Content uploaded by Nikolaus C Netzer
Author content
All content in this area was uploaded by Nikolaus C Netzer on Jan 02, 2014
Content may be subject to copyright.
DOI: 10.1378/chest.124.4.1406
2003;124;1406-1414 Chest
Rudolfo Alvarez-Sala and Kingman P. Strohl
Nikolaus C. Netzer, Josef J. Hoegel, Daniel Loube, Cordula M. Netzer, Birgit Hay,
Prevalence of Symptoms and Risk of Sleep Apnea in Primary Care
This information is current as of November 7, 2005
http://www.chestjournal.org/cgi/content/full/124/4/1406
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
ISSN: 0012-3692.
may be reproduced or distributed without the prior written permission of the copyright holder.
3300 Dundee Road, Northbrook IL 60062. All rights reserved. No part of this article or PDF
published monthly since 1935. Copyright 2005 by the American College of Chest Physicians,
CHEST is the official journal of the American College of Chest Physicians. It has been
by on November 7, 2005 www.chestjournal.orgDownloaded from
Prevalence of Symptoms and Risk of
Sleep Apnea in Primary Care*
Nikolaus C. Netzer, MD; Josef J. Hoegel, PhD; Daniel Loube, MD, FCCP;
Cordula M. Netzer, MD; Birgit Hay, BS; Rudolfo Alvarez-Sala, MD, PhD; and
Kingman P. Strohl, MD, FCCP; for the Sleep in Primary Care International
Study Group†
Background: To obtain prevalence estimates for key symptoms and features that can indicate the
presence of obstructive sleep apnea (OSA) in a broad range of primary care settings.
Design: Cross-sectional survey.
Setting: Forty offices and clinics in the United States, Germany, and Spain.
Participants: Consecutive patients who were > 15 years of age, regardless of the reason for the
visit.
Measurements: We collected demographic information, prevalence of self-reported chronic
snoring, sleepiness, obesity (body mass index [BMI] > 30), hypertension, and calculation of OSA
risk, and we also compared results between the United States and Europe.
Results: There was a 78% return rate for 8,000 surveys (mean age, 51 years; age range, 15 to 98
years; 52% women). One third of participants (32%) had a high pretest probability for OSA, with
a higher rate in the United States (35.8% of 3,915 participants) than in Europe (26.3% of 2,308
participants; p < 0.001; age-matched and sex-adjusted odds ratio [OR], 1.37; 95% confidence
interval [CI], 1.16 to 1.61). Sleepiness (32.4% vs 11.8%, respectively; p < 0.001) followed by
obesity and/or hypertension (44.8% vs 37.1%, respectively; p < 0.01) contributed to the OSA risk
difference between participants in the United States and Europe, as frequent snoring and
breathing pauses were similarly reported (44%). A high pretest probability for OSA was more
often present in men than in women (37.9% vs 27.8%, respectively; p < 0.005; OR, 1.96; CI, 1.59
to 2.88) and in those that were obese (ie, BMI, > 30 kg/m
2
), a condition that is generally more
common in the US population than in the European population (27.9% vs 17.2%, respectively;
p < 0.01).
Conclusions: Primary care physicians in the United States and Europe will encounter a high
demand for services to confirm or manage sleep apnea, sleepiness, and obesity.
(CHEST 2003; 124:1406–1414)
Key words: hypertension; obesity; questionnaire; sleep apnea; snoring
Abbreviations: BMI ⫽ body mass index; CI ⫽ confidence interval; OR ⫽ odds ratio; OSA ⫽ obstructive sleep apnea
S
leep-disordered breathing, in particular obstruc-
tive sleep apnea-hypopnea syndrome, is indepen-
dently associated with car crashes involving drivers
who fall asleep,
1
hypertension,
2
a 4-year risk of
developing hypertension,
3
myocardial infarction,
4
and cardiovascular events of all causes.
5
Variations in
the rates for obstructive sleep apnea-hypopnea syn-
drome depend on age, gender, and obesity.
6–8
The
community prevalence of symptoms and/or signs of
obstructive sleep apnea (OSA) vary by region
9
and by
country.
10,11
For instance, in one state in the United
States (ie, Wisconsin), an elevated apnea index along
with symptoms were present in 2 to 4% of a middle-
aged working population,
12
while in a similar popu-
lation in Spain it was present in 0.8 to 2.2%.
13
The recognition of and the demand for resources
to manage sleep apnea begins with a patient report
or physician questions about key symptoms and
signs, including persistent snoring, sleepiness, and
the presence of obesity and hypertension.
10,14
Most
often this occurs in primary care offices. There is
some evidence that the prevalence of OSA in pri-
mary care offices is higher than in the community.
15
For instance, among five Cleveland, OH, adult pri-
mary care offices, the prevalence rate for snoring,
sleepiness, and a high pretest probability for OSA
was approximately 30%.
16
This difference between
community-based and clinic prevalence occurs be-
cause primary care practices are enriched for obe-
sity, hypertension, and complaints like fatigue. Yet
many primary care patients with signs and symp-
toms, or findings of sleep apnea in community
1406 Clinical Investigations
by on November 7, 2005 www.chestjournal.orgDownloaded from
surveys are undiagnosed.
9,16,17
All of these clinical
prevalence reports were based on isolated US pop-
ulations, and used different instruments and meth-
ods. A uniform collection of key symptoms in a wider
spectrum of primary care practices would permit a
more general idea of the needs for diagnostic plan-
ning or physician education in the management of
sleep apnea, and of sleep disorders in general.
