for persistent severe OSA caused by incom-

plete response to other procedures. Although patients

with craniomaxillofacial abnormality, such as max-

illary or mandibular deficiencies, usually have

improved facial esthetics after surgery, the authors

found that many patients with normal cephalometric

measurements preoperatively also have an improved

facial appearance after MMA, because many patients

are middle-age adults who are already showing signs

of facial aging caused by soft tissue sagging. Skeletal

Fig. 3. Fiberoptic laryngoscopy demonstrating tongue base

obstruction before maxillomandibular advancement.

Fig. 4. Fiberoptic laryngoscopy demonstrating significant

lateral wall collapse duringMueller’s maneuver before maxil-

lomandibular advancement.

Fig. 5. Fiberoptic laryngoscopy demonstrating im-

proved tongue base obstruction after maxillomandibular

advancement.

Fig. 6. Fiberoptic laryngoscopy demonstrating improved

lateral wall collapse during Mueller’s maneuver after

maxillomandibular advancement.

K.K. Li / Clin Chest Med 24 (2003) 365–370 367

expansion of the maxilla and mandible enhances

appearance by improving soft tissue support.

Radiofrequency tissue reduction

Using temperature-controlled RF to reduce soft

tissue volume in the upper airway was first inves-

tigated in the animal tongue model [12]. After RF

treatment, tissue volume reduction results in a pre-

dictable pattern of wound healing, which consists of

coagulation necrosis that leads to fibrosis and tissue

contraction. The relationship of lesion size to total

RF energy delivery and the resultant volume reduc-

tion have been shown to be closely correlated, and

the application of RF to the human tongue in a serial

fashion was demonstrated to be the most effective

use of this technology in improving sleep-disordered

breathing (SDB) [15]. More importantly, the safety

parameters for temperature-controlled RF in the

human tongue were established in that speech and

swallowing were not affected based on barium

swallow, speech evaluation, and subjective question-

naires [15].

Oropharyngeal and hypopharyngeal

surgical outcomes

The authors’ surgical results were reported in

1992 [9]. Two hundred thirty-nine patients underwent

surgery, with most of the patients requiring interven-

tion at the pharyngeal and hypopharyngeal levels.

The overall cure rate was 61% (145/239 patients).

The surgical results were comparable to nasal CPAP

results. The mean preoperative respiratory distur-

bance index (RDI) was 48.3, with the postoperative

mean RDI of 9.5 (nasal CPAP RDI 7.2, P = NS). The

lowest oxygenation saturation (LSAT) improved from

75% to 86.6% (nasal CPAP LSAT 86.4%, P = NS).

There was a higher cure rate with mild to moderately

severe disease (approximately 70%) as compared

with severe disease (42%). Most of the nonrespond-

ers had severe OSA (mean RDI 61.9) and morbid

obesity (mean body mass index [BMI] 32.3 kg/m2).

The postoperative morbidity rate was low. The

mean hospital stay was 2.1 days. The complications

associated with genioglossus muscle and hyoid bone

advancement were infection ( < 2%), injury of tooth

roots that required root canal therapy ( < 1%), per-

manent paresthesia and anesthesia of the mandib-

ular incisors ( < 6%), and seroma ( < 2%). Major

complications, such as mandibular fracture, alteration

of speech, alteration of swallow, or aspiration, were

not encountered.

More than 350 patients underwent MMA with a

success rate of approximately 90%. An analysis of

175 patients who underwent MMA between 1988

and 1995 demonstrated that 166 patients had a

successful outcome, with a cure rate of 95%. The

mean preoperative RDI was 72.3. The mean post-

operative RDI was 7.2. The surgical results were

comparable to nasal CPAP results (nasal CPAP RDI

8.2, P = NS). The mean LSAT improved from 64%

to 86.7% (nasal CPAP LSAT 87.5%, P = NS).

86 patients who failed UPPP and genioglossus

muscle/hyoid bone advancement underwent MMA.

The mean age of patients was 43.5 years. The cure

rate in this group was 97% (83/86 patients). The

mean hospital stay for MMA was 2.4 days. The

surgical morbidity included transient anesthesia of

the lower lip, chin, and cheek in all of the patients.

