Evaluation of positive airway pressure treatment for sleep related breathing disorder in adults

University of Minnesota Duluth, Duluth, Minnesota, United States
Sleep (Impact Factor: 4.59). 04/2006; 29(3):381-401.
Source: PubMed
Positive airway pressure (PAP) is used to treat obstructive sleep apnea (OSA), central sleep apnea (CSA), and chronic hypoventilation. This document provides a systematic analysis and grading of peer-reviewed, published clinical studies pertaining to application of PAP treatment in adults. The paper is divided into 5 sections, each addressing a series of questions. The first section deals with whether efficacy and/or effectiveness have been demonstrated for continuous PAP (CPAP) treatment based on a variety of parameters and the level of OSA severity. Next, CPAP titration conducted with full, attended polysomnography in a sleep laboratory is compared with titration done under various other conditions. The third section investigates what can be expected regarding adherence and compliance with CPAP treatment as measured by subjective and objective methods and what factors may influence these parameters. Side effects and the influence of other specific factors on efficacy, effectiveness and safety of CPAP therapy are evaluated in the fourth section. Finally, the use of bilevel PAP therapy is reviewed for both patients with OSA and those with other selected nocturnal breathing disorders. Each section also contains a brief summary and suggestions for future research.


Available from: Terri E Weaver, Jun 30, 2014
SLEEP, Vol. 29, No. 3, 2006
sleep apnea (OSA), but may also be used for some patients with
central sleep apnea (CSA) and chronic hypoventilation. Types of
PAP treatment include continuous PAP (CPAP), bilevel PAP, and
automatic adjusting PAP (APAP). The literature evaluating APAP
was recently evaluated in a systematic review and will not be con-
sidered in this review.
The most widespread application of PAP is the treatment
of OSA with CPAP. Untreated OSA may result in a variety of
well-known symptoms including increased daytime sleepiness,
impaired neurobehavioral performance, decreased quality of
life and increased risk for cardiovascular disease (most studies
have focused on hypertension). A systematic review of the cur-
rent CPAP treatment outcome literature provides a basis for best
clinical practice and helps direct future research efforts. We also
reviewed the literature supporting various techniques to titrate
CPAP to the optimal pressure setting. We attempted to answer
questions such as, “How will the outcomes from a technologist
attended CPAP titration with a standard polysomnogram (PSG)
differ from unattended or partial montage PSG?”
We also asked,
“What factors impact CPAP treatment acceptance and adherence
including treatment delay, adherence monitoring, disease sever-
ity, equipment type, and side effects?”
Treatment of CSA and hypoventilation with PAP is less studied
and established than treatment of OSA. The prevalence of other
sleep related breathing disorders (SRBDs) in many sleep clinic
populations is less than OSA; this may explain the relative pau-
city of data. For this reason, treatment of SRBDs other than OSA
are not addressed in detail in this review except as they pertain to
the use of bilevel PAP treatment. The use of PAP in the treatment
of heart failure is also not included in this review.
Nocturnal bilevel PAP treatment is also used to treat adult pa-
tients with hypoventilation from chronic obstructive pulmonary
disease (COPD), neuromuscular diseases, and chest wall disease
such as kyphoscoliosis. Pressure titration methods, treatment ad-
herence, side effects and treatment outcomes will also be system-
atically reviewed for this heterogeneous group of patients with
A task force was assigned by the Standards of Practice Commit-
tee (SPC) of the American Academy of Sleep Medicine (AASM)
for the purpose of developing a review of the literature pertaining
to the treatment of SRBDs in adults with PAP. The project was
initiated in the fall of 2000 with the formulation of questions as
identified under each section, followed by the construction of an
extraction worksheet and the development of evidence table for-
mat. Although 6 abstractors were part of the initial effort, the final
document was written by 4 authors who were actively involved
Evaluation of Positive Airway Pressure Treatment for Sleep Related Breathing
Disorders in Adults
Peter Gay, MD
; Terri Weaver, RN, CS, FAAN
; Daniel Loube, MD
; Conrad Iber, MD
Mayo Clinic, Rochester, MN;
University of Pennsylvania, Philadelphia, PA;
Providence Medical Group, Portland, OR;
University of Minnesota,
Minneapolis, MN
Review of CPAP and Bilevel PAP
—Gay et al
Abstract: Positive airway pressure (PAP) is used to treat obstructive
sleep apnea (OSA), central sleep apnea (CSA), and chronic hypoven
tilation. This document provides a systematic analysis and grading of
peer-reviewed, published clinical studies pertaining to application of PAP
treatment in adults. The paper is divided into 5 sections, each addressing
a series of questions. The first section deals with whether efficacy and/or
effectiveness have been demonstrated for continuous PAP (CPAP) treat-
ment based on a variety of parameters and the level of OSA severity. Next,
CPAP titration conducted with full, attended polysomnography in a sleep
laboratory is compared with titration done under various other conditions.
The third section investigates what can be expected regarding adherence
and compliance with CPAP treatment as measured by subjective and ob-
jective methods and what factors may influence these parameters. Side
effects and the influence of other specific factors on efficacy, effectiveness
and safety of CPAP therapy are evaluated in the fourth section. Finally,
the use of bilevel PAP therapy is reviewed for both patients with OSA and
those with other selected nocturnal breathing disorders. Each section also
contains a brief summary and suggestions for future research.
Keywords: Sleep related breathing disorder, obstructive sleep apnea;
continuous positive airway pressure; CPAP; sleep disordered breathing;
bilevel positive airway pressure, BiPAP
Citation: Gay P; Weaver T; Loube D et al. Evaluation of positive airway
pressure treatment for sleep related breathing disorders in adults. SLEEP
A Review by the Positive Airway Pressure Task Force of the Standards of Practice Committee of the
American Academy of Sleep Medicine
Disclosure Statement
This was not an industry supported study. Dr. Gay has received research
support from Respironics and ResMed. Dr. Weaver has received research
equipment from Respironics and Protech; is a member of the scientific advi-
sory board for Sanofi-Aventis Pharmaceutical; is a consultant for Jazz Phar
maceutical; and has FOSQ License Agreements with Jazz Pharmaceutical,
Sanofi-Aventis Pharmaceutical, Organon NV, Sleep Solutions, Aspire Medi
cal, and InfluENT Medical. Drs. Loube and Iber have indicated no financial
conflicts of interest.
Address correspondence to
: Dr. Peter C. Gay, Mayo Clinic, E-18B, Roch-
ester, MN 55905; Tel: (507) 284-2957; Fax: (507) 266-7772; E-mail:
Page 1
SLEEP, Vol. 29, No. 3, 2006
throughout the entire project. This review provided the necessary
data for the development of the companion practice parameters by
the SPC.
The level of evidence for the data in each paper relevant to the
evaluation is listed in evidence tables specific for each question.
Each paper was analyzed independently by 2 task force members.
The level of evidence was rated using the AASM classification of
evidence for intervention studies (Table 1), an adaptation of the
Sackett criteria.
Disagreements between the 2 raters were adjudi-
cated by a vote of the task force members.
Searches in the English language literature (Medline 1966 - ear-
ly 2005) of major topics relevant to PAP treatment during SRBDs
were conducted. The initial literature search was done in April of
2001 followed by an update in April of 2002. A final literature
search for just Level I studies was done in January of 2005 in or-
der to keep the review as timely as possible and to avoid omission
of potentially high impact studies published in the interim. The
decision to limit the final search and some entire sections to Level
I or II evidence was decided upon by the Task Force for the pur-
poses of simplification and brevity. The Task Force did not feel
this would detract from the overall conclusions made within the
body of this review. The search focused on peer-reviewed clini-
cal studies, including case-series and controlled trials, which con-
tained information regarding PAP treatment outcomes, methods
for polysomnographic titration, factors affecting adherence and
side effects. Major search terms are included as Table 2. Review
papers, commentary, case reports, pediatric populations, and stud-
ies pertaining to APAP were excluded, except where parenthetical
comments are specifically noted. Comments are provided when
necessary to emphasize where lack of a power analysis, clearly
designated primary endpoint, or small sample size may have con-
founded the conclusions.
Some questions used all levels of evidence whereas others were
confined to specific levels of evidence as noted in each section.
OSA now has well-recognized associations with many other sys-
temic effects, especially cardiovascular effects; a comprehensive,
detailed discussion of all is beyond the scope of this paper. For
this reason, our review was confined to the best-studied cardio-
vascular issue associated with OSA (hypertension) and the effect
of PAP treatment. Five major questions are detailed and discussed
in this review paper; a summary and suggestions for future re-
search appear at the end of each section.
3.1 Has Efficacy and/or Effectiveness Been Demonstrated for CPAP
Treatment in Patients with OSAHS?
The literature search identified 342 articles that met the ex-
traction criteria discussed in Section 2.0. Only investigations that
were randomized controlled clinical trials and considered Level I
or II evidence, compared CPAP to placebo or conservative treat-
ment (nasal strips, weight loss, sleep hygiene, positional therapy),
and that employed generally accepted and validated endpoints
were included in the review. Of these 29 studies,
45% com-
pared CPAP to sham-CPAP,
31% used tablets
as placebos,
and 24% compared CPAP to conservative
Of the randomized controlled clinical trials
that evaluated adherence to both placebo and active CPAP inter-
ventions and reported hours of use, mean nightly active CPAP
use was 4.46 hrs
compared to 4.85 hrs on sham-
Most of the randomized controlled clinical
trials assessed multiple outcomes, but only 28% of the studies
specified a primary endpoint.
Based on Spilkers defi-
nition of a double blinded study (neither the investigator follow-
ing the participant nor the participant were aware of treatment
), few clinical trials were double blinded
which can possibly introduce unintended bias. There were a num-
ber of controlled clinical trials that published negative findings,
but failed to include a power analysis introducing the probability
of a Type II error.
Of the 29 clinical trials, 11 stud-
ies performed Intent-To-Treat analysis
and only 3
Table 1—AASM Classification of Evidence
Evidence Study Design
I Randomized well-designed trials with low alpha and
beta error*
II Randomized trials with high alpha and beta error*
III Nonrandomized concurrently controlled studies
IV Nonrandomized historically controlled studies
V Case series
Adapted from Sackett
*Alpha error refers to the probability (generally set at 95% or great
er) that a significant outcome (e.g., p<0.05) is not a result of chance
occurrence. Beta error refers to the probability (generally set at 80%
to 90% or greater) that a nonsignificant result (e.g., p > 0.05) is the
correct conclusion of the study or studies. The estimation of beta
error is generally the result of a power analysis. The power analy-
sis includes a sample size analysis to project the size of the study
population necessary to ensure that significant differences will be
observed if actually present.
Table 2—Literature Search Terminology and Keywords
Mesh Terms
Sleep apnea syndromes
Positive-pressure respiration
Congestive heart failure
Obstructive lung diseases
Text Words
Obstructive sleep apnea
Central sleep apnea
Cheyne-Stokes respiration
Sleep apnea-hypopnea syndrome
Upper airway resistance syndrome
Sleep-disordered breathing
Chronic hypoventilation
Chronic respiratory failure
Positive airway pressure
Nocturnal ventilation
Positive pressure therapy
Bilevel positive airway pressure
Ambulatory monitoring
Split-night study
Review of CPAP and Bilevel PAP
—Gay et al
Page 2
SLEEP, Vol. 29, No. 3, 2006
provided effect sizes to characterize the magnitude of the impact
of CPAP on the outcomes under consideration.
We examined the efficacy of CPAP for the following outcomes:
apnea/hypopnea index (AHI), considered synonymous with re-
spiratory disturbance index (RDI), sleep architecture, daytime
sleepiness, quality of life, neurobehavioral performance and psy-
chological effects, and cardiovascular morbidity (hypertension).
3.1.1 Reduction of Apnea/Hypopnea Index
In general, the literature documents that CPAP eliminates respi-
ratory disturbances, reducing the AHI. All of the 11 clinical trials
that studied this outcome demonstrated that CPAP was superior to
conservative management,
and positional
This effect was demonstrated during follow-up poly-
somnography (PSG) after 16
or 65 days
; 1,
weeks; or 3 months
of intervention.
3.1.2 Sleep Architecture
The results of controlled clinical trials do not provide support
that CPAP affects total sleep time when compared to placebo
or positional therapy.
The evidence supporting the effect of CPAP
on duration and proportion of stage 1 or 2 sleep is mixed with sev-
eral placebo controlled trials demonstrating a positive effect,
whereas other studies did not.
There was also no significant dif-
ference shown in stage 1 or 2 sleep between positional therapy
and CPAP.
Two placebo-controlled studies found a difference in
length of time in REM sleep
but this was not true when CPAP
efficacy was compared with positional therapy.
22, 23,31
the 5 placebo-controlled studies
reported improvements in
stage 3 or 4 sleep. As with the other stages of sleep, there was no
difference in stage 3 or 4 sleep when CPAP was compared with
positional therapy.
Several randomized clinical trials found that
active treatment was no better than placebo
or positional ther-
in affecting the amount of time awake or sleep efficiency.
There is inconsistent support for the effect of CPAP on arousal
index. However, several Level I and II studies provided evidence
of a lower arousal index with CPAP compared to placebo,
and only 1 study did not.
3.1.3 Daytime Sleepiness
There has been considerable study of the impact of CPAP on
subjective and objective daytime sleepiness. The majority of these
studies have evaluated subjective sleepiness, principally using the
Epworth Sleepiness Scale (ESS).
Of the placebo-controlled tri-
als employing the ESS,
most found that CPAP re-
duced subjective daytime sleepiness.
