ORIGINAL ARTICLE
Effect of a 2 week CPAP treatment on mood states in patients
with obstructive sleep apnea: a double-blind trial
Alexander Haensel & Daniel Norman & Loki Natarajan &
Wayne A. Bardwell & Sonia Ancoli-Israel &
Joel E. Dimsdale
Published online: 15 May 2007
#
Springer-Verlag 2007
Abstract Obstructive sleep apnea (OSA) is a common
disease with significant medical and psychiatric comorbid-
ities. The literature documenti ng the effects of continuous
positive airway pressure (CPAP) treatment on mood in
OSA patients is mixed. We previously observed that 1 week
of CPAP treatment did not result in improvements in mood
beyond those observed in a group treated with placebo–
CPAP. This study examined the effect of a 2 week CPAP
treatment on mood in a placebo-controlled design in OSA
patients. Fifty patients with untreated sleep apnea were
evaluated by polysomnography and completed the Profile
of Mood States (POMS) pre-/post-treatm ent. The patients
were randomized for 2 weeks to either therapeutic CPAP or
placebo–CPAP (at insufficient pressure). Both the thera-
peutic CPAP and the placebo–CPAP groups showed
significant improvements in POMS total score, tension,
fatigue, and confus ion. No significant time × treatment
effect was observed for either group. We could not show a
specific beneficial impact of CPAP treatment on mood in
OSA patients.
Keywords Obstructive sleep apnea
.
Mood disorders
.
Continuous positive airway pressure
.
POMS
Introduction
Obstructive sleep apnea (OSA) is a highly prevalent sleep
disorder, characterized by repeated disruptions of breathing
during sleep. The sleep fragmentation and the accompa-
nying hypoxemia have been linked to many negative
consequences including cardiac arrhythmias, nocturnal
hypertension, confusion, cognitive impairment, daytime
sleepiness, and mood disorders [1]. OSA is now recognized
as a common disorder that is associated with major
morbidity and mortality [2].
OSA pa tients regularly show a high prevalence of
psychiatric morbidity [3, 4]. Several studies have examined
the effect of continuous positive airway pressure (CPAP) on
mood in OSA patients, but findings have been inconclu-
sive. Derderian et al. showed in a relatively small group of
OSA patients (n =7) a significant drop in the Profile of
Mood States (POMS) total mood disturbance after 2 months
of CPAP treatment compared with a nonrandomized control
group [5]. Similarly, Engleman et al. examined moderate
[6](n =32) and mild [7](n =16) OSA patients in random-
ized, placebo-controlled, crossover studies and found an
Sleep Breath (2007) 11:239–244
DOI 10.1007/s11325-007-0115-0
A. Haensel
Department of General Internal Medicine,
University Hospital Berne,
Bern, Switzerland
A. Haensel
:
W. A. Bardwell
:
S. Ancoli-Israel
:
J. E. Dimsdale
Department of Psychiatry, University of California,
San Diego, USA
L. Natarajan
:
W. A. Bardwell
:
S. Ancoli-Israel
:
J. E. Dimsdale
Moores UCSD Cancer Center,
La Jolla, USA
D. Norman
Pulmonary and Critical Care Medicine, University of California,
San Diego, USA
A. Haensel (*)
USCD Psychiatry,
9500 Gilman Drive,
La Jolla, CA, USA
e-mail: ahaensel@ucsd.edu
improvement in mood after 4 weeks of CPAP treatment.
Recently, different research groups have examined the
effect of CPAP on the most common mood disorder,
depres sion, in ap nea patients. Means et al. showed a
significant decrease in the Beck Depression Inventory
(BDI) scores 3 months after CPAP treatment [8]. However,
the design of this study was not placebo-controlled and it is
interesting to note that the improvement in BDI was
independent of object ively observed CPAP compliance.
Schwartz et al. found that 38% of a sample of 50 OSA
patients were already being treated with antidepressants
before CPAP therapy started; these patients had higher BDI
scores than the patients without antidepressant medication
[9]. Both groups decreased their BDI after 4 to 6 weeks use
of CPAP, but the decrease was greater in the patients taking
antidepressants. Again, the design was not placebo-
controlled and the authors did not evaluate compliance to
antidepressant drug treatment.
