Positional vs Nonpositional Obstructive Sleep Apnea Patients: Anthropomorphic, Nocturnal Polysomnographic, and Multiple Sleep Latency Test Data

Article (PDF Available)inChest 112(3):629-39 · September 1997with97 Reads
DOI: 10.1378/chest.112.3.629 · Source: PubMed
Abstract
To compare anthropomorphic, nocturnal polysomnographic (PSG), and multiple sleep latency test (MSLT) data between positional (PP) and nonpositional (NPP) obstructive sleep apnea (OSA) patients. This is a retrospective analysis of anthropomorphic, PSG, and MSLT data of a large group of OSA patients who underwent a complete PSG evaluation in our sleep disorders unit. The patients were divided in two groups: the PP group, those patients who had a supine respiratory disturbance index (RDI) that was at least two times higher than the lateral RDI, and the NPP group, those patients in whom the RDI in the supine position was less than twice that in the lateral position. From a group of 666 consecutive OSA patients whose conditions were diagnosed in our unit from September 1990 to February 1995, 574 patients met the following criteria and were included in the study: RDI > 10; age > 20 years, and body mass index (BMI) > 20. Of all 574 patients, 55.9% were found to be positional. No differences in height were observed but weight and BMI were significantly higher in the NPP group, these patients being on the average 6.5 kg heavier than those in the PP group. The PP group was, on average, 2 years younger than the NPP group. Nocturnal sleep quality was better preserved in the PP group. In this group, sleep efficiency and the percentages of deep sleep (stages 3 and 4) were significantly higher while the percentages of light sleep (stages 1 and 2) were significantly lower than in the NPP group. No differences for rapid eye movement (REM) sleep were found. In addition, wakefulness after sleep onset and the number of short arousals (< 15 s) were significantly lower in the PP group. Apnea index and total RDI were significantly higher and the minimal arterial oxygen saturation in REM and non-REM sleep was significantly lower in the NPP. No differences in periodic limb movements data were found between the two groups. The average MSLT was significantly shorter in the NPP group. Univariate and multivariate stepwise logistic regression analysis showed that the most dominant variable that correlates with positional dependency in OSA patients is RDI, followed by BMI which also adds a significant contribution to the prediction of positional dependency. Age, although significant, adds only a minor improvement to the prediction of this positional dependency phenomenon. A severe, obese, and older OSA patient is significantly less likely to be positional than a mild-moderate, thin, and young OSA patient. In four obese OSA patients who lost weight, a much more pronounced reduction was seen in the lateral RDI than in the supine RDI, and three of these cases who were previously NPP became PP. In a large population of OSA patients, most were found to have at least twice as many apneas/hypopneas in the supine than in the lateral position. These so-called "positional patients" are on the average thinner and younger than "nonpositional patients." They had fewer and less severe breathing abnormalities than the NPP group. Consequently their nocturnal sleep quality was better preserved and, according to MSLT data, they were less sleepy during daytime hours. RDI was the most dominant factor that could predict the positional dependency followed by BMI and age. RDI showed a threshold effect, the prevalence of PP in those with severe RDI (RDI > or = 40) was significantly lower than in those OSA patients with mild-moderate RDI. BMI showed a major significant inverse relationship with positional dependency, while age had only a minor although significant inverse relationship with it. Body position during sleep has a profound effect on the frequency and severity of breathing abnormalities in OSA patients.
Positional vs Nonpositional Obstructive
Sleep Apnea Patients*
Anthropomorphic, Nocturnal Polysomnographic,
and Multiple Sleep Latency Test Data
Arie Oksenberg, PhD; Donald S. Silverberg , MD; Elena Arons, PhD; and
Henryk Radwan, MD
Study objectives: To compare anthropomorphic, nocturnal polysomnographic (PSG), and multiple sleep
latency test (MSLT) data between positional (PP) and nonpositional (NPP) obstructive sleep apnea (OSA)
patients.
Design: This is a retrospective analysis of anthropomorphic, PSG, and MSLT data of a large group of OSA
patients who underwent a complete PSG evaluation in our sleep disorders unit. The patients were divided
in two groups: the PP group, those patients who had a supine respiratory disturbance index (RDI) that was
at least two times higher than the lateral RDI, and the NPP group, those patients in whom the RDI in the
supine position was less than twice that in the lateral position.
Subjects: From a group of 666 consecutive OSA patients whose conditions were diagnosed in our unit from
September 1990 to February 1995, 574 patients met the following criteria and were included in the study:
RDI>10; age>20 years, and body mass index (BMI)>20.
Results: Of all 574 patients, 55.9% were found to be positional. No differences in height were observed but
weight and BMI were significantly higher in the NPP group, these patients being on the average 6.5 kg
heavier than those in the PP group. The PP group was, on average, 2 years younger than the NPP group.
Nocturnal sleep quality was better preserved in the PP group. In this group, sleep efficiency and the
percentages of deep sleep (stages 3 and 4) were significantly higher while the percentages of light sleep
(stages 1 and 2) were significantly lower than in the NPP group. No differences for rapid eye movement
(REM) sleep were found. In addition, wakefulness after sleep onset and the number of short arousals
(<15 s) were significantly lower in the PP group. Apnea index and total RDI were significantly higher and
the minimal arterial oxygen saturation in REM and non-REM sleep was significantly lower in the NPP. No
differences in periodic limb movements data were found between the two groups. The average MSLT was
significantly shorter in the NPP group. Univariate and multivariate stepwise logistic regression analysis
showed that the most dominant variable that correlates with positional dependency in OSA patients is RDI,
followed by BMI which also adds a significant contribution to the prediction of positional dependency. Age,
although significant, adds only a minor improvement to the prediction of this positional dependency
phenomenon. A severe, obese, and older OSA patient is significantly less likely to be positional than a
mild-moderate, thin, and young OSA patient. In four obese OSA patients who lost weight, a much more
pronounced reduction was seen in the lateral RDI than in the supine RDI, and three of these cases who
were previously NPP became PP.
Conclusions: In a large population of OSA patients, most were found to have at least twice as many
apneas/hypopneas in the supine than in the lateral position. These so-called “positional patients” are on the
average thinner and younger than “nonpositional patients.” They had fewer and less severe breathing
abnormalities than the NPP group. Consequently their nocturnal sleep quality was better preserved and,
according to MSLT data, they were less sleepy during daytime hours. RDI was the most dominant factor
that could predict the positional dependency followed by BMI and age. RDI showed a threshold effect, the
prevalence of PP in those with severe RDI (RDI>40) was significantly lower than in those OSA patients
with mild-moderate RDI. BMI showed a major significant inverse relationship with positional dependency,
while age had only a minor although significant inverse relationship with it. Body position during sleep has
a profound effect on the frequency and severity of breathing abnormalities in OSA patients.
(CHEST 1997; 112:629-39)
Key words: body posture; breathing disturbances; human sleep; obstructive sleep apnea; polysomnography; sleep disorders;
sleep position
Abbreviations: AI5apnea index; ANOVA5analysis of variance; A-P5anteroposterior; BMI5body mass index; Min
SaO
2
-REM5minimum SaO
2
level during REM sleep; Min SaO
2
-NREM5minimum SaO
2
level during non-REM sleep;
MSLT5multiple sleep latency test; nCPAP5nasal continuous positive airway pressure; NREM sleep5non-rapid eye
movement sleep; NPP5nonpositional patients; OSA5obstructive sleep apnea; PLM5periodic limb movements; PLM
I5periodic limb movements index; PLM AI5periodic limb movements arousal index; PP5positional patients;
PSG5polysomnographic; RDI5respiratory disturbance index; REM sleep5rapid eye movement sleep; SaO
2
5arterial
oxygen saturation; TST5total sleep time; UA5upper airway; UARS5upper airway resistance syndrome
CHEST / 112/3/SEPTEMBER, 1997 629
I
n patients with obstructive sleep apnea (OSA), the
level of respiratory distress during sleep, as judged
by the apnea/hypopnea index or respiratory distur-
bance index (RDI), is on average about 40 to 50%
lower when they lie on their side than when they lie
on their back (ie, in the supine position).
