ilure 9 (2007) 820–826www.elsevier.com/locate/ejheart

European Journal of Heart Fa

eurjhD

ownloaded from

Influence of cardiac resynchronisation therapy ondifferent types of sleep disordered breathing

Olaf Oldenburg a,⁎, Lothar Faber a, Jürgen Vogt a, Anja Dorszewski a,Florian Szabados b, Dieter Horstkotte a, Barbara Lamp a

a Department of Cardiology, Heart and Diabetes Center North Rhine-Westphalia, Ruhr University Bochum,Georgstrasse 11, D-32545 Bad Oeynhausen, Germany

b Institute of Laboratory and Transfusion Medicine, Heart and Diabetes Center North Rhine-Westphalia, Ruhr University Bochum,Georgstrasse 11, D-32545 Bad Oeynhausen, Germany

Received 10 August 2006; received in revised form 16 February 2007; accepted 22 March 2007Available online 27 April 2007

at Ruhr U

niversitaet Bochum

on f.oxfordjournals.org

Abstract

Aims: This study investigates the influence of cardiac resynchronisation therapy (CRT) on sleep disordered breathing (SDB) in patients withsevere heart failure (HF).Methods and results: Seventy-seven patients with HF (19 females; 62.6±10 years) eligible for CRT were screened for presence, type, andseverity of SDB before and after CRT initiation (5.3±3 months) using cardiorespiratory polygraphy. NYHA class, frequency of nycturia,cardiopulmonary exercise, 6-minute walking test results, and echocardiography parameters were obtained at baseline and follow-up.

Central sleep apnoea (CSA) was documented in 36 (47%), obstructive sleep apnoea (OSA) in 26 (34%), and no SDB in 15 (19%) patients. CRTimproved clinical and haemodynamic parameters. SDB parameters improved in CSA patients only (apnoea hypopnoea index: 31.2±15.5 to 17.3±13.7/h, pb0.001; SaO2min: 81.8±6.6 to 84.8±3.3%, p=0.02, desaturation: 6.5±2.3 to 5.5±0.8%, p=0.004). Daytime capillary pCO2 wassignificantly lower in CSA patients compared to those without SDB with a trend towards increase with CRT (35.5±4.2 to 37.9±5.7 mm Hg, ns).After classifying short term clinical and haemodynamic CRT effects, improved SDB parameters in CSA occurred in responders only.Conclusions: In patients with severe HF eligible for CRT, CSA is common and can be influenced by CRT, this improvement depends ongood clinical and haemodynamic response to CRT.© 2007 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

Jan

uary 17,

Keywords: Heart failure; Cardiac resynchronisation therapy; Sleep disordered breathing; Central sleep apnoea

2011

1. Introduction

Sleep disordered breathing (SDB) is common in patientswith heart failure (HF) and is associated with a further increasein mortality in these patients [1–5].

Continuous positive airway pressure (CPAP) ventilationin patients with obstructive (OSA) as well as central sleepapnoea (CSA) had positive effects on cardiovascular param-eters in various studies [6–9], however a recent controlled

⁎ Corresponding author. Tel.: +49 5731 97 1258; fax: +49 5731 97 2194.E-mail address: akohlstaedt@hdz-nrw.de (O. Oldenburg).

1388-9842/$ - see front matter © 2007 European Society of Cardiology. Publishdoi:10.1016/j.ejheart.2007.03.009

multicentre trial failed to demonstrate a positive effect onmortality in HF patients with CSA [10]. Atrial pacing hasalso been suggested to improve SDB in patients with brady-cardia [11], although this hypothesis has not been supportedby recent studies [12,13].

Cardiac resynchronisation therapy (CRT) has been shown toimprove haemodynamics, functional parameters, and mortalityin patients with advanced congestive heart failure (NYHA≥III) and a wide QRS complex [14,15]. Sinha et al. reported abeneficial effect of CRT on CSA and Cheyne–Stokes res-piration (CSR) in a total of 24 patients with chronic HF [16].However, the effects of CRT on OSA and patients withoutsignificant SDB were not investigated.

ed by Elsevier B.V. All rights reserved.