The purpose of this study was to perform a
standardized survey of primary care outpatients from
a diverse socioeconomic range of practice settings to
elicit the frequency of symptoms and risk factors for
sleep apnea, as well as a composite score for a high
pretest probability for finding OSA. The design for
the collection of data in both the United States and
Europe also permitted analyses that might indicate
geographic or cultural factors that affect the presen-
tation and prevalence of such risk factors in primary
care medicine.
Materials and Methods
The survey was conducted over a 2-year period (from 1997 to
1999) in 40 offices and clinics. Local Sleep in Primary Care Study
Group members with expertise in pulmonary and sleep medicine
identified one to two offices or clinics where a physician had
practiced adult general medicine for ⱖ 4 years (range, 4 to 12
years) and had handled 2,000 to 4,000 patient visits per year.
These local experts explained the aim and procedures to the
physicians and their staffs. Questionnaires were distributed in
batches of 200 per study site. Twenty-six US sites (Midwest, 7
sites; eastern seaboard, 19 sites) and 14 European sites (Germany,
8 sites; Spain, 6 sites) participated in the study. There was a wide
range of geographic, social, and ethnic profiles (data available on
request).
The intention was that office staff would hand out copies of the
questionnaire to consecutive patients ⱖ 15 years of age who
visited the physician for any reason. The patient was asked to
complete the questionnaire in the office. Each site kept the
original of the questionnaire and returned a copy of the responses
(without patient identifiers) to the local study group specialist,
who then sent the form for data entry (to Cleveland, OH). The
practitioner could contact the local specialist to address questions
or concerns about specific patients but was not obligated to do so.
To be considered for analysis, questionnaires had to be dated
within 3 weeks of distribution, and originals had to be returned to
the local study group member within 1 month.
The instrument, called the Berlin Questionnaire, was devel-
oped in 1996, and its origin and use in primary care has been
reported previously.
16,18
It is a self-report instrument that is
focused on a set of known symptoms and clinical features
associated with sleep apnea. One introductory question and four
follow-up questions concerned snoring, witnessed apneas, and
the frequency of such events. Three questions addressed daytime
sleepiness, with a sub-question about drowsy driving. One ques-
tion asked for a history of high BP. Patients were to provide
information on age, weight, height, and sex. Body mass index
(BMI) was calculated from the self-reported patient information
on weight and height. Bilingual physicians translated the Berlin
Questionnaire from its original English version into German and
Spanish. Translations were performed from the other languages
back into English by other bilingual physicians and were consis-
tent with the intent of the original version.
Prior to its use in this study, the questionnaire was piloted to 20
bilingual patients in Germany and in Spain, who, after filling out
the native language version, were also given the English version.
Symptom attribution and risk grouping were similar. In addition,
the reliability of self-reporting was tested in 142 subjects for age,
height, weight, the presence or absence of hypertension, and the
calculation of BMI for risk grouping. These self-reported data
were compared to those from medical chart reports. There was a
99% concurrence in age within 1 year. In 99% of surveys, there
was confirmation of the self-reported data on hypertension and a
concurrence with the office chart (within 5%) for data on height
(94%) and weight (93%). A 4% error also occurred in assignment
to the BMI ⬎ 30-group, and a 1% error occurred in the
assignment for risk grouping in the category for obesity/hyper-
tension (ie, category 3, see below).
Risk grouping for high risk and lower risk for OSA were based
on responses grouped into three categories. In category 1, a
positive score for risk was defined as frequent symptoms (ie,
“more than three to four times per week” or “almost every day”)
in the questions about snoring and witnessed apneas. In category
2, a positive score for risk was frequent symptoms in two or more
questions about awakening sleepy, waketime sleepiness, and/or
drowsy driving. In category 3, a positive score for risk was defined
as a self-report of high BP and/or of height/weight information
giving a BMI of ⬎ 30 kg/m
2
. To score “high” for OSA, an
individual’s questionnaire must have had positive scores in two of
the three categories, or in all three. Those patients who denied
having symptoms with such frequency, who did not report
symptoms to permit risk assessment (see “Results” section), or
who qualified in only one category were placed into a lower risk
group.
16,18
The relationship of risk grouping for a high pretest probability
was previously shown to have a positive predictive value of 89%
and a likelihood ratio of 3.79 for a subsequent finding of a
respiratory disturbance index ⬎ 5 on a sleep study.
16
The RDI of
⬎ 5 along with symptoms is the current threshold value for
initiating therapy for sleep apnea.
18,19
In a study of 350 patients
in Germany (RA-S; unpublished results) that was performed
prior to this study, results were similar to those previously
reported with regard to risk grouping and pretest probability for
an apnea-hypopnea index of ⬎ 5 using the German version of the
*From the The Center for Sleep Disorders Research (Drs. NC
Netzer, CM Netzer, and Strohl), Case Western Reserve Univer-
sity, Cleveland, OH; Department of Biometry and Medical
Documentation (Dr. Hoegel and Ms. Hay), University of Ulm,
Ulm, Germany; Swedish Medical Center (Dr. Loube), Seattle,
WA; and the Pulmonary Department (Dr. Alvarez-Sala), Hospital
de la Paz, Autonomous University of Madrid, Madrid, Spain.
†A list of other members of the Sleep in Primary Care Interna-
tional Study Group is located in the Appendix.