In contrast to these results, we previously observed no
significant difference in change of mood in a placebo-
controlled study that compared therapeutic CPAP with
placebo–CPAP delivered at insufficient pressure for 1 week
[10]. Similar to our results, Borak et al. did not observe any
improvement in anxiety, depression, and mental stress in a
nonplacebo-controlled study after 3 and 12 months of
CPAP and Munoz et al. found no improvement in BDI in
80 OSA patients matched with 80 healthy control subjects
after 12 months of CPAP [11, 12]. In summary, as recently
noted by Gay et al., results from placebo-controlled studies
are inconclusive with regard to the impact of CPAP on
mood [13].
In our previous study design, we wondered if 1 week of
treatment was insufficient to improve mood. In this study,
we attempted to replicate and extend our prior findings by
using a placebo-controlled study design for 2 weeks of
treatment. When this study was designed, we had reser-
vations about offering placebo–CPAP for an extensive
period of time to patients with moderate to severe sleep
apnea; thus we limited this study to a 2-week treatment trial.
Materials and methods
Patients
Fifty patients with a history of snoring and daytime
sleepiness suggestive of sleep apnea were recruited by
advertisement, word of mouth referral, referral from local
medical practitioners in the San Diego area, and referral from
previous participants to participate in a study evaluating
sympathetic nervous system physiology, cognitive dysfunc-
tion, and mood in obstructive sleep apnea. We limited
enrollment to subjects in the range of 30 to 65 years and with
weight 100–200% of the ideal body weight per Metropolitan
Life Insurance tables [14]. Participants >200% of the ideal
body weight were excluded because of the possibility of
confounding by other conditions associated with obesity.
Because of the prominent focus on sympathetic nervous
system physiology, the sample is somewhat limited in terms
of severe obesity, which has attendant effects on sympa-
thetic nervous system physiology and other medical
comorbidities such as diabetes, which could confound an
analysis of CPAP effects on mood. Potential participants
were also excluded if they had a history of heart, liver or
renal disease, diabetes, psychosis, narcolepsy, current
alcohol or drug abuse, severe asthma or cerebrovascular
disease.Patientswithahistoryofdepressionwere
excluded. Patients who were taking antihypertensive med-
ication (n =7) had their medication tapered slowly in 2–3
steps for 3 weeks before admission to enable more
meaningful studies of sympathetic nervous system physiol-
ogy. Tho se whose blood pressure at baseline exceeded 170/
105 mmHg off treatment were retur ned to antihypertensive
medication and not studied. The protocol was approved by
the University of California, San Diego, Human Subjects
Committee. After complete description of the study to the
subjects, written informed consent was obtained.
Polysomnography
Participants had their sleep monitored for an entire night in
the General Clinical Research Center (GCRC) Gillin
Laboratory of Sleep and Chronobiology (LSC) with
polysomnography measuring central and occipital EEG,
bilateral electrooculogram, submental and bilateral tibialis
electromyogram, naso-oral airflow thermistor and nasal
airflow pressure transducer, and thoracic and abdominal
excursions with Respitrace respiratory inductive plethys-
mography. Oxygen saturation was monitored with a pulse
oximeter (Biox 3740, Datex-Ohmeda, Louisville, CO) and
was analyzed using a computer software (Profox Asso-
ciates, Escondido, CA).
Sleep recordings were scored according to the criteria of
Rechtschaffen and Kales [15]. Apneas were defined as
decrements in airflow of ≧90% from baseline for ≧10 s.
Hypopneas were defined as decrements in airflow of ≧50%
but <90% from baseline for ≧10 s [16]. The majority of
subjects had solely obstructive type events with only a few
subjects showing evidence of central apneas. Those few
who had predominantly central apneas (>50% of total
apneas) were exclu ded. The number of apneas and
hypopneas per hour were calculated to obtain the apnea
hypopnea index (AHI). Obstructive sleep apnea was
defined an AHI ≧15.
240 Sleep Breath (2007) 11:239–244
Psychological assessment
Participants completed the POMS before randomization
and after 2 weeks of either CPAP or placebo treatment.