Table 2Baseline patient characteristics and medication

All patients(n=77)

CSA(n=36)

OSA(n=26)

noSDB(n=15)

pvalue

SexMale, n 58 (75%) 29 (81%) 21 (81%) 8 (53%) NSFemale, n 19 (25%) 7 (19%) 5 (19%) 7 (47%) NS

Age, years 62.6±9.6 63.6±9.1 64.5±7.9 56.9±11.5 NSWeight, kg 81.0±17.4 81.6±21.2 82.6±15.2 76.8±9.1 NSHeight, cm 170.5±

17.0171.3±18.7

169.0±17.4

170.7±11.2

NS

AetiologyICM, n 33 (43%) 17 (47%) 13 (50%) 3 (20%) NSDCM, n 41 (53%) 17 (47%) 13 (50%) 11(73%) NSICM+VCM, n 3 (4%) 2 (6%) – 1 (7%) NS

Diabetes, n 25 (33%) 14 (39%) 9 (35%) 2 (13%) NSRhythm

Sinus rhythm 62 (80%) 27 (75%) 21 (81%) 14 (93%) NSAfib 12 (16%) 7 (19%) 5 (19%) – NSPM 3 (4%) 2 (6%) – 1 (7%) NS

HF — medicationACE-inhibitors/AT1-blockers

76 (98%) 35 (97%) 26 (100%) 15 (100%) NS

β-Blockers 71 (92%) 35 (97%) 23 (88%) 13 (87%) NSDiuretics 74 (96%) 36 (100%) 24 (92%) 14 (93%) NSSpironolactone 57 (74%) 25 (69%) 22 (85%) 10 (67%) NSDigitalis 54 (70%) 27 (75%) 17 (65%) 10 (67%) NS

NYHA class 3.0±0.3 3.1±0.3 3.0±0.2 2.9±0.2 NS

CSA: central sleep apnoea, OSA: obstructive sleep apnoea, noSDB: no sleepdisordered breathing, ICM: ischaemic cardiomyopathy; DCM: dilated,nonischaemic cardiomyopathy; VCM: valvular cardiomyopathy; Afib: atrialfibrillation; PM: pacemaker; NS: not significant.

821O. Oldenburg et al. / European Journal of Heart Failure 9 (2007) 820–826

at Ruhr U

niversitaet eurjhf.oxfordjournals.org

Dow

nloaded from

This prospective trial was initiated to study the effects ofCRT on different types of SDB.

2. Methods

2.1. Patients

A total of 77 consecutive patients eligible for CRT weretested for the presence and type of SDB before CRT deviceimplantation (between January 2003 and January 2005) andafter a mean follow-up of 5.3±3.3 months. Data were col-lected in a registry. The study was conducted in accordancewith institutional guidelines, and all patients gave writteninformed consent. Centre-specific criteria for CRT have beenreported previously [17,18], in summary patients had to pres-ent with dyspnoea according to the NYHA class III or IV,a left bundle branch block (LBBB) with a QRS width of≥150 ms, a left ventricular enddiastolic diameter (LVEDD) of≥60mm, a left ventricular ejection fraction (LVEF) of≤35%,and a peak oxygen uptake (peak VO2) during standardisedcardiopulmonary exercise testing of ≤18 ml/kg/min. In ad-dition, during an initial testing of several LV-lead positions(posterolateral veins), RV-stimulation sites (apex vs. RVOT)and LV vs. biventricular pacing, pulse pressure as a surrogateparameter of haemodynamic acute response had to increase bymore than 10%. In some respects, these institutional criteria aremore conservative compared to international guidelines. Forexample, we chose a minimum QRS width of 150 ms toincrease the number of patients with asynchronous ventricularcontraction. In order to get a more reliable parameter fordyspnoea than just NYHA functional class, we included amaximum value of 18 ml/kg/min for peak VO2.

Table 1Scoring of CRT response in patients with CSA according to changes insymptoms, and haemodynamic and echocardiographic parameters

Parameter Change Score

NYHA class Improved≥1 class +1Unchanged 0Worsened≥1 class −1

Predicted VO2 peak ImprovedN15% +2Unchanged (±15%) 0WorsenedN15% −2

LVEDD Improved≥15% +2ImprovedN10% +1Unchanged (±10%) 0WorsenedN10% −1Worsened≥15% −2

LVEF ImprovedN5% +1Unchanged (±5%) 0WorsenedN5% −1

6-minute walk ImprovedN10% +1Unchanged (±10%) 0WorsenedN10% −1

Changes in continuous data are calculated as percent change compared to theindividual baseline value before CRT. LVEDD: left ventricular enddiastolicdiameter; LVEF: left ventricular ejection fraction; VO2 peak: peak oxygenconsumption during standardised cardiopulmonary exercise testing.

Bochum

on January 17, 2011

2.2. Tailored CRT

The method of “tailored CRT” was described before indetail by Vogt et al. [17,18]. In summary, before CRT deviceimplantation several LV-lead positions (posterolateral veins),RV-stimulations sites (apex vs. RVOT) and LV vs. biven-tricular pacing are tested in every patient to gain the optimalhaemodynamic response. In addition, AV-delay optimisationaccording to an optimal fusion of intrinsic conduction inpatients with sinus rhythm and stimulated excitation is per-formed. As a surrogate parameter, pulse pressure is used toverify haemodynamic acute response.

