12/30/2016

1

Neurosonology in the NSICUSebina Bulic, MD, RVT

Medical Director Cerebral Blood Flow Laboratory Keck USC

Education Director Vascular Neurology Fellowship USC

Sebina.Bulic@med.USC.edu

Disclosures

• No disclosures to report

Outline

• Transcranial Doppler• Vasospasm Monitoring

• Increased ICP

• Brain Death

• Emboli Detection

• Carotid Duplex• POCUS

• Optic Nerve Sheath Diameter• POCUS

Transcranial Doppler

• Insonation Windows• Transtemporal• Transorbital• Transforaminal

Doppler Facts:

Fixed transmission frequency (2 MHz)

Pulsed -wave Doppler probe:

• A Single crystal sends out and receives signal

• Range-gated depth

• Flow direction capabilities

Summary of findingsSubarachnoid Hemorrhage (SAH):

INDICATION SENSITIVITY

(%)

SPECIFICITY

(%)

REFERENCE

STANDARD

Vasospasm after

Spontaneous

Subarachnoid

Hemorrhage

Conventional

angiography

Intracranial

ICA

25-30 83-91

MCA 39-94 70-100

ACA 13-71 65-100

VA 44-100 82-88

Copyright 2004 American Academy of Neurology5

Summary of findingsSubarachnoid Hemorrhage (SAH) (continued)

INDICATION SENSITIVITY

(%)

SPECIFICITY

(%)

REFERENCE

STANDARD

BA 77-100 42-79

PCA 48-60 78-87

Recommendations: TCD is useful for the detection and

monitoring of angiographic VSP in the basal segments of the

intracranial arteries, especially the MCA and BA, following

sSAH (Type A, Class I-II evidence).

More data are needed to show if TCD affects clinical

outcomes in this setting (Type U).

Copyright 2004 American Academy of Neurology6

12/30/2016

2

Management approach to delayed cerebral ischemia

Suarez, CONTINUUM: Lifelong Learning in Neurology. 21(5, Neurocritical Care):1263-1287, October 2015.

3.Transcranial Doppler is

reasonable to monitor for

the development of arterial

vasospasm (Class IIa; Level

of Evidence B)

Is There an Influence of Routine Daily Transcranial Doppler Examination on Clinical Outcome in Patients After Aneurysmal Subarachnoid Hemorrhage?

50 January 2013 to December2013 received daily TCD measurements; 39 January 2014 to September 2014 received no TCD

Ehrlich at al, World Neurosurg. (2016)

Ehrlich at al, World Neurosurg. (2016) Ehrlich at al, World Neurosurg. (2016)

• NIS database

• January 1, 2000 - December 31, 2011

• 303,061 inpatient admissions with a primary diagnosis of aSAH

• 4576 (1.5%) of these patients had TCDs

• The Charlson Comorbidity Index (CCI) was similar between the two groups

• In terms of discharge allocation, patients who had TCDs had significantly lower mortality, more routine discharge and home health care than those who did not have TCDs.

Kim at al, ASN 2017

• 56yo female who presented with 5 days of headache, angio negative a-SAH HH2/F2

• EVD was placed

• Vasospasm was monitored transcranial Doppler (TCD)

• On HD #3, patient was found to have elevated velocities on TCD in the setting of normal clinical exam. Right MCA 199cm/s, left MCA 208cm/s

• CTA showed diffuse vasospasm and severe basilar artery spasm

• DSA right PICA aneurysm

• VA-left PICA bypass was planned

• Limited treatment options, IT nicardipineX15 doses

• HH therapy was initiated 2 days after the bypass

• MRI did not show evidence of stroke

• She was discharged to acute rehab facility

Burgos at al, ASN 2017

12/30/2016

3

• Up to 70 of SAH patients develop angiographic vasoconstriction

• Only 20–30 % develop DCI

• Cerebral infarction sometimes develops in the absence of demonstrable vasoconstriction, or in a vascular territory unaffected by vasospasm

• Prevailing hypotheses • large-vessel narrowing with subsequent low flow

• early brain injury (EBI) (multiple physiological derangements that are thought to occur in the first 72 hours after the ictus. The initial ICP crisis and global hypoperfusion trigger glial activation, endothelial dysfunction, and inflammatory pathway

• microcirculatory dysfunction with loss of autoregulation

• cortical spreading depolarization (CSD)

• microthrombosisFrancoeur and Mayer Critical Care (2016) Francoeur and Mayer Critical Care (2016)

• Mean maximal TCD values during SAH days 3–14 in patients who did or did not develop DCI

• Histogram shows the number of patients with new onset DCI between SAH days 3-14.