This research was supported in part by the Veterans Affairs
Medical Service and by a Sleep Academic Award (HL97015).
National Sleep Technologies Inc, Arnold, MD, provided logistic
support in the Washington DC area, and 3M Inc, Minneapolis,
MN, and 3M Medica GmbH, Neuss, Germany, provided an
unrestricted grant to print the various forms of the Berlin
questionnaire. The Berlin Questionnaire is held as a US copyright
by iONSLEEP LLC (Shaker Heights, OH) and may be used for
academic and research purposes without fee or license.
Received January 14, 2003; revision accepted May 14, 2003.
Reproduction of this article is prohibited without written permis-
sion from the American College of Chest Physicians (e-mail:
permissions@chestnet.org).
Correspondence to: Kingman P. Strohl, MD, FCCP, 111j(w),
Louis Stokes Cleveland VA Medical Center, Case Western Re-
serve University, 10701 East Blvd, Cleveland, OH 44106; e-mail:
KPSTROHL@aol.com
www.chestjournal.org CHEST / 124/4/OCTOBER, 2003 1407
by on November 7, 2005 www.chestjournal.orgDownloaded from
questionnaire. No testing against sleep studies was performed for
the Spanish questionnaire prior to this study, but subsequent data
suggest a positive predictive value of ⬎ 90% for an apnea-
hypopnea index of ⬎ 5 in that population as well (RA-S; unpub-
lished results).
Statistical Analysis
Data were entered into analysis files (SPSS, version 8.0 for
Windows; SPSS, Inc; Chicago, IL). Files were transferred via
secured Internet connection for analysis at the University of Ulm.
Statistical evaluations were carried out using a statistical software
package (SAS; SAS Institute; Cary, NC).
Frequency distributions and proportions were used to describe
categoric variables and the mean ⫾ SD of quantitative variables
among sites. This provided the primary data for the determina-
tion of prevalence rates for self-reported symptoms and risk
factors, as well as for risk scores. Another outcome was the
possible difference in the answers and risk scores between
Europe and the United States. Another related intent was to
analyze differences, if any, between genders in regard to risk
scores. For this purpose, logistic regression analysis was applied.
After proper dichotomization, the answers and scores formed the
outcome variables in separate logistic models. To facilitate inter-
pretation, some independent variables were transformed, as
follows: continent (Europe/United States); gender (male/female);
BMI (⬍ 30 kg/m
2
vs ⱖ 30 kg/m
2
); hypertension (yes/no); region
(north/south); and age. The age grouping was performed using
the following cutoffs: young persons (⬍ 35 years); middle-aged
persons (35 to 55 years); and older persons (⬎ 55 years). The age
groups were chosen arbitrarily to provide roughly equal groups
and yet capture climacteric events in women. When assessing the
adjusted influence of continent and gender on the answers and
scores, additional possible effect modification by the main covari-
ates (eg, age, gender, obesity, and hypertension) was accounted
for by entering corresponding interaction terms into the logistic
model, including interaction between continent and, when exam-
ined, gender itself.
Rate differences between continents and between genders
were calculated by odds ratios (ORs) and the corresponding 95%
confidence intervals (CIs). In some instances, analysis was re-
stricted to subgroups identified by significant interactions be-
tween the variables. Statistical significance meant that the p value
of the corresponding statistical test was ⬍ 0.05.
Results
Of the 8,000 distributed questionnaires, 6,223
forms (78%) were entered for analysis. The response
rate varied among sites, as follows: 3 of 26 US sites,
40 to 50%; 2 of 14 European sites, 48% and 54%; all
remaining sites, 70 to 98% (data available on re-
quest). The elimination of sites with a response rate
of ⬍ 50% did not significantly alter the results and so
were left in the final analysis. The major reason for a
low response rate was a failure to distribute/return
copies. The return rate did not significantly differ
with geographic or socioeconomic profiles of the
sites (p ⬎ 0.05). US and European populations re-
ported similar ages, gender distributions, and in-
stances of high BP, but self-reported obesity (ie,
BMI ⱖ 30 kg/m
2
) was significantly more common in
the US population than in the European population
(27.9% vs 17.2%, respectively; p ⬍ 0.01). Those in-
dividuals with inadequate data were assigned to a
negative-risk or low-risk category. This effect was
small, as only 1.1% of individuals did not provide
enough answers to the questions about category 1,
2% did not provide answers to questions about
sleepiness (category 2), and 3.2% gave no response
to the questions about history of high BP or provided
enough information about height and weight to
estimate BMI. This exclusion therefore resulted in
some minor underestimation of the prevalence of
categories and OSA risk.
Responses and Categories
The rates for categories 1 to 3 and OSA risk varied
among the sites (Table 1). Overall, category 2 (sleep-
iness), in particular, differed between continents.
The US population reported generally higher rates
than their European counterparts in frequency of
“not rested after sleep” (36% vs 16%, respectively;
p ⬍ 0.001) and in “waketime tiredness” (39% vs
14%, respectively; p ⬍ 0.001). US patients were
more likely than European patients to report drowsy
driving (17% vs 7%, respectively; p ⬍ 0.001), but the
differences diminished with increasing age, as fol-
lows: young patients: OR, 2.94; 95% CI, 2.01 to 4.29;
middle-aged patients: OR, 2.17; 95% CI, 1.65 to
2.86; and older patients: OR, 1.62; 95% CI, 1.12 to
2.34. In Europe, those in Spain reported less sleep-
iness than those in Germany (OR, 0.43; 95% CI, 0.31
to 0.59). Equally dividing the US population along a
north-south axis resulted in the finding that those in
the southern United States also reported somewhat
less sleepiness than those in the north (OR, 0.77;
95% CI, 0.58 to 1.03).