2.3. Cardiorespiratory polygraphy

Sleep studies were performed by in-hospital unattendedcardiorespiratory polygraphy (Embletta™, Embla, Amster-dam, The Netherlands) as described before [1]. In summary,nasal air flow, chest and abdominal effort, pulse oxymetry, andbody position were recorded continuously. Analyses wereperformed by Somnologica for Embletta software (Embla,Amsterdam, The Netherlands) and reviewed and corrected bytwo independent SDB specialists, not involved in furtherclinical treatment of these patients. Hypopnoea was defined asa ≥30% reduction in airflow in combination with a drop inoxygen saturation of at least 3%. Apnoea was defined as a

Table 3Symptoms, haemodynamic, echocardiographic, and laboratory parametersat baseline and follow-up

All patients(n=77)

CSA(n=36)

OSA(n=26)

NoSDB(n=15)

Follow-up,months

Months 5.3±3.3 5.0±2.6 5.8±4.0 5.1±3.7

NYHA class Pre 3.0±0.3 3.1±0.3 3.0±0.2 2.9±0.2CRT 2.2±0.6 2.4±0.6 2.0±0.4 2.2±0.6p pb0.001 pb0.001 pb0.001 pb0.001

Nycturia Pre 2.1±1.0 2.2±1.1 2.1±0.8 1.8±1.3CRT 1.4±0.9 1.3±0.8 1.4±0.7 1.7±1.3p pb0.001 pb0.001 pb0.001 NS

CPX duration,min

Pre 6.2±2.5 5.6±2.4† 6.6±2.6 7.0±2.7CRT 7.8±3.1 7.4±3.3 8.1±2.9 8.5±2.8p pb0.001 pb0.001 p=0.003 p=0.03

CPX workload,W

Pre 65.0±27.5 59.6±26.7 69.8±29.6 69.9±24.9CRT 77.4±28.2 72.0±30.2 82.2±28.9 81.9±20.5p pb0.001 p=0.003 p=0.001 pb0.05

CPX VO2 peak,ml/kg/min

Pre 12.8±3.5 11.9±3.2† 13.2±3.7 14.4±3.5CRT 15.1±4.1 14.1±4.6 16.0±0.7 16.1±3.5p pb0.001 p=0.006 pb0.001 p=0.05

CPX VO2-AT,ml/kg/min

Pre 11.0±2.7 10.4±2.3 11.5±3.6 11.9±1.8CRT 13.1±3.0 13.0±3.1 13.7±2.7 12.2±3.4p pb0.001 p=0.003 p=0.004 NS

CPX predictedVO2 peak, %

Pre 54.1±14.9 50.1±14.0 60.0±15.1 58.6±14.9CRT 63.3±18.1 58.4±16.7 70.0±17.9 68.5±18.6p pb0.001 p=0.006 pb0.001 p=0.01

6-minute walk,m

Pre 347±114 328±121 377±94 343±123CRT 414±106 390±108 443±125 423±47p pb0.001 p=0.001 p=0.01 p=0.02

LVEF, % Pre 25.5±5.9 25.2±6.1 26.3±5.7 24.9±5.9CRT 30.2±7.0 29.1±7.3 30.9±6.7 31.8±6.1p pb0.001 p=0.003 p=0.006 p=0.007

LVEDD, mm Pre 73.7±9.6 73.6±9.8 74.2±9.0 73.0±10.7CRT 70.0±10.6 72.1±11.1 69.1±9.2 66.6±11.0p pb0.001 NS p=0.001 p=0.007

CSA: central sleep apnoea, OSA: obstructive sleep apnoea, noSDB: no sleepdisordered breathing, CPX: standardised cardiopulmonary exercise testing;LVEF: left ventricular ejection fraction; LVEDD: left ventricular end-diastolic diameter; †pb0.05 vs. noSDB for baseline parameters; NS: notsignificant.

