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 <50% resting tidal volume for> 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-<
HOME MONITORING OF NOCTURNAL BREATHING
137
100
...
80
:J
o
.r:.
<II
c:
~
60
Qj
FIG. 3. Relationship between respiratory disturbance index calculated from
portable and polygraph tracings. Solid
line is line of identity: y = 0.6lx + 14.7,
r = 0.7, p < 0.01.
W
..J
III
~
40
a:
o0..
o
a:
20
o
20
40
60
80
100
RDI POLYGRAPH (events/hour)
sequent retrieval. The portable system often failed to record any change in %Sa02
during brief apneas when %Sa02 did not fall below 90%.
Arousals. Activity was sensed by the wrist movement detector in 650 of 1,057
epochs (61%) when EEG arousal to wakefulness occurred. The portable monitor regularly detected movement arousals but missed brief arousals defined by EEG criteria.
Few false positives were recorded, and these were largely confined to one subject.
Daytime studies
Tidal volume. The relationship between tidal volume measured by integrating inspiratory flow at the mouth and the portable system (calibrated by rebreathing a known
volume) is shown in Fig. 5. The portable system was most accurate when tidal volume
was between 501-1,000 ml. Comparison of grouped tidal volumes revealed the following: 0-500 ml, pneumotachograph 460 ± 10 ml versus portable 550 ± 40 ml (p <
0.05, paired Student's t test); 501-1,000 ml, pneumotachograph 800 ± 80 ml versus
portable 790 ± 90 ml (not significantly different); and 1,001-1,500 ml, pneumotachograph 1,230 ± 80 ml versus portable 1,070 ± 100 ml (p < 0.02). No differences in
accuracy were observed between measurements made in the supine and left lateral
TABLE 3. Classification of events by polygraph and portable systems
Polygraph system
Portable system
Obstructive apnea
Central apnea
Hypoventilation
No disturbance
Obstructive
apnea
Central
apnea
Hypoventilation
885
56
556
374
1,871
167
30
11
57
265
12
38
73
81
204
No
disturbance
375
143
492
N/A
1,010
Manufacturer's criteria was used for analysis of portable record.
= not analyzed.
N/A
Sleep, Vol. 10, No.2, 1987
S. GYULAY ET AL.
138
100
1
90
FIG. 4. Relationship between polygraph and portable recordings of percentage of arterial oxygen saturation
(%Sa02)' Line of regression with
95% confidence limits is shown. y =
0.87x + 14.5 %Sa02, r = 0.92, p <
0.001.
80
70
60
o
60
70
80
90
100
GRASS POLYGRAPH
positions. Manual calibration of the portable system (i.e., calibration based on an estimate of tidal volume according to body weight) proved inaccurate and inconsistent.
Paradoxical thoracoabdominal motion. The portable monitoring system detected 89
of the 92 isovolume maneuvers performed. The system appeared more sensitive when
calibrated by volume than when calibrated manually (98 vs. 89%) but this difference
was not statistically significant. Detection rate by the portable system was similar
whether isovolume maneuvers resulted in small or large thoracoabdominal excursions.
During quiet breathing, there was frequent random variation of the portable system's
paradox tracing up to 2 mm from baseline. During isovolume maneuvers, deflections in
the portable tidal volume tracing occurred frequently.
Heart rate. The relationship between heart rate calculated from the ECG and that
recorded by the portable system showed good agreement (y = 1.1x - 9.5; r = 0.95, p
< 0.01).
Activity monitor. The portable system detected 208 of 296 deliberate movements
(70%). Brisk movements were detected more frequently than slow movements.
Calibration bags. Various bags are used in the calibration of the portable monitor's
tidal volume tracing. The accuracy of bag size was tested by measuring the volume of
bags of varying sizes, each constructed according to the manufacturer's instructions,
and comparing measured volume with that predicted from the instructions. Twelve
bags were tested. Mean measured volume was 936 ± 110 ml; predicted volume was
850 ± 124 ml (p < 0.001, paired Student's t test).