The evidence
that CPAP is superior to conservative or positional therapy is
less compelling. Ballester and colleagues
noted improvements
with CPAP use compared to conservative therapy (sleep hygiene,
weight loss and diet). However, Monasterio and associates
a similar intervention and Redline and coworkers
added nasal
strips to conservative therapy and did not observe an improve-
The Level I and II evidence for objective daytime sleepiness
was more equivocal. For example, 2
of the 3 studies
employed sham-CPAP and evaluated objective sleepiness using
the Maintenance of Wakefulness Test found a greater effect for
CPAP compared to the placebo. It should be noted that the sample
in the negative study consisted of non-sleepy patients prior to in-
tervention and measured objective sleepiness employing the Mul-
tiple Sleep Latency Test. However, when a tablet was used as the
placebo to evaluate this effect, only 2
of 6 studies
showed that CPAP was the superior intervention. The 2 studies
that compared CPAP to conservative treatment failed to detect
significant differences between these treatments.
3.1.4 Neurobehavioral Performance and Psychological Effects
Of the 29 placebo-controlled trials, 10
the impact of CPAP on neurobehavioral performance. Perfor-
mance variables included cognitive processing, sustained atten-
tion, executive function, memory and mood. Only 2
of the
9 placebo-controlled studies that evaluated cognitive function-
found CPAP superior to placebo. The studies were
inconclusive with regard to the benefit of CPAP over placebo in
improving sustained attention.
The 2 studies comparing
conservative therapy to CPAP treatment produced conflicting re-
sults with regard to the impact on cognitive processing.
were too few studies to draw conclusions regarding differences
between conservative and CPAP therapy and change in sustained
attention. The 1 study that evaluated positional versus CPAP
therapy found that CPAP was superior to positional therapy for
cognitive processing, but not sustained attention.
CPAP therapy
was not superior to placebo
or positional therapy
for restor-
ing memory. However, it was more effective than conservative
therapy, such as sleep hygiene and weight loss.
Of the Level I
and II investigations evaluating executive functioning,
few of these studies provided support that CPAP was more effec-
tive than placebo
or positional therapy.
There was only
1 Level I study that compared CPAP to conservative therapy with
regard to executive functioning and it demonstrated a greater im-
pact for CPAP.
Results from placebo-controlled studies were
inconclusive with regard to the efficacy of CPAP in elevating
Inconsistent results in neurobehavioral perfor-
mance among studies may be related to the different measures
employed to evaluate these outcomes as well as the likelihood of
a beta error.
3.1.5 Quality of Life
A number of studies have compared the impact of CPAP rela-
tive to placebo,
conservative treatment
or posi-
tional therapy
on quality of life. These studies have employed
both generic (SF36, Nottingham Health Profile) and disease-spe-
cific (Functional Outcomes of Sleep Questionnaire) measures.
Among the Level I and II placebo-controlled investigations, the
findings are inconclusive with equal numbers of positive
and negative
conclusions regarding the superiority of
CPAP treatment. Three randomized studies compared CPAP with
conservative therapy.
One study found improvement in 2 of
the 6 subscales (social isolation and energy) of the Nottingham
Health Profile, a generic measure of quality of life.
A second
study that employed both a generic and a disease-specific mea-
and a third study that used only a generic measure
did not
find significant improvement. The 1 study that examined quality
of life in patients randomly assigned to CPAP or positional thera-
py did not find that CPAP produced greater gains than positional
therapy as measured by a generic quality of life measure.
Review of CPAP and Bilevel PAP
—Gay et al
Page 3
SLEEP, Vol. 29, No. 3, 2006
3.1.6 Cardiovascular Morbidity (Focus on Hypertension)
The effects of CPAP therapy on cardiovascular disease, espe-
cially hypertension, has been a target of several recent investiga-
The majority of placebo-controlled studies that
employed at least 19 hours of ambulatory blood pressure monitor-
ing did not find that CPAP improved mean arterial pressure.
However, several Level I and II studies found that CPAP did have
a greater impact on nocturnal blood pressure than placebo
and 2 studies showed lower mean diastolic pressure in patients on
CPAP compared to controls.
One Level I study demonstrated a
large reduction in mean arterial blood pressure following those on
anti-hypertensive medications,
and a Level II study suggested
that change in 24 hour systolic, diastolic and mean arterial blood
pressure following CPAP treatment was greater than placebo in
those with more than 20 desaturations of 4% or more per hour of
sleep. The 1 study that compared CPAP to conservative therapy
(weight loss, diet and sleep hygiene) found that blood pressure
measured by sphygmomanometer in CPAP patients was not dif-
ferent than in patients treated with conservative therapy.
sole study that examined change in blood pressure associated with
treatment in milder OSA using a tablet placebo failed to show
differences between CPAP treatment and placebo.
The 2 Level
II studies that evaluated the impact of CPAP versus placebo on
heart rate produced conflicting results.
Therefore, the impact
of CPAP treatment on cardiovascular risk and associated organ
dysfunction in milder OSA is unknown.
3.1.7 OSA Severity – Relationship to Efficacy and/or Effectiveness
Most of the Level I and II studies of the efficacy of CPAP treat-
ment have been conducted in patients with moderate (AHI 15 - 30)
and severe (AHI > 30) disease as defined by the AASM.
The 3
Level I studies
and 3 Level II studies
that were restricted
to patients with mild to moderate OSA found that CPAP reduced
but did not improve objective sleepiness
blood pressure.
Conflicting results were found for subjec-
tive measures of sleepiness,
neurobehavioral perfor-
and quality of life.
it remains unclear whether CPAP has utility across outcomes for
this level of disease severity.
Other Level I and II studies using a higher criterion of disease
severity (AHI > 30) have shown that, compared to placebo, CPAP
reduces apneas and hypopneas,
increases time in REM
and improves oxygenation.
There is incon-
sistent evidence for an effect of CPAP on other aspects of sleep
architecture, subjective and objective sleepiness, neurobehavioral
function, mood, quality of life and blood pressure.
There have been no published Level I or II investigations that
compared outcomes of CPAP treatment concurrently among mild,
moderate and severe OSA as defined by the AASM.
There are no Level I or II studies that have examined the ef-
ficacy or effectiveness of CPAP treatment in OSA patient with an
AHI < 5. There have been several Level III studies as described
in a large review paper
that have examined the use of CPAP in
Upper Airway Resistance Syndrome (with an AHI < 5) and in
subjects with an AHI < 10. There is insufficient evidence to draw
conclusions regarding the efficacy and/or effectiveness of CPAP
treatment in this population.
3.2 Summary
The studies reviewed for this section document that CPAP
eliminates respiratory disturbances, thereby reducing the AHI
compared to placebo, conservative management or positional
therapy. There was somewhat stronger evidence supporting im-
proved Stage 3 and 4 sleep and decreased EEG arousals with
CPAP vs placebo. However, whether CPAP yields significant
consistent improvement in overall sleep architecture or fragmen-
tation is less clear. There is equivocal evidence whether CPAP
improves objective daytime sleepiness, neurobehavioral perfor-
mance, psychological functioning and quality of life. The impact
of CPAP on cardiovascular risk, especially hypertension, is large-
ly mixed and the data for differences in the effectiveness of CPAP
based on various levels of OSA severity remains unknown.
3.3 Future Research
There is a need for double-blinded CPAP studies that have a
clearly defined primary outcome and include power analyses and
effect size calculations. Studies that specifically examine OSA
subgroups with respect to severity are particularly lacking.
4.1 If CPAP Titration is Done Under the Following Conditions, How
Will it Differ From Full, Attended Polysomnography (PSG, 4.1.1) in
a Sleep Laboratory?
There were 140 articles reviewed for this question. Ultimately,
28 articles were selected for review, including studies graded at
all evidence levels but excluding those with less than 10 patients.
There were 2 Level I, 2 Level II, no Level III, and 24 Level IV
studies included in this review. No studies were available for sec-
tions 4.1.4, 4.1.5, and 4.1.7. Thirteen studies had no comparative
findings but provided results regarding baseline condition, full
night (FN) diagnostic PSG data and a second night of CPAP ti-
tration. Some studies were described in relation to APAP. These
were the subject of an earlier AASM position paper,
and, for the
purposes of this review, these data were not considered. The larg-
est number of comparative studies (7) was for split night (SN)
attended PSG. Pertinent study endpoints for this question con-
sisted primarily of comparative effective CPAP levels required
to resolve sleep related breathing disorders, effect on sleep qual-
ity, and compliance or adherence to therapy at variable lengths of
time of treatment. The individual studies are described in the con-
text of each specific aspect of the question as indicated above.
4.1.1 Full, Attended PSG in a Sleep Laboratory
Since the full, attended PSG in a sleep laboratory is the standard,
we wanted to evaluate the reliability of this standard. We searched
for studies that examined the repeatability of CPAP titration, or
that examined issues that affect the variability of results of such
titrations (position, time on treatment, etc.) Of the 13 studies,
all were rated at Level IV evidence and no comparative patient
population for the purposes of this question was provided. Nine
studies provided data regarding compliance and adherence rates
to CPAP therapy. Nine studies discussed results of sleep quality
although not all included both diagnostic and CPAP titration sleep
staging. The studies ranged from 10 to 95 patients and either ex-
clusively enrolled men or men were the majority (71 to 93%) with
a mean age of 48.2 to 57 yrs.
Review of CPAP and Bilevel PAP
—Gay et al
Page 4
SLEEP, Vol. 29, No. 3, 2006
Very few studies involved patients with mild disease; the range
of the mean AHI from all of the studies was between 25 and 97.6
with most above 50. The effective CPAP pressure was between
8.1 to10 cm H
O with a reduction in mean AHI to a level of be-
tween 2 and 9.1 after CPAP titration. Adherence data were not al-
ways provided, but acceptance rates when available ranged from
74% to 93.7% at follow-up times of 21 to 784 days. Most studies
relied on objective machine counters and use ranged from 4.7 to
7.6 hours per night. Detailed complication data were rarely re-
Four reports focused on less commonly selected endpoints
rather than comprehensively describing the typical results of 2 FN
studies. One study looked at the identification of effective CPAP
pressure during 3 repeated titration PSGs and found a reduction
of 1 to 2 cm H
O after either 2 or 8 months but not at 20 months.
Another study evaluated the hysteresis of CPAP titration demon-
strating that repeat downward adjustment during FN PSG resulted
in lowering of the effective CPAP pressure from 9.5 to 8.9 cm
O; however, it did not provide adherence data, and informa-
tion regarding sleep staging was available only at baseline. Most
patients had severe OSA with mean baseline AHI of 40.5 that cor-
rected to 4.8.
A third study assessed the effect of body position
and rapid eye movement (REM) sleep state on the effective CPAP
level in patients with severe sleep apnea and found that supine
position was associated with the maximum CPAP requirements
in 86.7% of patients.
Body position effects were related to body
mass index (BMI), RDI, and REM sleep state. Lastly, a study of
the effect of CPAP pressure consistency over 2 consecutive full
nights of CPAP titration indicated no difference in a group of pa-
tients with moderate-severe OSAHS of AHI = 39.3, but these au-
thors provided no adherence or sleep stage data.
4.1.2 Split-Night Study Attended Full PSG
One of the earliest reports of successful SN CPAP titration was
published in 1984.
Of the 7 papers published since 1993 and re-
viewed for this section,
most (6) were rated as Level IV and
1 was rated as Level II. There were 5 studies that included adher-
ence rates to CPAP therapy and 4 that provided data describing
sleep quality, 1 with all PSG studies, 2 with the diagnostic PSG
only and 1 with the CPAP titration night alone. One study noted
the frequency of healthcare access and utilization following the
introduction of CPAP treatment.
The highest Level II evidence study had groups that were not
strictly comparable.
In this study involving 20 patients with se-
vere OSA, a SN PSG was followed by a CPAP titration FN PSG 2
weeks later, where CPAP was readjusted in group 1 (lowered from
an initial mean of 12 cm H
O to a mean 9.5) but was not changed
in group 2 (initial mean of 12 cm H
O was maintained). There
were no differences between groups in adherence (6.9 vs.6.4 hours
per night), sleep quality, and improvement in Epworth sleepiness
scale (ESS) at 2 to 4 weeks after titration. These findings might
have been confounded by the 2-week delay between SN and FN
CPAP titration and by a lowering of the effective CPAP pressure
rather than being a true comparison of FN and SN full PSG CPAP
Long term adherence (22 to 27 months) was assessed in anoth-
er cohort study that matched 2:1, FN:SN PSG for age (50 yr.), sex
(80% men), AHI (49/hr), and ESS (15).
There was a reduction
in TST, percent REM, and slow wave sleep during the diagnostic
study with SN PSG compared to a FN PSG, but no difference in
mean prescribed CPAP (8.5 vs. 9 cm H
O) was noted between
the methods. Five of 46 SN vs. 6 of 92 FN PSG patients that
required retitration for effective CPAP levels to achieve an AHI
< 10. There were no differences in symptom relief, clinic visits,
nurse interventions per year, initial acceptance (78 vs. 79%) or
ultimate adherence rates (6 vs. 6.2 hours per night) between the 2
titration techniques. An unusually long time transpired from time
of diagnosis to initiation of CPAP therapy in both groups, averag-
ing over 1 year. The overall CPAP use correlated with low total
sleep time (TST) during the diagnostic PSG, and a high ESS at
baseline regardless of titration method.