Table 4Sleep study and blood gas analysis results at baseline and follow-up

All patients(n=77)

CSA(n=36)

OSA(n=26)

NoSDB(n=15)

AHI Pre 21.2±17.0

31.2±15.5†,‡

18.2±13.3†

2.5±1.8

CRT 13.7±12.2 17.3±13.7 14.6±9.8 3.3±3.2p pb0.001 pb0.001 NS NS

Mean nocturnalSaO2, %

Pre 93.2±2.6 93.1±2.5 93.3±2.2 93.7±3.4CRT 94.1±2.1 93.8±2.2 93.2±2.0 95.4±1.7p NS NS NS NS

Minimal nocturnalSaO2, %

Pre 82.9±6.1 81.8±6.6† 81.9±5.4 87.3±3.8CRT 84.8±4.8 85.5±3.2 81.2±5.3 88.7±2.9p NS p=0.004 NS NS

Mean nocturnaldesaturation, %

Pre 5.8±1.9 6.5±2.3† 5.5±1.0 4.5±0.8CRT 5.3±0.9 5.3±0.8 5.6±1.0 4.6±0.9p NS p=0.003 NS NS

Maximal apnoeaduration, s

Pre 36.1±14.4 36.6±15.7 37.4±11.9CRT 32.0±12.8 30.0±12.8 35.8±12.9p NS NS NS

Maximalhypopnoeaduration, s

Pre 39.0±11.2 41.0±10.4 40.6±11.4CRT 35.7±10.5 35.4±7.8 39.3±12.3p NS p=0.008 NS

Daytime capillarypCO2, mm Hg

Pre 37.8±6.3 35.5±4.2† 39.3±8.9 40.3±4.1CRT 38.6±6.0 37.9±5.7 37.6±4.2 41.7±8.2p NS NS NS NS

†pb0.05 vs. noSDB for baseline parameters; ‡pb0.05 vs. OSA for baselineparameters; NS: not significant. AHI: apnoea hypopnoea index, CSA:central sleep apnoea, OSA: obstructive sleep apnoea, noSDB: no sleepdisordered breathing.

822 O. Oldenburg et al. / European Journal of Heart Failure 9 (2007) 820–826

at Ruhr U

niversitaet Bochum

on January 17, 2011eurjhf.oxfordjournals.org

Dow

nloaded from

cessation of airflow for ≥10 s, in case of CSA without anyabdominal or thoracic breathing efforts and in case ofOSAwithtypical efforts. Classification intoOSAorCSAwas done on thebasis of the predominant type of SDB. The apnoea hypopnoeaindex (AHI) describes the number of apnoea and hypopnoeaepisodes per hour sleeping time, and is an established markerof SDB severity. According to current recommendations,an AHI of N5/h was chosen as a pathological cut-off [19].Graduation of SDB was performed according to AHI values,SDBwas consideredmild if AHI was 6–14/h, moderate if AHIwas 16–29/h, and severe if AHI was at least 30/h.

2.4. Echocardiography

Two-dimensional echocardiography was performed toevaluate LV function in all patients. LVEF was determinedusing apical 4- and 2-chamber views and the Simpsonmethod.All recordings were performed on Vingmed/GE ultrasound

systems. The echocardiographers were blinded to the patient'ssleep study results.

2.5. Spiroergometry

Symptom-limited bicycle exercise testing with spirom-etry (CPX) was used to evaluate exercise tolerance, peak oxy-gen consumption, and oxygen consumption at the individualaerobic–anaerobic threshold (ZAN Ferraris, Germany). Exer-cise testing started with a workload of 0–10 W with a con-tinuous increase of 10 W every minute. Maximum workloadand total exercise time were recorded, predicted VO2 peak wascalculated automatically taking sex and age into account.

2.6. 6-minute walk testing

A standardised hallway 6-minute walk test was performedon days with no other strenuous activities or exercise testing.

2.7. Determination of haemodynamic response

Changes in NYHA class, CPX testing, echocardiographicparameters and 6-minute walking distance were obtained andscored as described in Table 1. A score of 1 or less wasconsidered to reflect no relevant improvement after CRT, ascore of 2 or 3 was considered to be a moderate improve-ment, and 4 or 5 to be a good response.

Fig. 1. Apnoea hypopnoea index (AHI) in patients with central sleep apnoea(CSA), obstructive sleep apnoea (OSA) or patients without significant sleepdisordered breathing (noSDB) at baseline and follow-up. Significant changesoccurred in CSA patients only ( pb0.001).

Fig. 3. Changes in daytime capillary pCO2 in patientswith central sleep apnoea(CSA) according to their response to cardiac resynchronisation therapy (CRT).