DISCUSSION
The microprocessor-based portable monitoring system tested appears sufficiently
sensitive to be used as a screening instrument for detection of breathing disorders
during sleep. In 12 sleep apnea patients, 78% of episodes of disturbed breathing were
detected by the portable system, but the recorded information was not sufficient to
Sleep, Vol. 10. No.2, 1987
HOME MONITORING OF NOCTURNAL BREATHING
1500
.
.
. ""
139
/
",/
...
/
• o.~
1000
.
e. ,.
...'"'
500
•
0
0
0
..
a
1000
1500
TIDAL VOLUME (PNEUMOTACHOGRAPH) MLS.
1500
FIG. 5. Relationship between pneumotachograph and
portable recordings of tidal volume. Closed symbols,
supine; open symbols, left lateral position. Upper panel:
Comparison of mean tidal volumes computed from 30-s
epochs of breathing. Each point is one epoch. Regression line is shown: y = 0.67x + 280 ml, r = 0.84, p <
0.01. Lower panel: Comparison of mean tidal volume
computed after each breath was grouped according to
size (described in text). Different symbols show different experimental runs. Regression line is shown: y =
0.68x + 237 ml, r = 0.96, p < 0.01.
1000
500
1000
1500
TIDAL VOLUME (PNEUMOTACHOGRAPH) MLS.
classify events with confidence. Adoption of the manufacturer's criteria to define
apneas and hypoventilation optimized the sensitivity of the system to identify episodes
of OSA but led to misclassification of central apnea and low positive predictive value,s.
Application of stricter criteria to define breathing disturbances improved the ability of
the system to identify central apneas correctly and reduced false-positive d~tection
rates but led to many episodes of obstructive apnea being missed. Incorrect classification of apneic episodes was caused by two problems: small «2 mm) changes in the
paradox baseline, in part due to random noise, leading to misclassification of central
apneas; and ,errors in measurement of rib cage and abdominal volume during episodes
of upper airway occlusion leading to the recording of apparent tidal volume change
during apneas. This latter error, which was evident in awake studies of paradoxical
thoracoabdominal motion, may relate either to the operator's imperfect calibration of
rib cage and abdominal signals such that during an isovolume maneuver the sum of the
signals is not zero or to an error in the algorithm that is used to detect and measure
thoracoabdominal movement. This error is presumably the reason for the manufacturer's recommendation to record episodes during which tidal volume is less than onethird of baseline as "apnea." Nevertheless, respiratory disturbance indices calculated
Sleep, Vol. 10, No.2, 1987
140
S. GYULAY E1 AL.
from the portable record correlated well with those calculated from the standard polysomnogram.
The portable monitor misdiagnosed two subjects as having a mild disorder but correctly identified all 10 patients with sleep apnea. One must be cautious when generalizing findings to the population at large, however. First, our sample was biased. We
sought to study patients known to have repeated episodes of complete occlusion of the
upper airway during sleep and we did not have many episodes of partial upper airway
obstruction leading to hypoventilation to analyze. Similarly, our subjects did not have
many central apneas, and we studied only one subject with pure central sleep apnea
syndrome. None of our subjects was massively obese, and all subjects were used to
sleeping with monitoring equipment in place. Second, thoracoabdominal movement
was measured by two different systems. Two inductance systems could not be used
because of electrical interference. Differences per se in sensitivity between magnetometry and inductance plethysmography may explain, in part, differences in detection
rates of episodes of disturbed breathing, although this seems unlikely given the findings
of our awake studies. Third, we studied most of our subjects with nCPAP in place for
half the night to examine the false-positive detection rate of the portable monitor. Because nCPAP may have altered lung volume (albeit only slightly), false-positive detection rates may have differed in the absence of nCPAP. Last, we used polysomnography
as our "gold standard," yet errors may occur in distinguishing between central and
obstructive events using this technique (20). Although our findings show no difference
between RDI computed from the portable record and routine polysomnography, a
much larger study of subjects with a wide range of respiratory disturbance index (from
zero to severe) is needed before the sensitivity and specificity of the portable monitoring system in the diagnosis of sleep apnea can be determined.