Comparing category scores between continents
(Fig 1), only category 1 (snoring) was different when
examined with regard to age groups. The rate for US
respondents was higher in the younger age grouping,
but the rate for European respondents was higher in
the older age grouping. As a history of high BP is part
of category 3, it followed that those reporting high
BP were much more likely to be in the high pretest
probability for OSA than those who did not report it
(OR, 5.60; 95% CI, 4.68 to 6.71). However, when the
data set was adjusted for age and obesity, a report of
high BP was positively correlated with symptom
category 1 (snoring: OR, 1.30; 95% CI, 1.11 to 1.52)
and with a minor effect on symptom category 2
(sleepiness: OR, 1.17; 95% CI, 0.97 to 1.43).
Gender Effects
Male and female respondents did not differ with
respect to age (men, 51.3 ⫾ 17.0 years; women,
1408 Clinical Investigations
by on November 7, 2005 www.chestjournal.orgDownloaded from
50.2 ⫾ 17.7 years) or reporting a history of high BP
(men, 29.6%; women, 26.4%). The mean BMI in
men was 28.0 ⫾ 5.4 kg/m
2
, and in women,
27.1 ⫾ 6.4 kg/m
2
. There were no gender differences
with regard to BMI ⬎ 30 (men, 24.8%; women,
24.0%).
Differences in symptom frequencies and OSA risk
between men and women are presented in Tables 2
and 3. Men were more likely to report frequent
snoring and breathing pauses (OR, 1.92; 95% CI,
1.61 to 2.29). Men more often reported drowsy
driving (OR, 2.08; 95% CI, 1.61 to 2.69), but women
were more likely to complain about fatigue in the
morning and during the daytime. Women were more
likely than men to qualify as being positive for
symptom category 2 (OR, 1.83; 95% CI, 1.48 to
2.25). For those with a BMI of ⬎ 30, US men had a
higher likelihood for OSA than women (OR, 1.66;
95% CI, 1.22 to 2.27). However, not considering age,
for those with a BMI of ⬍ 30 kg/m
2
, the OR was 1.13
(95% CI, 0.90 to 1.43).
High OSA Risk
In the United States, a high score was present
more often than in Europe, an observation that was
true for both men and women (Fig 2). Young US
patients, both men and women, had the highest
likelihood of qualifying as being positive compared
with their European counterparts. In those patients
⬎ 55 years of age, there was no continental differ-
ence (Table 3). A higher probability of finding OSA
in the US population was present if the analysis was
restricted to being positive in both category 1 (snor-
ing) and category 2 (sleepiness) [United States,
16.4%; Europe, 6.7%; p ⬍ 0.001), without regard to
category 3.
Discussion
This is the first large data set providing informa-
tion collected by a standardized protocol on snoring,
sleepiness, and other features associated with sleep
Table 1—Risk Factors and Functional Sleepiness Among Study Regions*
Geographic Region
Patients at High
Risk for OSA
Patients Qualifying
for Symptom
Category 1
Patients Qualifying
for Symptom
Category 2
Patients Qualifying
for Symptom
Category 3
Patients Who
Report Drowsy
Driving
% No. % No. % No. % No. % No.
United States (n ⫽ 3,915) 35.8 1,403† 43.3 1,695‡ 32.4 1,267§ 44.8 1,755㛳 17.7 693¶
Ashland, OH (n ⫽ 99) 42.4 42 41.4 41 32.3 32 48.5 48 17.2 17
Cleveland, OH (n ⫽ 736) 37.9 279 42.0 309 31.5 232 50.3 370 18.6 137
Louisville, KY (n ⫽ 81) 66.7 54 63.0 51 53.1 43 64.2 52 24.7 20
Milton-Quincy, MA (n ⫽ 148) 55.4 82 54.1 80 42.6 63 57.4 85 18.2 27
Naples, FL (n ⫽ 143) 36.4 52 42.7 61 35.7 51 39.9 57 21.7 31
Springfield, MA (n ⫽ 216) 19.9 43 46.3 100 33.3 72 34.7 75 12.5 27
Stuart, FL (n ⫽ 375) 43.5 163 45.1 169 21.1 79 68.8 258 4.0 15
Washington, DC (n ⫽ 2,037) 32.4 661 41.8 851 33.2 676 38.0 773 20.0 407
Wilmette, IL (n ⫽ 80) 33.8 27 41.3 33 23.8 19 46.3 37 15.0 12
Europe (n ⫽ 2,308) 26.3 607 43.5 1,005 11.8 272 37.1 855 7.1 163
Germany
Leipzig (n ⫽ 393) 23.2 91 35.6 140 7.9 31 44.0 173 4.8 19#
Mindelheim (n ⫽ 141) 30.4 43 45.4 64 22.0 31 31.9 45 19.9 28
Wuerzburg/Darmstadt
(n ⫽ 625)
27.2 170 48.8 305 17.4 109 35.0 219 8.0 50
Ulm (n ⫽ 96) 44.8 43 55.2 53 30.2 29 46.9 45 8.3 8
Spain
Madrid (n ⫽ 1,053) 24.7 260 42.1 443 6.8 72 35.4 373 5.5 58**
Total (n ⫽ 6,223) 32.3 2,010 43.4 2,700 24.7 1,539 41.9 2,610 13.8 856††
*Patients who could not be categorized because of missing data were considered not to qualify for a particular symptom category.