Table 5Sleep disordered breathing parameters in CRT patients with CSA accordingto their clinical and haemodynamic response

Response to CRT in CSA patients

None or mild(n=11)

Moderate(n=13)

Good(n=12)

823O. Oldenburg et al. / European Journal of Heart Failure 9 (2007) 820–826

at Ruhr U

niversitaet Bochum

on Janueurjhf.oxfordjournals.org

Dow

nloaded from

2.8. Statistics

Continuous data are expressed as the mean value±SD.Statistical analyses were performed with SigmaStat™ soft-ware (SPSS Inc., Chicago, Illinois, USA). In continuous data,paired t-tests orWilcoxon signed-rank tests were used to checkfor differences before and after treatment. Analysis of dif-ferenceswithin the CSA groupwas done byANOVA. The chi-square test was used for nominal variables. Avalue of pb0.05was considered significant for all comparisons.

3. Results

Patients' characteristics are given in Table 2. There were nostatistically significant differences in the baseline character-istics of patients with CSA, OSA, or those without SDB:Patients without SDB tended to be younger, more frequentlyfemale, with lower body weight, were less often diabetic, and

Fig. 2. Changes in apnoea hypopnoea index (AHI) in patients with centralsleep apnoea (CSA) according to their response to cardiac resynchronisationtherapy (CRT).

presented less often with atrial fibrillation. Mean follow-upwas 5.3±3.3 months (range: 3–9months) after CRT initiation,without any statistically significant difference between thegroups (Table 3). CRT significantly improved almost all heartfailure parameters within this period of time. Patients withCSA presented with a more impaired cardiopulmonary func-tion, seen especially in CPX duration and peak oxygen con-sumption, when compared with patients without SDB.

Significant changes in SDB parameters only occurred inpatients with preexisting CSA (Table 4). There was a sub-stantial decrease in total apnoea hypopnoea episodes (Fig. 1),

AHI Pre 33.3±19.8 27.4±11.5 34.1±16.6CRT 27.0±18.1 13.7±10.0 12.4±8.2p NS p=0.001 pb0.001

Mean nocturnalSaO2, %

Pre 92.9±2.4 93.5±2.3 92.5±3.1CRT 92.6±2.2 94.5±2.2 94.7±1.6p NS NS NS

Minimum nocturnalSaO2, %

Pre 82.8±5.1 82.2±5.0 80.0±9.8CRT 84.2±3.5 85.9±3.5 86.9±2.5p NS pb0.05 pb0.05

Mean nocturnaldesaturation, %

Pre 5.8±1.4 6.2±1.4 7.5±3.8CRT 5.6±1.0 5.1±0.6 5.0±0.8p NS p=0.009 p=0.03

Maximum apnoeaduration, s

Pre 30.1±13.8 37.6±9.4 38.0±20CRT 29.1±12.5 32.5±10.7 25.7±12.3p NS NS NS

Maximum hypopnoeaduration, s

Pre 41.4±9.1 42.8±13.6 38.7±8.0CRT 36.5±6.6 34.0±7.6 33.7±4.0p NS NS NS

There is no statistically significant difference in the magnitude of changes inAHI, minimum nocturnal SaO2, or mean nocturnal desaturation betweenpatients with moderate or good response to CRT. AHI: apnoea hypopnoeaindex; NS: not significant.

ary 17, 2011

824 O. Oldenburg et al. / European Journal of Heart Failure 9 (2007) 820–826

at Ruhr U

niversitaet Bochum

on January 17, 2011eurjhf.oxfordjournals.org

Dow

nloaded from

as well as in maximum apnoea and hypopnoea duration,and an increase in minimal oxygen saturation. Baseline AHIin CSA was higher than in OSA, both, per definition, werehigher than in patients without SDB. Daytime pCO2 con-centration tended to be lowest in CSA patients, reachingsignificance only when compared to patients without SDB.

Further analysis of changes in SDB parameters in CSApatients revealed a positive influence depending on clinicaland haemodynamic response to CRT. AHI improved only inpatients with moderate or good response to CRT (Fig. 2) anddaytime pCO2 concentration increased only in patients withgood response to CRT (Fig. 3). In addition, minimum noc-turnal oxygen saturation and oxygen desaturation improvedonly in moderate or good CRT responders (Table 5).

4. Discussion

This study confirms a high prevalence of SDB in HFpatients eligible for CRT. Potential positive effects of CRTonSDB were seen in HF patients with preexisting CSA, but notin those with OSA, and were limited to those with a goodclinical and haemodynamic response to CRT. Patients with-out preexisting SDB did not show a substantial change intheir sleep study results.