• Number (in parentheses) represents the number of TCD examinations performed for each corresponding SAH day.

• Increased TCD flow velocities imply only a mild incremental risk of DCI after SAH, with maximal sensitivity by day 8.

Francoeur and Mayer Critical Care, 2016Carrera at al, Neurosurgery, 2009Kumar at al, J Neurosurg 2016

Kumar and Alexandrov, J UltrasoundMed 2015

TBI1. Primary Injury- direct damage due to mechanical forces including diffuse

axonal injury

2. Secondary Injury- cascade of cellular and molecular processes including damage due to hypotention, hypoglycemia, hypoxia, and high intracranial pressure

• Hallmark of severe traumatic Brain Injury

• Differential Movement of Adjacent regions of Brain during acceleration and Deceleration.

• DAI is major cause of prolonged COMA after TBI, probably due to disruption of Ascending Reticular connections to Cortex.

• Angular forces > Oblique/ SagitalForces

12/30/2016

4

Martin Neurotrauma 1997

Rovegno et al 2011

Traumatic Brain Injury and Raised Intracranial Pressure• TBI may lead to hypoperfusion (day 0), hyperemia (days 1–3), vasospasm

(days 4–15), and raised ICP

• TCD can avoid use of invasive CBF measurement techniques and provide similar prognostic information.

• MCA MFV of < 35 cm/s within 72 hours of head injury has been shown to predict unfavorable outcome at 6months

• However, on multivariate analysis, this association was significantly less with initial GCS being a stronger predictor of outcome

• 50 patients with TBI monitored in the first 7 days, the vasospasm and hyperemia groups experienced a poor outcome at 6 months (GOS 1–3)

• A significant correlation between PI and ICP (correlation coefficient 0.938 𝑃 < 0.0001 ) was demonstrated in a group of 81 patients who underwent TCD MCA PI measurements combined with invasive ICP measurements

• ICP = (11.1 X PI)−1.43, which could determine an ICP via the PI within 4.2mmHg of the actual ICP, which is reasonably accurate.

• ICP of > 20mmHg could also be determined with 89% sensitivity and 92% specificity

• PI ≥ 1.56 predicted 83% of patients who had a poor outcome at 6 months, whereas a PI ≤ 1 identified 71% of patients with a good outcome (GOS 4–5)

ICP Monitoring Techniques

• Types of Monitoring Devices• Fluid-filled transduced

ventriculostomy

• Fiberoptic sensors

• Microchips (internal strain-gauge devices)

• Air pouch technologies

• Locations of Monitoring Devices• Ventricular

• Parenchymal

• Subdural

Complications

• Infection

• Hemorrhage

• Breakage

• Malfunction of the device

• Difficulty with placement

• Difficult to assess the true rate of some of these issues• Variability of definitions used in

the literature when reporting infections and hemorrhage

Summary of findingsIncreased Intracranial Pressure (ICP) and Cerebral Circulatory Arrest

INDICATION SENSITIVITY

(%)

SPECIFICITY

(%)

REFERENCE

STANDARD

Cerebral

Circulatory Arrest

and Brain Death

91-100 97-100 Conventional

angiography, EEG,

clinical outcome

Recommendation: TCD is a useful adjunct test for the

evaluation of cerebral circulatory arrest associated with brain

death (Type A, Class II evidence).

Copyright 2004 American Academy of Neurology23

What is brain death?

• Case of Melanie Bacchiochi, Connecticut

• Transcripts of the hearing

“when the soul leaves the body”

“the loss of our ‘personal identity’”

“Loss of vital fluid flow”

“Putrefaction is the only sure sign of death”

“disintegration of individual organ and tissue”

12/30/2016

5

1968 Harvard Criteria of Brain Death

• Defined as the irreversible loss of function of the whole brain, including the brainstem.