The portable system overestimated Sa02, particularly at lower levels. The error was
not large and amounted to <10% of the true reading at Sa0 2 values >60%. Because
both recording systems received a common oximeter signal, an error must lie in the
processing, storage, and/or retrieval of data by the portable system. It is relevant that
the portable system frequently missed brief episodes of arterial de saturation when
Sa02 did not fall below 90%. There were no major false-positive or false-negative
findings with the oximetry data, however, and when the oximetry data were considered
with the breathing measurements, a clear picture of the severity of the ventilatory
disturbance was given by the portable system in all patients. It should be emphasized
that we did not test the Vitalog oximeter and thus cannot assess its accuracy.
The portable monitoring system we tested appears to have accuracy in the detection
of breathing disorders during sleep similar to that of the portable analogue-recording
device described by Ancoli-Israel et al. (13). For example, the correlation between
apnea index computed from polysomnogram records and simultaneous portable
records reported by Ancoli-Israel et al. (13) is very similar to the correlation between
respiratory disturbance indices found in the present study (r = 0.80 and 0.70, respectively). In addition, neither study showed a significant difference between indices computed from the laboratory and portable records.
Our findings also support those of Nino-Mercia et al. (16) and Walker et al. (17), who
compared polysomnography with the portable monitoring system used in the present
study and showed RDI calculated from polygraph and portable records to be well correlated. Our finding of low sensitivity of the portable monitor for central apnea sup-
Sleep, Vol. 10, No.2, 1987
HOME MONITORING OF NOCTURNAL BREATHING
141
ports the findings of Walker et al. (17), but, in contrast to the present study, those
authors also found that the portable system overestimated obstructive apneas and RDI.
We did not assess the portable monitoring system's accuracy in the scoring of sleep
apart from showing that the wrist sensor underestimated the number of arousals to
wakefulness. Previous studies using a similar wrist device have shown that sleep parameters derived from measurement of wrist activity are well correlated with parameters derived from standard EEG records, although total sleep time may be overestimated (21) and wakefulness after sleep onset underestimated (13) when wrist activity is
used to score sleep.
Comparison of tidal volume measurements by the portable system with measurements made by integrating airflow at the mouth revealed that the portable monitor
provided only approximate data. Accuracy was best when tidal volume was <900 ml,
and no significant differences were found between portable monitor and pneumotachograph measurements of tidal volumes between 501 and 1,000 ml. Although no differences were found between measurements made in the supine and left lateral positions,
others using a similar inductance plethysmographic method to measure tidal volume
have shown changes in calibration with changes in position (22). These findings, together with the differences found between measured and predicted volumes of the reference calibration bags supplied with the portable system, suggest that absolute measurements of tidal volume should be interpreted cautiously.
The portable monitor and equipment is easy to apply and comfortable to wear. Body
netting worn over ECG electrodes, respiration bands, and leads keep them firmly in
place. The pouches that contain the monitors can be slipped along the belt to accomodate different sleeping positions without causing discomfort. The device is also fairly
simple to maintain. A record of battery usage, both for rechargeable and nonrechargeable batteries, must be kept. Respiration bands may be washed, as may the pouches in
which the monitor is placed. Set-up time is -20 min, including application of the device and calibration. The instructions supplied with the system are clear and easy to
read. The menu-driven programming is friendly and has not caused us any problems. It
is easy to learn, even for a computer novice.
We conclude that this microprocessor-based portable monitoring system is sufficiently sensitive to allow detection of patients with disordered breathing during sleep,
but further developments are needed before the system can be relied on to classify
apneas and hypoventilation accurately. The development of precise analysis software
will greatly facilitate the study of large population samples.
Acknowledgment: This work was supported by grants from the National Health and Medical
Research Council and the Royal Australian College of General Practitioners. Joy Peate and
Brenda Whyte provided expert secretarial assistance.
REFERENCES
1. Lugaresi E, Cirignotta F, Coccagna G, Piana C. Some epidemiological data on snoring and cardiocirculatory disturbances. Sleep 1980;3:221-4.
2. Carskadon MA, Dement WC. Respiration during sleep in the aged human. J Gerontol 1981 ;36:420-3.
3. Franceschi M, Zamproni P, Crippa D, Smime S. Excessive daytime sleepiness: a one-year study in an
un selected inpatient population. Sleep 1982;5:239-47.