1-9
Cart-
wright
2
and Lloyd and Cartwright
4
defined “posi-
tional patients” (PP) as those OSA patients in whom
the RDI was at least twice as high in the supine
position as in the lateral position. In fact, the degree
of severity of OSA in these patients is mostly related
to the sleep time spent or not spent in the supine
position. Those patients in whom the RDI in the
supine position was less than twice that in the lateral
position were called “nonpositional patients” (NPP).
The percentage of PP in OSA patients varies in
different reports
2,4,10-14
from 9%
11
to 60%.
4
This
variation is probably due to the small numbers and
the different types of OSA patients studied. Some
patients have succeeded in lowering their total RDI
to normal by merely sleeping on their sides
2,10,14-19
and it has been estimated that this type of therapy
alone could be successful in treating about 50% of all
OSA cases.
4,20
Since OSA is present in about 9% of
men and 4% of women in the middle-age population
(by using a RDI$15),
21
the implications of such an
approach are obvious. Little information exists about
the relationship of positional dependency to the
physical characteristics of the OSA patients, the
quality of nocturnal sleep, and the level of daytime
somnolence.
The aim of this report is to compare anthropomor-
phic, nocturnal polysomnographic (PSG), and mul-
tiple sleep latency test (MSLT) data between the PP
and NPP group in 574 consecutive OSA patients
whose conditions were diagnosed in our Sleep Dis-
orders Unit.
Materials and Methods
All the patients were referred to the Sleep Disorders Unit at
the Loewenstein Hospital-Rehabilitation Center because of snor-
ing complaints and/or a suspicion of OSA from September 1990
to February 1995. During this period, 666 consecutive patients
were diagnosed as having OSA (RDI.10). Of these, 574 patients
who were older than 20 years, had a body mass index (BMI).20,
RDI.10, and slept more than 30 min in either the supine or the
lateral position were included in the analysis. These patients were
divided into PP and NPP according to the criterion of Cart-
wright.
2
For that purpose, in addition to the overall RDI, supine
and lateral RDI values were calculated for each patient. The
supine and lateral RDI data define to which group (PP or NPP)
each patient belongs.
The patients arrived at the sleep unit around 8 pm and the PSG
recordings usually began between 10 pm and midnight.
The PSG recordings were carried out using polygraphs (Nihon
Kohden models 4321 and 4414; Tokyo, Japan) and included the
following parameters: electro-oculogram (two to four channels);
EEG (four to six channels); electromyogram of submental mus-
cles (one to two channels); ECG (one channel); electromyogram
of the anterior tibialis muscle of both legs (two channels); and
airflow (with a nasal/oral thermistor; Nihon Kohden). chest and
abdominal effort (two channels) was recorded using inductive
plethysmography (Respitrace; Ambulatory Monitoring Inc; Ard-
sley, NY; or Resp-Ez breathing belts; Tel Aviv, Israel); arterial
oxygen saturation (SaO
2
) levels (one channel) by pulse oximetry
(Ohmeda 37000e; Boulder, Colo) with a finger probe, and audio
(one channel) by a microphone located above the patient’s head
at a distance of 1 m and connected to a sound level meter (Quest
Electronics model 2700; Oconomowoc, Wis), were recorded. The
output from the sound level meter was also recorded in parallel
on a calibrated (40 to 80 dB) chart recorder at a paper speed of
10 cm/h.
The recordings were carried out at a paper speed of 10 mm/s
and sleep stages were scored according to the standard criteria of
Rechtschaffen and Kales.
22
The PSG technician who followed the patient’s behavior
through a closed-circuit 21-inch TV monitor marked the changes
in body position in two places simultaneously, on the polygraph
and on the chart recorder which registered the output of the
pulse oximeter data.
The PSG technician was responsible for the monitoring of one
or two sleeping patients. The two TV monitors were placed side
by side to allow easy visualization of all patients’ body move-
ments. Since our unit is especially interested in the effect of body
position on sleep-related breathing disturbances, our PSG tech-
nicians are encouraged to pay special attention to this issue.
Apnea was defined as an episode of a complete breathing
cessation of $10 s. Hypopneas were considered as such if a
partial breathing cessation (.20% reduction in oral/nasal airflow
compared with the level of the previous five breaths) occurred,
accompanied by a drop of SaO
2
of at least 3%.
Apnea index (AI) and RDI were calculated as the number of
apneas per sleep hour and the number of apneas1hypopneas per
sleep hour, respectively.
23
Arousals were divided as shorter or
longer than 15 s and were scored according to accepted defini-
tions.
24
Periodic limb movements (PLM), PLM index (PLM I), and
PLM arousal index (PLM AI) were scored and calculated
according to Coleman.
25
PLM events associated with breathing
abnormalities were not taken for the analysis.
The MSLT was carried out on the basis of published guide-
lines
26
and included four naps at 9am,11am,1pm, and 3pm on
the day after the nocturnal PSG evaluation. The MSLT was
performed in all patients who complained about daytime sleep-
iness. BMI is weight (kg)/height (m)
2
. In order to assess the effect
of weight loss on positional dependency, four obese OSA patients
(average BMI533.6) who refused nasal continuous positive
airway pressure (nCPAP) treatment had a PSG evaluation before
and after they lost weight in a dietary weight reduction program.
Data Analysis
For the comparison of the different anthropomorphic, sleep,
and breathing parameters between the PP and the NPP group,
the data were analyzed using the two-sample Student ttest and
Bonferroni correction was used for multiple ttests. For the
MSLT data, two-way analysis of variance (ANOVA) with re-
peated measurements was performed.
*From the Sleep Disorders Unit (Drs. Oksenberg, Arons, and
Rodwan), Loewenstein Hospital Rehabilitation Center,
Raanana, Israel, and the Department of Nephrology (Dr. Sil-
verberg), Tel-Aviv Medical Center, Tel-Aviv, Israel.
Manuscript received May 31, 1996; revision accepted February
27, 1997.
Reprint requests: Arie Oksenberg, PhD, Sleep Disorders Unit,
Loewenstein Hospital Rehabilitation Center, POB 3 Raanana,
Israel
630 Clinical Investigations
To estimate how RDI, BMI, and age were related to the body
position dependency, the statistical analysis was performed in two
steps; first, a univariate analysis (x
2
test for categorical variables)
was carried out, and subsequently a stepwise multivariate logistic
regression analysis was performed to assess the simultaneous
contribution of all three parameters on positional dependency.
All statistical analyses were performed with a statistical software
package (SPSS version 6.08; SPSS, Inc; Chicago).
Results
Of the 574 OSA patients, 321 (55.9%) were PP
and the other 253 (44.1%) were NPP.
Anthropomorphic Data
The mean age difference between the two groups
was 2 years and was significant (p50.02), the PP
group being younger than the NPP group (Table 1).
The mean height was not significantly different in
the two groups.
The BMI was significantly greater in the NPP
group (p50.001) and this was due to the significantly
higher weight in this group (p50.003); the average
weight in this group was 6.5 kg more than in the PP
group.
Nocturnal PSG Data
Table 2 shows data on the comparison of various
nocturnal sleep parameters in the two groups of
patients. After Bonferroni correction, the signifi-
cance level became p50.003 instead of p50.05.
No significant differences between both groups
were seen for total recording time (TRT) (p50.007).