Recently, Sinha et al. reported a beneficial effect of CRTon CSA and CSR in patients with chronic heart failure [16].Definition of CSA and the cut-off AHI of 5/h were com-parable to ours and as recommended by guidelines [19]. Intheir study, CSA and CSR were found in 14 of 24 (58%)consecutive patients, in whom CRT was initiated. The prev-alence of CSA in the present study was 47% which is lessthan Sinha reported, but still a high number compared topublished prevalence data. In a previous prevalence study,we found SDB to be present in 76% of patients with symp-tomatic heart failure (LVEF≤40%, NYHA class≥ II), 36%had OSA and 40% CSA [1]. Javaheri et al. reported a CSAprevalence of 40% in a cohort of 81 outpatients with stableheart failure [2]. In a retrospective analysis of 450 patientswith HF, using an AHI cut-off of 10/h, Sin et al. found SDBto be present in 72% of all patients, 33% presenting withCSA and 38% presenting with OSA [3]. Assuming that CSAand CSR reflect heart failure severity at least to some extent,a higher prevalence can be expected in patients qualifying forCRT, because heart failure is usually more advanced in thesepatients.

Sinha et al. demonstrated an improvement in AHI inevery patient with CSA after CRT (n=14) after a follow-upof 17±7 weeks. Patients without SDB at CRT initiation didnot develop SDB during follow-up, however further analysison OSA is not presented in this study [16]. The present studyallows further insights into SDB and CRT. Tailored CRT inour patients led to an excellent improvement in cardiacfunction after a mean follow-up of 5.3 months. However,recovery of cardiac function is not homogeneous, and, atleast in the present study, improvement of SDB did notmanifest in every patient.

First, improvement in SDB was seen in the group of CSApatients only. There was no significant change in AHI, meanor minimal nocturnal oxygen saturation, maximal apnoea orhypopnoea duration in OSA patients. In accordance withSinha et al. [16], patients without SDB at baseline did notdevelop SDB after CRT. Second, improvement in CSA wasdependent on clinical response to CRT. For this purpose, wemodified a scoring system introduced by Packer [20]. Thisscore is not designed to verify reverse remodelling, but isintroduced to capture clinical, morphological, and functionaldata during CRT. Considerable effects on SDB parameterswere observed only in CSA patients with a moderate or goodresponse to CRT. Exclusively in these patients AHI, min-imum nocturnal oxygen saturation, and desaturation im-proved, and daytime capillary pCO2 increased.

In a recently published study, Gabor et al. screened 28patients eligible for CRT for the presence and type of SDB[21]. By using an AHI cut-off value of 10/h, no relevant SDBwas found in 12 patients (43%), OSA in 4 patients (14%),and CSA in 12 patients (43%). They were able to follow 10of these 12 patients with CSA over a period of 27±7 weeksof CRT. In this small cohort of CSA patients, the AHIdecreased from 42.7±9.1 at baseline to 30.8/h±18.7/h ( pb0.05) at follow-up. This decrease in AHI was restricted to 6of these patients in whom central AHI decreased; the fourother patients did not experience a decrease. By analysingchanges in cardiac function, no significant improvement inLVEF, LV size, mitral regurgitation, and 6-minute walkingdistance was found in these four patients. The present studyand the data from Gabor indicate that an improvement inCSA under CRT depends on the haemodynamic response toCRT. However, this hypothesis still needs to be proven bycontrolled studies.

In an early study on the effect of atrial pacing on SDB,Garrigue et al. claimed a remarkable influence of atrial over-drive pacing on CSA and OSA without reduction in totalsleep time [11]. In 13 out of 15 patients with symptomaticsinus bradycardia, atrial pacing at a rate of 15 beats perminute faster than the mean nocturnal heart rate resulted in asignificant reduction in the number of episodes of all types ofapnoea. In the present study, heart rate before and during CRTwas not systematically investigated, but CRT pacing was notintended in an overdrive manner and recent studies have notconfirmed Garrigue's results [22,23]. In addition, Garriguedid not document significant effects of atrial overdrive pacingon OSA in a recent study performed in 17 unselected patientswith symptomatic bradyarrhythmia [13]. The results pre-sented by Garrigue might be divergent, but further analysismight resolve the problem. In the first study, Garrigue [11]showed a greater reduction in the apnoea index in CSApatients when compared to OSA patients. Like others [24],we favour the hypothesis that atrial overdrive pacing in pa-tients with symptomatic bradycardia led to an improvementin cardiac output. In terms of the present study this may alsoimply that these beneficial effects are missing in patientsshowing only mild or even no haemodynamic response to

825O. Oldenburg et al. / European Journal of Heart Failure 9 (2007) 820–826

at Ruhr U

niversitaet Bochum

on January 17, 2011eurjhf.oxfordjournals.org

Dow

nloaded from

CRT and therefore no effect on CSA can be observed. Mostlikely, the improvement of CSA is not specifically due to CRTbut to a general effect of improved heart failure.