Two other large studies (> 50 patients) tested whether an ade-
quate CPAP prescription could be established with a SN design as
compared to a single subsequent FN CPAP titration PSG, but no
adherence data was examined.
Patients were predominately
male (over 80%), middle-aged (near 50 years), with either mild
or very severe OSA (mean AHI = 23.6 or 76.7). In the study of se-
vere OSA patients, no change in CPAP pressure was noted in 62%
of patients, 22% with 2.5 cm H
O higher, and 6% with 5 cm H
higher CPAP pressure needs with the FN titration.
There was
no difference in the mean CPAP levels (13 vs. 14 cm H
O) when
the AHI was corrected to less than 5. As expected, TST on CPAP
was much less with SN vs. FN (132.4 vs. 257 min.) but both sleep
efficiency (SE) (89.4 vs. 95.6%) and percent REM sleep (24.1 vs.
22%) were similar. The other study of mild OSA patients showed
a significantly lower CPAP at the end of the SN vs. FN CPAP
titration (8.8 vs. 10.3 cm H
O) but this differed significantly only
for patients with AHI < 20.
The relative effects of sleep stages,
SE, and TST on CPAP were essentially similar for the SN and FN
CPAP titration as in the prior study above.
Two smaller retrospective matched cohort studies with FN di-
agnostic PSG looked at adherence at 1 to 2 months. The study
involving more severe OSA patients (AHI approximately 65) re-
vealed a high acceptance rate of near 85% with mean use of 4.8
hours per night at 41 to 55 days and a marked reduction in mean
arousals from about 44 to near 10 per hour.
The study with fewer
patients in each group (12), consisted of more women than most
other studies (33%), and with milder OSA (AHI = 27) showed
one of the lowest acceptance rates of 63% and overall adherence
at 5.2 hours per night for SN vs. 3.8 hours per night for FN al-
though this difference was not significant (p = 0.29).
The last paper in this group prospectively followed 27 long-
term patients with severe OSA (AHI = 63.6) after CPAP titration
with SN PSG, and compared results to unmatched historical FN
CPAP titration study patients. An effective SN CPAP level (AHI
< 5) was obtained in 87% patients with an initial acceptance rate
of 78%. At a mean of 285 days later, the acceptance remained
high at 80% and an adherence rate of 6.7 hours per night that
compared favorably to their FN historical controls with 77% ac-
ceptance at over 3 months and 51% of patients self-reporting use
> 7 hours per night.
4.1.3 Daytime Study Attended Full PSG
There were only 2 studies reporting on daytime studies, and
both were Level IV using a cohort design matched for age, sex,
BMI, and OSA severity.
The larger study of 32 patients in each
group (82% male; age 50 years) had more severe OSA (mean
AHI > 60) and the nighttime diagnostic study was performed for
Review of CPAP and Bilevel PAP
—Gay et al
Page 5
SLEEP, Vol. 29, No. 3, 2006
all “night” patients and half of the “day” patients, which were se-
lected for variable reasons that were not clearly specified.
comparing the CPAP titration techniques, there were no signifi-
cant differences in the effective CPAP level, number of failures,
or complete or partially successful titrations (88%) although there
was a trend towards more complete success in the night vs. day
titration group (84 vs. 69%). There was an unusually high mean
AHI while on CPAP (day = 29.4 vs. night = 20.3) due to the rela-
tively smaller overall success rates seen in other FN titration stud-
ies. The day titration group had significantly lower sleep time at
effective CPAP (1 vs. 2.5 hours), lower and total sleep time (2.1
vs. 4.8 hours), higher sleep efficiency (75.8 vs. 65.2%) and a simi-
lar percentage of REM sleep. After at least 12 weeks, there was no
difference in the number of patients never on CPAP (day = 22 vs.
night = 31%) or in those with self-reported CPAP use > 5 nights
per week (day = 66 vs. night = 59%). There were no clear predic-
tors of successful daytime titration including AHI, lowest oxygen
saturation, or BMI although there was a trend with higher baseline
ESS (successful = 15.9 vs. unsuccessful = 10.2; p = .07).
The other CPAP day titration study was smaller with 14 pa-
tients matched similarly as above with very severe OSA (AHI
> 80). All patients underwent a FN diagnostic study resulting in
CPAP levels of 12 cm H
O for a treated mean AHI < 10 for both
day and night CPAP titration. The time in bed was shorter for
the day vs. night group during CPAP titration (386 vs. 488 min)
but sleep efficiency (≥ 80%) was similar as were the sleep time
and percentage REM sleep at effective CPAP. The authors did
not indicate how many patients required repeat titration studies
to achieve an effective CPAP level but patients were followed up
after 1 week of CPAP use at home. About 85% of patients from
each group used their CPAP showing significant improvement in
daytime sleepiness. The objectively measured number of nights
used was similar with a mean period near 4.5 hours per night. The
authors of both studies concluded that this method was a viable
alternative to nighttime CPAP delivery but neither offered factors
that predicted higher likelihood of success with day titration.
4.1.4 Home Study Attended Full PSG
No studies were available for review in these categories as data
from portable monitoring equipment were excluded. The reader is
referred to a recent position paper on portable sleep monitors for
4.1.5 Home Study Unattended Full PSG
No studies were available for review in these categories as data
from portable monitoring equipment were excluded. The reader is
referred to a recent position paper on portable sleep monitors for
4.1.6 Sleep Laboratory Study Attended Partial PSG
There were 2 studies evaluated for this category with 1 rated as
Level IV and the other as Level I.
The higher level evidence
paper utilized a FN hospital laboratory PSG for all subjects fol-
lowed by a randomized FN PSG and compared respiratory moni-
toring alone on a respiratory ward to complete PSG in a sleep
This study was designed to assess identification of
optimal CPAP level by both methods but did not provide sleep
data, compliance or adherence data. This group of moderately se-
vere OSA patients (mean AHI approximately 50) showed no sig-
nificant difference in effective CPAP levels (9.3 vs. 9.7 cm H
In 1 patient full PSG monitoring led to a pressure 5 cm H
O less
than with respiratory monitoring; in other patients pressures were
within 2.5 cm H
The smaller study of 11 patients was done in a sleep laboratory
and was attended by a technologist to assess the utility of forced
oscillation (FO) technique for CPAP titration.
The authors did
not report acceptance or adherence rates nor was actual sleep
stage data provided. They concluded that the FO measurements
did not disturb sleep and provided a quantitative index of airway
obstruction during CPAP titration.
4.1.7 Sleep Laboratory Study Unattended Partial PSG
No studies were available for review in these categories as data
from portable monitoring equipment was excluded. The reader is
referred to a recent position paper on portable sleep monitors for
4.1.8 Home Study Attended Partial PSG
One Level IV study reported the feasibility of home attended
partial PSG for CPAP titration after an in-laboratory FN PSG di-
agnostic study.
This appeared to be a non-consecutive retrospec-
tive case series of 17 patients with moderately severe OSA (AHI
= 52.1) many of whom were selected because they did not wish
to return to the laboratory for another PSG. CPAP was titrated
based on respiratory monitoring and snoring as determined by the
home-based technologist. A mean effective CPAP level of 10.3
cm H
O was obtained and acceptance of CPAP was reported at
76% (13 of 17) with a very high adherence rate of 7.2 hours per
night after a mean of 13.4 months.
4.1.9 Home Study Unattended Partial PSG
There were 4 papers available for review in this section al-
though none of them actually used any form of home PSG equip-
ment. The true home study unattended partial PSG studies all used
portable monitors which as noted earlier were excluded from this
review having been reviewed by another task force.
For this re-
view, there was 1 Level I study using patient adjustment but no
home monitoring equipment and the 3 others were Level IV with
2 studies using predicted equations and 1 using empiric CPAP
titration based on bed partners observations.
The Level I study
followed 18 primarily middle-aged males
with more severe apnea (age = 50 years, AHI = 40) in a random-
ized crossover trial each of 5 weeks comparing standard in-labo-
ratory CPAP titration to patient CPAP home self-adjustment.
Patients were instructed to alter the CPAP pressure in response
to issues suggesting excessive or inadequate pressure levels with
no sleep monitoring. The primary outcome was optimal CPAP
pressure level with high patient adherence rate for relief of apnea
and hypopnea, improved sleep quality, and resolution of symp-
toms. The optimal CPAP levels were identical near 10 cm H
with maximum deviation of > 2 cm H
O in only 3 patients and
similar mean adherence near 6.5 hours. There was equal and sig-
nificant improvement in AHI (both means were 6 or less) as well
as equivalent response in objective and subjective symptom and
performance parameters (see evidence table for details). The au-
thors concluded that in selected cases, home CPAP titration might
Review of CPAP and Bilevel PAP
—Gay et al
Page 6
SLEEP, Vol. 29, No. 3, 2006
be feasible and allow elimination of a second laboratory-based
PSG. The study could be criticized for a high dropout rate (6 of
24 or 25% of enrolled patients) and small patient numbers. There
was also no improvement in sleep architecture for either treatment
arm although baseline sleep efficiency was at or above 80% and
percent of REM sleep was near or above 20% during the labora-
tory based PSGs.
The smallest study of 11 patients used an unattended home di-
agnostic study (baseline RDI = 41) to titrate beginning with CPAP
of 5 and increasing by 2.5 cm H
O until snoring and other symp-
toms improved.
Repeat unattended home study indicated resolu-
tion to a normal mean AHI = 2.4. After 18 months, all patients
reported good adherence with therapy although no details were
A much larger study used an initial cohort of 38 patients to
derive a prediction equation for effective CPAP levels and then
applied it prospectively to a group of 208 consecutive patients af-
ter a full-night laboratory-based diagnostic PSG, which revealed
moderate to severe OSA (AHI = 50).
The CPAP derived from
the prediction equation agreed closely with that obtained during
an additional laboratory PSG in a confirmation group of 129 pa-
tients (8 vs. 8.1 cm H
O). When separating groups into high (> 10
cm H
O) and low (< 5 cm H
O) CPAP needs, the most important
determining factors included AHI, BMI, and neck circumference.
Although no sleep stage or adherence data were given, the authors
concluded that prediction equations might be used to simplify em-
piric determinations of best CPAP level to use in the laboratory or
at home and perhaps reduce the time to obtain the effective CPAP
This same approach was later tested by 1 of the authors as he
and others prospectively studied 329 predominately male patients
with diagnostic FN lab PSG (AHI > 10 [mean AHI = 47]).
ter a laboratory-based FN PSG CPAP titration, they segregated
the patients into those who had a successful effective CPAP level
predicted by the equation with both pressures within 2 cmH
(84%), were over-predicted (13%), or were under-predicted (3%).
The group that had an over-prediction of effective CPAP tended
to be significantly less overweight and had milder sleep apnea.
Patients that were under-predicted included those with significant
central sleep apnea, poor sleep at higher CPAP pressures or major
mask leaks. The investigators did not provide any sleep stage or
adherence data.
4.2 Summary
The studies reviewed for this section do not negate the assump-
tion that the FN PSG CPAP titration can be regarded as a standard.
We do realize, however, that since all papers decided upon the FN
PSG as the comparison standard on an a priori basis, showing that
nothing else proved superior is not the same as proving FN PSG
CPAP titration is a quality gold standard. The data supporting
the utility of the SN PSG CPAP titration was generally of lower
evidence level and usually showed feasibility rather than clear
equivalence. Similar adherence, ESS, and optimal CPAP pressure
in many of the SN studies supports this as an option in selected
cases. The same could be said of daytime PSG CPAP although the
body of evidence is much more scant. Partial PSG CPAP titration
data whether attended by a technologist or not was very limited
but again, portable monitoring equipment and auto-titration tech-
niques were not evaluated for this evidence review. There was
only 1 high evidence level study supporting home self-adjusted
CPAP titration making this approach less easily acceptable.
4.3 Future Research
Although the most evidence was available for comparative as-
sessment of split-night sleep studies, there were still too few Lev-
el I investigations to accept that future research in this area is not
needed. As noted in the summary above, virtually all areas of this
question require further assessment. Nearly 86% of the studies
were Level IV. Although not directly reviewed in this paper, the
role of portable monitoring with differing levels of sophistication
may also need strong reconsideration and study to clarify many of
the issues raised above.
5.1 On Initiating CPAP Treatment in a Patient:
5.1.1 Does Immediate or Near-Immediate Initiation of Therapeutic
CPAP Change Acceptance or Adherence Compared to a Delay of
Days or Weeks?
There were 51 articles reviewed for this question with the vast
majority meeting criteria; only 4 were excluded for a total of 47.
There were 23 Level I, 14 Level II, 8 Level III, 1 Level IV, and 1
Level V studies reviewed.
One study
randomly imposed CPAP after PSG within 2 weeks
(Group 1 -- 82 patients) or a 6 month delay for CPAP after PSG
(Group 2 -- 89 patients). New patients suspected of mild-moder-
ate (AHI = 10 to 30) or more severe (AHI > 30) OSA were then
assessed for objective CPAP adherence by machine time on coun-
ter but they did this only for the group receiving more immediate
CPAP (Level IV for this issue). The adherence rate in these Group
1 patients with AHI = 10 to 30 was 5.2 ± 2.1 hours per night at
3 months and somewhat less at 4.8 ± 2.3 hours per night after
months. In the Group 1 patients with an AHI > 30, adherence was
5.6 ± 2.0 hours per night at 3 months and 5.5 ± 2.2 hours per night
after 6 months. There was no statistically significant difference in
adherence between these
2 subgroups of different severity apnea
but unfortunately, no adherence comparison was made between
Group 1 and Group 2 patients to provide insight to our question.