Nevertheless, total sleep time (TST) and sleep effi-
ciency (S EFF) were significantly higher in the PP
group (p50.001) for both parameters. Sleep laten-
cies (both to stage 1 and to persistent sleep) were not
significantly different between the groups. REM
sleep latency from onset of persistent sleep (REM
LAT) and REM latency without intervening wake
time (REM LAT w/o AW) were not significantly
different between the groups (p50.004 and
p50.057, respectively). The number of REM peri-
ods and the average length of REM periods were
also not different between the groups (p50.005 and
p50.110, respectively). However, the percentages of
all non-REM sleep stages (out of TST) differed
significantly in the two groups. The PP group had
significantly lower percentages of the lighter sleep
stages (stage 1, p50.001, and stage 2, p50.001) and
higher percentages of deeper sleep (stages 3 and 4,
p50.001). For REM sleep percentages, no statistical
differences were obtained (p50.03). Also, the dura-
tion of wakefulness after sleep onset (WASO) and
the number of long arousals (.15 s) were not
significantly different between the two groups. How-
ever, the number of short arousals (,15 s) was
significantly greater in the NPP group (p50.001).
Table 3 shows a comparison of breathing param-
eters in the two groups of patients. All these param-
eters (AI, RDI, Min SaO
2
-REM, Min SaO
2
-NREM)
were significantly different (p50.001) in the two
groups. In the NPP group, all four parameters
showed a greater degree of abnormality.
Figure 1 shows an example of the SaO
2
and heart
rate recordings during sleep in a typical PP. The
differences in the frequency of SaO
2
desaturations
and bradycardia/tachycardia episodes while sleeping
on the back compared to the absence of these events
while sleeping on the right side is clearly evident.
Table 4 shows a comparison of PLM data in the
two groups of patients. In the PP group, 115 patients
(35.8%) had PLM as a secondary diagnosis com-
pared to 79 (31.2%) of the NPP group. This differ-
ence was not statistically significant (p50.25). No
significant differences were found between the two
groups, either in the total number of PLMs, the
number of PLMs causing arousals, the PLM I, or in
the PLM AI.
Table 1—Anthropomorphic Data*
PP
(n5321)
NPP
(n5253) p Value
Age, yr 52.9 (10.4) 54.9 (10.1) 0.020
Weight, kg 85.7 (14.0) 92.2 (15.8) 0.003
Height, cm 170.4 (8.8) 170.1 (8.5) 0.701
BMI 29.4 (4.1) 31.9 (4.9) 0.001
*Values are mean (SD).
Table 2—Nocturnal PSG Data*
PP (n5321) NPP (n5253) p Value
TRT, min 422.4 (56.8) 407.4 (76.9) 0.007
TST, min 353.6 (66.4) 327.6 (78.1) 0.001
S EFF, % 83.4 (11.0) 80.1 (12.4) 0.001
LAT STG 1, min 12.9 (18.7) 11.7 (15.5) 0.410
LAT Perst Sleep, min 16.0 (21.3) 16.1 (19.7) 0.940
REM LAT, min 89.6 (46.6) 104.1 (63.7) 0.004
REM LAT w/o AW, min 78.0 (35.3) 85.5 (47.8) 0.057
No. of REMs 3.6 (1.5) 3.3 (1.4) 0.005
REM length, min 25.7 (20.8) 23.1 (15.3) 0.110
% STG 1 5.4 (4.2) 7.5 (7.0) 0.001
% STG 2 55.0 (9.9) 61.3 (13.3) 0.001
% STG 3 5.2 (3.1) 4.2 (3.0) 0.001
% STG 4 12.9 (8.3) 9.7 (8.6) 0.001
% STG 314 18.3 (10.1) 14.0 (10.1) 0.001
% REM 19.1 (7.4) 17.5 (8.5) 0.030
WASO, min 53.6 (35.0) 64.1 (45.7) 0.004
No. of arousals .15 s 33.3 (20.3) 38.1 (48.7) 0.160
No. of arousals ,15 s 159.2 (92.2) 209.6 (139.5) 0.001
*TRT5total recording time; S EFF5sleep efficiency; REM
LAT5REM latency from onset of persistent sleep (the first 10 min
in which at least 8 of them were sleep); REM LAT w/o AW5REM
latency without intervening awake time; STG5stage;
WASO5wakefulness after sleep onset. Values are mean (SD).
Significant differences after Bonferroni correction.
CHEST / 112/3/SEPTEMBER, 1997 631
MSLT Data
In the PP group, 194 patients (60.4%) had an
MSLT compared to 175 patients (69.2%) of the NPP
group (Table 5). This difference was statistically
significant (p50.03). Two-way ANOVA analysis
showed that the MSLT data difference between the
two groups was of borderline significance (p50.054).
Nevertheless, the average sleep latency for all four
naps was significantly shorter in the NPP group than
in the PP group (p50.01). In addition, it should be
noted that for each of the four naps, the sleep latency
was consistently longer in the PP than in the NPP
group.
Effect of RDI on Positional Dependency
In order to estimate the influence of OSA severity,
as expressed by RDI, on the positional dependency,
the entire group of OSA patients was first divided
into four different RDI categories (10 to 19.9, 20 to
29.9, 30 to 39.9, .40) and the prevalence of PP in
each category was calculated. But before that, since
the sleep time spent in the supine position is a major
factor correlating with RDI, we evaluated the sleep
time in the supine position for the four RDI catego-
ries in a random sample of 20 patients in each group.
No significant differences for sleep time in the
supine position were found among the four groups
(p50.35). The mean sleep time spent (min6SD)
in the supine position for the four RDI categories
was 131.0672.1, 165.4695.8, 148.2682.6, and
174.4681.4 min, respectively. The PP prevalence
remained high and fairly steady (between 65.1% and
69.0%) in the mild-moderate categories (RDI 10 to
19.9, 20 to 29.9, and 30 to 39.9), but showed a
marked and significant reduction to 32.4% in the
most severe category (RDI.40), (x
2
558.8, df53,
p50.001; Table 6). Although a positive trend toward
an inverse relationship was obtained (Kendall test),
this result suggests that rather than an overall inverse
relationship with positional dependency, RDI
showed a threshold effect on positional dependency
with a significant decrease in the prevalence of PP in
the most severe RDI category. A test for the identi-
fication of the threshold was carried out and the RDI
threshold point that maximized the x
2
test for the
positional dependency was RDI540. The entire
group was then divided into a nonsevere category
(RDI#40) and a severe category (RDI.40). In the
nonsevere category, the PP prevalence was 66.6%
compared with only 32.4% in the severe category
(x
2
558.38, df51, p50.0001; Fig 2).
Figure 1. Effect of body position on obstructive sleep apnea. Heart rate (HR), bottom, and SaO
2
tracing, top, on a chart recorder at paper speed of 10 cm/h in a typical positional OSA patient (PP). In
the SaO
2
tracing, each peak represents an episode of decreased oxygen saturation (desaturation) and
the return to baseline (resaturation) as a consequence of apneas and/or hypopneas. Note the frequent
desaturation-resaturation episodes in parallel with bradycardia/tachycardia changes in the HR tracing
while sleeping on the back and their absence while sleeping on the right side. This patient achieved
complete relief of breathing abnormalities during sleep by merely avoiding the supine position.
Table 3—Breathing Abnormalities Data*
PP
(n5321)
NPP
(n5253) p Value
AI 13.7 (15.1) 26.5 (29.4) 0.001
RDI 27.8 (17.7) 44.0 (29.7) 0.001
Min SaO
2
REM 81.1 (11.0) 72.7 (15.8) 0.001
Min SaO
2
NREM 84.7 (6.2) 81.5 (9.7) 0.001
*Values are mean (SD).
632 Clinical Investigations
Thus, an OSA patient with a severe RDI is less
likely to be positional than an OSA patient with a
mild to moderate RDI.
Effect of BMI on Positional Dependency
To estimate the correlation of BMI with positional
dependency in OSA patients, the entire group was
first divided into five different categories (20 to 24.9;
25 to 29.9; 30 to 34.9; 35 to 39.9; and $40) and the
percentage of PP in each category was calculated. A
steady, marked, and significant reduction was ob-
served in the prevalence of PP with the increase in
BMI in the five categories (70.5, 67.6, 46.3, 34.8, and
33.3%, respectively); x
2
542.2, df54, p50.001; Ta-
ble 6). In addition, when the total group was divided
into two categories, nonobese (BMI#30) and obese
(BMI.30), the PP prevalence in the nonobese
group was 68.0% compared to only 42.2% in the
obese group (x
2
538.61, df51, p50.0001; Fig 2).