• Brain death is a clinical diagnosis• 1) Coma/unresponsiveness• 2) Absence of all brainstem reflexes• 3) Apnea

• Exclusion• of complicating medical condition ( severe electrolyte, acid base or endocrine

disturbance)• drug intoxication or poisoning• hypothermia (core temperature > 35 C or 90 F

Pathophysiology of Brain Death

• Loss of respiratory function

• Hemodynamic instability• Cardiac arrhythmias

• Loss of vasomotor control

• Loss of temperature regulation

• Endocrine imbalances • ADH, T3/T4, insulin, cortisol

• Electrolyte disturbances

DI DIC

arrhythmias

pulmonary edemaacidosis hypothermia

hypotension

Consequences of Brain Death Brain Death

• Brain Death interval increased

%Donor decreased 57% -> 45%

%Cardiac Arrest 3.9% -> 7.4%

%Family decline 23% -> 36%

Organ Donation Statistics

Transplantable organs: heart, kidneys, liver, lungs, pancreas, small intestine

Transplantable tissues: blood, blood vessels, bones, bone marrow, cartilage, connective tissues, eyes, heart valves, skin

• 123,000 adults and children are on the waiting list

• 3% of all deaths are brain dead

• 21 people on average die each day waiting for an organ

• In 2013, 14,257 Donors resulted in 28,953 transplants

• One donor can save up to 9 lives. The same donor can save/improve the lives of hundreds by donating tissues.

• Success rate is now 80-90%

“Confirmatory” test for brain death

Brain death is a clinical diagnosis These tests are not mandatory & are options

• Cerebral angiography

• Electroencephalography (EEG)

• Transcranial Doppler Ultrasonography

• Technetium 99 brain scan

12/30/2016

6

• TCD patterns of CCA detected in 56 pts (9%)

• Anatomic locations by brain CT• 41 supra-infratentorial lesions• 13 supratentorial lesions • 2 infratentorial lesions.

• 26/56 showed hemodynamic patterns of BD at the first TCD examination

• 30 pts were evaluated by daily TCD until the detection of TCD patterns of BD

• 33 TCD examinations showed an oscillating flow pattern

• 23 examinations showed a systolic spikes patternMarinoni at al, Neurol Sci, 2011

TCD patterns diagnostic of irreversible CCA

• reverberating flow with retrograde flow in diastole

• short systolic spikes

• complete absence of any TCD signal.

Marinoni at al, Neurol Sci, 2011

• 28 ICP monitoring (mean ICP 83.9 mmHg ± 26,3; range:15–136 mmHg)• 17 (60%) had an intraparenchymal probe

• 8 (30%) an endoventricular probe

• 3 (10%) a subdural device.

• On comparing mean ICP values with the two TCD patterns of BD, no significant differences were found between pts with reverberating flow signal and those with systolic spike pattern.

• 91% and a specificity of 100%

• In 5 pts, false-negative TCD patterns were detected, showing diastolic flow in at least one artery• 1 was affected by carotid-cavernous fistula

• 4 were treated by therapeutical decompressive craniotomy.

Marinoni at al, Neurol Sci, 2011

Summary of findingsDetection of Cerebral Microemboli

INDICATION SENSITIVITY

(%)

SPECIFICITY

(%)

REFERENCE

STANDARD

Cerebral

Microembolization

Experimental model,

pathology, magnetic

resonance imaging,

neuropsychological

tests

Recommendation: TCD is probably useful to detect cerebral

microembolic signals in a wide variety of cardiovascular/

cerebrovascular disorders/procedures (Type B, Class II-IV

evidence).

However, data at present do not support the use of TCD for

diagnosis or for monitoring response to antithrombotic therapy in

ischemic cerebrovascular disease in these settings(Type U).Copyright 2004 American Academy of Neurology34

TCD Embolus DetectionClinical Applications in Ischemic Stroke

• Diagnosis• Cardioembolic sources

• Atrial fibrillation• Myocardial infarction• Prosthetic heart valves• Native valve disease

• Parodoxical sources – patent foramen ovale• Arterial sources

• Carotid atherosclerosis: asymptomatic/symptomatic• Dissection

• Treatment• Guide to intensity of antithrombotic therapy• Ultrasound to guide and enhance enzymatic fibrinolysis

Identification Criteria

• Transient (usually < 300 milliseconds)