4. Lavie P. Incidence of sleep apnea in a presumably healthy working population: a significant relationship
with excessive daytime sleepiness. Sleep 1983;6:312-8.
5. Koo KW, Sax DS, Snider GL. Arterial blood gases and pH during sleep in chronic obstructive pulmonary disease. Am J Med 1975;58:663-70.
Sleep, Vol. 10, No.2, 1987
142
S. GYULAY ET AL.
6. Wynne JW. Block AJ, Hemenway J, Lynn A, Hunt LA, Flick M. Disordered breathing and oxygen
de saturation during sleep in patients with chronic obstructive lung disease (COLD). Am j ,'vIed
1979;66:573-9.
7. Douglas NJ, Calverley PMA, Leggett RJE, Brash HM, Flenley DC, Brezinova V. Transient hypoxaemia
during sleep in chronic bronchitis and emphysema. Lancet 1979;1:1-4.
8. Guilleminault C, Cummiskey J, Motta J. Chronic obstructive airflow disease and sleep studies. Am Rev
Respir Dis 1980;122:397-406.
9. Muller NL, Francis PW, Gurwitz 0, Levison H, Bryan AC. Mechanism of hemoglobin desaturation
during rapid eye movement sleep in normal subjects and in patients with cystic fibrosis. Am Rev Respir
Dis 1980;121:463-9.
10. Mezon BL, West P, Israels J, Kryger M. Sleep breathing abnormalities in kyphoscoliosis. Am Rev Respir
Dis 1980;122:617-21.
11. Guilleminault C, Kurland G, Winkle R, Miles LE. Severe kyphoscoliosis, breathing and sleep. Chest
1981 ;79:626-30.
12. Bye PT, Issa F, Berthon-Jones M, Sullivan CEo Studies of oxygenation during sleep in patients with
interstitial lung disease. Am Rev Respir Dis 1984;129:27-32.
13. Ancoli-Israel S, Kripke D, Mason W, Messin S. Comparisons of home sleep recordings and polysomnograms in older adults with sleep disorders. Sleep 1981;4:283-91.
14. Miles L, Rule B, Benson K, Herakan S, Dement WC. Screening outpatients for sleep apnea: continuous
measurement of tidal volumes and other parameters using the Vitalog portable microcomputer [Abstract]. Sleep Res 1982;11:204.
15. Benson K, Miles LE, Rule RB, Miles SC, Dement We. Monitoring tidal volume, respiratory effort,
ECG, blood oxygen saturation and body movement with the Vitalog PMS-8 portable microcomputer
[Abstract]. Sleep Res 1983;12:342.
16. Nino-Mercia G, Bliwise D, Keenan S, et al. Respiration monitoring in sleep: comparison of judgements
based on conventional polysomnography (PSG) and an ambulatory, microprocessor-derived recording
(AMR) [Abstract]. Presented at ASDC/SRB Meeting, Seattle, 1985.
17. Walker LE, Walker JM, Farney RJ, Kramer J: A comparison of polysomnography with a portable home
monitoring system in the detection of sleep apnea (Abstract). Am Rev Respir Dis 1986;133:A54.
18. Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring system
for sleep stages of human subjects. Los Angeles: University of California, Brain Information Servicel
Brain Research Institute, 1968.
19. Guilleminault C, Cummins key J, Dement WC. Sleep apnea syndrome: recent advances. Adv Intern Med
1980;26:347-74.
20. Bliwise 0, Bliwise NG, Kraemer HC, Dement W. Measurement error in visually scored electrophysiological data: respiration during sleep. J New'osci Methods 1984;12:49-56.
21. Mullaney DJ, Kripke OF, Messin S. Wrist-actigraphic estimation of sleep time. Sleep 1980;3:83-92.
22. Zimmerman PV, Connellan SJ, Middleton HC, Tabona MV, Goldman MD, Pride N. Postural changes in
rib cage and abdominal volume-motion coefficients and their effect on the calibration of a respiratory
inductance plethysmograph. Am Rev Respir Dis 1983;127:209-14.
Sleep, Vol. 10, No.2, 1987