Thus, BMI showed an inverse relationship to
positional dependency and a nonobese OSA patient
is more likely to be positional than an obese one.
Effect of Weight Loss on Positional Dependency
The data from four severely obese OSA patients
who refused nCPAP treatment but successfully lost
weight by changing their eating habits are summa-
rized in Table 7. All three NPP cases (patients 1
through 3) were converted into PP cases by weight
loss. In all four cases, after weight reduction, the
RDI fell to normal (RDI,10) while sleeping in the
lateral position. However, the RDI still remained
elevated after weight loss in all four while sleeping in
the supine position, and in one case (case 3), the
supine RDI actually increased despite the weight
loss. The weight reduction was associated with a
91.1% reduction in the RDI in the lateral position
but with only a 38.9% reduction in the supine
Figure 2. Effect of RDI, BMI and age on the prevalence of OSA positional patients (PP). The percent
of PP is significantly higher in OSA patients with RDI ,40, with BMI ,30, and in OSA patients
younger than 60 years old. Asterisk indicates p50.0001; two asterisks, p50.081.
Table 4PLM Data*
PP (n5115) NPP (n579) p Value
Total PLM 129.3 (134.9) 136.3 (122.0) 0.71
Arous PLM 62.8 (71.8) 71.0 (75.0) 0.44
PLM I 23.0 (25.0) 25.7 (24.0) 0.44
PLM AI 11.5 (13.7) 13.9 (14.4) 0.25
*All values are mean (SD). Arous PLM5number of PLM causing
arousals.
Table 5—MSLT Data*
PP
(n5194) NPP (n5175) p Value
Nap 1 9.9 (6.3) 8.6 (6.1)
}
Nap 2 9.0 (5.9) 7.9 (5.8)
Nap 3 8.0 (5.2) 7.4 (5.6) 0.054
Nap 4 11.6 (6.4) 10.6 (14.0)
Av MSLT 9.6 (4.5) 8.4 (4.6) 0.01
*The values refer to sleep latency time in minutes (SD). The Av
MSLT is the average sleep latency time for the four naps. The naps
were carried out at 9am,11am,1pm, and 3pm.
CHEST / 112/3/SEPTEMBER, 1997 633
position. Thus, weight loss causes a much more
striking improvement in the lateral RDI than in the
supine RDI.
Effect of Age on Positional Dependency
To estimate the correlation of age with positional
dependency in OSA patients, the entire group was
first divided into three different age categories. The
two youngest categories (age 20 to 39.9 years and 40
to 59.9 years) showed an equal prevalence of PP
(59.2%), while in the 601group, the PP prevalence
decreased to 48.6%. These differences were found to
be only of borderline statistical significance
(x
2
55.58, df52, p50.06; Table 6). The entire group
was then divided into a younger (age#60 years) and
older (age.60 years) group. In the younger group,
the prevalence of PP was 59.2% compared with
48.6% in the older one (x
2
55.58, df51, p50.01; Fig
2). Thus, age was a contributing factor of only
borderline significance for positional dependency
but older OSA patients were still less likely to be
positional than younger ones.
Stepwise Multivariate Logistic Regression Analysis
To estimate the relative influence of each of these
independent variables on the positional dependency,
a stepwise multivariate logistic regression model was
built and analysis was performed. Table 8 summa-
rizes the results of this analysis. As can be seen, the
variable that most significantly predicted positional
dependency was RDI which had a greater improve-
ment goodness of fit (x
2
558.38, df51, p50.0001).
The second variable included into the model was
BMI, also with an improvement goodness of fit
(x
2
538.61, df51, p50.0001). The last variable in-
cluded into the model was age (x
2
55.58, df51,
p,0.018).
Discussion
This study has shown that 55.9% of the 574 adult
OSA patients whose conditions were diagnosed in
our sleep unit have at least twice as many apneas/
hypopneas in the supine than in the lateral position.
These PP were found to be younger and weigh less
than the NPP. In addition, they had fewer and less
severe breathing abnormalities than the NPP group.
Consequently, their nocturnal sleep quality was bet-
ter preserved and, according to MSLT data, they
were less sleepy during daytime hours. This high
prevalence of the positional dependency phenome-
non found in our OSA patients is similar to that
found by other investigators,
4,13,20
but our study is
based on a much larger sample than those previous
ones. This high percentage of PP underlines the
major contribution played by body position during
sleep on the occurrence and severity of OSA.
The most dominant variable that correlates with
positional dependency in OSA patients is RDI, fol-
lowed by BMI which also adds a significant contri-
bution to the prediction of positional dependency.
Age, although significant, adds only a minor im-
provement to the prediction of this positional depen-
dency phenomenon.
PP were found to be younger than NPP. The
difference in the mean age between the PP group
and the NPP group was small (2 years) but signifi-
cant. Support for this result was obtained when the
entire OSA patient group was divided into two age
categories. The prevalence of PP in the youngest
category (age#60 years) was found to be significantly
higher (59.2%) than the 48.6% observed in the older
group (age.60 years), and when age was included in
the multivariate stepwise regression analysis, it
showed a small but still significant contribution to
the prediction of positional dependency. These re-
sults differ from those of others.
13,27
RDI showed a strong relationship to positional
dependency. First of all, by the univariate analysis, it
was found that the prevalence of PP in the severe
OSA patients is significantly lower than that found in
the mild to moderate OSA patients. Second, the
multivariate analysis showed that RDI is the most
dominant factor that could predict the positional
dependency. Thus, by knowing the RDI of an OSA
patient, the chances of predicting correctly if the
patient is positional or not are quite high, and higher
than if only the BMI is known and certainly if only
age is known. We, as others,
28
found that most of the
adult mild-moderate OSA patients are positional or,
Table 6Prevalence of PP According to Different
Categories of RDI, BMI, and Age*
Variable No. of OSA Patients No. of PP (%)
RDI
10-19.9 215 140 (65.1)
20-29.9 109 74 (67.9)
30-39.9 71 49 (69.0)
$40 179 58 (32.4)
BMI
20-24.9 44 31 (70.5)
25-29.9 262 177 (67.6)
30-34.9 175 81 (46.3)
35-39.9 69 24 (34.8)
$40 24 8 (33.3)
Age, yr
20-39.9 49 29 (59.2)
40-59.9 348 206 (59.2)
$60 177 86 (48.6)
*Effect of RDI, BMI, and age on the prevalence of OSA PP. The
number and percent of PP for each category are presented. The
prevalence of PP shows a significant reduction as RDI (x
2
558.8,
df53, p50.001) and BMI (x
2
542.2, df54, p50.001) increased. For
increasing age, these differences were found to be only of borderline
statistical significance (x
2
55.58, df52, p50.06).
634 Clinical Investigations
expressing this in another way, that most of the
breathing abnormalities occur when they sleep in the
supine position. Consequently, these OSA patients
are the ones in whom positional therapy (avoiding
the supine posture) should play an important role in
their treatment. Some of them will eliminate all the
breathing abnormalities merely by only avoiding the
supine posture. On the contrary, patients with severe
OSA are less likely to be positional. Their condition
is so severe that they have breathing abnormalities in
all different body postures. For them, nCPAP treat-
ment is certainly the treatment of choice.