There was an 87% resolution rate between 6 and

12 months. There was no postoperative bleeding or

infection. Mild malocclusion encountered in some

patients was treated satisfactorily with dental occlusal

adjustment. No major skeletal relapse occurred.

To date, 59 patients (49 men) have had long-term

follow-up results [25]. The mean age was 47.1 years.

The mean BMI was 31.1 kg/m2. 19 patients had only

subjective (quality of life) results. These patients

refused long-term polysomnography for various reas-

Table 1

Polysomnography results

Parameter Baseline Posttreatment Follow-up P valuea

RDI 39.5 F 32.7 17.8 F 15.6 28.7 F 29.4 0.29

Apnea index 22.1 F 33 4.1 F 6.2 5.4 F 10.3 0.88

Hypopnea index 17.4 F 11.9 13.6 F 11.5 22.9 F 23.1 0.20

Total sleep time (min) 337 F 89 346 F 75 337 F 97 0.66

Sleep efficiency index (%) 80 F 10 80 F 10 80 F 10 0.80

Oxygen saturation nadir (%) 81.9 F 11.6 88.1 F 5.3 85.8 F 6.6 0.18

REM sleep (%) 11.4 F 7.5 17.6 F 8.9 14.5 F 7.8 0.16

a Paired student’s t tests were performed on the change scores between posttreatment and follow-up.

K.K. Li / Clin Chest Med 24 (2003) 365–370368

ons, including inconvenience, time, and cost. Sixteen

of the 19 patients continued to report subjective

success with minimal to no snoring, no observed

apnea, and no recurrence of excessive daytime

sleepiness. All patients reported stable (unchanged)

weight to mild weight gain ( < 5 kg). Three patients

reported recurrence of snoring and excessive daytime

sleepiness. Long-term polysomnography data were

available in 40 patients (33 men). The mean age was

45.6 years. The mean BMI was 31.4 kg/m2. The

preoperative RDI and LSAT were 71.2 and 67.5,

respectively. The 6-month postoperative RDI was

9.3, and the LSAT was 85.6. The mean follow-up

period was 50.7 months, and long-term RDI and LSAT

were 7.6 and 86.3, respectively. The mean weight at

the long-term follow-up was 32.2 kg/m2 (P = 0.002).

4 patients had recurrent OSA. The 6-month postoper-

ative RDI in these 4 patients was 10.5, but the long-

term RDI (61 F 24.7 months) was 43. The LSAT

decreased from 87.5% to 81.8%.

Radiofrequency treatment outcomes

The initial RF tongue base reduction study con-

sisted of 18 patients (17 men). All had the diagnosis

of SDB and reported symptoms of daytime sleepi-

ness. The mean age was 44.9 F 8.7 years. The mean

pretreatment BMI was 30.2 F 5.5 kg/m2, and the

mean posttreatment BMI was unchanged at 30.2 F5.8 kg/m2 [15].

All of the patients had serial RF tongue base

reduction under local anesthesia to minimize risks.

The mean number of treatment sessions was 5.5 per

patient. The mean overall total number of joules

administered per patient was 8490 F 2687 J with

1543 J per treatment session. The mean duration from

the completion of treatment to the final PSGwas 2.6F0.7 months. The mean RDI improved from 39.5 F32.7 to 17.8 F 15.6 (P = 0.003). The mean apnea

index improved from 22.1 F 33.0 to 4.1 F 6.2

(P = 0.023), and the mean hypopnea index improved

from 17.4 F 11.9 to 13.6 F 11.5 (P = 0.326). The

mean LSAT improved from 81.9F 11.6 to 88.1F 5.3

(P = 0.03). The mean Epworth Sleepiness Scale

improved from 10.4 F 5.6 to 4.1 F 3.2 (P = 0.0001),

and the speech and swallowing visual analog scale did

not change from baseline.