Several experimental studies have demonstrated deleteri-ous effects of OSA [25–27] and CSA [28–30] on cardiacfunction and prognosis leading to or worsening HF. The pres-ent data further support the theory that CSA is also a markerof HF severity: The prevalence of CSA in our cohort of severeHF patients qualifying for CRT is high and can be reduced byimproving cardiac function.

Pulmonary congestion and stimulation of pulmonary vagalirritant J-receptors may be a consequence of severe HF [31].Experimental studies demonstrate hyperventilation with sub-sequent decrease in blood pCO2 as a consequence. In com-bination with an enhanced receptor sensitivity to CO2 [32], analtered apnoea threshold [33] at night and an increased cir-culatory delay [34], this may lead to CSA and CSR.Moreover,the hyperventilatory state may be maintained while patientsare awake and be expressed by a progressive ventilatory in-stability [29,35]. As a consequence, the presence of low pCO2

during the day should raise the suspicion for the presence ofCSR and CSA, which is clearly associated with a worseprognosis [5], and should result in an intensified treatment.

5. Limitations

Categorisation of haemodynamic response was basedon a novel scoring system not yet prospectively validated.Nevertheless, each parameter used in this scoring system hasbeen proven to be a prognostic marker or a reliable parame-ter of CHF progression, remodelling or reverse remodelling.Another limitation is that we prospectively followed ourCRT patients without calculating the statistical power neededto show results for patients without SDB, those with OSA, orCSA in advance. Therefore, as mentioned above, presentedeffects of CRT on SDB are indicative but not yet proven.

6. Conclusion

SDB is common in HF patients eligible for CRT. Bothtypes of SDB, OSA and CSA, may worsen HF and its prog-nosis, but CSA especially seems to be a marker of heartfailure severity. Improvement of CSA depends on a positiveclinical and haemodynamic response to CRT. OSA is notinfluenced by CRT.

References

[1] Oldenburg O, Töpfer V, Lamp B, Faber L, Teschler H, Horstkotte D.Sleep disordered breathing in patients with symptomatic heart failure.Eur J Heart Fail 2007;9:251–7.

[2] Javaheri S, Parker T, Liming JD, et al. Sleep apnea in 81 ambulatorymale patients with stable heart failure. Circulation 1998;97:2154–9.

[3] Sin D, Fitzgerald F, Parker J, Newton G, Floras J, Bradley T. Riskfactors for central and obstructive sleep apnea in 450 men and womenwith congestive heart failure. Am J Respir Crit Care Med 1999;160:1101–6.

[4] Lanfranchi P, Somers V, Braghiroli A, Corra U, Eleuteri E, GiannuzziP. Central sleep apnea in left ventricular dysfunction. Prevalence andimplications for arrhythmic risk. Circulation 2003;107:727–32.

[5] Hanly P, Zuberi-Khokhar N. Increased mortality associated withCheyne–Stokes respiration in patients with congestive heart failure.Am J Respir Crit Care Med 1996;153:272–6.

[6] Naughton M, Liu P, Benard D, Goldstein R, Bradley T. Treatment ofcongestive heart failure and Cheyne–Stokes respiration during sleepby continuous positive airway pressure. Am J Respir Crit Care Med1995;151:92–7.

[7] Sin D, Logan A, Fitzgerald F, Liu P, Bradley T. Effects of continuouspositive airway pressure on cardiovascular outcomes in heart failurepatients with and without Cheyne–Stokes respiration. Circulation2000;102:61–6.

[8] Kaye D, Mansfield D, Aggarwal A, Naughton M, Esler M. Acuteeffects of continuous positive airway pressure on cardiac sympathetictone in congestive heart failure. Circulation 2001;103:2336–8.

[9] Javaheri S. Effects of continuous positive airway pressure on sleep apneaand ventricular irritability in patients with heart failure. Circulation2000;101:392–7.

[10] Bradley T, LoganA,Kimoff R, et al. Continuous positive airway pressurefor central sleep apnea and heart failure. N Engl J Med 2005;353:2025–33.

[11] Garrigue S, Bordier P, Jais P, et al. Benefit of atrial pacing in sleepapnea syndrome. N Engl J Med 2002;346:404–12.

[12] Krahn A, Yee R, Erickson MK, et al. Physiologic pacing in patientswith obstructive sleep apnea. J Am Coll Cardiol 2006;47:379–83.

[13] Pepin J, Defaye P, Garrigue S, Poezevara Y, Levy P. Overdrive atrialpacing does not improve obstructive sleep apnea syndrome. Eur RespirJ 2005;25:343–7.