No other papers specifically addressed this question and most
papers did not include this issue for discussion at all. However,
our committee suggested that this question could be investigated
in 2 other potential ways. One approach would be to assess the
impact on acceptance/adherence of the time between referral and
CPAP home treatment initiation. Several papers imply that CPAP
treatment is initiated sooner after a SN protocol but this did not
improve adherence (see Section 4.1.2 for further details). The SN
PSG and CPAP use topic was more specifically addressed in a
study conducted in the United Kingdom
in which the median
time from referral to beginning CPAP was significantly less with
a SN (13 months) vs. a 2 full-night protocol (22 months). This
study found no difference in the nightly CPAP time on (CPAPto)
for SN vs. FN PSG groups by objective measures (6 hours vs. 6.2
hours per night); however the 1 to 2 year wait for CPAP initiation
was extraordinary compared to most other studies.
Our second approach to investigate question 5.1.1 was to as-
sess the impact of the time between diagnostic testing and/or ther-
apeutic trial polysomnograms and the beginning of CPAP home
treatment. Most papers do not specifically report the lag time
between diagnostic testing and treatment initiation. One study
Review of CPAP and Bilevel PAP
—Gay et al
Page 7
SLEEP, Vol. 29, No. 3, 2006
(Level IV for single arm active CPAP data only) specifically se-
lected patients with moderately severe apnea (AHI 30) but no
daytime sleepiness (ESS < 10) and started sham or therapeutic
CPAP at home the day after titration. There was no significant
difference in the CPAPto for CPAP vs. sham (5 ± 0.4 vs. 4 ± 0.5
hours per night) after 6 weeks in this group of new CPAP users.
This gives some indication of what to expect for adherence and
compliance when CPAP treatment is initiated the day after CPAP
titration PSG.
5.1.2 What is the Expected Acceptance and Adherence When Mea
sured Objectively, Subjectively?
One of the best known studies for this question was a prospec-
tive investigation of CPAP adherence by Kribbs and coworkers
(Level I) which found that subjective and covertly monitored ob-
jective CPAP adherence were discordant and that OSA patients
in aggregate overestimate subjective CPAP adherence compared
with objective adherence measurements obtained by microproces-
sor. Adherence was arbitrarily defined as 4 hours of CPAP us-
age for ≥ 70% of the nights monitored. Although 60% of patients
subjectively reported nightly use of CPAP for a mean of 106.9
days, only 16 of 35 (46%) were objectively using CPAP at least
4 hours per night on 70% of the nights. Patients over-estimated
actual CPAP use by 69 ± 110 min. The regularity of use was de-
termined by the first month use with an overall mean use of 66 ±
37% of the days used for mean CPAPto = 4.88 ± 1.93 hrs/night
with the percentage of time that patient’s were on the effective
CPAP pressure (CPAPeff) = 91 ± 14.7% of the CPAPto time. They
concluded that longer-term CPAP use fell far short of optimal for
a large percentage of patients. A secondary analysis of the Kribbs
study by Weaver, et al,
(Level III) also suggests that patterns of
CPAP usage are manifest within the first week of therapy. Con-
sistent users were redefined as those applying CPAP > 90% of the
nights while the intermittent users skipped CPAP use 1 or more
nights each week. Once patterns of usage were established after
the first week, these patterns again remained stable at 1 and 3
months for the 2 groups.
There were 10 Level I papers
that addressed home
CPAP use over various periods of time looking at different popu-
lations including new and established CPAP users. Objective use
was reported in terms of CPAPto and/or CPAPeff. Studies describ-
ing CPAP acceptance and adherence in the context of other inter-
ventions such as mask interface type, humidification, or follow-up
and education plans are discussed under Section 5.1.3. Only 1 pa-
included both subjective and objective reports in established
users. In Rauschers group of 63 moderate OSA patients CPAP use
of at least 1 year (mean = 539 ± 44 days), the CPAPto of 4.9 ± 0.3
hours per night differed notably from that reported by the patients
at 6.1 ± 0.3 hours per night. Another smaller study
of 17 mod-
erately severe OSA patients with established CPAP use for 820 ±
262 days showed that objective CPAPto in the first year (prior to
participation in the study) was not significantly different during
the subsequent study period (7.1 ± 1.1 vs. 6.9 ± 1.3 hrs/night).
The third and largest study of this group
prospectively measured
long-term objective adherence and adherence rates. Of 233 new
OSA patients enrolled, 19 initially refused CPAP (8.2%) and after
874 ± 48 days, 181 patients continued (84.6%) with a CPAPto =
5.6 ± 0.1(SEM) hours per night (range 0.9-10.3 hours).
Four more Level I studies examined objective adherence rates
during randomized controlled trial (RCT) studies that included a
placebo tablet
or sham CPAP.
In the crossover design placebo
tablet trial,
the CPAPto was lower at 3.7 ± 0.4 (SE) hours per
night with CPAPeff at 89 ± 3% after 4 weeks of CPAP in these
newly diagnosed, moderately severe OSA patients (AHI = 49 ±
1.5). Another RCT study using a placebo tablet arm revealed a
low mean use of CPAP use-3.5 ± 2.1 but showed a bi-modal use
pattern of CPAP with 12 patients using the device 1.7 hours per
night and 11 pts using it for 5.5 hours per night while placebo use
was at 93% of nights.
The first active vs. sham CPAP study
in 60 patients with se-
vere OSA (mean AHI near 60 in both groups) reported objective
CPAPto near 5.5 hours per night in both groups. Results of the
second sham CPAP study are somewhat lower as noted under sec-
tion 5.1.1.
Two of the remaining 3 Level I studies in new CPAP users
looked only at longer-term objective CPAP use. In a study con-
ducted by Reeves-Hoche,
38 of an initially enrolled 47 severe
OSA patients showed a CPAPto of 4.7 hours per night (range 0
- 10.2) and CPAP use at effective pressure was 68% of the stated
sleep time. All but 5 patients (87%) in this study reported all night
CPAP use. Objective measures of CPAP use (time on per hours of
sleep) in these patients did not greatly change after 2 weeks of use
and up to 28 weeks of follow-up. The second large study of 121
new CPAP users with severe OSA
noted a high CPAPeff/CPAP-
to ratio ranging from 94-98% throughout the study measurement
points of 1, 2 and 3 months use. They were able to distinguish
compliant vs. non-compliant groups of patients by using objec-
tive measurements at follow-up. The Kribbs study
has already
been addressed above and compared both objective and subjec-
tive adherence.
There were 13 Level II papers
that related to this
question category with only 2 papers clearly reporting both ob-
jective and subjective adherence. In 1 study
of 204 established
CPAP users (mean 632 days, range 16 - 2921 days) self-reported
CPAP use was 5.8 ± 2 hours per night. In a subgroup of 62 pa-
tients, however, objective (run-time adherence) vs. self-report of
CPAP use was significantly different (p = 0.003) at 5.1 ± 2.5 vs.
6 ± 1.9 hours per night. In the only study of newly diagnosed pa-
tients with very mild apnea,
the overall group CPAPto and CPA-
Peff was not different from oral placebo although notably low at
3.2 ± 2.4 and 2.8 ± 2.1 hours per night. There was, however, a
significant difference (p < 0.001) for the 24 patients in this study
with both subjective CPAP data, (4.5 ± 2.5 hours) vs. objective
CPAPeff (3.5 ± 2 hours). Another paper was especially difficult
to interpret,
enrolling 33 new and established CPAP patients and
randomizing to 3 different forms of follow-up. Patients were also
allowed to read their own meters at 1 to 2 months reporting either
CPAPeff or CPAPto (range 4.4 to 7.1 hours per night) (see 5.1.3
below). The reported objective adherence for SN vs. FN PSG is
noted in response to question 5.1.1 above.
Three papers
reported subjective adherence alone. The
largest of these studies examined 300 consecutive patients
unreported OSA severity established on CPAP for over 6 months.
Eighty-three percent reported nightly use (mean = 7.8 ± 8.1 hours
per night). In a study using a symptom questionnaire,
tively reported continued use of CPAP was 82% (79/96 respon-
dents) in established CPAP users at 3 ± 4 months. The third study
focused on new CPAP users
; 85% of their 96 patients with se-
vere OSA agreed to take home CPAP initially, and 76% reported
Review of CPAP and Bilevel PAP
—Gay et al
Page 8
SLEEP, Vol. 29, No. 3, 2006
continued use at 14.5 ± 10.7 months.
There were 6 more Level II papers
that described only
objective measures of adherence. Three of these papers examined
newly diagnosed CPAP users. The largest study of 54 patients
with a broad OSA severity range, reported a CPAPto = 4.7 ± 0.4
hours per night at 1 to 3 months with the CPAPeff = 89 ± 3% of
CPAPto. In a different study, 39 of 44 more severe, new OSA
were still using CPAP after a mean of 9 months: Thirty-
eight were regular users (97%) at CPAPeff = 5.9 ± 1.1 hours per
night. During the first 9 weeks of new CPAP use in 32 more se-
vere patients,
only 53% were consistent users (> 90% nights)
with CPAPeff = 6.2 ± 1.2 hours per night vs. inconsistent users
who had a CPAPeff of 3.5 ± 1.9 hours per night. In the 3 other
studies with only objective use data, all were established CPAP
users. The biggest study of 117 patients (76% of the original 155
who agreed to try CPAP at home) showed a mean adherence rate
of 5 hours per night at 784 ± 366 days that did not differ from a
smaller subgroup measured at 2 consecutive yearly time points.
In another long-term study,
81% of patients agreed to take home
CPAP and at a mean of 14 months (range 8-39), 68% of these 44
patients with severe OSA had a CPAPto of more than 5 hours per
night. The last study
of 106 patients showed that between 3 and
4 months, patients used CPAP 88% of the days (range 16-100%)
with CPAPto = 4.9 vs. CPAPeff = 4.5 hours per night.
A total of 5 papers met Level III criteria
with one study
simply stating that 63 of 103 OSA patients continued with CPAP
use for 18 months to 10 years but provided no specific daily or
hourly use rate. Another 2 studies
only had objective adher-
ence data available from 12 patients. The 2 remaining studies
both reported subjective adherence rates in established CPAP
users after at least 6 months at over 6.5 hours nightly use. One
of these studies used group education sessions every 6 months
and compared both subjective and objective use.
They noted
the CPAPto use in 25 patients increased from 5.2 ± 0.5 hours per
night at a mean of 463 ± 69 days to 6.3 ± 0.6 hours per night at a
mean of 212 ± 22 days later during which time they subjectively
reported the nightly use to be 6.6 ±.3 hours per night.
5.1.3 What Pre-Initiation Factors Predict or Potentially Affect Accep
tance or Adherence (e.g., Severity of OSA, Pressure Level, Interface
Type, Having a Plan for Treatment of Side Effects, etc.)?
Thirty-four trials evaluated the outcomes of various interven-
tions affecting CPAP adherence and side effects. The remainder of
the studies were case-control or case-series designs. Acceptance
and adherence issues regarding the effect of various interventions
such as humidification, mask interfaces, and education or follow-
up plans are reviewed here but other treatment-effect issues are
discussed under Section 6.1. Studies with CPAP adherence not
designated as the primary endpoint that were reviewed mainly
provided documentation of CPAP-related side effects and in a
number of studies measurement of specific side effect prevalence
(see Section 6.1).
The influence of OSA severity or AHI level on CPAP adher-
ence has been evaluated in 12 studies
(5 Level
I, 5 Level II, 2 level III) with variable results, but the preponder-
ance of evidence (8 positive and 4 negative impact studies) favors
a positive influence of OSA severity on subsequent adherence
with CPAP. An early RCT study
(Level I) which was the largest
of this group (181 of 233 new OSA patients continued CPAP more
than 2 years) reported excellent long-term objective adherence
and compliance rates and hours of CPAP use did correlate with
baseline AHI, but it was expected to be a positive correlation and
that was not the case in this study (r = -0.18, p < .02). Popescu
(Level III) evaluated data derived during
a 2 week home CPAP
trial in 209 patients to identify factors associated with more com-
pliant longer
term use of CPAP. The 153 patients (73.2%) who
accepted home CPAP had a higher baseline AHI (40.4 ± 23.4 vs.
31.8 ± 20.6,
p = 0.016) and 1 year later, 128 (68.5% on an inten-
tion to treat analysis) continued to use
the machine with a mean
use of 5.0 ± 2.4 hours per night. McArdle and colleagues
III) also examined determinants of objective
CPAP use prospec-
tively in 1,155 patients with a median follow-up of 22
The AHI was an independent predictor (AHI 15; p < 0.001) and
was a significant determinant of the hours per night that the CPAP
was used (p = 0.004). Eighty six percent of patients with ESS > 10
and an AHI ≥ 30 were still using CPAP at 3 years. Lack of benefit
and side effects were the most frequent reasons for discontinuing
CPAP. In another long-term study,
high CPAP adherence and
CPAPto was associated with both higher baseline AHI (R = 0.37,
p = 0.013) and the difference in AHI after treatment (R = -0.34,
p = 0.025). In a study of 106 severe OSA patients by Noseda
and colleagues
after 3 to 4 months of use showed good CPAP
adherence rates (CPAPto near 5 hours per night) and a weak but
insignificant correlation with baseline AHI (R = 0.15, p > 0.05).
(Level I) studied 16 consecutively recruited patients
with mild
OSA (AHI = 5 ± 14.9 and 2 or more symptoms
of OSA)
in a prospective RCT with 4 weeks on either placebo or CPAP.