†Denotes significantly different from European population (p ⬍ 0.001); however, there are significant interactions with gender (p ⬍ 0.05), obesity
(p ⬍ 0.001), and report of hypertension (p ⬍ 0.001).
‡Denotes significantly different from European population (p ⬍ 0.05); however, there are significant interactions with age (p ⬍ 0.001).
§Denotes significantly different from European population (p ⬍ 0.001); however, there are interactions with region (north/south; p ⬍ 0.01).
㛳Denotes significantly different from European population (p ⬍ 0.01).
¶Denotes significantly different from European population (p ⬍ 0.001); however, there are interactions by region (p ⬍ 0.005) and history of
hypertension (p ⬍ 0.005).
#A total of 23.6% did not answer that question.
**A total of 10.1% did not answer the question.
††A total of 5.5% did not answer the question.
www.chestjournal.org CHEST / 124/4/OCTOBER, 2003 1409
by on November 7, 2005 www.chestjournal.orgDownloaded from
apnea across many primary care sites. The results
confirmed a widespread distribution of patients with
a surprisingly high rate of reported frequent sleepi-
ness and drowsy driving, behaviors that pose individ-
ual and societal risks. Approximately one third of
patients reported symptoms and risk factors with a
high likelihood for finding sleep apnea if referred for
testing. The major feature of the difference in a
high-risk profile for OSA between Europe and the
United States was sleepiness, the second feature was
obesity, while the self-report of frequent snoring was
common (approximately 4 in 10 patients) and was
similarly present at US and European sites. In this
population, the rate of high risk for OSA was high in
both men (approximately 37%) and women (approx-
imately 27%).
A self-report of snoring was generally as common
in the United States as in Europe, and by itself this
is emerging as a health issue.
11
The occurrence of
snoring is a predictor of a subsequent diagnosis of
hypertension
3
and of diabetes 10 years later,
20,21
even after adjustment for previously well-described
risk factors such as obesity. In our cross-sectional
study, aging qualitatively and quantitatively altered
differences in self-reports of snoring between conti-
nents, independent of daytime tiredness or obesity,
but we collected no information that permits us to
gain insight into the greater European rate when
patients were older (ie, ⬎ 55 years of age). Possibil-
ities include an age-related decline in informative
bed partners, European “survivors” who have the
opportunity to snore, or differences in the reporting
of this herald of sleep.
Men had a higher composite score for OSA, and
this score resulted from reports of snoring and
observed apnea, rather than of sleepiness. In regard
to sleepiness, women reported twice as much sleep-
iness as men, but men reported more experience
with drowsy driving. Gender differences previously
have been attributed to disease expression or report-
ing bias.
22,23
Of interest, between genders there was
no difference in the rate of OSA risk (approximately
one in four), when analyses were restricted to indi-
viduals with a BMI of ⬍ 30. Hence, gender and
obesity may not be definitive features on which a
primary care practitioner should base recognition
profiles, especially if one considers the real possibil-
ity of underreporting by women.
22,23
Excessive daytime sleepiness is an important fea-
ture in the diagnosis of obstructive sleep apnea-
hypopnea syndrome, but OSA is not the only cause
for this symptom. A report
24
from a large Japanese
population, but without reference to latitude or sleep
disorders, identified factors of unhealthy lifestyle,
poor general health, and urban living as predictors of
daytime sleepiness. In our study, sleepiness was not
just an urban phenomenon, as it increased with age.
Among US and European sites, there were differ-
ences in sleepiness and drowsy driving. North-south
regions differed in the frequency of reported sleep-
Figure 1. OR comparisons between US and European patients according to category. Age showed a
significant effect only on reports in category 1 (ie, snoring and sleep behavior). An OR ⬎ 1 indicates
that the US population is greater, and an OR ⬍ 1 indicates that the European population is greater.
Bars represent CIs.
1410 Clinical Investigations
by on November 7, 2005 www.chestjournal.orgDownloaded from
Table 2—Distribution of Responses According to Gender and Continent
Questions
Men Women
United States Europe United States Europe
No. % No. % No. % No. %
Category 1
Do you snore? 1,690 1,042 2,011 1,218
Yes 1,144 67.7 684 65.6 921 45.8 553 45.4
No 387 22.9 282 27.1 769 38.2 493 40.5
Do not know 159 9.4 76 7.3 321 16.0 172 14.1
Snoring loudness 1,148 689 933 570
Loud as breathing 408 35.5 203 29.5 506 54.2 253 44.4
Loud as talking 363 31.6 251 36.4 243 26.1 212 37.2
Louder than talking 158 13.8 136 19.7 96 10.3 68 11.9
Very loud 219 19.1 99 14.4 88 9.4 37 6.5
Snoring frequency 1,207 720 1,022 616
Almost every day 474 39.3 295 41.0 285 27.9 177 28.7
3–4 times/wk 215 17.8 137 19.0 135 13.2 82 13.3
1–2 times/wk 266 22.0 167 23.2 226 22.1 184 29.9
1–2 times/mo 127 10.5 72 10.0 139 13.6 91 14.8
Never or almost never 125 10.4 49 6.8 237 23.2 82 13.3
Does your snoring bother other people? 1,321 765 1,221 712
Yes 905 68.5 514 67.2 558 45.7 329 46.2
No 416 31.5 251 32.8 663 54.3 383 53.8
How often have your breathing pauses been
noticed?