The POMS is a well-established, factor-analytically de-
rived measure of psychological distress for which high
levels of reliability and validity have been documented
[17]. The POMS consists of 65 adjectives rated on a 5-point
scale (0=not at all to 4=extremely) that can be consolidated
into 6 subscales. Patients are questioned: How often have
you been feeling ...during the past week including today?
Thus, stating ‘extremely’ means high fatigue–inertia,
depression–dejection, tension–anxiety, anger–hostility, con-
fusion–bewilderment; vigor–activity is scored in reverse—
i.e., high scores indicate beneficial symptoms as opposed to
troubling symptoms. The total score adds the first five
subscales and subtracts the last subscale.
In addition, a total mood score is calculated by summing
scores for the first five subscales and subtracting the score
for vigor– activity. The POMS has been used in a variety of
chronically ill and well populations, [17–24] including
obstructive sleep apnea patients [10, 25–27].
Treatment groups
Patients with OSA were randomized double-blind into two
different treatment groups: therapeutic CPAP and placebo–
CPAP. On the second night of GCRC admission, the
patients randomized to the therapeutic CPAP treatment
group underwent a standard CPAP titration study. CPAP
was started at a pressure of 4 cm H
2
O and was increased by
1–2cmH
2
O increments based on the pressure of apnea,
hypopnea, snoring or respiratory effort related arousals. The
titration was ended and considered successful when all sig-
nificant respiratory events were abolished while the patient
was supine after the second or third rapid eye movement
sleep period, and the patient had spent at least 15 min of
sleep in the final CPAP level. Patients randomized to the
placebo–CPAP group underwent a mock titration night,
using room air, via modified CPAP equipment that de-
livered <1 cm H
2
O pressure. The equipment used in both
groups appeared identical, consisting of a CPAP unit (Aria
LX CPAP System, Respironics, Murrysville, PA) and a
heated humidifier (Fisher and Pykel HC100, Aukland New
Zealand). In the placebo–CPAP group, the CPAP un it was
modified as follows: the CPAP machine was set at 3 cm
H
2
O pressure; however, a 3 mm-diameter pressure restrictor
was placed downstream in the CPAP tubing and a special
nasal mask with 10 one-quarter-inch drilled holes was used
to allow for adequate air exchange with the environment.
Mask press ure was 0.5 cm H
2
O at end-expiration and 0 cm
H
2
O on inspiration. The noise level of both treatment con-
ditions was not perceptibly different. After the “ titration”
night, the patients were sent home with their assigned treat-
ment equipment. Subsequent telephone calls and home
visits ensured proper equipment setup and use. All of the
home CPAP units had a hidden compliance clock, which
determined the amount of time the equipment was used at
the designated pressu re. The compliance data were d own-
loaded at the end of the treatment period to determine
average hours of use per night. Upon completion of the
2-week treatment period, the patients returned to the
GCRC–LSC for a final attended polysomnogram, per-
formed with the patients on their assigned therapy. When
the study was initially desig ned in 2000, we had ethical
Table 1 Patient characteristics before treatment
CPAP treatment group
(n =25)
Placebo–CPAP group
(n =25)
Age (±SD) 48.2±10.2 49.0±10.6
Sex
Male 20 20
Female 5 5
Race/ethnicity
White 14 16
Black 5 1
Hispanic 3 3
Asian 2 1
Other 1 4
BMI (±SD) 33.1±8.2 33.7±7.5
AHI (±SD) 63.6±29.1 58.4±30.4
Mean oxygen
saturation
(±SD)
92.9±4.4 92.8±4.15
There was no significant difference between the two groups in any of
these variables.