If mild-moderate OSA patients are mainly PP, it
could also mean that “positionality” is perhaps a
characteristic of the natural development of the OSA
entity, and as the severity increases (as occurs with
increase in weight), the positional OSA patient may
convert into a nonpositional one. The reverse ap-
pears also to be true and this was demonstrated in
those obese OSA patients who after losing weight
were converted from NPP into PP (Table 7). With
this reasoning in mind, it is also possible that alcohol
intake and sleep deprivation could convert a PP into
a NPP, but this has yet to be proven. A similar
parallelism seems to exist in the natural development
of snoring. The spouse/partner of a typical snorer
patient often notes that initially the patient snored
only when sleeping on the back but as the severity of
the snoring increased (often related to a gain in
weight), snoring is also present when sleeping on the
sides.
Some authors
2,5
have found that PP is weight-
dependent, being more common in the less obese.
Others have found no differences in BMI between
the PP and NPP group.
13
The anthropomorphic data
in our study demonstrated a marked inverse relation-
ship between the prevalence of PP and the degree of
severity of obesity.
Whereas 68% of those with a BMI #30 were PP,
this was the case for only 42.2% of those with a BMI
.30. The finding in our study that NPP were on
average 6.5 kg heavier than PP suggests that approx-
Table 7—The Effect of Weight Loss on Positional Dependency*
Case No.
12 3 4
Age, yr/sex 45/M 40/M 58/M 45/M
Duration of follow-up, mo 7 20 3 3
Initial weight, kg 103.0 121.5 98.5 92.0
Final weight, kg 87.0 89.0 82.0 82.0
weight 216.0 232.5 216.5 210.0
%weight 215.5 226.7 216.8 210.9
Initial BMI 30.7 38.4 34.5 30.7
Final BMI 25.9 28.8 28.7 27.4
RDI total
Initial 80.1 85.9 58.4 83.1
Final 21.1 9.0 15.7 25.2
RDI supine
Initial 90.2 94.5 65.6 105.2
Final 33.7 15.6 83.7 65.9
RDI 256.5 278.9 118.1 239.3
%RDI 262.6 283.5 127.6 237.4
RDI lateral
Initial 70.0 63.7 55.7 47.3
Final 0 3.5 6.3 9.0
RDI 270.0 260.2 249.4 238.3
%RDI 2100 294.5 288.7 281.0
*Total RDI is the RDI for the total sleep period. Initial refers to the values obtained at the first PSG evaluation. Final refers to the values obtained
at the PSG evaluation after weight loss. Duration of follow-up is the time in months between the first and the second PSG evaluation.
Table 8Multivariate Stepwise Logistic Regression*
Variable bSE WALD df OR (95% CI) p Value
RDI.40 0.6493 0.0988 43.1522 1 4.16 (2.9-5.9) 0.0001
BMI.30 0.4811 0.0917 27.5186 1 2.91 (2.1-4.0) 0.0001
Age.60 yr 0.2254 0.0983 5.2521 1 1.53 (1.1-2.1) 0.0219
*The variable that most significantly explains positional dependency in OSA patients was RDI, followed by BMI and age. OR5odds ratio;
CI5confidence interval; WALD5statistical test used in the regression.
CHEST / 112/3/SEPTEMBER, 1997 635
imately this amount of weight would have to be lost
to convert a NPP into a PP. These results also imply
that in a population of OSA patients who are on
average more obese than those seen by us, the
prevalence of PP could be significantly less than that
found in the present study.
Lloyd
5
found that the degree of obesity correlates
better with the RDI in the lateral position than in the
supine position, suggesting that weight loss is more
effective in improving breathing abnormalities in the
lateral than in the supine position. In the four obese
patients with severe OSA presented in Table 7, we
clearly verified this dominant relationship between
weight loss and the improvement of breathing ab-
normalities during sleep in the lateral position com-
pared to the supine position. It should be noted that
all four patients achieved a normal RDI (,10) after
weight reduction when they avoided the supine
position.
The fact that the PP group had a higher sleep
efficiency shows that they enjoy a better sleep quality
than the NPP group. A better preserved sleep
architecture in the PP group is also noted as judged
by higher percentages of deep (stages 3 and 4) and
lower percentages of light (stages 1 and 2) sleep
stages. Two other important PSG features demon-
strating the more disturbed sleep architecture in the
NPP group were the increased number of short
arousals (,15 s) and the increased time spent awake
after sleep onset. The difference in the sleep archi-
tecture between the two groups is not surprising.
Since PP had fewer and less severe breathing abnor-
malities as judged by the AI and RDI values and by
the min SaO
2
values in both REM and NREM sleep,
their sleep was less disturbed. Only one other study
13
in a smaller sample of patients also found similar
differences in sleep architecture and severity of
breathing abnormalities between the PP and NPP
groups.
No differences in either the total number of limb
movements or limb movements causing arousal or in
the indexes per sleep hour of both parameters were
observed between the two groups. This suggests that
the difference in the number of short arousals (,15 s)
between these two groups is mainly accounted for by
the respiratory disturbances and not by the PLMs.
We found no information in the literature about
daytime sleepiness in the PP group compared to the
NPP group. The percentage of patients who had an
MSLT was significantly higher in the NPP group.
Since all patients were selected to carry out the
MSLT based on the presence of complaints about
daytime sleepiness before the nocturnal PSG was
carried out (consequently before knowing if they
belong to the PP or NPP group), and before the
MSLT test was performed, this result shows that
patients in the NPP group complained more fre-
quently about daytime sleepiness than patients in the
PP group. Although the two-way ANOVA for our
MSLT data results in a borderline (p50.054) signif-
icant difference between the two groups, for each of
the naps, the differences were consistent. In addi-
tion, our MSLT data show that the NPP group had a
significantly shorter average MSLT than the PP
group. Thus, these data are consistent with the
higher prevalence of the daytime sleepiness com-
plaints in the NPP group and also show that patients
of the NPP group are objectively sleepier than those
in the PP group. These MSLT data also correlate
with the nocturnal sleep data. The NPP group had
higher RDI values and more severe breathing ab-
normalities, all of which cause a more disturbed
sleep. Nevertheless, because the PP group had a
longer TST than the NPP group, the differences
obtained in the MSLT could be due to a combination
of better sleep quality in addition to an increased
TST in the PP group.
Cephalometric data have shown that many OSA
patients have a narrower pharynx than non-OSA
control subjects.
29-31
Adopting the supine position
from the sitting position causes a further narrowing
of all segments of the pharynx during the awake state
in normal
32-34
as well as in OSA patients.
35,36
This
pharyngeal cross-sectional narrowing is probably due
mainly to the effect of gravity, which increases the
apposition of the soft palate, uvula, tongue, and
epiglottis to the posterial pharyngeal wall.
32-36
This
gravity effect also causes an increase in the tongue
cross-sectional area,
33
uvular width,
36
and soft palate
thickness,
33
which would also contribute to the re-
duced pharyngeal cross-sectional area. These ana-
tomic changes, which occur in the upper airway (UA)
by adopting the supine position, are followed by one
major physiologic change: an increase in the UA
resistance. The direct consequence of this is that
breathing during sleep becomes more difficult, in-
creasing the probability of the occurrence of epi-
sodes of partial UA obstruction (manifested as snor-
ing and/or hypopneas) or complete UA obstruction
(manifested as apneas).
Since our results demonstrate that the body posi-
tion effect is more dominant in the mildest forms of
OSA, we suggest that the recently described upper
airway resistance syndrome (UARS),
37
which ap-
pears to be the mildest form of UA disturbance
during sleep, occurring even in nonsnoring sleepy
patients, may be caused mainly by sleeping in the
supine position. If this is true, avoiding the supine
position during sleep might be enough to prevent the
increase of UA resistance, and as a consequence,
avoid the need for nCPAP treatment, which has
been shown to be associated with very poor compli-
ance in these UARS patients.
38
Most important
perhaps, at least for some patients, avoiding the
supine position could prevent the gradual progres-
sion over time from partial to complete pharyngeal
636 Clinical Investigations
obstruction during sleep. This proposed relationship
between UARS and the supine position needs to be
evaluated.
One limitation of our study is that we studied the
body position effect without taking into consider-
ation the effect of head/neck position on the occur-
rence and severity of breathing abnormalities during
sleep. Future studies should consider this aspect,
which might also be of significance. Currently data
on this point are limited and conflicting.