1,473 865 1,560 913
Almost every day 75 5.1 32 3.7 32 2.1 7 0.8
3–4 times/wk 42 2.9 36 4.2 11 0.7 5 0.5
1–2 times/wk 42 2.9 53 6.1 23 1.5 17 1.9
1–2 times/mo 72 4.9 50 5.8 30 1.9 19 2.1
Never or almost never 1,242 84.3 694 80.2 1,464 93.8 865 94.7
Category 2
Are you tired after sleeping? 1,665 1,034 1,993 1,215
Almost every day 329 19.8 94 9.1 571 28.7 145 11.9
3–4 times/wk 183 11.0 49 4.7 264 13.3 70 5.8
1–2 times/wk 289 17.4 109 10.6 381 19.1 128 10.5
1–2 times/mo 279 16.7 117 11.3 294 14.7 139 11.5
Never or almost never 585 35.1 665 64.3 483 24.2 733 60.3
Are you tired during waketime? 1,676 1,029 1,998 1,209
Almost every day 369 22.0 85 8.3 572 28.6 132 10.9
3–4 times/wk 207 12.3 50 4.9 315 15.8 78 6.5
1–2 times/wk 358 21.4 124 12.0 426 21.3 143 11.8
1–2 times/mo 325 19.4 135 13.1 364 18.2 153 12.7
Never or almost never 417 24.9 635 61.7 321 16.1 703 58.1
Have you ever fallen asleep while driving? 1,671 989 1,989 1,060
Yes 383 22.9 119 12.0 279 14.0 43 4.1
No 1,288 77.1 870 88.0 1,710 86.0 1,017 95.9
Asleep driving frequency? 835 324 783 282
Almost every day 11 1.3 2 0.6 15 1.9 2 0.7
3–4 times/wk 12 1.5 1 0.3 16 2.0 5 1.8
1–2 times/wk 37 4.4 7 2.2 28 3.6 2 0.7
1–2 times/mo 102 12.2 29 8.9 76 9.7 10 3.5
Never or almost never 673 80.6 285 88.0 648 82.8 263 93.3
Category 3
Do you have high blood pressure? 1,663 1,028 1,993 1,212
Yes 511 30.7 304 29.6 534 26.8 330 27.2
No 1,035 62.3 676 65.7 1,361 68.3 843 69.6
Do not know 117 7.0 48 4.7 98 4.9 39 3.2
BMI ⬎ 30 kg/m
2
?
1,578 1,022 1,858 1,184
Yes 513 32.5 170 16.6 565 30.4 221 18.7
No 1,065 67.5 852 83.4 1,293 69.6 963 81.3
Has your weight changed? 1,684 1,040 2,009 1,208
Increased 813 48.3 211 20.3 1,157 57.6 271 22.4
Decreased 291 17.3 102 9.8 353 17.6 149 12.4
No change 580 34.4 727 69.9 499 24.8 788 65.2
www.chestjournal.org CHEST / 124/4/OCTOBER, 2003 1411
by on November 7, 2005 www.chestjournal.orgDownloaded from
iness (category 2), suggesting differences in health,
perception, attitude, or sleep behavior in northern
and southern latitudes. The frequency of reports of
drowsy driving in the present study (range across all
sites, 4 to 22%) are cause for concern because such
active sleepiness is not only a major proximate cause
of car crashes,
1
but also is a marker for personal
accidents and workplace errors.
25
The study design permitted intercontinental com-
parisons. Compared to European patients, US pa-
tients reported an overall higher frequency of day-
time sleepiness, as well as a higher composite score
for OSA. Different among genders and independent
of continent were sleepiness (women more than
men) and score for OSA (men more than women).
US women reported symptoms for OSA more than
their European counterparts, while rates for US and
European men were more similar. Difference be-
tween genders and a higher European rate of fre-
quent snoring/breathing pauses diminished with age.
Reports of drowsy driving were not uncommon, but
these also diminished with age and were more
common in US patients. We found obesity and
recent weight gain to be more prevalent in the
United States than in Europe, which is consistent
with the current literature.
26,27
However, obesity was
second place to sleepiness as a factor in the generally
high rate for OSA risk.
The Berlin Questionnaire does not capture all
information that a physician might want or seek, nor
Figure 2. OR comparisons between US and European patients with regard to an OSA score for high
pretest probability. An OR ⬎ 1 indicates that the US population is greater; and an OR ⬍ 1 indicates
that the European population is greater. Bars represent CIs.
Table 3—Risk Factors and Functional Sleepiness Between Genders*
Patients
Patients at High
Risk for OSA
Patients Qualifying for
Symptom Category 1
Patients Qualifying for
Symptom Category 2
Patients Qualifying for
Symptom Category 3
Patients Who Report
Drowsy Driving
% No. % No. % No. % No. % No.
Men (n ⫽ 2,750) 37.9 1,043† 57.5 1,580‡ 20.7 569§ 44.8 1,232㛳 18.3 502¶
Women (n ⫽ 3,272) 27.8 911 32.0 1,046 28.0 915 40.0 1,308 9.8 322
Total (n ⫽ 6,223)# 32.3 2,010 43.4 2,700 24.7 1,539 41.9 2,610 13.8 856
*Patients who could not be categorized because of missing data were considered not to qualify for a particular symptom category.