The objective CPAPto was 2.8 ± 0.7 hours per night and 10 of 16
patients preferred CPAP (i.e., opting to continue treatment) but
this proportion was non-significant vs. the placebo CPAP group
(p > 0.4). Those who complied better with CPAP therapy did have
a higher
average AHI (p
= 0.02) than the poorer compliers. In the
(Level I) with 38 of 47 patients agreeing to use CPAP for
severe apnea, the CPAPto was higher than the preceding study
at 4.7 hours per night and although the baseline AHI did not
relate with CPAPto adherence, a high AHI did correlate with
CPAPeff use (R =
0.27048, p = 0.0006). Similarly
(Level I) in
63 established CPAP users after 1 year, severity of baseline AHI
level distinguished patients who objectively used CPAP for > 4
hours per night vs. those using CPAP < 1.5 hours per night (p =
0.049), although higher AHI levels did not identify patients using
CPAPto more or less than 80% of reported sleep time.
There were 4 negative studies beginning with Engleman
(Level II) who in 1 of the later studies reported no correlation
between AHI and CPAP adherence in a larger RCT. Thirty-four
patients with mild AHI (5 to 15) and daytime sleepiness spent 4
weeks on each treatment arm. CPAP use was much lower than
other studies and AHI showed no significant
regression with treat-
ment effects. The lack of correlation between adherence and AHI
was also noted in 1 of their studies described earlier
(Level II).
In another study randomizing patients to having a covert CPAP
use monitor or not as discussed earlier
(Level I), the authors re-
ported generally good adherence over the first 3 months of CPAP
use but they did not find any correlation with CPAPto and AHI
at baseline or after CPAP initiation. Lastly, as noted above un-
der section 4.1, a study that reviewed 125
OSA patients
II), revealed there were no statistically significant differences
tween the compliant and noncompliant patients in baseline AHI.
They concluded that compliant and
noncompliant patients who
Review of CPAP and Bilevel PAP
—Gay et al
Page 9
SLEEP, Vol. 29, No. 3, 2006
have equally severe sleep apnea could have a good initial
to nasal CPAP.
The effect of the necessary CPAP pressure level on CPAP ad-
herence has also been evaluated in 7 studies
(5 Level
I and 1 level III) but most authors make only vague reference
or supply limited details relevant to this issue. The data suggests
there is little to no dependence of CPAP pressure level on subse-
quent adherence. Mixed results were reported in 4 trials compar-
ing therapeutic vs. sub-therapeutic or sham CPAP treatment lev-
els. The first of the 2 parallel RCT trials
(Level I) of therapeutic
vs. sham CPAP was done in 107 men with OSA and sleepiness.
CPAP adherence was 5.4 hours per night (therapeutic) and 4.6
hours per night (subtherapeutic) with therapeutic being superior
to subtherapeutic CPAP in all primary outcome measures. The
use of CPAP by the therapeutic group was 48 minutes per night
longer than that of the subtherapeutic group (p = 0.035). Another
parallel RCT study
(Level I) also compared nasal CPAP set at
therapeutic or subtherapeutic levels of pressure after 1 month in
101 men who had OSA and were sleepy. All outcome measures
also improved significantly in the therapeutic group, compared to
the subtherapeutic group.
The third RCT study that included a sham CPAP arm
ated 60 consecutive patients with moderate to severe OSA who
were randomly assigned to either effective or subtherapeutic nasal
CPAP for 9 weeks on average. Although apneas and hypopneas
were reduced by approximately 95% vs. 50% in the therapeutic
vs. subtherapeutic groups, respectively there was no difference
in adherence. The final study comparing therapeutic and sham
(Level I), as described earlier in section 5.1.1, also showed
there was no difference in CPAP adherence after 6 weeks. These
patients with moderate apnea were also specifically selected so
as not to have any daytime sleepiness or other major OSA symp-
Three other studies
(2 Level I and 1 Level III) assessed ef-
fect of CPAP pressure level on CPAP adherence. The first of these
(Level III) characterized patients who were either able or
unable to tolerate CPAP treatment (non-complaint vs. complaint).
The patients who continued CPAP actually had slightly but sig-
nificantly higher CPAP level requirements (p < 0.01), perhaps re-
flecting a more adequate therapeutic titration. The Kribbs
I) study detailed in several sections above found no difference in
CPAP pressure requirements between regular and irregular CPAP
users. The last study
also showed no difference between the de-
gree of prescribed CPAP pressure and adherence.
There were 4 Level I studies
that utilized an RCT, cross-
over or parallel design and there was 1 Level IV study
compared different nasal, oral, or full-face masks. A parallel RCT
was performed with full-face vs. nasal mask CPAP therapy
for 4 weeks in 20 OSA patients with or without previous uvulo-
palatopharyngoplasty (UPPP). Nightly CPAPto adherence after 1
year was higher with a nasal vs. full-face mask (5.3 ± 0.4 vs. 4.3
± 0.5 hours per night, p = 0.01) for OSA patients but in patients
with previous UPPP and OSA, adherence was only marginally
higher with nasal vs. full-face masks (CPAPto = 5.1 ± 0.7 vs. 4 ±
0.8 hours per night, p = 0.07). During this RCT trial showing bet-
ter adherence with the nasal vs. full-face mask, nasal masks were
rated more comfortable by 19 of 20 patients (p < 0.001) despite
more mouth leak related symptoms. The authors concluded that
adherence is greater with a nasal vs. full-face mask because the
overall comfort is better and compensates for increased symptoms
associated with mouth leakage. The Level IV study
that also
studied full-face CPAP mask in 10 males who could not toler-
ate nasal CPAP due to nasal
congestion compared the effects of
a therapeutic
level of CPAP pressure applied through a nasal or
full-face mask in patients with moderate-severe OSA. The
response (reduced to < 8 events hour) on CPAP nights were simi-
lar regardless of mask type so the authors concluded that the full-
face mask may be a reasonable treatment alternative in patients
who cannot
tolerate nasal CPAP but they did not assess CPAP
adherence rate.
In 1 of 2 studies
comparing a nasal vs. oral mask (Oracle
mask, Fisher-Paykel, Auckland, NZ) was done in 21 newly di-
agnosed severe OSA patients (mean AHI = 85 per hour) for 4
weeks and found there was no significant difference on objective
CPAP adherence CPAPto and % pts using at least 3 hours per
night. Their values however were notably low compared to other
studies (nasal vs. oral CPAPto = 3.8 vs. 3.5 hours night; percent-
age of patients using 3 hours per night = 62 vs. 57%). Another
nasal vs. Oracle mask study
with a parallel RCT design of 38
patients with severe OSA (mean oral vs. nasal RDI = 58.5 vs.
63 per hour) reported objective and subjective adherence rates
at 1 and 2 month follow-up. There was no significant difference
for either mask at either follow-up periods for objective average
hours per night [oral vs. nasal month 1 and month 2 CPAPto (with
percentage of patients using 4 hours per night) = 4.6 ± 2.1 (52%)
vs. 4.3 ± 2.6 (47%) at 1 month and 5.5 ± 2.6 (73%) vs. 4.6 ±
2.5 (67%) at 2 months]. Twenty-nine percent of patients in each
group dropped out by 2 months. The subjectively reported aver-
age hours per night tended to be somewhat higher than objective
reports (subjective oral vs. nasal month 1 and month 2 CPAP time
= 5.8 ± 1.4 vs. 5.8 ± 1.7 at 1 month and 5.8 ± 1.7 vs. 5.7 ± 2.6 at
2 months).
In the last interface study, an RCT crossover trial
was done
on 39 new patients with OSA comparing nasal pillows (Breeze;
Mallinckrodt Corporation; Minneapolis, MN) and a nasal mask
(Contour; Respironics; Murrysville, PA) after a 3-week treatment
period. The percentage of days CPAP was utilized favored the
nasal pillows (94.1% vs. 85.7%; p = 0.02), but nightly minutes of
use were similar (nasal pillows, 223 min; nasal mask, 288 min).
Fewer adverse effects, less sleep difficulty and air leak occurred
with nasal pillows (p < 0.04). The authors concluded that nasal
pillows were associated with fewer adverse effects and better
sleep quality.
Six studies
are published evaluating the use of humidifi-
cation to augment CPAP adherence and also evaluated other inter-
ventions effects. As noted earlier, Massie et al,
(Level I) found
that a heated humidifier increased adherence compared to either
room temperature humidifiers or no humidification and specific
side effects such as dry mouth or throat and dry nose were reported
less frequently when CPAP was used with heated humidity com-
pared to CPAP use without humidity (p < 0.001). Subjective pa-
tient satisfaction with treatment was equally improved for heated
or room temperature humidifiers compared to no humidification.
A similar study by Neill, et al,
(Level I) demonstrated a small
improvement in adherence with heated humidification, but no dif-
ference in subjective sleepiness or treatment satisfaction. A Level
II study
examined initial preferences in new CPAP users over
2 consecutive nights either with or without humidification. They
concluded that the use of humidity during the initiation phase of
CPAP treatment was associated neither with an initial improve-
Review of CPAP and Bilevel PAP
—Gay et al
Page 10
SLEEP, Vol. 29, No. 3, 2006
ment in comfort nor with greater initial treatment acceptance.
Three lower evidence level studies
have also evaluated
heated humidification use with CPAP. One study
(Level III)
of 24 OSA patients complaining of serious CPAP-related upper
airway dryness were randomized to 6 weeks of either heated hu-
midification or oily nose drops demonstrated that heated humidi-
fication was more effective than non-heated humidification in
increasing measured absolute humidity. No patient that received
heated humidification discontinued CPAP therapy but only 5 of
12 patients (42%) in the oily nose drops group reported their de-
gree of upper airway dryness to be improved (P =0.003) while
all 3 of the patients who intended to terminate CPAP did so. A
second study
(Level III) evaluated how nasal CPAP therapy in-
fluences the relative humidity (rH) of inspired air; and how the
addition of a heated humidifier or a full-face mask may effect the
rH in 25 OSA patients receiving long-term nasal CPAP therapy
and complaining of nasal discomfort. When comparing the values
obtained with CPAP alone, heated humidification significantly
increased rH during the sleep recording, both when the mouth
was closed and during mouth leaks but they did not assess the ef-
fect on humidity on CPAP adherence. A retrospective, case-series
design study
(Level V) looked at the effects of CPAP on infec-
tious complications by analyzing the kinds and rate of infections
of the upper airway in 246 consecutive patients with or without
a heated humidifier and compared them with OSA patients who
did not receive CPAP therapy. Forty patients received conserva-
tive therapy and 206 CPAP treatments, 36 of them with a heated
humidifier with a mean follow-up of 165.4 ± 92.1 weeks. Patients
using CPAP without the humidifier had significantly more upper
airway infections than controls (42.9 vs. 25%; p < 0.05), and more
patients on CPAP therapy with humidifier than controls (22.2 vs.
2.5%; p < 0.01) reported an increased rate of upper airway in-
fections since initiation of CPAP therapy or diagnosis of OSAS.
Patients who did not adequately clean their heated humidifier de-
vices had significantly more upper airway infections since diag-
nosis (57.1 vs. 20%; p < 0.05) or during the past 6 months (52.4
vs. 13.3%; p < 0.05) than patients who regularly cleaned CPAP
machines, humidifiers and ventilatory circuits. They concluded
that humidification, especially if daily cleaning of the humidifier
is not facilitated, is an increased risk of infection but the effect of
this on adherence was not assessed.
Six studies, all prospective, (4 Level I,
and 1 each, Level
and Level III
) have evaluated the impact of enhanced pa-
tient education programs on CPAP adherence. Three out of the 4
studies found that increased intensity of patient education or fre-
quency of health provider contact resulted in improved adherence
rates. Chervin, et al,
found that additional printed patient educa-
tion materials or weekly health provider telephone calls increased
adherence compared to the control group. Another RCT study
of 80 new CPAP patients on more intensive vs. standard follow-
up reported a significant difference for each group in time at ef-
fective CPAP level of 5.4 hours per night vs. 3.8 hours per night
at 6-month follow-up. Hoy et al,
combined 3 nights of initial
observation in a sleep laboratory with weekly home visits from a
nurse, with a resultant increase in mean usage from 3.8 hours per
night to 5.4 hours per night after 6 months (p = 0.003). This inten-
sive approach to promoting CPAP adherence increased hours of
nightly CPAP usage at 1 and 3 months of CPAP treatment. Patient
reported CPAPto in the Palmer study was assessed serially and
baseline and 3 months after randomization to either nurse home
visits or office consultant review.
The study was conducted in
139 long-term established CPAP users of unknown OSA severity
(AHI not provided) whose baseline nurse vs. consultant CPAPto
of 4.9 vs. 5.2 hours per night increased significantly (p < 0.04) in
both groups to 5.9 vs. 5.6 hours per night but there was no dif-
ference related to the follow-up person. In contrast to the other
5 studies, Hui et al,
found no change in CPAP adherence with
weekly phone calls from a nurse but did demonstrate improved
quality of life for patients attempting to use CPAP.
A very recent study
(Level I) assessed the impact of a com-
puter-based telephone system (TLC) designed to improve initial
CPAP adherence compared to usual care in patients with moder-
ately severe OSA (mean AHI near 40). TLC is a computer-based
system that provides structured education and reinforcement for
CPAP use. New CPAP users were enrolled and at the 2-month
follow-up, CPAPeff for the TLC group was significantly differ-
ent at 4.4 hours per night compared with 2.9 hrs in the usual care
group and there was a more significant (p = 0.047) reduction in
the sleep symptom scores for the TLC group. They did not report
the lag time between diagnosis and initiation of CPAP therapy or
the difference in adherence between the
2 subgroups of different
apnea severity.