34,39-41
Although several studies have already shown a
marked improvement in breathing function by sleep-
ing in the lateral position,
1-9
the anatomic and phys-
iologic mechanisms responsible for this phenomenon
have not yet been clarified. In normal awake sub-
jects, Jan et al
34
recently showed that even though
pharyngeal areas were smaller in the supine than in
the sitting position, no differences were found be-
tween the lateral and supine postures. However, the
anatomic and functional aspects of the pharynx
change markedly during sleep.
29-31,37,42
Unfortu-
nately, studies comparing the pharyngeal cross-sec-
tional areas in the lateral vs supine position in normal
subjects and especially in OSA patients during sleep
are lacking and are urgently needed. Shepard and
Thawley,
43
in OSA patients, studied the effect of
body position and sleep stage on the regions over
which the UA collapses. They found that in most
patients, the site and extent of the UA collapse were
similar in the supine and lateral position indepen-
dent of the sleep stage. This suggests that although
body position plays an important role in determining
whether UA collapse occurs, when it does, the
anatomic location and the extent of the collapse are
similar in the supine and lateral position.
Two studies have examined the anatomic changes
in the UA in a PP group and compared them to
either unselected OSA patients
44
or to an NPP
group.
45
Kovacevic-Ristanovic et al
44
found that the
PP group had a significantly larger posterior airway
space, a less elongated soft palate, and somewhat
more prominent retrognathia than unselected OSA
patients. Pevernagie and Shepard
45
have recently
described for the first time the differences in the size
and shape of the UA in the PP and NPP groups while
subjects were awake. These differences were found
mainly in the velopharyngeal segment of the UA (the
retropalatal area) where the minimal cross-sectional
area is normally located. The size of the UA was
significantly different in the two groups; the minimal
cross-sectional area of the PP group was almost twice
that of the NPP group in both the supine and right
lateral positions. The UA shapes of the two groups
were also different—elliptical in the PP group and
circular in the NPP group. The differences in shape
were due predominantly to the significantly greater
lateral diameter in the PP group, the anteroposterior
(A-P) diameter being no different in the two groups.
The PP group, due in part to the gravity effect on the
soft tissues, which reduces the A-P diameter signif-
icantly,
45
presents breathing abnormalities during
sleep while adopting the supine position. These data
also suggest that when the lateral position is adopted
by these patients, the A-P diameter is increased, and
the lateral walls are far enough apart that they will
not come together and block the pharyngeal lumen.
Thus, sufficient airway space is preserved to avoid a
complete collapse of the UA. In the NPP group,
however, the pharyngeal cross-sectional area is re-
duced to about half of that of the PP group due
primarily to a much reduced lateral diameter, and
changing to the lateral position cannot prevent the
pharyngeal collapse.
45
Despite the high prevalence of positional depen-
dency in OSA observed in this study and others, and
despite the encouraging preliminary results of using
positional therapy of Cartwright et al
15
many years
ago, it is surprising how few investigations have been
carried out on the therapeutic efficacy of avoiding
the supine position during sleep in PP. Moreover,
despite the standards of practice recommendation
that require monitoring of body position in PSG
studies,
46
it appears that only a few sleep disorders
units are recording and analyzing their sleep-related
breathing abnormalities data, including RDIs, by
body position. In addition, how precise would be the
estimation of the results of any surgical or other
intervention procedure in OSA patients without
taking into account the sleep time spent in the
supine position in the pre- and post-PSG? An inex-
plicable worsening after surgery for OSA in a posi-
tional OSA patient could be simply due to the much
longer sleep time spent in the supine position in the
postintervention PSG, compared to that time in the
presurgery PSG and not to a real failure of the
surgery. The crucial issue is that in positional OSA
patients, the severity of the disease is mainly related
to the sleep time spent or not spent in the supine
posture.
In some recent reviews on OSA, the role of
positional therapy has not even been mentioned.
47,48
One possible explanation is that many PP still con-
tinue to snore, and sometimes loudly, when they
sleep in either the lateral or prone position. Conse-
quently, their spouses continue to complain about
snoring. Thus, even though in many cases this ther-
apy may clearly produce a major improvement to the
breathing function during sleep, the fact that snoring
continues in the lateral or prone positions may limit
the long-term effectiveness of this approach. An-
other possible explanation is that this positional
therapy is not a radical solution for the most severe
OSA patients who are the ones that urgently search
for treatment. For these patients, nCPAP is at
present the treatment of choice. Nevertheless, some
of those obese patients with severe OSA who refuse
CHEST / 112/3/SEPTEMBER, 1997 637
nCPAP treatment adopt a sitting posture during
sleep, achieving an improvement in their breathing
function during sleep. This type of postural therapy
has also been objectively shown to improve the
breathing function during sleep in these patients.
49
In addition, if those patients lose enough weight to
become PP, they can be successfully treated by
positional therapy as was demonstrated in the four
obese OSA patients in Table 7.
We place a tennis ball into a pocket of a wide cloth
band or belt attached around the abdomen so that
the ball lies in the center of the back. When the
patient rolls onto the back he feels the pressure of
the ball and instinctively rolls back onto his side
again. Several other methods have been used to train
patients to avoid the supine position. Some may use
a T-shirt with a long vertical pocket holding three or
four tennis balls along the back. This is perhaps less
likely to slip out of place during sleep. Cartwright et
al
15
used an alarm system that momentarily woke the
patients whenever they lay on their backs. The time
spent in the supine position while using this alarm
system decreased from 51.4 to 2.1% of TST. After
the patient went home and practiced avoiding supine
sleep without an alarm for 3 months, a repeat PSG
evaluation, however, revealed that the actual sleep
time spent in the supine position was 24.1%. Al-
though some recent preliminary studies
50
have pro-
vided some data on the long-term efficacy of posi-
tional therapy, more studies investigating this issue
are urgently needed. This is particularly the case
because of the difficulties often experienced by OSA
patients in complying with current therapeutic mo-
dalities such as nCPAP
51
and weight reduction,
52
and
the uncertain results of surgical interventions
53
and
prosthetic intraoral devices.
54
In a recent study
55
we investigated the effect of
avoiding the supine position during sleep for a 1
month period on 24-h blood pressure (BP) in 13
positional OSA patients. In all the patients (hyper-
tensive and normotensive patients) there was a re-
duction in the 24-h mean BP values. A significant
reduction was observed for the mean 24-h, for the
mean awake BP, and mean asleep BP. BP variability
and BP load also fell significantly. Since, as shown in
the present study, the majority of OSA patients have
supine-related breathing abnormalities, and since
about a third of all hypertensive patients have OSA,
avoiding the supine position during sleep, if con-
firmed by future larger studies, could become a new
nonpharmacological form of treatment for many
hypertensive patients.
In conclusion, this study has shown, in a large
population of OSA patients, that the majority
(55.9%) were found to have at least twice as many
apneas/hypopneas in the supine than in the lateral
position. These PP are on the average younger and
weigh less than NPP. They also had fewer and less
severe breathing abnormalities than the NPP group.
Consequently, their nocturnal sleep quality was bet-
ter preserved and, according to MSLT data, they
were less sleepy during daytime hours. The likeli-
hood of being an OSA PP is correlated with RDI,
BMI, and age, generally in a reverse relationship, ie,
an OSA patient with a severe RDI, and who is obese
and older than age 60 years, is significantly less likely
to be positional than a OSA patient with a mild-
moderate RDI, who is nonobese and younger than
60 years of age. The above data stress the profound
effect of body position during sleep on the frequency
and severity of breathing abnormalities in OSA
patients. The data also reinforce the crucial necessity
of not only monitoring body position during PSG
evaluation of every suspected OSA patient, but also
of reporting the severity of OSA (RDI) according to
body position.