• High amplitude (> 3 dB higher than background blood flow

• Unidirectional velocity spectrum

• Audible output – “chirp,” “snap, “moan”

--Consensus Committee of the Ninth International Cerebral Hemodynamics Symposium, Stroke 1996

. Characteristics of Artifacts

• Rough, broad, nonharmonic sound

• Bidirectional: signal intensity increases simultaneously in both directions

• Coincident with probe impacts, sudden motion, or electrical switching transients

• Appear simultaneously at two different sites

12/30/2016

7

Techniques

• Types of monitoring• Unilateral vs. bilateral

• Duration of monitoring• Minutes vs. hours

• Single gated vs multigated• Power Mmode(PMD)

Mechanical Heart Valves

• Generate microbubbles via microcavitation

• Micro aeroemboli of no pathogenic consequence

5 patients, 30 minutes monitoring

Median MES

Resting 9 (per 30 min)

100% O2 0 (per 30 min)

Mechanical Heart Valves

• More MES in those with a history suggestive of stroke

• More MES with valve obstruction

• Majority of MES are gaseous

• Overall not a good indicator of stroke risk

Left Ventricular Assist Devices

• MES varies from 20% to 100%

• Novacor N100• MES correlate with stroke risk

• Antiplatelet and anticoagulation reduce MES

• Reduction in MES correlated with stroke risk

• Debakey• MES not correlated with stroke risk

• Oxygen inhalation reduces numbers

POCUS CarotidRespiratory variation in carotid peak systolic velocity predicts volume responsiveness in mechanically ventilated patients with septic shock: a prospective cohort study

• Single-center, prospective, cohort study.

• mechanical ventilation, septic shock, and hemodynamic instability for which the

• MICU/SICU tertiary academic hospital from May-October 2014.

• ACVC mechanical ventilation was performed with TV 6 ml/kg

• Fluid challenges with NS at a 7 mL/kg dose over a 30-min period

• Thermodilution before and after each challenge.

Ibarra-Estrada et al. Critical Ultrasound Journal (2015)

12/30/2016

8

• CCA 2 cm from the bifurcation• Maxi and min PSV obtained in a single

respiratory cycle• ΔCDPV=(MaxCDPV − MinCDPV) /

[(MaxCDPV +MinCDPV) / 2] × 100• In addition

• TTE for stroke volume• Fem A: Pulse pressure variation PPV (%) = 100 ×

(Pp max − Pp min) / [(Pp max + Pp min) / 2] • The passive leg raising (PLR) test was Inferior

vena cava diameter (IVC-d) measurement was• IVC CI • Transpulmonary thermodilution to obtain an

automated SVI, stroke volume variation (SVV), and other variables.

• Patients with an increase of more than 15 % in the SVI after the fluid challenge were classified as “responders”, and those with an increase of less than 15 % in the SVI or those with no increase were classified as “non-responders.”

Ibarra-Estrada et al. Critical Ultrasound Journal (2015) Ibarra-Estrada et al. Critical Ultrasound Journal (2015)

Areas under the receiver operating characteristic curve ofpredictors of fluid responsiveness. The p value indicates comparisonbetween respiratory variation in carotid peak velocity and strokevolume variation (SVV) with the Hanley–McNeil test

Correlation between variation in respiratory carotid peak systolic velocity and fluid challenge-induced changes in the stroke volume index

Feasibility of common carotid artery point of care ultrasound in cardiac output measurements compared to invasive methods• Single urban tertiary care academic center

• SICU and CT ICU

• All patients had an existing invasive monitoring device, either a PA catheter or PCA

• Carotid Doppler waveforms and the time average velocity 3-5 cardiac cycles

• Volume flow = Cross-sectional diameter X Time average velocity.

• Physiology studies quote a range of 15–26 % for the proportion of CO that is dedicated to cerebral perfusion

• cerebral flow comprised 20 % of resting CO

• 80 % of cerebral perfusion was supplied via the carotid arteries, 40 % via each carotid

• CO = Carotid Doppler flow volume X 10.