The second Level I augmented education/follow-up RCT
reported both objective and subjective adherence rates
in 108 new patients with moderate-severe OSA (AHI = 48) at 4
and 12 weeks follow-up. There was no significant difference in
objective CPAPto at both follow-up periods for both the basic and
augmented support groups at all time periods averaging 6.3-6.5
hours per night with over 70% of patients using CPAP ≥ 4 hours
per night for at least 70% of the nights of the week.
6.1 Once CPAP Treatment Has Been Initiated and Excluding Accep
tance and Adherence Issues From Question 3, How and How Often
are the Following Documented to Determine Efficacy, Effectiveness
and Safety?
The literature search identified 497 studies that met the extrac-
tion criteria for this question. After an initial review excluding
studies with only peripheral relevance to this body of literature,
64 studies were fully extracted and referenced in the evidence
tables. There were 17 Level I, 7 Level II, 14 Level III, 15 Level
IV, and 11 Level V studies utilized. Case studies and articles re-
lated to comments are not included in the evidence table but are
6.1.1 Adequacy of CPAP Treatment Including Pressure Settings and
Interface Type
The occurrence of patient complaints related to pressure in-
tolerance or interface problems are the most commonly reported
CPAP-related side effects (See section 6.1.3). As CPAP usage pat-
terns are established very early after the initiation of treatment,
pressure or interface related side effects are often promptly ad-
dressed (
Level III,
Level IV). Recommendations for longer-
term provider follow-up are not certain, but equipment obsoles-
cence, including mask, headgear, tubing and other items suggest
yearly health care provider follow-up reasonable as a method
choice for many studies. The need for serial adjustment of CPAP
settings in the clinical setting is not known, but studies do sug-
gest that initial laboratory specified pressures are usually a few
cm H
O higher than that specified on repeat titration study a few
Review of CPAP and Bilevel PAP
—Gay et al
Page 11
SLEEP, Vol. 29, No. 3, 2006
weeks later (see details below). The use of APAP is not included
in this review. In addition, the possible use of bilevel PAP to ad-
dress pressure intolerance is addressed in Section 7.1. The need
for repeat or serial long-term (for example, every year) CPAP ti-
tration studies was not well supported, although logic encourages
recheck for persistent adherence problems or the recurrence of
symptoms of untreated OSA.
Three studies revealed that discontinuation of CPAP either
1 night (
Level III) or even half of the night (
Level V)
resulted in the recurrence of obstructive respiratory events and
the clinical sequelae of untreated OSA including hypersomnia.
Despite adherence with therapy, OSA usually persists with un-
changed severity over 2 years after diagnosis (
Level III). How-
ever, pressure requirements vary with time and other studies sug-
gested that CPAP level is decreased approximately 2 cm within 2
to 4 weeks of initiating treatment (
Level II). Initial titration in
the sleep lab was higher based on a 2-month interval evaluation.
Bureau showed that subsequent downward titration of CPAP with
introduction of CPAP during PSG can reduce the optimal CPAP
recommendation (
Level IV).
No studies were found to determine whether a fixed CPAP
level specified soon after diagnosis was adequate after 3 months
or longer after the start of treatment. There have been no studies
demonstrating the need for routine, serial CPAP titration studies
such as on an annual or bi-annual basis. Specifically, no study has
directly evaluated whether an initially adequate CPAP level might
result in an elevated AHI at a later time but it is known that the
prevalence of OSA increases with aging.
In the face of significant weight loss, CPAP pressure levels may
need to be adjusted. Obesity is a well-recognized risk factor for
the occurrence of OSA and studies have shown that the severity
of OSA is reduced in many patients who lose weight. In fact, the
need for CPAP treatment may be negated for some patients with
weight loss.
Changes in CPAP level requirements with serial
titration studies at various weights for individual OSA patients
have been cited in case reports and case-series with small num-
bers (fewer than 10 subjects) such that these were not included in
this literature review. No larger studies have evaluated the pres-
sure requirement over time with changes in weight.
Four studies evaluated the use of humidification to augment
CPAP adherence. Massie et al,
(Level I) determined that a
heated humidifier increased objectively measured adherence
compared to either room temperature humidifiers or no humidi-
fication. However, subjective patient satisfaction with treatment
was equally better for heated or room temperature humidifiers
compared to no humidification. A similar study by Neill, et al,
(Level I) demonstrated a small improvement in adherence with
heated humidification, but no difference in subjective sleepiness
or treatment satisfaction. Two lower evidence level studies have
also evaluated heated humidification use with CPAP. A study by
Martins de Araujo et al
(Level III) demonstrated the effective-
ness of heated humidification in decreasing nasal discomfort even
in the face of persistent mouth leak. However, a retrospective,
case-series design study (
Level V) suggested an association be-
tween CPAP use and upper respiratory infections which is com-
pounded by the use of heated humidification, especially if daily
cleaning of the humidifier is not facilitated.
Several controlled studies are available addressing the advan-
tages or disadvantages of specific interfaces for CPAP as a prima-
ry endpoint. Described interfaces include nasal masks, oronasal
masks, nasal prongs, and oral only masks. Mask fit and com-
fort, as well as the presence or absence of mask leak and mouth
leak, are specific issues that are assessed in the studies evaluat-
ing CPAP side effects as discussed below. During the RCT trial
showing better adherence with the nasal vs. full-face masks (
Level I), nasal masks were rated more comfortable by 19 of 20
patients (p < 0.001) despite more mouth leak related symptoms.
This study determined that pressure requirements are higher for
patients using oronasal CPAP masks. One study of nasal versus a
novel oral interface found no difference in CPAP adherence at up
to 2 months follow-up with equivalent CPAP pressures (
I). Massie et al found improved adherence and fewer overall ad-
verse effects (p < 0.001) with nasal prongs vs. one type of nasal
interface but the pressure requirements were the same (
No studies evaluated the impact of a ramp mechanism used
with the initiation of CPAP on adherence or side effects. This
mechanism is intended to facilitate a gradual increase in pressure
over time during the initiation of sleep. One case report cites re-
peated patient activation of the ramp feature which led to marked
recurrence of apneas and hypopneas.
6.1.2 Treatment Effect Such as Continued Reduction in AHI or Re-
spiratory events, Improvement in Sleepiness, Psychological Ben-
efit and Quality of Life, or Systemic Blood Pressure
The discussion for Section 3.1 provides a comprehensive re-
view of the physiologic and performance-based outcomes that oc-
cur with effective CPAP treatment. A preponderance of evidence
demonstrates the resolution of the nocturnal respiratory events
with standardized methods of CPAP titration. Section 6.1.1 noted
the reoccurrence of disordered breathing with abrupt withdrawal
as well as the persistence of AHI elevation even
after 2 years of CPAP use so there is no clear evidence of a treat-
ment effect on baseline AHI or other respiratory events.
A recent literature review on the clinical indications of the
included many studies assessing the utility of this test
in assessing response to CPAP treatment in OSA. The majority of
these studies demonstrated an increase in the mean sleep latency
with CPAP treatment. However, one should note that both pre-
treatment and post-treatment means were within 1 standard de-
viation of normal control means. There were 13 studies reporting
MSLT results cited in the evidence table for this section with 5
Level I studies showing a positive effect from CPAP
1 of these studies showed the CPAP benefit after 4 weeks with
mean use of < 4 hours per night
. Two studies showed no change
in the MSLT with CPAP use between 2 weeks and 4 months.
Two more reports noted no difference on the MSLT when com-
paring intensive follow-up to routine care (
Level V,
Level I).
Three lower evidence level studies also supported a significant
improvement in the MSLT with CPAP treatment of 24 hours
to 6 months or more (
Level III,
Level IV). Two other lower
evidence level studies showed a relation between adherence rate
and MSLT benefit (
Level IV,
Level III) and 1 of these dem-
onstrated that the MSLT change was sensitive to even a 1 day
withdrawal of CPAP treatment.
None of these investigations
revealed a direct impact of the MSLT changes on the management
of CPAP pressure adjustment or need for reassessment.
As discussed in question one, the Epworth Sleepiness Scale
(ESS) has been evaluated as an indicator for OSA severity and
Review of CPAP and Bilevel PAP
—Gay et al
Page 12
SLEEP, Vol. 29, No. 3, 2006
used in clinical settings to assess treatment response.
were 23 studies reviewed for this section that utilized the ESS
or some other measure of subjective sleepiness to assess CPAP
response (13 Level I, 1 Level II, 4 Level III, and 5 Level IV). A
poor correlation between ESS, pre-treatment AHI and measured
mean sleep latency has been reported (
Level II). The ESS value
also did not show an effect from humidity (
Level I), relation-
ship with adherence or age (
Level III) or with more intensive
follow-up (
both Level I). Two more studies did not show any
effect of CPAP treatment on the ESS after either 1 week (
I) or up to 2 months of treatment (
Level I). The preponderance
of studies however did show improvement in the ESS with CPAP
treatment (9 Level I
; 2 Level III,
and 5 Level
). Two of the Level I studies reported improved ESS
scores at 4 weeks with less the 4 hours per night adherence
2 other studies stated either nasal prongs
or a full-face mask
showed a positive effect on the ESS.
Sleepiness from all causes, especially OSA, is well known to
negatively affect driving performance. In OSA patients (untreated
or with increasing OSA severity as measured by AHI), 3 of 4 stud-
ies (
Level IV,
Level II,
Level IV) have shown a posi-
tive effect on reduction in motor vehicle accident rates. Kribbs,
et al
(Level I) did not find an increase in self-reported, sleepi-
ness related near-miss motor vehicle accidents associated with de-
creased, objectively measured CPAP use. However, re-analysis of
these data
did demonstrate that consistent CPAP users reported
fewer accidents than intermittent users. The Level IV studies
showed reduced self-reported driving accident rates with initia-
tion of CPAP and a relationship with severity of OSA. Self-re-
ported motor vehicle accidents may not be an accurate end-point
due to the potential for its under-reporting in non-compliant CPAP
patients who could be legally liable when an accident occurs.
A positive effect with CPAP treatment on objective driving per-
formance tests was demonstrated in 4 of 5 studies (
Level I,
Level III,
Level IV,
Level III,
Level IV). CPAP treatment
in a high evidence level study improved driving performance
measured by a simulator after only 4 weeks of treatment (
I). Although daytime sleepiness decreased with improvement of
other vigilance tests, driving simulation by ‘steer-clear testing
did not change with CPAP up to 12 months later in one lower
evidence level study.
The Functional Outcomes Sleep Questionnaire (FOSQ) is a
self-administered survey which evaluates the impact of sleepiness
on the ability to perform activities of daily living (
Level II).
The FOSQ score improves with CPAP treatment after 6 weeks
compared to sham CPAP(
- Level I) and up to 6 months (
Level I) but it was not sensitive in confirming the advantage seen
by other parameters for nasal prongs over a conventional CPAP
The concordance between continued, objectively mea-
sured CPAP use or other parameters and the FOSQ has not been
evaluated over longer time intervals.
In the study that randomly imposed immediate CPAP after PSG
within 2 weeks or a 6 month delay
for PSG and CPAP, the pri-
mary objective was to look at the effect of delayed treatment on
daytime sleepiness, cognitive function, and healthcare expendi-
tures. Many studies have included quality of life, performance,
and mood measures before and after treatment with CPAP. In un-
treated severe OSA patients, quality of life was often assessed with
the Medical Outcomes Study Short Form 36 (SF-36) with several
high evidence level studies (
- Level I;
- Level III;
- Level
IV). At baseline, SF-36 was decreased in a number of domains
and all studies showed general improvement with CPAP.
The relationship between quality of life, mood or depres-
sion, cognition and other neuropsychiatric variables in patients
with obstructive sleep apnea syndrome has also been assessed
with a wide variety of investigative tools. There were 17 addi-
tional studies beyond those utilizing the SF-36 that investigated
these areas (
Level I;
Level II;
Level IV,
Level V). Although the vast majority of
studies showed improvement with selected testing after CPAP,
several did not. One Level III investigation
demonstrated that
although most neuropsychiatric deficits normalized with treat-
ment, planning abilities and manual dexterity did not normalize
after 6 months of CPAP use. The authors speculated that since
the latter parameters have been found to highly correlate with the
severity of nocturnal hypoxemia, patients with more severe apnea
may have irreversible deficits. Two Level I studies found either
that CPAP improved both objective and subjective sleepiness, but
not psychometric parameters after 4 weeks,
or that there was
no improvement in daytime function at just 2 weeks of CPAP
Slight but positive benefit was seen in cognitive tests but
not in neuropsychiatric tests (
Level II) after 3 and 12 months,
but memory and concentration tests were not sensitive in showing
the benefit of humidification with CPAP use.
revealed that delayed initial use of CPAP also did not change cog-
nition despite benefits seen in quality of life. The selection of
testing technique, degree of standardization, and the characteris-
tics of the baseline population as well as the intervention must be
considered since all these issues can have an effect on the find-
ings of individual studies.
Section 3.1 addresses the immediate and shorter term impact
of CPAP on systemic hypertension; 6 of the studies are included
for our discussion here (
Level I;
Level II;
Level IV;
Level V). Although treatment with CPAP does decrease day
and/or night blood pressure after variable treatment times, the
treatment effect may be relatively modest and it is often the case
that therapy with antihypertensive medication must be continued.