ACKNOWLEDGMENTS: We would like to thank Yael Vila for
her advice and assistance in the statistical analysis, Danna Gal for
preparing the figures and tables, and the technical team of the
Sleep Disorders Unit at the Loewenstein Hospital-Rehabilitation
Center for the dedicated and responsible work that they per-
formed.
References
1 Lerner SA, Cecil WT. The effect of sleeping posture on
obstructive sleep apnea [abstract]. Chest 1984; 86:327
2 Cartwright RD. Effect of sleep position on sleep apnea
severity. Sleep 1984; 7:110-14
3 Phillips BA, Okeson J, Paesani D, et al. Effect of sleep
position on sleep apnea and parafunctional activity. Chest
1986; 90:424-29
4 Lloyd SR, Cartwright RD. Physiologic basis of therapy for
sleep apnea [letter]. Am Rev Respir Dis 1987; 136:525-26
5 Lloyd SR. The sleep position effect in sleep apnea as a
continuous variable [abstract]. Sleep Res 1988; 17:14
6 Miki H, Hida W, Kikuchi Y, et al. Effect of sleep position on
obstructive sleep apnea. Tohuku J Exp Med 1988; 156(suppl):
143-49
7 Demirozu MC, Elsasser S, Gazeroglu HB, et al. The effect of
body weight, posture and sleep stage on obstructive sleep
apnea syndrome [abstract]. Sleep Res 1990; 19:319
8 Demirozu MC, Razzetti A, Gazeroglu HB, et al. Sleep body
position influences apnea frequency and duration in obese
and non-obese patients with obstructive sleep apnea [ab-
stract]. Sleep Res 1991; 20A:302
9 Diaz T, Norman SE, Kiel M, et al. A comparison of apnea/
hypopnea index and its relationship to sleep stages and body
posture during the first and second half of a diagnostic
polysomnogram in patients with sleep apnea syndromes
[abstract]. Sleep Res 1991; 20:233
10 Kavey NB, Blitzer A, Gidro-Frank S, et al. Sleeping position
and sleep apnea syndrome. Am J Otolaryngol 1985; 6:373-77
11 George CF, Millar TW, Kryger MH. Sleep apnea and body
position during sleep. Sleep 1988; 11:90-99
12 Miles L, Bailey A. Evaluation of sleep apnea treatment must
be related to sleeping position [abstract]. Sleep Res 1990;
19:256
13 Pevernagie DA, Shepard JW. Relations between sleep stage,
posture and effective nasal CPAP levels in OSA. Sleep 1992;
15:162-67
638 Clinical Investigations
14 Braver HM, Block J. Effect of nasal spray, positional therapy
and the combination thereof in the asymptomatic snorer.
Sleep 1994; 17:516-21
15 Cartwright RD, Lloyd S, Lilie J, et al. Sleep position training
as treatment for sleep apnea syndrome: a preliminary study.
Sleep 1985; 8:87-94
16 Cartwright RD, Ristanovic R, Diaz F, et al. A comparative study
of treatments for positional sleep apnea. Sleep 1991; 14:546-52
17 Jackson EI, Schmidt HS. Modification of sleeping position in
the treatment of obstructive sleep apnea [abstract]. Sleep Res
1982; 11:149
18 Chaudhary BA, Chaudhary TK, Kolbeck RC, et al. Therapeu-
tic effect of posture in sleep apnea. South Med J 1986;
79:1061-63
19 Katz A, Dinner DS. The effect of sleep position in the
diagnosis of obstructive sleep apnea: a word of caution
[abstract]. Cleve Clin Q 1992; 59:634-36
20 Dyonzak J, Cartwright RD. Prevalence of positional differ-
ences in obstructive sleep apnea. Sleep Res 1993; 22:191
21 Young T, Palta M, Dempsey J, et al. The occurrence of
sleep-disordered breathing among middle-aged adults.
N Engl J Med 1993; 238:1230-35
22 Rechtschaffen A, Kales A, eds. A manual of standardized
terminology techniques and scoring system for sleep stages of
human subjects. Los Angeles: Brain Information Service/
Brain Research Institute, University of California at Los
Angeles, 1968
23 Guilleminault C. Sleep and breathing in sleeping and waking
disorders: indications and techniques. Menlo Park, Calif:
Addison-Wesley, 1982; 155-82
24 American Sleep Disorders Association. EEG arousals: scoring
rules and examples, Sleep 1992; 15:173-84
25 Coleman RM. Periodic movements in sleep (nocturnal my-
oclonus) and restless legs syndrome. In: Guilleminault C, ed.
Sleeping and waking disorders: indications and techniques.
Menlo Park, Calif: Addison-Wesley, 1982; 265-95
26 Association of Sleep Disorders Centers Task Force on Daytime
Sleepiness. Guidelines for the multiple sleep latency test
(MSLT): a standard measure of sleepiness. Sleep 1989; 9:519-24
27 Nichols CD. Differential effects of obesity and body position
in older vs younger patients with obstructive sleep apnea
[abstract]. Sleep Res 1994; 23:298
28 Swieca J, Westbrook PR. Relationship between body position
dependence of apnea and hypopnea and overall severity of
sleep-disordered breathing [abstract]. Sleep Res 1994; 23:333
29 Fleetham JA. Upper airway imaging in relation to obstructive
sleep apnea. Clin Chest Med 1992; 3:399-416
30 Hudgel DW. The role of upper airway anatomy and physiol-
ogy in obstructive sleep apnea. Clin Chest Med 1992;
3:383-98
31 Pepin JL, Levy P, Veale D, et al. Evaluation of the upper
airway in sleep apnea syndrome. Sleep 1992; 15:S50-55
32 Fouke JM, Strohl KP. Effect of position and lung volume on
upper airway geometry. J Appl Physiol 1987; 63:375-80
33 Pae E-K, Lowe AA, Sasaki K, et al. A cephalometric and
electromyographic study of upper airway structures in the
upright and supine positions. Am J Orthod Dentofacial
Orthop 1994; 106:52-59
34 Jan MA, Marshall I, Douglas NJ. Effect of posture on upper
airway dimensions in normal human. Am J Respir Crit Care
Med 1994; 149:145-48
35 Brown IB, McClean PA, Boucher R, et al. Changes in
pharyngeal cross-sectional area with posture and application
of continuous positive airway pressure in patients with ob-
structive sleep apnea. Am Rev Respir Dis 1987; 138:628-32
36 Yildirim N, Fitzpatrick MF, Whyte KF, et al. The effect of
posture in upper airway dimensions in normal subjects and in
patients with the sleep apnea/hypopnea syndrome. Am Rev
Respir Dis 1991; 144:845-47
37 Guilleminault C, Stoohs R, Clark A, et al. A cause of excessive
daytime sleepiness: the upper airway resistance syndrome.
Chest 1993; 104:781-87
38 Guilleminault C, Kim Y-D, Stoohs R. Upper airway resistance
syndrome. Oral Maxillofac Sur Clin North Am 1995; 7:243-56
39 Safar P, Esscaraga LA, Chang F. Upper airway obstruction in
the unconscious patient. J Appl Physiol 1959; 14:760-64
40 Rubinstein I, McClean PA, Boucher R, et al. Effect of mouth-
piece, noseclips and head position on airway area measured by
acoustic reflections. J Appl Physiol 1987; 63:1469-74
41 Wilson SL, Thach BT, Brouillette RT, et al. Upper airway
potency in the human infant: influence on airway pressure
and posture. J Appl Physiol 1980; 48:500-04
42 Hudgel DW, Brooks BI, Harasick TM. Measurements of
awake upper airway caliber do not predict upper airway
resistance during sleep [abstract]. Am Rev Respir Dis 1989;
139:A374
43 Shepard JW Jr, Thawley SE. Localization of upper airway
collapse during sleep in patients with obstructive sleep apnea.