Sixteen of the original 18 patients completed a

long-term follow-up study [16]. 2 patients (both

men) were lost to follow-up. The mean follow-up

period was 28 F 4 months. There was a mean

weight increase of 3.1 F 7.9 kg. The follow-up PSG

data showed a persistent improvement of the mean

apnea index; however, there was a trend of worsen-

ing hypopnea index, which resulted in a trend of

worsening RDI (Table 1). There was also a trend of

worsening LSAT.

Table 2

Short Form 36 scores

Domain Posttreatment Follow-up Mean change P valuea

Physical functioning 91 F 13.08 92 F 15.67 1 F 20.79 0.44

Role-physical 95 F 10.54 92.5 F 23.72 � 2.5 F 27.51 0.61

Bodily pain 87.3 F 18.37 80.7 F 19.31 � 6.60 F 27.58 0.77

General health 74.6 F 16.53 79.1 F 11.59 4.5 F 13.01 0.15

Vitality 60 F 23.57 71 F 13.5 11 F 17.76 0.05

Social functioning 81.3 F 20.58 92.5 F 16.87 11.2 F 15.91 0.03

Role emotional 86.6 F 28.25 96.7 F 10.44 10.1 F 31.71 0.17

Mental health 76 F 13.73 82 F 7.83 6 F 15 0.12

Physical component 54 F 4.08 52.39 F 7.89 � 1.61 F 9.5 0.69

Mental component 48.99 F 8.34 54.73 F 4.06 5.74 F 8.14 0.03

a Paired student’s t tests were performed on the change scores.

Table 3

Questionnaire visual analog scale results

Parameter Baseline Posttreatment Follow-up P valuea

Epworth Sleepiness Scale 10.4 F 5.7 4.1 F 3.2 4.5 F 3.4 1

Snoring 4.7 F 3.5 2 F 1.4 3.5 F 2.7 0.01

Speech 1.2 F 1.9 0.6 F 1.1 2.5 F 2.9 0.02

Swallowing 1.1 F 1.9 0.3 F 0.5 1.3 F 2.2 0.09

a Paired student’s t tests were performed on the change scores between posttreatment and follow-up.

K.K. Li / Clin Chest Med 24 (2003) 365–370 369

The quality-of-life measurements by Short Form

36 (Table 2) and excessive daytime sleepiness by the

Epworth Sleepiness Scale (Table 3) demonstrated

persistent improvement compared with baseline, and

no differences were found compared with posttreat-

ment results. Although no changes in swallowing or

speech were reported, the visual analog scale mea-

surement did increase significantly (see Table 3).

Summary

Nasal CPAP is and should be the first-line treat-

ment for OSA. Any physician who uses nasal CPAP

undoubtedly recognizes that this treatment modality

has limitations, however. The authors believe that

surgery offers a viable alternative to nasal CPAP in

patients who are intolerant of nasal CPAP. Potential

risks and complications must be explained fully to any

potential surgical candidate. The selection of surgical

procedure(s) should be determined based on a

patient’s airway anatomy, medical status, severity of

sleep apnea, and his or her desire and preference.

References

[1] Kuhlo W, Doll E, Frank MD. Erfolgreiche behandlung

eines Pickwick-syndroms durch eine dauertrachealka-

nule. Dtsch Med Wochenschr 1969;94:1286–90.

[2] Ikematsu T. Study of snoring. 4th report. J Jpn Otol

Rhinol Laryngol Soc 1964;64:434–5.

[3] Fijita S, Conway W, Zorick F, Roth T. Surgical correc-

tion of anatomic abnormalities of obstructive sleep

apnea syndrome: uvulopalatopharyngoplasty. Oto-

laryngol Head Neck Surg 1981;89:923–34.

[4] Riley RW, Guilleminault C, Powell NB, Derman S.

Mandibular osteotomy and hyoid bone advancement

for obstructive sleep apnea: a case report. Sleep 1984;

7:79–82.

[5] Riley RW, Powell NB, Guilleminault C. Obstructive

sleep apnea and the hyoid: a revised surgical proce-

dure. Otolaryngol Head Neck Surg 1994;111:717–21.

[6] Powell NB, Guilleminault C, Riley RW. Mandibular

advancement and obstructive sleep apnea syndrome.

