Evaluation of a Microprocessor-Based Portable Home Monitoring ...

Sleep 10(2): 130-142, Raven Press, New York © 1987, Association of Professional Sleep Societies

Evaluation of a Microprocessor-Based Portable Home Monitoring System to Measure Breathing During Sleep Stephen Gyulay, Deborah Gould, Beverley Sawyer, Dimity Pond, Andrea Mant, and Nicholas Saunders Department of Medicine, Royal Newcastle Hospital, the University of Newcastle, NSW, and General Practice and Primary Care Research Unit, Royal Australian College of General Practitioners, Sydney, Australia

Summary: Study of the epidemiology of disturbances of breathing during sleep was hampered until recently by the need to conduct studies in the laboratory, with attendant inconvenience and limited sample sizes. We assessed the accuracy of a microprocessor-based portable monitoring system (Vitalog PMS-8, Vitalog Corp., CA) to detect and classify episodes of disturbed breathing during sleep in 14 patients with sleep apnea by simultaneously recording oxygenation and thoracoabdominal motion on the portable system and a polygraph. Each patient slept in the laboratory for 1 night. In two subjects, the portable system failed to record thoracoabdominal signals. In the remaining subjects, the portable system detected 78% of 2,340 episodes of disturbed breathing, but the recorded information was not sufficient to allow confident classification into central or obstructive events. The positive predictive value of disturbed breathing detected by the portable system was 64%, Respiratory disturbance indices (RDI) computed from the polygraph and portable records were correlated (r = 0.70; p < 0.01), and all patients with sleep apnea were correctly diagnosed by the portable system. The portable system overestimated arterial oxygen saturation (Sa0 2) recorded by an ear oximeter (Biox IIA, Ohmeda, CO) but the error was < 10% of the true value at Sa02 > 60%. Seven normal subjects were studied while awake to examine the accuracy of volume measurements made by the portable system and the system's ability to detect paradoxical thoracoabdominal motion of various degrees. Absolute measurement of tidal volume was inaccurate, but detection rate of paradoxical thoracoabdominal motion was excellent (97%). We conclude that the portable system is sufficiently sensitive to allow detection of patients with breathing disorders during sleep, but further developments are necessary before the system can be relied on for accurate classification of apneas and hypoventilation. Key Words: Sleep apnea-Home monitoring-Oximetry.

Although the exact incidence and prevalence of the obstructive sleep apnea syndrome (OSA) are uncertain, it is clear that the problem is widespread in the community Accepted for publication September 1986. Address correspondence and reprint requests to Dr. Andrea Mant, Director at General Practice and Primary Care Research Unit, Royal Australian College of General Practitioners, 43 Lower Fort Street, Sydney NSW 2000, Australia.

130

HOME MONITORING OF NOCTURNAL BREATHING

131

(1-4). Disturbances of ventilation during sleep also occur in patients with chronic airflow limitation (5 -8) and other forms of lung and chest wall disease (9-12). Thus, the community is likely to have a substantial burden of illness related to abnormalities of breathing during sleep. A major difficulty encountered to date in study of the epidemiology of such breathing disturbances has been the inconvenience and artificiality of conducting studies in a laboratory. These problems have led to studies being limited to small population samples, potentially biased by selection procedures and inadequate to examine the natural history of nighttime breathing disorders. Recently, home monitoring devices have been developed that have the potential to overcome these problems. An analogue recording system has been shown to have acceptable sensitivity for the diagnosis of OSA (13). The present study evaluated the ability of a portable microprocessor system (14-17) to measure accurately breathing, oxygenation, and arousals during sleep. METHODS Two series of experiments were conducted: nocturnal studies of breathing, oxygenation and arousal in patients with sleep apnea, and daytime studies of breathing and heart rate in normal subjects. Series 1: Nocturnal studies in patients with sleep apnea Subjects. Fourteen male patients, whose ages ranged from 50 to 73 years (mean ± SD, 58.5 ± 7.9 years) were recruited for study. All patients had been studied on at least one occasion previously. Eleven patients had moderate to severe obstructive apnea (mean respiratory disturbance indices, RDI = 48.4 ± 27.2). Two patients had Deen diagnosed previously as having OSA but had an RDI < 5 at the time of the present study. Eleven patients were receiving treatment with nasal continuous positive airway pressure (nCPAP) at the time of study. One patient had central apnea (apnea index 41.8). Nine of the 14 patients were overweight. Apparatus. Simultaneous recordings were made with a microprocessor-controlled home monitoring system (Vitalog PMS-8, Vitalog, CA) (Fig. 1) and a Grass Polygraph Recorder (Model 78, Grass Instruments, Quincy, MA) routinely used in the sleep laboratory. The portable system measured tidal volume by inductance plethysmography, summing the calibrated signals from sensor bands placed around the rib cage and abdomen; paradoxical breathing by examining the phase relationship between rib cage and abdominal signals; percentage of arterial oxygen saturation (%Sa0 2) with a Biox IIA oximeter (Ohmeda, CO) interfaced with the Vitalog system; heart rate, determined by the R-R interval of the electrocardiogram (ECG); and body movement with an activity transducer strapped to the subject's wrist. On the polygraph, electroencephalogram (EEG), electromyogram (EMG) and electroocculogram (EOG) were recorded using standard electrode placements (I8). Rib cage and abdominal motion were measured with linearized magnetometers; no attempt was made to quantitate the magnetometer signals. Airflow at the nose and mouth was measured with thermocouples. Oxygenation was measured with a Biox IIA oximeter; the oximeter signal was split and recorded simultaneously by the portable recorder and polygraph. Protocol. Each patient was studied in the laboratory for 1 night. After attachment of all electrodes, the portable monitor rib cage and abdominal signals were calibrated Sleep, Vol. 10, No.2, 1987