†Difference from the opposite gender depends on obesity (p ⬍ 0.01), continent (p ⬍ 0.05), and history of hypertension (p ⬍ 0.05).
‡Difference from the opposite gender (p ⬍ 0.001).
§Difference from the opposite gender (p ⬍ 0.001).
㛳Difference from the opposite gender depends on age (p ⬍ 0.001).
¶Difference from opposite gender (p ⬍ 0.001); modified (p ⬍ 0.005) by region (p ⬍ 0.005).
#Includes 201 patients who gave no information on their gender.
1412
Clinical Investigations
by on November 7, 2005 www.chestjournal.orgDownloaded from
does it substitute for direct measurements of breath-
ing during sleep.
28,29
The self-report format had
advantages of convenience, simplicity, and cost, and
may be more uniform in presentation compared to
face-to-face screening or a physician-obtained pa-
tient history. We used questions in the three lan-
guages to reduce the variations in meaning further.
The study did not include information on the
utility of the questionnaire in regard to patient
outcome. Also, we did not require specialist referral
or sleep studies, which might have created some
barriers and bias, including that of access and cost,
and a more limited sample. The data justify a more
detailed look at risk stratification and clinical deci-
sions on the diagnosis and treatment of sleep disor-
ders in both the United States and Europe.
There are other potential limitations to consider.
This study was not randomized by the choice of the
practice site or the patients surveyed in each prac-
tice. As a result, there may be bias with regard to
interest or other effects on the reporting of symp-
toms. The diversity of practice plans, styles of prac-
tice, interests of the individual practitioner, patient
utilization practices, and time of the survey, alone or
in combination, might have influenced patient re-
ports or produced variability among sites, but the
similarities across continents and the correspon-
dence to other community surveys of health in terms
of such features as obesity mitigate these concerns to
some extent. We did not independently confirm or
refute reports of snoring or sleepiness. However, the
concordance among patient reports and bed-partner
reports is reported to be sufficiently high
30
to believe
that there might not be overreporting of these
symptoms.
In summary, this study is the first large survey of
primary care practices that has used a standardized
approach and has obtained rather high prevalence
values for key symptoms and features that might
result in a referral for evaluation of OSA. These
international cross-sectional data find a considerable
rate for risk factors such as sleepiness and obesity, for
which one might envision behavioral interventions
that would reduce the composite risk for OSA. The
data appear to justify more attention to issues related
to cost-effective case finding, severity of disease and
prognosis, and the need for and efficacy of the
treatment of sleep disorders in primary care systems
in both the United States and Europe.
ACKNOWLEDGMENTS: We thank the primary care physi-
cians for their participation in the study. We acknowledge Ed
Schuck, Alpha One Foundation, Wayzata, MN, for his continuous
support to the Sleep in Primary Care International Study Group
since 1995. We also thank Mansour Mustafa, MD, and Susan
Redline, MD, MPH, for providing advice and internal review of
the manuscript.
Appendix
The other members of the Sleep in Primary Care International
Study Group are as follows: Jose Alvarez Sala, Madrid, Spain;
Cherryl Carlucci, Stuart, FL; Martin Cohn, Naples, FL; Michael
Coppola, Springfield, MA; Eugene C. Fletcher, Louisville, KY;
James Mooney, Ashland, OH; Rainer Morawa, Kitzingen, Ger-
many; Annette Neumann, Leipzig, Germany; Wolfgang Pirsig,
Ulm, Germany; James O’Brien, Milton, MA; and Peter Werner,
Wilmette, IL.
The authors thank these members for acting as tutors for the
participating primary care practitioners and for helping to make
this survey possible.
References
1 Teran-Santos J, Jimenez-Gomez A, Cordero-Guevara J. The
association between sleep apnea and the risk of traffic
accidents: Cooperative Group Burgos-Santander. N Engl
J Med 1999; 340:847–851
2 Nieto FJ, Young TB, Lind BK, et al. Association of sleep-
disordered breathing, sleep apnea, and hypertension in a
large community-based study: Sleep Heart Health Study.
JAMA 2000; 283:1829 –1836
3 Peppard PE, Young T, Palta M, et al. Prospective study of the
association between sleep-disordered breathing and hyper-
tension. N Engl J Med 2000; 342:1378–1384
4 Marin JM, Carrizo SJ, Kogan I. Obstructive sleep apnea and
acute myocardial infarction: clinical implications of the asso-
ciation. Sleep 1998; 21:809–815
5 Shahar E, Whitney CW, Redline S, et al. Sleep-disordered
breathing and cardiovascular disease: cross-sectional results
of the Sleep Heart Health Study. Am J Respir Crit Care Med
2001; 163:19 –25
6 Redline S, Strohl KP. Recognition and consequences of
obstructive sleep apnea hypopnea syndrome. Clin Chest Med
1998; 19:1–19
7 Foresman BH, Gwirtz PA, McMahon JP. Cardiovascular
disease and obstructive sleep apnea: implications for physi-
cians. J Am Osteopath Assoc 2000; 100:360–369
8 Lindberg E. Snoring and sleep apnea: a study of evolution
and consequences in a male population; minireview based on
a doctoral thesis. Ups J Med Sci 1998; 103:155–202
9 Kapur V, Strohl KP, Redline S, et al. Underdiagnosis of sleep
apnea syndrome in United States communities. Sleep Breath
2002; 6:49 –54
10 Vgontzas AN, Kales A. Sleep and its disorders. Annu Rev
Med 1999; 50:387– 400
11 Young T, Peppard PE, Gottlieb DJ. Epidemiology of obstruc-
tive sleep apnea: a population health perspective. Am J Respir
Crit Care Med 2002; 165:1217–1239
12 Young T, Palta M, Dempsey J, et al. The occurrence of
sleep-disordered breathing among middle-aged adults.
