Multicenter Validation of an ECG and OxygenSaturation Based Sleep Diagnostic System

Clement Cahan1, Michael J. Decker2,3, Shuli Eyal4, Zvika Shinar4, Armanda Baharav1,4, William C. Reeves3

(1) Shaare Zedek Medical Center, Sleep Disorders Clinic, Israel; (2) Fusion Sleep – Program in Sleep Disorders, Suwanee, GA, USA;

(3) Centers for Disease Control & Prevention, Atlanta, GA, USA; (4) HypnoCore, Netanya, Israel.

Methods The HC1000P :

Detects respiratory events by oximetry and ECG

morphology (ECG Derived Respiration-EDR).4,5

In addition the system can identify sleep architecture,

arousals and awakenings. This is based on the connection

between sleep and differential autonomic nervous system

modulation of instantaneous heart rate6 during different

sleep stages.7

In this blinded, multi-center retrospective study we compared

the automated application of HC1000P to results obtained by

the gold standard, manual method using Rechtschaffen & Kales

and ASDA scoring criteria.

104 recordings were used in this study:

54 were randomly selected from a cohort Chronic Fatigue

Syndrome study for the CDC in Wichita, Kansas. The patients

were screened to exclude OSA.

50 were consecutive studies of patients referred for PSG for

suspected OSA at Shaare Zedek Medical Center in

Jerusalem.

Age: 48.6 ± 14.0 (Range: 19 – 79)

BMI: 29.3 ± 5.8 (Range: 18 – 48)

Male/Female: 47/57

The ECG and pulse oximetry signals were separately analyzed

by the HC1000P, to determine the respiratory disturbance index

(RDI), sleep time, sleep efficiency, the time spent awake, and

NREM, REM sleep, awakenings and arousals.

Inclusion criteria: recording time of at least 5 hours, normal

sinus rhythm and good quality ECG and oximetry signals.

Recordings of poor signal quality and ECG with extensive

arrhythmia over 20% of the recording time were excluded.

Results Respiratory evaluation:

The desaturations found by the HC1000P system and those found by manual gold standard scoring correlated highly, with R = 0.97, as can be seen in Figure 1(a). The Bland-Altman plot in Figure 1(b) shows the good agreement between results obtain by the two methods.

ConclusionsThe HC1000P allows for an accurate, specific and highly sensitive method to diagnose Obstructive Sleep Apnea (OSA).

As opposed to other partial tests, the HC1000P gives insight into sleep architecture.

Results are accurate even at low RDIs, i.e. mild cases of OSA or normal individuals. This implies it can be used as a true diagnostic tool, not just a method of screening.

The simplicity of the method allows for easy follow up studies which are not affordable by standard methods. Thus, compliance with CPAP might increase.

It is a cost-effective measure for mass diagnosis, of great importance for a disorder with a high prevalence of undiagnosed cases.

Introduction Sleep Related Breathing Disorders (SBD) affect 2-4% of the

adult population and are treatable, yet 82% of men and 93%

of women are believed to remain undiagnosed.1,2

SBD have been associated with systemic hypertension,

significant cardiovascular morbidity and increased mortality.3

Simple, reliable, user friendly and cost effective diagnostic

procedures represent a burning need.

The purpose of this study is the validation of an alternative

diagnostic system that relies on ECG and pulse oximetry

signals to diagnose Obstructive Sleep Apnea (OSA), the

HC1000P System by HypnoCore.

Fig 2: (a) Correlation of Respiratory Disturbance Index (RDI) as calculated by gold standard methods and HC1000P. (b) Bland-Altman showing level of agreement between RDI by the two methods.

Number of Desaturations

R = 0.966

0

100

200

300

400

500

600

700

800

900

0 100 200 300 400 500 600 700 800 900

Gold Standard

HC

10

00

P

Number of Desaturations

-150

-100

-50

0

50

100

150

200

250

0 100 200 300 400 500 600 700 800 900

Average

Dif

fere

nc

e

+2SD

-2SD

Average

(a) (b)

Respiratory Disturbance Index - RDI

R = 0.901

0

10

20

30

40

50

60

70

80

90

100

0 20 40 60 80 100

Gold Standard

HC

10

00

P

Respiratory Disturbance Index - RDI

-20

-10

0

10

20

30

40

50

0 10 20 30 40 50 60 70 80 90 100

Average

Dif

fere

nc

e +2SD

-2SD

Average

(a) (b)

Sleep architecture evaluation:

Results of the analysis show good overall agreement with regards to total sleep time (TST), sleep efficiency, percentage of time spent in REM and NREM sleep as seen in Figure 3. The number of arousals detected by the two methods was also highly correlated.

The total number of respiratory events highly correlated between the two methods with R = 0.92.

RDI (apnea/hypopnea events per hour of sleep) highly correlated with R = 0.90 as can be seen in Figure 2.

When RDI values are divided into two groups, low and high RDI with the cutoff at 15 events per hour, the results have a 98.1% agreement between results obtained by gold standard scoring and those calculated by HC1000P.

Cohen’s Kappa is 0.96, both specificity and sensitivity are 0.98.

Subjects had a wide range of the disorder, allowing to divide them s into 4 clinically relevant groups with regard to the degree of respiratory compromise during sleep: RDI ≤ 5 (normal), RDI between 5 and 15 (mild), RDI between 15 and 30 (moderate) and RDI ≥ 30 (severe), as can be seen in table 1.

80.8% of recordings show an exact match in the category assigned by gold standard methods and by HC1000P.

The remaining recordings show disagreement by no more than one category.

Fig 1: (a) Correlation of number of desaturations detected by gold standardmethods and by HC1000P. (b) Bland-Altman plot showing level of agreement indetecting desaturations by the two methods.

References

1. Young, T., et al., The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med, 1993. 328(17): p. 1230-5.

2. Young, T., et al., Estimation of the clinically diagnosed proportion of sleep apnea syndrome in middle-aged men and women. Sleep, 1997. 20(9): p. 705-6.

3. Peppard, P.E., et al., Prospective study of the association between sleep-disordered breathing and hypertension. N Engl J Med, 2000. 342(19): p. 1378-84.

4. Furman, G., et al., Electrocardiogram Derived Respiration during Sleep.5. Cooper, B.G., et al., Value of nocturnal oxygen saturation as a screening test for sleep apnoea. Thorax,

1991. 46(8): p. 586-8.6. Akselrod, S., et al., Power spectrum analysis of heart rate fluctuation: a quantitative probe of beat-to-beat

cardiovascular control. Science, 1981. 213(4504): p. 220-2.7. Baharav, A., et al., Fluctuations in autonomic nervous activity during sleep displayed by power spectrum

analysis of heart rate variability. Neurology, 1995. 45(6): p. 1183-7.

Fig 3: Comparison of sleep parameters as calculated by gold standard andHC1000P methods: Total Sleep Time (TST), and time in each sleep stagepost-sleep onset (in minutes).

Sleep Architecture Comparison

0

50

100

150

200

250

300

350

400

450

500

TST Wake REM NREM SWS LS

Gold Standard

HC1000P

RDI<5 5≤RDI<15 15≤RDI<30 30≤RDI Total

RDI<5 28 2 0 0 30

5≤RDI<15 8 21 1 0 30

15≤RDI<30 0 1 19 4 24

30≤RDI 0 0 4 16 20

Total 36 24 24 20 104

Gold Standard

HC1000P

Table 1: Correlation between HC RDI and Gold Standard RDI.