Gassner at al, J Ultrasound (2015)

Gassner at al, J Ultrasound (2015)Scatter plot diagram showing the correlation betweencardiac output measurements using the pulse contour analysis vs. ultrasound. CO cardiac output. PCA pulse contour analysis. b Scatter plot diagram showing the correlation between cardiac output measurements using the pulmonary artery catheter vs. ultrasound. CO cardiac output. PAC pulmonary artery catheter

Scatter plot diagram showing the overall correlationbetween cardiac output measurements using invasive modalities versus ultrasound. b The Bland–Altman plot indicating that for the most part, the overall cardiac output measurements tend to agree with a mean difference close to zero and there is no statistical difference between the cardiac output measurement by invasive modality and ultrasound. P = 0.262. CO cardiac output in L/min

Arterial trauma during central venous catheterinsertion

• Case series 11pt

• Case series 12 pt

• Case series 13 pt

• Approximately 7 million of CVC are installed each year in the US

• Small needle puncture 5% of cases

• Arterial misplacement of large caliber cannula have an incidence of 0.1% to 0.8%

Shah at al, j.jamcollsurg.2004

• Carotid Artery Puncture/Cannulation

• Of the 16 claims for CA puncture/cannulation• 5 resulted in a stroke• 4 resulted in airway obstruction

due to a hematoma• 3 involved extra surgery with

arterial repair• 4 resulted in case cancellation• In two claims, the arterial position

of the catheter was unrecognized for 22 h or more

• Neither ultrasound guidance nor pressure waveform monitoring was used for vessel localization in any of these claims

Domino at al, Anesthesiology 2004;

Bechara,at al, JAMA Surg. 2013

Guilbert at al, J Vasc Surg 2008

Table 2 injuries

POCUS ONSD

12/30/2016

9

• Patient supine • The linear probe was

placed lightly over the closed upper eyelid

• The structures of the eye visualized to align the optic nerve directly opposite the probe

• A single optic nerve sheath diameter measured 3.00 mm behind the globe in each eye

• Measurements from each eye were averaged to create a binocular optic nerve sheath diameter measurement

1968, Hayreh established the presence of a constant communication between the subarachnoid space of the optic nerve sheath and the intracranial cavity Normal ONSD measured by ultrasound ranges from 4.7 to 5.0 mm in previous

optic nerve attaches to the globeposteriorly and is wrapped in a sheath

that contains fluid

•

Hansen at al, Acta Ophthalmologica 2011

Ultrasound series of the same optic nerve preparation with increasing pressure loadTop row: pressure 15 mmHg, optic nerve sheath diameter (ONSD) 5.5 mmMiddle row: pressure 45 mmHg, ONSD 6.4 mmBottom row: pressure 55 mmHg, ONSD 7.0 mm.

Results

ONSD behaviour following stepwise pres-

sure increase

Elevations of pressure in the SAS

were invariably followed by an

increase in the ONSD (Fig. 3A). The

pressure-induced dilatation (difference

of ONSD, or DONSD) ranged

between 0.9 and 2.5 mm (mean:

1.97 mm, SD: 0.52). With baseline lev-

els ranging between 3.8 and 5.1 mm

(mean: 4.69 mm, SD: 0.40), the mean

increment of the ONSD was 40.2%.

Concerning the degree of dilatation

(DONSD), no difference was found

between high and low baseline diameters

(ONSD at baseline < 4.6 mm: DONSD

1.84; ONSD at baseline > 4.6 mm;

DONSD 1.95 mm).

The relation of step magnitude and

corresponding ONSD changes was

nearly linear within a wide range.

The initial pressure increase from

0 mmHg (baseline) to 5 mmHg

resulted in a rather large increase in

diameter (DONSD mean 0.7 mm).

Later, this DONSD became somewhat

smaller (diameters of 5.46, 5.76, 5.85,

6.09, 6.30, 6.45 and 6.66 mm after

exposure to 5, 15, 25, 35, 45, 55 and

65 mmHg, respectively). The mean

slope of the diameter–pressure plot

was calculated as 0.025 mm per

mmHg (r = 0.94, p < 0.01, two-

tailed Pearson correlation). Thus on

average, an additional ONS dilatation

of 0.25 mmHg resulted from every

10-mmHg pressure increment.

ONSD behaviour following stepwise pres-

sure decrease

All reductions of pressure in the SAS

likewise resulted invariably in a

decrease in the dilated ONSD. How-

ever, thediameter reduction wasmostly

incomplete as it did not reach baseline

again. The mean relative decreases of

theONSD ranged around 81.4% (90.9–

74.6 %). The residual deviation from

baseline ranged between 0.1 and

1.1 mm (mean: 0.50 mm, SD: 0.52).