Bull Eur Physiopathol Respir 1983;19:607–10.

[7] Bear SE, Priest JH. Sleep apnea syndrome: correction

with surgical advancement of the mandible. J Oral

Surg 1980;38:543–9.

[8] Kuo PC, West RA, Bloomquist DS, et al. The effect of

mandibular osteotomy in three patients with hypersom-

nia and sleep apnea. Oral Surg Oral Med Oral Pathol

1979;48:385–92.

[9] Riley RW, Powell NB, Guilleminault C. Obstructive

sleep apnea syndrome: a review of 306 consecutively

treated surgical patients. Otolaryngol Head Neck Surg

1993;108:117–25.

[10] Li KK, Riley RW, Powell NB, Gervacio L, Troell RJ,

Guilleminault C. Obstructive sleep apnea surgery: pa-

tients’ perspective and polysomnographic results.

Otolaryngol Head Neck Surg 2000;123:572–5.

[11] Waite PD, Wooten V. Maxillomandibular advancement

surgery in 23 patients with obstructive sleep apnea

syndrome. J Oral Maxillofac Surg 1989;47:1256–61.

[12] Powell NB, Riley RW, Troell RJ, et al. Radiofrequency

volumetric reduction of the tongue. Chest 1997;111:

1348–55.

[13] Powell NB, Riley RW, Troell RJ, et al. Radiofrequency

volumetric tissue reduction of the palate in subjects

with sleep-disordered breathing. Chest 1998;113:

1163–74.

[14] Li KK, Powell NB, Riley RW, et al. Radiofrequency

volumetric reduction for treatment of turbinate hyper-

trophy: a pilot study. Otolaryngol Head Neck Surg

1998;119:569–73.

[15] Powell NB, Riley RW, Guilleminault C. Radiofre-

quency tongue base reduction in sleep-disordered

breathing. Otolaryngol Head Neck Surg 1999;120:

656–64.

[16] Li KK, Powell NB, Riley RW, Guilleminault C. Tem-

perature-controlled radiofrequency tongue base re-

duction for sleep-disordered breathing: long-term

outcomes. Laryngoscope 2002;127:230–4.

[17] Riley RW, Powell NB. Maxillofacial surgery and ob-

structive sleep apnea syndrome. Otolaryngol Clin

North Am 1990;23:809–26.

[18] Powell N, Riley R, Guilleminault C, et al. A reversible

uvulopalatal flap for snoring and sleep apnea syn-

drome. Sleep 1996;19:593–9.

[19] Sher AE, Schechtman KB, Piccirillo JF. The efficacy

of surgical modifications of the upper airway in adults

with obstructive sleep apnea syndrome. Sleep 1996;

19:156–77.

[20] Lowe A, Gionhaku N, Tadeuchi K, et al. Three dimen-

sional reconstructions of the tongue and airway in adult

subjects with obstructive sleep apnea. Am J Orthod

1986;90:364–74.

[21] Rojewski TE, Schuller DE, Clark RW, et al. Video-

endoscopic determination of the mechanism of

obstruction in obstructive sleep apnea. Otolaryngol

Head Neck Surg 1984;92:127–31.

[22] Van de Graaf WB, Gottfried SB, Mitra J, et al. Respi-

ratory function of hyoid muscles and hyoid arch. J Appl

Physiol 1984;57:197–204.

[23] Patton TJ, Thawley SE, Water RC, et al. Expansion

hyoid-plasty: a potential surgical procedure designed

for selected patients with obstructive sleep apnea syn-

drome. Experimental canine results. Laryngoscope

1983;93:1387–96.

[24] Li KK, Riley RW, Powell NB, Troell RJ. Obstructive

sleep apnea surgery: genioglossus advancement revis-

ited. J Oral Maxillofac Surg 2001;59:1181–4.

[25] Riley RW, Powell NB, Li KK, Troell RJ, Guilleminault

C. Surgery and obstructive sleep apnea: long-term clin-

ical outcomes. Otolaryngol Head Neck Surg 2000;

122:415–21.

K.K. Li / Clin Chest Med 24 (2003) 365–370370