Am Rev Respir Dis 1990; 141:1350-55
44 Kovacevic-Ristanovic R, Alder G, Lloyd S, et al. Cephalomet-
ric analysis in positional sleep apneics [abstract]. Sleep Res
1989; 18:249
45 Pevernagie DA, Stanson AW, Sheedy BK, et al. Effects of
body position on upper airway size of patients with obstruc-
tive sleep apnea. Am J Respir Crit Care Med 1995; 152:
179-85
46 Martin RJ, Chairman; Block AJ, Cohn MA, Conway WA, et
al. Indications and standards for cardiopulmonary sleep stud-
ies. Sleep 1985; 8:371-79
47 Kryger MH. Management of obstructive sleep apnea. Clin
Chest Med 1992; 13:481-92
48 Polo O, Berthon-Jones M, Douglas NJ, et al. Management of
obstructive sleep apnea/hypopnea syndrome. Lancet 1994;
344:656-60
49 Mcevoy RD, Sharp DJ, Thornton AT. The effect of posture
on obstructive sleep apnea. Am Rev Respir Dis 1986; 133:
662-66
50 Freebeck P, Stewart D. Positional training while you sleep
[abstract]. Sleep Res 1995; 24:235
51 Kribbs NB, Pack AI, Kline LR, et al. Objective measurements
of patterns of nasal CPAP use by patients with obstructive
sleep apnea. Am Rev Respir Dis 1993; 147:887-95
52 Strobel RJ, Rosen RC. Obesity and weight loss in obstructive
sleep apnea: a critical review. Sleep 1996; 19:104-15
53 Rodenstein DO. Assessment of uvulopalatopharyngoplasty for
the treatment of sleep apnea syndrome. Sleep 1992; 15:556-62
54 Clark GT, Arand D, Chung E, et al. Effect of anterior
mandibular positioning in obstructive sleep apnea. Am Rev
Respir Dis 1993; 147:624-29
55 Berger M, Oksenberg M, Silverberg DS, et al. Avoiding the
supine position during sleep in obstructive sleep apnea (OSA)
patients: the effect on 24 hr blood pressure. J Hum Hyper-
tension 1997 (in press)
CHEST / 112/3/SEPTEMBER, 1997 639
    • "It has been shown that the cross-sectional area of upper airways (UA) is reduced in OSA patients [18] and that supine position causes further narrowing of the pharynx [19]. The gravity causes anatomical changes in the UA during the supine position resulting in higher UA resistance leading to breathing difficulties during sleep increasing the probability of UA obstruction [20]. These anatomical changes might explain the elevated proportion of apnea events in supine position. "
    [Show abstract] [Hide abstract] ABSTRACT: Positional obstructive sleep apnea (OSA) is common among OSA patients. In severe OSA, the obstruction events are longer in supine compared to nonsupine positions. Corresponding scientific information on mild and moderate OSA is lacking. We studied whether individual obstruction and desaturation event severity is increased in supine position in all OSA severity categories and whether the severity of individual events is linked to OSA severity categories. Polygraphic recordings of 2026 patients were retrospectively analyzed. The individual apnea, and hypopnea durations and desaturation event depth, duration, and area of 526 included patients were compared between supine and nonsupine positions in different OSA severity categories. Apnea events were 6.3%, 12.5%, and 11.1% longer ( p < 0.001 ) in supine compared to nonsupine position in mild, moderate, and severe OSA categories, respectively. In moderate and severe OSA categories desaturation areas were 5.7% and 25.5% larger ( p < 0.001 ) in supine position. In both positions the individual event severity was elevated along increasing OSA severity category ( p < 0.05 ). Supine position elevates apnea duration in all and desaturation area in moderate and severe OSA severity categories. This might be more hazardous for supine OSA patients and therefore, estimation of clinical severity of OSA should incorporate also information about individual event characteristics besides AHI.
    Full-text · Article · Mar 2016
    • "The same was found to hold true in our present investigation. Generally, patients are stratified into positional and non-positional OSAHS [6,7,19202122. Studies have shown that up to 60% of sleep apnea patients are classified as positional [21], which has been in agreement with our findings. "
    [Show abstract] [Hide abstract] ABSTRACT: Background: Supine sleep position has been shown to induce or aggravate sleep-breathing disorders, namely obstructive sleep apnea. In patients suffering from moderate-to-severe Obstructive Sleep Apnea-Hypopnea Syndrome (OSAHS), sleeping supine may significantly rise the Apnea-Hypopnea Index (AHI) scores, often causing profound oxygen desaturation. As such, OSAHSrelated complications would emerge secondary to chronic hypoxemia. In a fraction of cases where the condition is even more-strongly dependent to sleeping position, the term positional-OSAHS would apply. In this study, we examined the sleep position and other polysomnographic bioparameters in sleep apnea patients, in relation to their AHI and the incidence of concurrent medical conditions. This report also discussed the emerging evidence regarding the potential effects of sleep position and OSAHS on the brain. Materials and Methods: A retrospective analysis was done on 78 patients (mean age of 43.6 years), demonstrating increased AHI in supine compared to other positions, which referred to our sleep disorders unit and underwent polysomnography during 2013-2015. They were divided into mild (n=31) and moderate-to-severe (n=47) groups depending on their AHI scores (5≤AHI≤15 and AHI>15, respectively). Results: There found to be a significantly higher prevalence of hypertension, coronary artery disease and cerebrovascular disease in moderate-to-severe OSAHS patients. Likewise, Epworth Sleepiness Score (ESS) was notably higher and the mean oxygen saturation was lower in moderate-to-severe compared to mild OSAHS patients. Moderate-to-severe OSAHS cases were spending more time in supine rather than non-supine positions during sleep. AHIs and arousal index were found to be higher in supine position. Conclusion: Sleeping supine seems to worsen OSAHS and increase AHI, subsequently contributing to medical comorbidities. In case of positionalOSAHS, treatments should be individualized to prevent position-dependent airway collapse during sleep. Patients’ awareness on the significance of sleep position in preventing OSAHS-related complications needs to be improved, and medically-proven interventions to maintain non-supine position during sleep should be advised in such patients.
    Full-text · Article · Feb 2016 · Sleep And Breathing
    • "This implies that the majority is explained by variables not included in this analysis. Although earlier studies reported the AHI as negative predictor for POSA [12, 13,202122, this was not found in the current study. Additionally, neck circumference appeared no predictor in current analysis, whereas contrasting results were published by Mador et al. and Teerapraipuk et al. [12, 22]. "
    [Show abstract] [Hide abstract] ABSTRACT: Up to 80 % of the bariatric surgery (BS) patients suffer from obstructive sleep apnea (OSA). BS patients with moderate to severe OSA (apnea-hypopnea index (AHI) ≥15) are usually treated with continuous positive airway pressure (CPAP). This is not indicated in mild OSA patients (AHI <15). However, >50 % of patients with mild OSA have positional OSA (POSA); their AHI is at least twice as high in supine sleeping position than in other positions. Since many patients sleep in supine position for surgical safety reasons after BS, evaluating the AHI in this position might be more relevant in this group. The aim of this study is to evaluate the postoperative cardiopulmonary complication rate in mild OSA patients with and without POSA. Secondary aim is to evaluate predictive factors for POSA. A single-institute retrospective analysis was achieved with all consecutive patients who underwent primary laparoscopic Roux-en-Y gastric bypass or laparoscopic sleeve gastrectomy between 2006 and 2014. All patients with an AHI between 5 and 15 were included. Postoperative complications were compared between POSA and non-POSA patients. Predictive factors were evaluated through univariate and multivariable logistic regression analysis. A total of 277 patients, 153 with and 124 without POSA, were included. After BS, three patients (1.1 %) experienced severe cardiopulmonary complications. No significant difference was found between POSA and non-POSA patients. In multivariate analysis, age and BMI were found to be negative predictors for POSA. In terms of 30-day postoperative cardiopulmonary outcome, CPAP therapy is not indicated in mild (P)OSA patients scheduled for BS.
    Full-text · Article · May 2015
Show more