S. GYULAY ET AL.

132

---

---ECG Electrode

..... " Respiraflon

--

L~Jl1mf\:1Sfi _

.>

Bands

_ _Vitalog

Monitor

FIG. 1. Components of the portable recording system.

according to the manufacturer's instructions. This required the subject first to perform a series of isovolume thoracoabdominal movements and second, to rebreathe briefly from a bag of known volume. Recording sessions began at ~ 11 :00 p.m. and ended at ~6:30 a.m. Eleven of the patients were awakened after 3.6 ± 0.6 h of sleep (mean ± SD) that included at least one period of REM sleep and fitted with their nasal CPAP device. The subjects then slept with nCPAP for the rest of the recording session (mean duration of sleep with nCPAP was 2.6 ± 0.3 h). The study of patients with nCPAP allowed the portable system to be tested for false-positive recording of apneas. Analysis. At the end of the recording session, the data that had been continuously processed and stored by the portable system was recovered, displayed, and printed by an IBM-PC microcomputer (Fig. 2). Both polygraph and portable records were analyzed by hand; we did not evaluate analysis software associated with the portable recorder in this study. In all analyses of the portable records, the scorer was unaware of the findings on the polygraph. To analyze breathing pattern, the polygraph was scored for central and obstructive apneas (19) and episodes of hypoventilation lasting> 15 s. Mixed apneas were recorded as obstructive. We chose a duration of 15 s because the manufacturer of the portable recorder recommended this duration. The portable record was analyzed in two ways: first an analysis according to the manufacturer's criteria was performed for all subjects. An apnea was defined as tidal volume being less than one-third of the resting tidal volume for> 15 s. Paradoxical thoracoabdominal motion was judged present whenever the "paradox" tracing deflected above zero baseline. Hypopneas were scored as present whenever tidal volume was one-third to two-thirds of resting tidal volume for> 15 s. Second, the portable record was analyzed according to stricter criteria. This analysis was performed for only five subjects. Apneas were defined as zero deflection on the tidal volume tracing for> 15 s. Paradoxical thoracoabdominal motion was deemed present only if the "paradox" tracing showed a deflection >2 mm above the baseline; pilot studies in supine healthy subjects breathing quietly had shown random fluctuaSleep, Vol. 10, No.2, 1987

HOME MONITORING OF NOCTURNAL BREATHING

133

'k-h

~--Fl •••.•• ' .............

FIG. 2. Tracings obtained with the portable recording system in a patient with obstructive sleep apnea (OSA) (upper panel) and central sleep apnea (lower panel). Tidal volume in milliliters; Pardx, paradoxical thoracoabdominal motion; Oxy, percentage of arterial oxygen saturation (scale in upper panel refers to oxygenation; heart rate scale not shown); heart rate in beats/min (scale in lower panel refers to heart rate; oxygenation scale not shown); Act, wrist movement.