This remaining DONSD was related

to the pressure level previously

applied (Fig. 3B). Therefore, no linear

correlation was present between pres-

sure drops and ONSD decline.

Comparing lower to higher levels of

previously applied pressure, the

ONSD level recovered only to base-

line coming from SAS pressure lower

than 35 mmHg. In contrary, following

decompression from higher pressure

levels (45 mmHg and above), a clear

residual dilatation remained (above

0.34 mm, see Table 1).

Comparison among the 10 prepara-

tions revealed some variation of their

threshold values for this irreversible

dilatation. The loss of complete diam-

eter recovery always occurred after

previous exposure to a pressure level

of 55 mmHg or more. In three prepa-

rations, this phenomenon was already

present after exposure to lower levels

of pressure in the SAS (45 mmHg,

n = 2 and 35 mmHg, n = 1).

DiscussionIn this paper, we describe for the first

time the pressure-dependent behaviour

of the ONSD in vitro after controlled

application of incremental pressure

steps in the SAS surrounding the

human optic nerve. To create nearly

physiological conditions, we kept the

pressure outside the optic nerve sheath

at about 2–3 mmHg equivalent to the

physiological orbital tissue pressure

(Møller 1955) simply by using a con-

stant depth within the experimental

tank. I ts temperature was also kept

constant at 37°C. Interferences result-

ing from autolysis were ruled out by

controlling of the preparation time

after autopsy (Anders et al. 1979).

Under these conditions, our ONSD

baseline data correspond nicely to ana-

tomical measurements reported from

the anterior portion of the optic nerve

and its sheath (Lang & Reiter 1985).

Dilatation of the optic nerve sheath

was confirmed in all 10 preparations.

(A) (B)

Fig. 3. (A) Optic nerve sheath response to pressure increase: optic nerve sheath diameter

(ONSD) values (open circles) after pressure increase (‘+ p’) in n = 10 preparations, depicted as

mean and standard deviation. Experimentally applied increasing pressure levels (open bars) in

the SAS result in ONSD elevations, even after exposure to pressures as low as 5 mmHg. Refer-

ring to baseline values (baseline, square), there is a continuous elevation of the ONSD with

increasing pressure burden. Optic nerve sheath diameter has gained more than 2 mm at

65 mmHg. (B) Optic nerve sheath response to pressure decrease: After experimental pressure

decreases (‘) p’) in the SAS, a residual dilatation of the ONSD (full circles) remains above the

baseline value (ONSD baseline, square). The decompression after increasingly higher pressure

levels (open bars), particularly when reaching 45 mmHg or more, leads to increasingly higher

residual ONSD (here given as mean and standard deviation of n = 10 preparations). After

decompression from SAS pressure 65 mmHg, a mean residual dilatation of 0.5 mm remains.

Table 1. Averaged ONSD data from 10 preparations measured before and after negative pres-

sure steps. Baseline pressure level is equivalent to pressure 0 mmHg at the beginning of the

experiment. Decompression from higher levels leads to increasing residual dilatations. The

resulting ONSD difference to baseline (DONSD) increases above 0.24 mm once pressure steps

exceed 35 mmHg. The decrement was calculated as ONSD ratio before and after decompres-

sion, and it gradually decreases because of a decline in elasticity.

Negative pressure step [mmHg] 0 5 15 25 35 45 55 65

Residual mean ONSD [mm] 4.69 4.76 4.86 4.89 4.93 5.03 5.12 5.19

DONSD to baseline [mm] 0 0.07 0.17 0.20 0.24 0.34 0.43 0.50

Decrement (%) 90.9 84.1 82.8 82.9 78.9 75.6 74.6

ONSD, optic nerve sheath diameter.

Act a Opht hal mol ogica 2011

e530

Zeiler at al, Crit Ultrasound J. 2016

As part of a previously

published study, a single

operator measured the ONSD

in 120 healthy volunteers over

a 6-month period. Utilizing the

assumption that the four

measurements made on each

subject during this study

should be equal, the

relationship of within-subject

variance was described using

a quadratic-plateau model as

assessed by segmental

polynomial (knot) regression.