N Engl J Med 1993; 328:1230 –1235
13 Marin JM, Gascon JM, Carrizo S, et al. Prevalence of sleep
apnoea syndrome in the Spanish adult population. Int J
Epidemiol 1997; 26:381–386
14 Smith R, Ronald J, Delaive K, et al. What are obstructive
sleep apnea patients being treated for prior to this diagnosis?
Chest 2002; 121:164 –172
15 Dement WC, Netzer NC. Primary care: is it the setting to
address sleep disorders? Sleep Breath 2000; 4:1–6
16 Netzer NC, Stoohs RA, Netzer CM, et al. Using the Berlin
Questionnaire to identify patients at risk for the sleep apnea
syndrome. Ann Intern Med 1999; 131:485– 491
17 Young T, Evans L, Finn L, et al. Estimation of the clinically
www.chestjournal.org CHEST / 124/4/OCTOBER, 2003 1413
by on November 7, 2005 www.chestjournal.orgDownloaded from
diagnosed proportion of sleep apnea syndrome in middle-
aged men and women. Sleep 1997; 20:705–706
18 Reprinting of the Berlin questionnaire. Sleep Breath 2000;
4:187–192
19 Department of Health and Human Services. Medicare
coverage issues manual, transmittal 151, header section
60-17. Available at: http://cms.hhs.gov/manuals/pm_trans/
R151CIM.pdf. Accessed September 17, 2003
20 Al-Delaimy WK, Manson JE, Willett WC, et al. Snoring as a
risk factor for type II diabetes mellitus: a prospective study.
Am J Epidemiol 2002; 155:387–393
21 Elmasry A, Lindberg E, Berne C, et al. Sleep-disordered
breathing and glucose metabolism in hypertensive men: a
population-based study. J Intern Med 2001; 249:153–161
22 Redline S, Kump K, Tishler PV, et al. Gender differences in
sleep disordered breathing in a community-based sample.
Am J Respir Crit Care Med 1994; 149:722–726
23 Young T, Finn L. Epidemiological insights into the public health
burden of sleep disordered breathing: sex differences in survival
among sleep clinic patients. Thorax 1998; 53:S16 –S19
24 Ohida T, Kamal AM, Uchiyama M, et al. The influence of
lifestyle and health status factors on sleep loss among the
Japanese general population. Sleep 2001; 24:333–338
25 Lyznicki JM, Doege TC, Davis RM, et al. Sleepiness, driving,
and motor vehicle crashes: Council on Scientific Affairs,
American Medical Association. JAMA 1998; 279:1908 –1913
26 Philipson T. The world-wide growth in obesity: an economic
research agenda. Health Econ 2001; 10:1–7
27 Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, et al.
CDC growth charts: United States. Adv Data 2000; 1–27
28 American Academy of Sleep Medicine. Sleep-related breath-
ing disorders in adults: recommendations for symptom defi-
nition and measurement techniques in clinical trials. Sleep
1999; 22:667– 689
29 Systematic review of the literature regarding the diagnosis of
sleep apnea. Evid Rep Technol Assess (Summ) 1999; i-viii:1–
154
30 Kump K, Whalen C, Tishler PV, et al. Assessment of the
validity and utility of a sleep-symptom questionnaire. Am J
Respir Crit Care Med 1994; 150:735–741
1414 Clinical Investigations
by on November 7, 2005 www.chestjournal.orgDownloaded from
DOI: 10.1378/chest.124.4.1406
2003;124;1406-1414 Chest
Rudolfo Alvarez-Sala and Kingman P. Strohl
Nikolaus C. Netzer, Josef J. Hoegel, Daniel Loube, Cordula M. Netzer, Birgit Hay,
Prevalence of Symptoms and Risk of Sleep Apnea in Primary Care
This information is current as of November 7, 2005
& Services
Updated Information
http://www.chestjournal.org/cgi/content/full/124/4/1406
figures, can be found at:
Updated information and services, including high-resolution
References
L
http://www.chestjournal.org/cgi/content/full/124/4/1406#BIB
free at:
This article cites 28 articles, 15 of which you can access for
Citations
articles
http://www.chestjournal.org/cgi/content/full/124/4/1406#other
This article has been cited by 2 HighWire-hosted articles:
Permissions & Licensing
http://www.chestjournal.org/misc/reprints.shtml
tables) or in its entirety can be found online at:
Information about reproducing this article in parts (figures,
Reprints
http://www.chestjournal.org/misc/reprints.shtml
Information about ordering reprints can be found online:
Email alerting service
sign up in the box at the top right corner of the online article.
Receive free email alerts when new articles cite this article
Images in PowerPoint format
article figure for directions.
teaching purposes in PowerPoint slide format. See any online
Figures that appear in CHEST articles can be downloaded for
by on November 7, 2005 www.chestjournal.orgDownloaded from