tions of the paradox baseline of up to 2 mm. Thus, an apnea was scored as obstructive if there was associated paradox >2 mm; it was scored as central if the paradox tracing showed 0-2 mm deflection. Hypopneas were scored present if the tidal volume tracing showed deflections 15 s. To compare Sa02 values recorded by the two systems, I-min epochs were identified every 10 minutes of record. Minimum and maximum %Sa0 2 recorded during these epochs by each system were noted. The sensitivity of the portable monitor's algorithm to detect arousals was assessed by relating activity recorded by the wrist sensor to periods of wakefulness on the EEG record, scored by standard criteria (18). Arousals that were too brief to be scored awake were also noted to ensure that activity recorded by the portable system during such a brief arousal would not be interpreted as a false-positive response. Series 2: Daytime studies in healthy subjects Subjects. Seven subjects (5 men and 2 women aged 35.0 ± 11.5 years) were recruited. None was obese. Methods. Recordings of tidal volume were obtained with the portable system while the subject breathed through a mouthpiece connected to a Fleisch no. 2 pneumotachograph. The flow signal was integrated (Grass 7PIO) to obtain volume. The integrated flow signal was calibrated at the beginning and end of each experimental session. Calibration of the portable system was performed at the beginning of each experimental session according to the manufacturer's instructions either by the subjects rebreathing a known volume from a bag or by a manual signal related the the subject's body weight. The accuracy of each system of calibration was tested separately. Measurements were made in the supine and left lateral positions. Subjects were asked to vary their tidal volumes within each experimental run, which lasted from 5 to 25 min (10.3 ± 7.5 min). Breaths> 1,500 ml were excluded from analysis since pilot studies had Sleep, Vol. /0, No, 2, 1987

134

S. GYULAY ET AL.

shown the portable system to be very inaccurate at volumes larger than this. Records were analyzed in two ways. First, polygraph and portable records were divided into matched 30-s epochs, mean tidal volume was computed for each epoch, and the correlation between polygraph and portable measurements was examined. Second, every breath was assigned to one of three groups (0-500, 501-1,000, and 1,001-1,500 ml) according to its size measured by the pneumotachograph system. Differences between portable recorder and pneumotachograph measurements were examined in each group. To assess ability of the portable system's algorithm to detect paradoxical thoracoabdominal motion, each subject performed a number of isovolume maneuvers in the supine and left lateral positions. Rib cage and abdominal motion was monitored by the portable monitor's sensor bands and by two pairs of magnetometers placed at the level of the nipples and umbilicus respectively. Subjects performed a series of isovolume maneuvers for 20 severy 2 min. The subject was instructed to use the same effort within any given series but to vary the effort from one series to the next. Thus,' isovolume maneuvers resulting in small, medium, and large thoracoabdominal excursions were obtained for analysis. The accuracy of the portable system's heart rate record was assessed by simultaneous recording of the electrocardiogram at rest and after exercise. Heart rate was calculated from the EeG record for 60-s epochs and compared with the portable record. The ability of the portable monitor's algorithm to detect body movement was tested by having the supine subject perform a predetermined sequence of wrist, arm, and whole body movements at fast and slow speeds. RESULTS Studies in patients with sleep apnea. Apnea detection and classification. All patients slept comfortably with the portable monitor sensors and recording devices in place, and none reported disturbed sleep because of the new devices they were wearing. In two subjects, the portable recorder failed to record respiratory signals. These two subjects are not considered further in the analysis. The reason for the failures was not apparent, but because no further technical problems were encountered it was assumed to be due to operator error. The sensitivity of the portable system to detect episodes of disordered breathing during sleep regardless of type is shown in Tables 1 and 2. The polygraph system detected 2,340 episodes of disordered breathing lasting> 15 s. Seventy-eight percent of these episodes were detected by the portable system when the manufacturer's criteria were used to analyze the portable record (Table O. The positive predictive value of disturbed breathing detected by the portable system (true positive/all positives detected) was 64% (1,828 of 2,838). When stricter criteria were applied to analysis of the portable record (Thble 1), the sensitivity of the portable system to detect episodes of disturbed breathing decreased (59%) but the positive predictive value increased (82%). The sensitivity of the portable system to detect episodes of disturbed breathing ranged from 40 to 94% among the 12 subjects for whom data were available (Table 2). Positive predictive values ranged from 12 to 84%. Tables 2 and 3 show the accuracy of the portable system to classify breathing disturbances during sleep. Use of the manufacturer's criteria to define apneas and hypoventilation (Table 3) resulted in correct identification of 47% of episodes of complete upper

Sleep, Vol. 10, No.2, 1987

HOME MONITORING OF NOCTURNAL BREATHING

135

TABLE 1. Comparison of polygraph and portable system in the detection of episodes of disturbed breathing during sleep regardless of type of disturbance Polygraph disturbed breathing Portable disturbed breathing Manufacturer's criteria (n Yes No Strict criteria (n Yes No