Table 2 Mean and standard deviations of respiratory variables, sleep time, and sleep efficiency before and after 2 weeks of CPAP treatment
CPAP treatment group (n =25) Placebo–CPAP treatment group (n =25)
>Day 1 Day 14 Day 1 Day 14 Time Time × treatment group
AHI 65.9±28.6 3.5±3.4 57.5±32.1 53.4±32.9 <0.001 <0.001
%O
2
<90 4.3±7.5 0.00±0 2.9±5.2 2.4±3.5 0.017 0.061
Mean O
2
saturation 93.2±3.9 92.9±15.8 92.9±4.4 91.8±3.9 0.733 0.819
Total sleep time (min) 346±9 366±6 340± 8 352±8 0.019 0.704
Sleep efficiency (%) 80.3±2.3 87.5±1.5 82±1.6 85.9±1.6 0.002 0.180
Sleep Breath (2007) 11:239–244 241
reservations about imposing a placebo treatment longer than
2 weeks. Instead, we reasoned that slow advance of the du-
ration of placebo treatment would be necessary before one
would be comfortable with a month-long placebo treatment,
even if that month-long treatment would be the most com-
pelling test of the hypothesis.
Statistical analysis
Pre-/post-treatment data were analyzed with repeated
measures analysis of variance using the SPSS 12.0 software
(SPPS, Chicago, IL). Significance was set at α =0.05.
Results
Demographic characteristics
Characteristics of patients are presented in Table 1. The
therapeutic CPAP and placebo–CPAP groups did not differ
significantly in age, sex, race/ethnicity or BMI. Mean age
overall was 48.6±10.3 years, mean BMI was 33.4±7.8. Of
the participants in both treatment groups, 80% were male.
Average nighttime CPAP use was 6.6±1.2 h for the
therapeutic CPAP group compared to 6.0±1.3 h in the
placebo–CPAP group. This difference was also not statis-
tically significant.
Sleep apnea characteristics
After 14 days of treatment, AHI and %SaO
2
decreased
significantly in the therapeutic CPAP group, whereas the
placebo group did not show any significant imp rovement
(see Table 2).
Mood
Table 3 and Fig. 1 show pre-/post-treatment POMS scores
for both groups. There was no significant time×treatment
interaction for any of the POMS subscales. However, a
significant time effect was observed for POMS total score,
tension, fatigue, and confusion.
Table 3 Pre-/post-treatment POMS scores for therapeutic CPAP and placebo–CPAP groups (mean±standard error)
POMS scores Therapeutic CPAP group (n =25) Placebo–CPAP group (n =25)
Day 1 Day 14 Day 1 Day 14 Time Time × treatment group
Tension 8.0±1.3 5.7±1.2 8.1±1.2 5.2±0.9 0.012 0.967
Depression 7.2±2.1 5.4±2.1 7.0±1.6 6.2±2.3 0.226 0.287
Fatigue 9.9±1.7 5.4±1.2 10.4±1.5 7.5±2.1 0.004 0.211
Confusion 6.9±0.9 4.3±0.7 7.0±0.8 5.5±1.1 0.008 0.150
Vigor 14.5±1.3 16.2±2.1 13.6±2.1 15.0±2.1 0.275 0.340
Anger 5.1±0.9 3.4±1.0 6.9±1.4 5.5±1.2 0.256 0.291
Total 22.6±6.2 7.9±6.5 25.8±7.3 14.8±8.0 0.013 0.141
Fig. 1 POMS sc ores in the
therapeutic and placebo–CPAP
groups at baseline and day 14
242 Sleep Breath (2007) 11:239–244
Discussion
In our sample, only the group that received therapeutic
CPAP showed a significant impact of treatment on RDI.
However, patients on both therapeutic CPAP and placebo–
CPAP showed significant improvement on four of the seven
POMS mood ratings; there were no time×treatment inter-
actions. These data suggest a nonspecific effect of CPAP
treatment on mood, i.e., the beneficial effects of CPAP on
mood may be the result of a placebo effect. This result
replicates and extends our previous findings in a different
sample of OSA patients; the previous study found that
1 week of CPAP treatment did not have a specific effect on
mood [10].
This result should not be reviewed as a criticism of
CPAP. CPAP has clearly been shown to be an effective
treatment of sleep apnea [6]. However, our data suggest that
an intervention consisting of a caring physician and
believable but not functioning CPAP has an enormous
beneficial effect on patient mood symptoms—an effect that
is independent of CPAP per se. It might be possible that a
longer trial of CPAP would demonstrate a beneficial effect
on mood.