[14] Abraham WT, Fisher WG, Smith AL, et al. Cardiac resynchronizationin chronic heart failure. N Engl J Med 2002;346:1845–53.

[15] Cleland J, Daubert J, Erdmann E, et al. The effect of cardiac resyn-chronization on morbidity and mortality in heart failure. N Engl J Med2005;352:1539–49.

[16] Sinha AM, Skobel EC, Breithardt OA, et al. Cardiac resynchronizationtherapy improves central sleep apnea and Cheyne–Stokes respiration inpatients with chronic heart failure. J Am Coll Cardiol 2004;44: 68–71.

[17] Vogt J, Lamp B, Hansky B. The bad Oeynhausen experience. EurHeart J Suppl 2004;6:D122–3.

[18] Vogt J, Heintze J, Lamp B, Hansky B, Horstkotte D. Standardhaemodynamic measurements. Eur Heart J Suppl 2004;6:D29–34.

[19] Sleep-related breathing disorders in adults: recommendations forsyndrome definition and measurement. Techniques in clinical research.Sleep 1999;22:667–89.

[20] Packer M. Proposal for a new clinical end point to evaluate the efficacyof drugs and devices in the treatment of chronic heart failure. J CardiacFail 2001;7:176–82.

[21] Gabor JY, Newman DA, Barnard-Roberts V, et al. Improvement inCheyne–Stokes respiration following cardiac resynchronisation ther-apy. Eur Respir J 2005;26:95–100.

[22] Simantirakis EN, Schiza SE, Chrysostomakis SI, et al. Atrial overdrivepacing for obstructive sleep apnea–hypopnea syndrome. N Engl J Med2005;353:2568–77.

[23] Unterberg C, Lüthje L, Szych J, Vollmann D, Hasenfuß G, Andreas S.Atrial overdrive pacing compared to CPAP in patients with obstructivesleep apnoea syndrome. Eur Heart J 2005;26:2568–75.

[24] Floras JS, Bradley TD. Atrial overdrive pacing for sleep apnea. A doornow closed? Am J Respir Crit Care Med 2005;172:1–2.

[25] Alonso-Fernandez A, Garcia-Rio F, Arias M, et al. Obstructive sleepapnea–hypopnea syndrome reversibly depresses cardiac response toexercise. Eur Heart J 2006;27:207–15.

[26] Usui K, Parker J, Newton G, Floras J, Ryan C, Bradley T. Leftventricular structural adaptions to obstructive sleep apnea in dilatedcardiomyopathy. Am J Respir Crit Care Med 2006;173:1170–8.

[27] Gami A, Howard D, Olson E, Somers V. Day–night pattern of suddendeath in obstructive sleep apnea. N Engl J Med 2005;352:1206–14.

826 O. Oldenburg et al. / European Journal of Heart Failure 9 (2007) 820–826

[28] Lanfranchi P, Braghiroli A, Bosimini E, et al. Prognostic value ofnocturnal Cheyne–Stokes respiration in chronic heart failure. Circu-lation 1999;99:1435–40.

[29] Corra U, Pistono M, Mezzani A, et al. Sleep and exertional periodicbreathing in chronic heart failure. Prognostic importance and interde-pendence. Circulation 2006;113:44–50.

[30] Solin P, Kaye D, Little P, Bergin P, RichardsonM, NaughtonM. Impactof sleep apnea on sympathetic nervous system activity in heart failure.Chest 2003;123:1119–26.

[31] Solin P, Bergin P, Richardson M, Kaye D, Walters E, Naughton M.Influence of pulmonary capillary wedge pressure on central apnea inheart failure. Circulation 1999;99:1574–9.

[32] Javaheri S. A mechanism of central sleep apnea in patients with heartfailure. N Engl J Med 1999;341:949–54.

[33] Xie A, Skatrud J, Puleo D, Rahko P, Dempsey J. Apnea–hypopneathreshold for CO2 in patients with congestive heart failure. Am JRespir Crit Care Med 2002;165:1245–50.

[34] Tkacova R, Niroumand M, Lorenzi-Filho G, Bradley T. Overnightshift from obstructive to central apneas in patients with heart failure.Circulation 2001;103:238–43.

[35] Guazzi M, Reina G, Tumminello G, Guazzi M. Exercise ventilationinefficiency and cardiovascular mortality in heart failure: the criticalindependent prognostic value of the arterial CO2 partial pressure. EurHeart J 2005;26:472–80.

at Ruhr U

niversitaet Bochum

on January 17, 2011eurjhf.oxfordjournals.org

Dow

nloaded from