=

=

Yes

No

1,828 512 2,340

1,010

12) N/A

5) 913

203

630 1,543

N/A

Details of criteria used for analyses are described in text. N/A = not analyzed.

airway occlusion (885 of 1,871) and 36% episodes of partial upper airway occlusion (73 of 204). Only 11 % of central apneas (30 of 265) were correctly identified, however, due mainly to the presence of small deflections «2 mm) on the paradox tracing, leading to misclassification of central apneas as obstructive. Positive predictive rates for obstructive and central apneas were 62% (885 of 1,439) and 11 % (30 of 267) respectively. When stricter definitions of apneas and hypoventilation were applied to analysis of the portable record, the sensitivity of the portable system to identify obstructive apneas correctly decreased to only 18% (239 of 1,283 episodes); the majority of obstructive apneas were either not detected (39%) due to continuing fluctuations on the tidal volume trace >50% resting tidal volume, were misclassified as central apnea (21%) due to only small «2 mm) deflections on the paradox trace, or were misclassified as hypoventilation (21%) due to small tidal volume deflections during apneas. Positive predictive rate for identification of episodes of upper airway occlusion was improved, however (86%), and use of the stricter criteria improved the sensitivity of the portable system in the identification of episodes of central apnea from 10 to 51%. Table 2 shows the range of sensitivities and positive predictive values obtained among subjects for obstructive and central apneas. Misclassification of apneas was not confined,to a few individuals. Despite this, the 10 patients with sleep apnea would have been correctly diagnosed as having respiratory disturbance during sleep by the portable system (Table 2). Respiratory disturbance index calculated from the polygraph record (mean ± SD, 43.8 ± 27.5) did not differ from the index calculated from the portable record when the manufacturer's criteria were applied to analysis (40.3 ± 24.7), and the measures were significantly correlated (r = 0.70, p < 0.01, Fig. 3). The portable system misclassified two patients as having significant respiratory disturbance during sleep (subjects EA. and J.S., Table 2). Sa02' The relation between %Sa0 2 recorded by the Grass polygraph and the portable system is shown in Fig. 4. Because both recording systems received the same signal from the Biox IIA oximeter, the tendency of the portable system to overestimate %Sa0 2 must reflect the way the system processes and stores the signal and/or its sub-

Sleep, Vol. 10, No.2, 1987

"'Uv

'";;;-

~

0\

~

--

""~ '"--

TABLE 2. Results of portable monitor in individual subjects

~

"

Disturbed ventilation No. of events a

Sensitivity

Pos. predict.

(%)

(%)

292 203

71 45 94 90 94 66 77 89 74 40 76 64

RDl Patient 1.S. D.S. E.M. G.P. S.C.

K.M.A.C EA. K.R.c D.C. 1.S. EL.c C.M.

BMI

Poly

Portable

25.2 35.0 25.9 25.0 24.4 34.2 28.1 25.6 36.0 24.0 22.3 29.0

77 55 26 62 74 15 5 56 79 3 42 31

25 24 40 71 82 16 15 41 75 13 54 27

Obstructiv,e apnea

72

303 262 77 22 468 262 20 216 143

82 72

25 74 82 48 27 84 76 12 58 38

Central apnea

No. of events a

Sensitivity

Pos. predict.

(%)

(%)

291 144 17 276 262 24 15 4S6 247 8 7 124

24 Ob

12 80 92

21 Ob

S6 24 13

43 23

86

N/A 17 68 83 25

N/A 77 78 7 I

44

No. of events a I I

6 27 0 4 0 12 2 3 209 0

Sensitivity

Pos. predict.

(%)

(%)

N/A N/A

N/A N/A

67 19

4 21

N/A

N/A

25

4

N/A N/A N/A

N/A N/A N/A N/A

10

37

N/A

N/A

0

BMI, body mass index «25 normal, >25 overweight, >30 obese); RDI, respiratory disturbance index (apneas and hypopneas per hour) calculated from sleep period without continuous positive airway pressure (CPAP); Pos predict., Positive predictive value (true positives/all positives recorded by portable system). Analysis of portable record according to manufacturer's criteria. In 2 of 14 subjects, monitor did not record. a Polygraph record. b All episodes detected by the portable monitor were classified as hypopneas. CStudied without CPAP. N/A = not applicable.

V:l

Cl

~

~

;:t.-

~

~

....,

;:t.t-