However, Becker et al
found significant improvement in mean
arterial blood pressure both day and night with effective CPAP
(95% reduction in AHI) vs. subtherapeutic CPAP (50% reduction
in AHI). Another higher evidence level study
demonstrated a
decrease in day and nighttime blood pressure after only 3 nights
of CPAP with restoration of the expected circadian effect with a
nighttime ‘dip’ in blood pressure. Two studies showed a predomi-
nant effect on diastolic pressure after only 2 weeks
or at 12 - 14
of CPAP use. Two final lowest evidence level studies
showed either a day and night blood pressure reduction only in
compliant CPAP users at 8 weeks or a positive response in as
little as 8 days. Labile hypertension is well described in “case-se-
ries” design studies in untreated OSA, but very few studies have
evaluated this as a definite marker for non-adherence with CPAP
6.1.3 Adverse Events Including Equipment Failure, Interface Prob
lems and Other Side Effects or Complications From CPAP Usage
Early adherence monitoring seems critical as the literature sug-
gests many side effects can occur during the first few weeks of
CPAP use and may lead to discontinuation of treatment.
sible side effects are listed in Table 3. Overall, there were 24 stud-
Review of CPAP and Bilevel PAP
—Gay et al
Page 13
SLEEP, Vol. 29, No. 3, 2006
ies entered into the evidence table for specifically commenting on
complications or side effects. These were comprised of both high-
er and lower evidence levels (7 Level I studies-
; 3
Level II-
; 4 Level III-
; 4 Level IV-
; and
6 Level V-
). Less bothersome side effects may persist
long-term (up to 2 years) but are less likely to impact on adher-
; however, decreased adherence or complete refusal did
correlate with more side effects.
Studies designed with
frequent patient interactions suggest > 70% adherence after 6
months despite persistence of side effects for many patients.
Many studies evaluating CPAP treatment side effects as a pri-
mary endpoint are retrospective in design. Several lower evidence
level studies reported CPAP side effects as secondary endpoints
with general descriptions or lists of these side effects
were relatively consistent and dominated by upper airway symp-
toms. The percentages of OSA patients experiencing various
side effects are contained in a number of these prospective stud-
Early retrospective studies suggested CPAP side effects oc-
curred in > 50% of OSA patients and were persistent even for
compliant patients.
Airway dryness was the most frequent
complaint (> 40%
) for studies that were completed prior to
the availability of humidification systems. Nasal interface side ef-
fects were noted to occur in > 50% of all CPAP patients in a number
of studies and included skin abrasion, and mask leak.
More recent studies suggest similar frequencies of CPAP side ef-
fects within the first few weeks of treatment, but the development
and application of airway humidifiers, and a multitude of nasal
and oronasal interfaces may have led to increased longer term ad-
herence rates.
Unfortunately, there are no studies show-
ing benefit when comparing adherence rates serially in groups of
newly treated patients from the same clinic or study population
before and after implementation of some of these improvements.
During a nasal vs. oronasal mask trial,
there was more nasal
congestion or dryness and air leaks with the nasal mask and more
oral dryness and gum pain with the oral mask, but these issues
had no impact on adherence or other aspects of CPAP treatment.
There was no significant difference in CPAP pressure, side effect
scores, or mask preferences during this crossover study.
6.2 Summary
The frequency with which efficacy and safety of CPAP needs
to be assessed is not specifically clear but benefit and adherence
appears to be determined within the first few weeks of treatment.
Available evidence does not direct the need or timing for serial
repeat titration studies and does not show a strong effect of CPAP
pressure settings. Although there is clear support that CPAP re-
solves most respiratory events and improves sleepiness, there was
no evidence for a continued change over time and resolution of
the underlying disorder without the use of CPAP. The resolution
of sleepiness was coupled with improved driving performance
and the majority of studies revealed a positive benefit on psy-
chometric or vigilance, neurobehavioral, and quality of life mea-
sures. However, the large variation in testing methods, population
selection, and interventions made it difficult to form firm conclu-
sions. The effect of CPAP on blood pressure appears to be vari-
ably significant during the nighttime or daytime and on systolic
versus diastolic readings. Other cardiovascular parameters were
not assessed in this review. Finally, there were a myriad of report-
ed side effects and complications of treatment with some impact
seen from the use of humidification and different interfaces.
6.3 Future Research
The utility of continuously monitoring the multitude of effects
with CPAP usage is unknown. Research might best focus on more
precise description of the patient population and selection tech-
niques as well as an attempt to standardize testing techniques.
Clarification of high impact interventions for the most prominent
adverse events would also seem worthwhile. Certainly, the persis-
tence or recurrence of the complaint of drowsiness while driving
in OSA patients requires additional clinical evaluation and inter-
Table 3—Possible CPAP-Related Side Effects
Mask leak
Skin abrasion/ulceration (pain)
Mask allergy
Conjunctivitis/Sore eyes
Dermatitis/facial irritation
Pressure-Related (Airway)
Otitis/Ear pain
Air swallowing/Aspiration
Mouth leak (dry mouth) or mask leak
Pressure intolerance
Sense of suffocation or difficulty exhaling
Central sleep apnea
Prolonged oxyhemoglobin desaturations
Equipment Related
Tubing condensation
Cumbersome equipment
Spousal intolerance/less intimacy
Ramp abuse
Equipment maintenance and cleaning
Equipment Failure
Lifespan of machine, tubing and mask
Recurrence of OSA
Periodic limb movements
Fatigue/Feeling tired
Chest discomfort
Review of CPAP and Bilevel PAP
—Gay et al
Page 14
SLEEP, Vol. 29, No. 3, 2006
vention. Widely applicable methods to monitor and detect OSA
patients at particularly increased risk for driving accidents are yet
to be established. A more in-depth assessment of other cardio-
vascular responses such as arrhythmia and cardiac function after
PAP treatment would be valuable. Lastly, continued efforts to de-
sign and assess new interfaces and PAP delivery modes should be
strongly supported.
7. 1 Bilevel Positive Airway Pressure
Bilevel PAP may be used as an alternative therapy to CPAP in
OSA or may be used to treat nocturnal breathing disorders other
than OSA. Only 88 of 752 articles cited by the PAP search criteria
in the years 1991-2003 addressed the use of bilevel PAP. Often,
studies of bilevel PAP in OSA include patients with coexisting
respiratory conditions. Four studies
addressed the use of
bilevel PAP in patients who have OSA without comorbid respi-
ratory conditions such as daytime hypercapnia or restrictive or
obstructive lung disease.
The limited number and variable design of studies employing
bilevel PAP required a modification of the criteria for accepting
studies for clinical evidence grading. Studies employing bilevel
PAP for nocturnal breathing disorders other that OSA often used
a sample size of less than 10, failed to use a sham positive pres-
sure treatment group, and frequently identified outcome measures
such as awake or sleeping gas exchange rather than polysomno-
graphic sleep parameters. Limiting evidence grading to studies
employing polysomnography or studies with ≥ 10 subjects would
have eliminated many of the 88 articles reviewed. On the other
hand, including studies of negative pressure ventilation, volume
cycle ventilation, or those study designs documenting only the
immediate awake responses to bilevel PAP is beyond the scope
the question being addressed. In order to allow a sufficient num-
ber of studies for review and yet ensure some consistency of study
design, the following inclusion criteria for evidence grading were
used: a) 5 subjects, b) any of 4 outcome measures (preference or
use of PAP, sleep parameters, measures of daytime sleepiness, or
measures of gas exchange), and c) use of bilevel PAP during sleep
periods. Polysomnography was not a requirement for evidence
grading. Of 88 studies using bilevel PAP, 18
met these
criteria for evidence grading, 6 studies
employed a
randomized controlled trial design and 4 used sham PAP in the
control arm.
Each of these 18 articles was evaluated us-
ing the evidence grading listed in Table 1.
7.2 When Should Bilevel PAP Be Used Instead of CPAP in OSA
7.2.1 Adherence, Comfort, and Preference.
Improvement in patient adherence to the use of PAP or im-
provement in mask comfort is a potential benefit of bilevel PAP
in OSA patients. Two Level I studies
of bilevel PAP vs. CPAP
for OSA patients without coexisting daytime respiratory disease
demonstrated no difference in effectiveness or long term adher-
ence with CPAP. In 1 of these Level I studies, adherence as mea-
sured either by the average hours that CPAP was used per night
or the time during which pressure was applied per night was not
different with bilevel PAP as compared to CPAP.
In this study,
patients related no difference in complaints in the rate of mask
discomfort or nasal symptoms between CPAP and bilevel PAP.
The second Level I study of OSA patients without coexisting
daytime respiratory disease demonstrated no difference in the
percentage of nights with at least 4 hours of use between bilevel
There are no Level I studies comparing adher-
ence of CPAP and bilevel PAP in OSA patients with coexisting
disease and limited evidence for patient preference in this popula-
tion. Two Level III studies suggest a subset of patients with OSA
and comorbidity prefer bilevel PAP as compared to CPAP.
these Level III studies, patients with OSA who preferred bilevel
PAP were more obese and had more resting hypercapnia and arte-
rial desaturation during the day
or demonstrated more spi-
rometric evidence of airflow obstruction as measured by FeV1/
7.2.2 Efficacy
The 2 Level I studies comparing CPAP and bilevel PAP in OSA
patients without coexisting daytime respiratory disease demon-
strated no difference in the improvement in RDI with PAP.
Subjective assessment of sleepiness (ESS) and sleep quality
(FOSQ) were not different with CPAP as compared bilevel PAP
in 1 of these studies.
Similarly, 2 Level III studies demonstrated
no difference in RDI between CPAP and bilevel PAP in OSA pa-
tients with coexisting COPD, obesity, or hypoventilation.
1 Level III study of OSA patients who exhibited a nocturnal rise
in PCO
of > 19 mmHg, subjective reports of morning headaches,
insomnia, and daytime hypersomnolence resolved with bilevel
There was no comparison to the effect of CPAP in this
7.3.1 What are the Indications for Bilevel PAP Treatment of Noctur
nal Breathing Disorders in Other than OSA Patients?
Eleven studies have examined the effect of bilevel PAP in
patients with obstructive and restrictive lung disease
with the intent to support ventilation rather than to im-
prove RDI. Because treatment studies have been targeted to these
specific categories of lung disease and because of the limited
number of studies and the heterogeneity of chosen primary end-
points, the discussion will focus on outcome measures according
to disease category.
7.3.2 Efficacy of Bilevel PAP in COPD Patients.
The most persuasive precedent for using bilevel PAP in breath-
ing disorders derives from 7 randomized controlled trials in the
setting of acute respiratory failure complicating COPD.
studies often employed relatively high initial inspiratory pres-
sures (15 - 20 cm H
0) that were based on a preliminary study
demonstrating improvement in objective measures of respiratory
distress and gas exchange in COPD patients with acute respira-
tory failure.
Higher ventilatory pressures might be required
in COPD as compared to neuromuscular disease because of the
higher respiratory impedance in COPD. Though the studies of
bilevel PAP in COPD patients with acute respiratory failure show
a reduction in mortality rate and a reduction in the need for inva-
sive ventilatory support,
the role of bilevel PAP as therapy dur-
ing sleep in chronic respiratory failure is less defined. Theoretical
benefits from nocturnal ventilatory support might include im-
provements in awake and sleeping gas exchange due to increased
sleeping ventilation, deeper and less interrupted sleep, and in-
Review of CPAP and Bilevel PAP
—Gay et al
Page 15
SLEEP, Vol. 29, No. 3, 2006
creased awake respiratory muscle strength after resting fatigued
respiratory muscles during sleep.
Four Level II studies
have addressed the use of bilevel
PAP during sleep in clinically-stable hypercapnic COPD patients
> 44 or 45 mmHg). These studies were somewhat hetero-
geneous with some differences in patient populations and inten-
sity of treatment. A wide range of inspiratory pressures (10 - 22
0) was employed and the range of sample size completing
the protocols on treatment was only 6 to 14 patients. One small
randomized study with sham PAP in a control group addressed ad-
herence and acceptance.
Adherence determined by timers was
not different between sham and bilevel PAP. However, more than
50% of patients dropped out at 3 months in the bilevel PAP arm
without attrition in the sham PAP group. This raises substantial
concerns regarding long-term acceptance of bilevel PAP in COPD
patients. In another study,
patients with mild sleep-disordered
breathing were included (average AHI =10). Comparison of the
AHI in the patient group in this study with the 3 other studies is
problematic because variable or unspecified definitions of hypop-
nea were used. Changing the threshold for definition of hypop-
neas can substantially change the numeric value of AHI.
methods developed for adjusting bilevel PAP to improve objective
measures of respiratory distress in acute respiratory failure during
wakefulness, none of the Level II studies employing bilevel PAP
during sleep employed titration of pressures to a targeted level of
nocturnal gas exchange or sleep parameters.