• 35 patients were academic ER

• 14 CT results consistent with EICP.

• ONSD > 5 mm on US

• The mean ONSD for the 14 patients with CT evidence of EICP was 6.27 mm (95% CI 1⁄4 5.6 to 6.89)

• The mean ONSD for the others was 4.42 mm (95% CI 1⁄4 4.15 to 4.72)

• The difference of 1.85 mm (95% CI 1⁄4 1.23 to 2.39 mm) yielded a p 1⁄4 0.001.

• The sensitivity and specificity for ONSD, when compared with CT results, were 100% and 95%, respectively

• PPV, NPV were 93% and 100%, respectively

Blavias a al, Academic Emergency Medicine 2003

April 2006 -January 2007Cranial CT findings were used as a reference 100 participants median GCS=11Mean binocular ONSD (5.8 0.57 mm) raised ICP Mean ONSD (3.5 0.75 mm) without

Goel at al, Injury, Int. J. Care Injured (2008

Goel at al, Injury, Int. J. Care Injured (2008

Ohle at al, Ultrasound Med 2015;

ONSU vsCT

12/30/2016

10

Kimberly at al, Academic Emergency Medicine 2008

38 ocular USs 15 individual patientsSpearman rank correlation coefficient of ONSD and ICP was 0.59 (p < 0.0005) ROC curve was created to assess the ability of ONSD to distinguish an abnormal ICP greater than 20 cm H2OThe area under the ROC curve was 0.93 (95% confidence interval [CI] = 0.84 to 0.99)ONSD > 5 mm performed well to detect ICP > 20 cm H2O Sensitivity of 88% Specificity of 93%

Dubourg at al, Intensive Care Med 2011

ONSUvsICP

6 studies231 patient pooled sensitivity 0.90 specificity was 0.85 The area under the summary receiver-operating charac-teristic (SROC) curve was 0.94 (95% CI 0.91–0.96).

Rajajee at al, Neurocrit Care (2011)

(94–99%), PPV 92% (95% CI 81–98%), NPV 97%

(93–99%). For patients not on mechanical ventilation (320

measurements in 38 patients), ROC analysis revealed

AUC = 0.97 (95% CI 0.94–0.99, P < 0.0001) and opti-

mal ONSD cutoff C0.48 cm with sensitivity 99% (95% CI

93–100%), specificity 95% (91–97%), PPV 85% (95% CI

76–92%), NPV 100% (98–100%).

CT/MR Versus Invasive ICP

Fifty seven of ninety seven (59%) measurement clusters

were performed in the same 24 h period as a CT or MR

scan (54 CT scans and 3 MR scans). The scans were per-

formed a mean of 7 h from ONSD measurement (SD 7 h,

range 15 min to 18 h). Overall, 44 of 57 (77%) scans

demonstrated at least one of the previously designated

signs (Table 1) of elevated intracranial pressure. Thirteen

of sixteen (81%) scans performed within the same 24 h

period of a measurement cluster with mean invasive

ICP > 20 mmHg versus 31 of 41 (76%) scans performed

within the same 24 h period of measurement clusters with

mean invasive ICP recording B 20 mmHg demonstrated at

least one radiological sign of raised ICP (Fisher Exact

2-tail probability P = 0.74). Sensitivity was 77% (95% CI

46–95%), specificity 29% (17–45%), PPV 24% (95% CI

12–40%) and NPV 81% (95% CI 54–96%) for the detec-

tion of invasive ICP > 20 mmHg.

Discussion

A reliable non-invasive means to detect intracranial

hypertension is an important unmet need in the field of

neuro-intensive care. While papilledema is a long-recog-

nized clinical sign of intracranial hypertension, it can take

many hours-to-weeks to develop and cannot be relied upon

in the critical care setting for the detection of acutely ele-

vated ICP [5]. While our study was not specifically

designed to analyze the predictive value of CT and MR

imaging for ICP, the analysis of CT and MR images per-

formed within 24 h of measurement revealed no

correlation between signs of raised ICP on imaging and the

presence of invasive ICP > 20 mmHg in this mixed group

of neuro-ICU patients. Other studies, in patients with TBI,

have also shown that CT imaging cannot always be relied

upon to identify ICP [14, 15]. Several other non-invasive

techniques of ICP measurement have been evaluated,

including transcranial doppler sonography, ophthalmic

doppler flow measurement and near-infrared spectroscopy

among others [16]. Most of these techniques have had

mixed results or are not easily available to the clinician at

the bedside.