Our study had several limitations. The PO MS was
developed initially as a method for assessing changes in
psychiatric outpatients; however, it has been widely used in
various nonpsychiatric populations. Our sample showed
relatively low distress values on most POMS subscales at
baseline [17]. This floor effect on total POMS score and
subscores could have contributed to our negative finding.
The literature reports that in healthy subjects, there is a
standard deviation of approximately nine in the depression
subscale. Such a large standard deviation in normal values
for this score make our observed drop of 1.8 points and 0.8
points on CPAP and placebo, respectively, look rather small
in comparison. Another potential limitation in our study is
the differential distribution of ethnicity in the treatment
groups. We had five African Americans on CPAP and only
one on placebo. It is possible that ethnicity affects response
differently to therapy; unfortunately, our small numbers do
not allow meaningful subgroup analysis.
However as expected, our sample reported higher fatigue
and lower vigor scores than in the normative sample.
Nonetheless, even in those subscores where our patients
were highly symptomatic, we failed to observe a specific
effect of CPAP on these symptoms. Is this a function of
sample size? While a larger sample size would be desirable,
note the powerful effects of the placebo treatment. Table 3
and Fig. 1 demon strate improvement in all POMS scores in
both the active and the placebo treatmen t groups, although
the degree of that improvement was significant only for
tension, fatigue, confusion, and total scores. Our interaction
analysis suggested no significant difference in the degree of
this improvement in the active vs placebo groups. Perhaps
with a larger sample we would have been able to perceive
subtler specific effects of CPAP on mood, but the tentative
conclusion from this study would suggest that the “im-
provement” in mood symptoms may be a regression to the
mean, as has been observed in previous studies of mood
[28]. Another limitation could rise from the neurobiological
basis of mood change. Response to antidepres sant medica-
tion typically requires 3 weeks, i.e., underlying changes in
the brain in response to antidepressant treatment take time
to develop and mature [29]. Perhaps our treatment period of
14 days was too short, i.e., possibly a specific effect of
treatment would be evident with a longer treatment.
One might also criticize our exclus ion of patients with
major medical illnesses other than OSA and hypertension.
These criteria resulted in a population that may not be
representative of OSA patients normally seen in a sleep
clinic. Our population might thus not be considered as
“typical” OSA patients because our exclusion criteria
served to limit the impact of several potential cofounders
that could affect mood. Regardless, our patients suffered
from significant sleep apnea verified by AHI (i.e., AHI>50)
and significant obesity (BMI>33).
Despite these limitations, this study is one of only a few
to examine CPAP effects on mood using a double-blind
placebo-controlled randomized trial. We found equal
improvements in mood after 2 weeks of therapeutic CPAP
or placebo–CPAP treatment, suggesting that any improve-
ments in mood are a result of a nonspecific treatment effect.
Acknowledgments This research was supported by NIH grants
AG08415, HL44915, HL36005, and RR00827.
References
1. Cheshire K, Engleman H, Deary I et al (1992) Factors impairing
daytime performanc e in patients with sleep apnea/hypopnea
syndrome. Arch Intern Med 152:538–541
2. Newman AB, Nieto FJ, Guidry U et al (2001) Relation of sleep-
disordered breathing to cardiovascular disease risk factors: the
Sleep Heart Health Study. Am J Epidemiol 154:50–59
3. Engleman HM, Cheshire KE, Deary IJ et al (1993) Daytime
sleepiness, cognitive performance and mood after continuous
positive airway pressure for the sleep apnoea/hypopnoea syn-
drome. Thorax 48:911–914
4. Douglas NJ (1998) Systematic review of the efficacy of nasal
CPAP. Thorax 53:414 –415
5. Derderian SS, Bridenbaugh RH, Rajagopal KR (1998) Neuro-
psychologic symptoms in obstructive sleep apnea improve after
treatment with nasal continuous positive airway pressure. Chest
94:1023–1027
6. Engleman HM, Martin SE, Deary IJ et al (1994) Effect of
continuous positive airway pressure treatment on daytime function
in sleep apnoea/hypopnoea syndrome. Lancet 343:572–575
Sleep Breath (2007) 11:239–244 243