The results of these 4 heterogeneous Level II studies were in-
consistent and employed different outcome measures. Three of
these studies
show no effect of bilevel PAP on daytime
gas exchange whereas 1 study
demonstrated improvement in
both nocturnal and daytime PaO
and PaCO
(daytime PaO
+ 5.9
mmHg [effect size .9] and PaCO
- 4.5mmHg [effect size .8]) as
well as an improvement in quality of life scores. Total sleep time
and sleep efficiency were improved in 2 studies
that used rel-
atively high inspiratory pressures (18 and 22 cm H
0) and either
failed to improve or deteriorated in 2 studies
employing rela-
tively low inspiratory pressures (10 and 12 cm H
0). Since none
of the studies with negative findings included a power analysis,
there is a reasonable probability of a Type II error. It is unclear
whether the inconsistent results of these RCTs of bilevel PAP dur-
ing sleep reflect differences in patient selection, differences in the
methods of treatment that were employed, or the possibility that
COPD patients with stable hypercapnic respiratory failure derive
less benefit than patients with acute respiratory failure. In addi-
tion, the relatively small sample sizes employed and the failure of
the protocols to adjust pressures to a targeted effect during sleep
significantly increases the likelihood of false negative results in
these studies. However, improved sleep quality in positive studies
suggest that higher inspiratory pressures may be more effective
than lower pressures.
7.3.3 Efficacy of Bilevel PAP in Patients with Restrictive Disease
Eight Level III studies
have addressed the
use of bilevel PAP with a wide variety of restrictive lung disor-
ders. Most of these studies have heterogeneous patient popula-
tions with a mixture of neuromuscular diseases and chest wall
restriction. With chronic nocturnal bilevel PAP, improvement
in daytime hypercapnia was noted in 3 of 4 studies that evalu-
ated daytime gas exchange.
Daytime respiratory muscle
strength increased in both studies which evaluated this param-
The only study evaluating objective daytime sleepiness
demonstrated an improvement in MSLT after chronic nocturnal
bilevel PAP in neuromuscular disease.
In an analysis of re-
sidual effects of prolonged ventilatory support on sleep and gas
exchange, 1 study demonstrated that nocturnal gas exchange and
distribution of sleep stages were improved (as compared to base-
line) after 6 and 12 months of chronic ventilatory support even on
a night when bilevel PAP was withheld.
The same author has
shown that patients with restrictive disease have improvements
in respiratory muscle and exercising muscle endurance following
3 months chronic nocturnal ventilatory support
and improve-
ments in mean pulmonary artery pressure after 1 year of chronic
nocturnal ventilatory support,
though these studies were not
included in the evidence grading because a mixture of volume
cycle and bilevel PAP was employed. One Level III study demon-
strated a substantial improvement in 1 year mortality in patients
with amyotrophic lateral sclerosis with a vital capacity of 40 ±
21% of predicted who were treated with undisclosed levels of
bilevel PAP as compared to patients treated with oxygen and “pal-
liative” measures.
7.3.4 Summary
There is no evidence that bilevel PAP improves efficacy or ad-
herence in the management of OSA in first time users of PAP but
the evidence thus far at least supports equivalency for efficacy
and adherence. There is limited evidence that patients with co-
existing lung disease or hypercapnia prefer and show some gas
exchange benefit with bilevel PAP as compared to CPAP.
Bilevel PAP improves gas exchange and sleep in patients with
restrictive lung disease based on studies with Level III evidence.
Current evidence regarding efficacy of PAP employed during
sleep in COPD patients is limited to a small number of conflict-
ing studies with Level II and III evidence employing different
outcome measures. The practice of employing arbitrary levels of
PAP for treatment of hypercapnic COPD patients, particularly at
relatively low levels of inspiratory pressure, may not improve
sleep quality or gas exchange.
7.3.5 Future Research
Additional larger RCTs with cross-over design should be per-
formed to substantiate the 2 negative outcome studies comparing
efficacy and patient adherence with bilevel PAP vs. CPAP in pa-
tients with OSA who do not have coexisting respiratory disease.
The use of bilevel PAP in patients with respiratory disease needs
to be better defined. Randomized sham-controlled trials of bilevel
PAP in patients who have OSAHS in the setting of coexisting
respiratory disease should be performed to determine whether
the positive outcomes of the existing Level III studies could be
substantiated. Patient stratification or selection criteria could be
designed to help develop guidelines for specific patient groups
such as hypercapnic and non-hypercapnic patients or for specific
diagnoses such as COPD, neuromuscular disease, or restrictive
rib cage disease.
In chronic respiratory failure complicating COPD, the pres-
sure prescription in randomized controlled trials of bilevel PAP
during sleep needs to parallel more closely the designs of bilevel
PAP in acute respiratory failure with higher inspiratory pressures
or adjustment of pressures to a desired effect. In addition, adher-
Review of CPAP and Bilevel PAP
—Gay et al
Page 16
SLEEP, Vol. 29, No. 3, 2006
ence and immediate and delayed outcomes should be examined.
Randomized sham-controlled trials of bilevel PAP in patients who
have restrictive lung disease could help to determine whether the
existing Level III studies with positive outcomes can be substanti-
1. Littner M, Hirshkowitz M, Davila D, et al. Practice parameters for
the use of auto-titrating continuous positive airway pressure devices
for titrating pressures and treating adult patients with obstructive
sleep apnea syndrome. An American Academy of Sleep Medicine
report. Sleep. 2002;25:143-7.
2. Sackett DL. Rules of evidence and clinical recommendations for the
management of patients. Can J Cardiol. 1993;9:487-9.
3. Ballester E, Badia JR, Hernandez L, et al. Evidence of the effec
tiveness of continuous positive airway pressure in the treatment
of sleep apnea/hypopnea syndrome. Am J Respir Crit Care Med.
4. Barbe F, Mayoralas LR, Duran J, et al. Treatment with continuous
positive airway pressure is not effective in patients with sleep apnea
but no daytime sleepiness. a randomized, controlled trial. Ann Intern
Med. 2001;134:1015-23.
5. Bardwell WA, Ancoli-Israel S, Berry CC, Dimsdale JE. Neuropsy
chological effects of one-week continuous positive airway pressure
treatment in patients with obstructive sleep apnea: a placebo-con
trolled study. Psychosom Med. 2001;63:579-84.
6. Barnes M, Houston D, Worsnop CJ, et al. A randomized controlled
trial of continuous positive airway pressure in mild obstructive sleep
apnea. Am J Respir Crit Care Med. 2002;165:773-80.
7. Becker HF, Jerrentrup A, Ploch T, et al. Effect of nasal continuous
positive airway pressure treatment on blood pressure in patients
with obstructive sleep apnea. Circulation. 2003;107:68-73.
8. Dimsdale JE, Loredo JS, Profant J. Effect of continuous positive
airway pressure on blood pressure : a placebo trial. Hypertension.
9. Engleman HM, Gough K, Martin SE, Kingshott RN, Padfield PL,
Douglas NJ. Ambulatory blood pressure on and off continuous
positive airway pressure therapy for the sleep apnea/hypopnea syn
drome: effects in “non-dippers”. Sleep. 1996;19:378-81.
10. Engleman HM, Kingshott RN, Wraith PK, Mackay TW, Deary IJ,
Douglas NJ. Randomized placebo-controlled crossover trial of con
tinuous positive airway pressure for mild sleep Apnea/Hypopnea
syndrome. Am J Respir Crit Care Med. 1999;159:461-7.
11. Engleman HM, Martin SE, Deary IJ, Douglas NJ. Effect of continu
ous positive airway pressure treatment on daytime function in sleep
apnoea/hypopnoea syndrome. Lancet. 1994;343:572-5.
12. Engleman HM, Martin SE, Deary IJ, Douglas NJ. Effect of continu-
ous positive airway pressure treatment on daytime function in sleep
apnoea/hypopnoea syndrome. Thorax. 1997;52:114-9.
13. Engleman HM, Martin SE, Kingshott RN, Mackay TW, Deary IJ,
Douglas NJ. Randomised placebo controlled trial of daytime func-
tion after continuous positive airway pressure (CPAP) therapy for
the sleep apnoea/hypopnoea syndrome. Thorax. 1998;53:341-5.
14. Faccenda JF, Mackay TW, Boon NA, Douglas NJ. Randomized
placebo-controlled trial of continuous positive airway pressure on
blood pressure in the sleep apnea-hypopnea syndrome. Am J Respir
Crit Care Med. 2001;163:344-8.
15. Hack M, Davies RJ, Mullins R, et al. Randomised prospective paral
lel trial of therapeutic versus subtherapeutic nasal continuous posi-
tive airway pressure on simulated steering performance in patients
with obstructive sleep apnoea. Thorax. 2000;55:224-31.
16. Henke KG, Grady JJ, Kuna ST. Effect of nasal continuous posi
tive airway pressure on neuropsychological function in sleep apnea-
hypopnea syndrome. A randomized, placebo-controlled trial. Am J
Respir Crit Care Med. 2001;163:911-7.
17. Jenkinson C, Davies RJ, Mullins R, Stradling JR. Comparison of
therapeutic and subtherapeutic nasal continuous positive airway
pressure for obstructive sleep apnoea: a randomised prospective
parallel trial. Lancet. 1999;353:2100-5.
18. Jokic R, Klimaszewski A, Crossley M, Sridhar G, Fitzpatrick MF.
Positional treatment vs continuous positive airway pressure in pa-
tients with positional obstructive sleep apnea syndrome. Chest.
19. Kajaste S, Brander PE, Telakivi T, Partinen M, Mustajoki P. A cog-
nitive-behavioral weight reduction program in the treatment of ob-
structive sleep apnea syndrome with or without initial nasal CPAP:
a randomized study. Sleep Med. 2004;5:125-31.
20. Lojander J, Maasilta P, Partinen M, Brander PE, Salmi T, Lehtonen
H. Nasal-CPAP, surgery, and conservative management for treat-
ment of obstructive sleep apnea syndrome. A randomized study.
Chest. 1996;110:114-9.
21. Lojander J, Kajaste S, Maasilta P, Partinen M. Cognitive function
and treatment of obstructive sleep apnea syndrome. J Sleep Res.
22. Loredo JS, Ancoli-Israel S, Dimsdale JE. Effect of continuous posi-
tive airway pressure vs placebo continuous positive airway pressure
on sleep quality in obstructive sleep apnea. Chest. 1999;116:1545-
23. McArdle N, Douglas NJ. Effect of continuous positive airway
pressure on sleep architecture in the sleep apnea-hypopnea syn
drome: a randomized controlled trial. Am J Respir Crit Care Med.
24. Monasterio C, Vidal S, Duran J, et al. Effectiveness of continuous
positive airway pressure in mild sleep apnea-hypopnea syndrome.
Am J Respir Crit Care Med. 2001;164:939-43.
25. Montserrat JM, Ferrer M, Hernandez L, et al. Effectiveness of CPAP
treatment in daytime function in sleep apnea syndrome: a random
ized controlled study with an optimized placebo. Am J Respir Crit
Care Med. 2001;164:608-13.
26. Pepperell JC, Ramdassingh-Dow S, Crosthwaite N, et al. Ambu-
latory blood pressure after therapeutic and subtherapeutic nasal
continuous positive airway pressure for obstructive sleep apnoea: a
randomised parallel trial. Lancet. 2002;359:204-10.
27. Profant J, Ancoli-Israel S, Dimsdale JE. A randomized, controlled
trial of 1 week of continuous positive airway pressure treatment on
quality of life. Heart Lung. 2003;32:52-8.
28. Redline S, Adams N, Strauss ME, Roebuck T, Winters M, Rosen
berg C. Improvement of mild sleep-disordered breathing with CPAP
compared with conservative therapy. Am J Respir Crit Care Med.
29. Yu BH, Ancoli-Israel S, Dimsdale JE. Effect of CPAP treatment
on mood states in patients with sleep apnea. J Psychiatr Res.
30. Ziegler MG, Mills PJ, Loredo JS, Ancoli-Israel S, Dimsdale JE. Ef
fect of continuous positive airway pressure and placebo treatment
on sympathetic nervous activity in patients with obstructive sleep
apnea. Chest. 2001;120:887-93.
31. Barnes M, McEvoy RD, Banks S, et al. Efficacy of positive airway
pressure and oral appliance in mild to moderate obstructive sleep
apnea. Am J Respir Crit Care Med. 2004;170:656-64.
32. Engleman HM, Martin SE, Douglas NJ. Compliance with CPAP
therapy in patients with the sleep apnoea/hypopnoea syndrome.
Thorax. 1994;49:263-6.
33. Spilker B. Guide to Clinical Trials Philadelphia: Lippincott Wil
liams and Wilkins; 1991.
34. Engleman HM, McDonald JP, Graham D, et al. Randomized cross-
over trial of two treatments for sleep apnea/hypopnea syndrome:
continuous positive airway pressure and mandibular repositioning
splint. Am J Respir Crit Care Med. 2002;166:855-9.
35. Johns MW. Sensitivity and specificity of the multiple sleep latency
test (MSLT), the maintenance of wakefulness test and the epworth
sleepiness scale: failure of the MSLT as a gold standard. J Sleep
Review of CPAP and Bilevel PAP
—Gay et al
Page 17
SLEEP, Vol. 29, No. 3, 2006
Res. 2000;9:5-11.
36. Jokic R, Klimaszewski A, Sridhar G, Fitzpatrick MF. Continuous
positive airway pressure requirement during the first month of
treatment in patients with severe obstructive sleep apnea. Chest.
37. Redline S, Adams N, Strauss ME, Roebuck T, Winters M, Rosen
berg C. Improvement of mild sleep-disordered breathing with CPAP
compared with conservative therapy.[see comment]. American Jour
nal of Respiratory & Critical Care Medicine. 1998;157:858-65.
38. Sleep-related breathing disorders in adults: recommendations for
syndrome definition and measurement techniques in clinical re-
search. The Report of an American Academy of Sleep Medicine
Task Force. Sleep. 1999;22:667-89.
39. Exar EN, Collop NA. The upper airway resistance syndrome. Chest.