Our study demonstrates that ON-US with measurement

of ONSD—when performed by an experienced operator—

is a highly accurate non-invasive technique for the detec-

tion of intracranial hypertension (whether defined as

Fig. 5 ROC curves of ONSD

for the detection of invasive

ICP > 20 mmHg (a) and

invasive ICP > 25 mmHg (b).

Dotted tracing represents 95%

confidence bounds

Table 3 Accuracy of different ONSD criteria for the detection of intracranial hypertension

High ICP

criterion (mmHg)

ONSD

criterion (cm)

Sensitivity

(95% CI)

Specificity

(95% CI)

Positive predictive

value (95% CI)

Negative predictive

value (95% CI)

> 20 C0.48 96% (91–99%) 94% (92–96%) 84% (77–89%) 99% (97–100%)

> 20 C0.50 86% (79–92%) 98% (96–99%) 92% (86–96%) 96% (94–98%)

> 20 C0.52 67% (58–75%) 98% (97–99%) 93% (86–97%) 91% (88–93%)

> 20 C0.59 19% (13–27%) 100% (99–100%) 96% (80–100%) 80% (76–84%)

> 25 C0.52 98% (89–100%) 91% (88–94%) 53% (42–64%) 100% (99–100%)

512 Neurocrit Care (2011) 15:506–515

123

The ONSD retained high accuracy for the detection of ICP > 20 mmHg Optimal ONSD cutoff between the two groupsMechanical ventilation (216 measurements in 27 patients) revealed AUC = 0.97 and optimal ONSD cutoff 0.50 cm No mechanical ventilation (320 measurements in 38 patients), ROC analysis revealed AUC = 0.97 optimal ONSD cutoff 0.48 cm with sensitivity

• Important and unanswered question is whether ONUS remains accurate in the setting of acutely fluctuating ICP

• Authors identified significant acute ICP fluctuation (SAIF) in which individual invasive ICP measurements both above and below 20 mmHg were recorded among the six attempted measurements.

• All ICP spikes to >30 mmHg that persisted for at least five minutes, the lowest ICP in this period and the highest ICP in this period were recorded. The accuracy of ONSD measurements performed within four hours of an ICP elevations to >30 mmHg was compared to the accuracy of ONSD measurements without any preceding spike

• Specificity and PPV of ONSD for ICP >20 mmHg are substantially decreased in patients demonstrating acute fluctuation of ICP between high and normal. This may be because of delayed reversal of nerve sheath distension.

Rajajee et al. Critical Care 2012

Rajajee et al. Critical Care 2012

Comparing sensitivity and specificity of ONSD value in Trauma and Non-Trauma Group

Sensitivity Specificity

Trauma Group

ONSD - 5.22 mm 94.40% 90.50%

ONSD - 5.47 mm 94.40% 95.20%

Non-Trauma Group

ONSD - 5.21 mm 100% 76.70%

ONSD - 5.48 mm 83.30% 93.30%

Raffiz at al, The American Journal of Emergency Medicine 2017

12/30/2016

11

Vaiman M et al. J Clin Neurosci (2016),

• Eleven patients (7 males, 4 females) with traumatic brain injury (TBI) and GCS<9

• All patients had EVD monitors

• A total of 29 ONUS obtained with invasive ICP measurements, pulsatility indices (PI) using middle cerebral artery (MCA) and ONSD.

• 14 ONSD was <5.0 mm, while ICPs ranged 1-18 mmHg

• 11 ONSD one or both ONS diameters ranged 5.0-5.7mm, while ICP ranged 10-19mmHg

• 1 ONSD 5.7 mm, with ICP 13mmHg, however, later patient developed ICP 37mmHg within 24 hours

• 1 ONSD 5.7mm with ICP 16mmHg, however, later developed ICP of >30mmHg within 24 hours.

• 2 ONSD > 3.8mm with corresponding ICP >20mmHg

• There was no correlation between PI on TCD, thus TCD was not useful in this dataset.

Bulic at al, NCS 2015