Obstructive Sleep Apnea


Data Source: 

Dr. Sharon Chung of the Department of Psychiatry at Toronto Western Hospital and her co-investigators Negar Ahmadi and Colin Shapiro


Alison Gibbs, Department of Statistics, University of Toronto


The Berlin Questionnaire for sleep apnea in a sleep clinic population: relationship to polysomnographic measurement of respiratory disturbance.

Obstructive sleep apnea (OSA) is characterized by the cessation of breathing in sleep lasting more than ten seconds and resulting in total or partial arrest in respiration. If left untreated, sleep apnea has been shown to result in an elevated mortality rate. This is primarily due to the causal relationship between sleep apnea and coronary artery disease, congestive heart failure, myocardial infarction, stroke, systemic hypertension and pulmonary hypertension. The most serious socio-economic impact of untreated sleep apnea is the three to seven-fold greater prevalence of motor vehicle accidents in drivers with untreated sleep apnea. Further consequences of undiagnosed and untreated sleep apnea include a reduction in the quality of life, cognitive dysfunction, decreased vigilance, disturbed concentration and memory, increased mental stress, depression, fatigue, anxiety, nocturnal panic attacks, general mood disorders and male sexual dysfunction.

The estimated prevalence of OSA varies with respect to gender and age. In middle-aged adults, clinically significant sleep apnea (Respiratory Disturbance Index (RDI) > 10) affects approximately 2-5% of males and 2% of females. It has been estimated that about 24% of primary care patients have OSA. About one out of 3 patients referred to the sleep specialists at the Toronto Western Hospital are suspected of having sleep apnea.

Overnight polysomnography with recording of respiratory disturbance is the gold standard for the diagnosis of sleep apnea. In the past, it had been preferable to administer two nights of polysomnographic testing. The reason for this is the first night effect wherein the first night of sleep in a laboratory (or any other strange environment) is characterized by a reduction in the quantity and quality of sleep. Further, there has been shown to be a night-to-night variability in respiratory disturbance index, the respiratory variable used to measure the presence and severity of OSA. Based on one night only of laboratory sleep study, it has been reported that between 11-22% of sleep apnea patients remain undiagnosed. Home-based sleep studies, where one would assume that the first night effect would not be a factor, have reported similar instances of missed diagnosis of OSA based on only one night of recordings.

However, recently the healthcare system has mandated that having two nights of diagnostic sleep studies should be discontinued in favour of a single night. Considering that an additional night is required for determining the appropriate adjustments for patients requiring treatment, there is a greater pressure on sleep specialists to make their diagnoses based on only a single night of recording. Given the greater chance of a missed diagnosis based on only one night of recording, the clinical interview will now play a greater role in identifying patients with sleep apnea. As an aid to the clinical interview process, all clinic patients are asked in advance to complete a battery of questionnaires that is reviewed by their sleep specialist at the time of the medical consultation. The questionnaire consists of a number of scales and instruments that screen for a variety of sleep disorders.

One of the instruments in the clinic questionnaire battery is the Berlin Questionnaire (BQ), a scale for sleep apnea. This 10-item scale assesses the risk (high or low) of having sleep apnea based on the responses to the questions. The questionnaire takes about 5 minutes to complete and for patients with clinically significant sleep apnea the positive predictive value is 0.97, the specificity 0.97 and the sensitivity 0.54. Given the reduction in the number of nights of sleep studies, the Berlin Questionnaire has the potential of playing a greater role in deciding whether or not a patient is at high risk of having OSA. However, only a couple of studies in the literature have compared the scores on the Berlin scale to the respiratory disturbance index (or apnea-hypopnea index), the gold standard for sleep apnea diagnosis. Further, none of these studies used a sleep clinic population where the presence of other sleep disorders may alter the sensitivity and specificity of the Berlin Questionnaire for sleep apnea.

Patients with sleep apnea generally present with excessive daytime sleepiness and, women, in particular, seem to complain more of insomnia or fatigue. Clinically, the Epworth Sleepiness Scale (a widely used measure of excessive daytime sleepiness) can be a useful additional tool to help identify those who might be suspected of having sleep apnea. Given the gender differences, the Fatigue Severity Scale (an instrument that has been used to measure fatigue in a number of patient populations) and the Athens Insomnia Scale (an instrument developed to screen for insomnia) would be especially useful to help identify women who may be missed by the usual criteria. No studies in the literature have looked at whether the predictive power of the Berlin Questionnaire in correctly identifying those with sleep apnea can be improved if augmented by information from the Epworth Sleepiness Scale, the Athens Insomnia Scale and the Fatigue Severity Scale.

A screening tool for sleep apnea has significance beyond sleep medicine. In a similar manner as occurs during sleep, reduction or cessation of airflow can occur in sleep apnea patients with central nervous system depression resulting from administration of sedatives or anesthetic agents. However, unlike normal sleep, the defense mechanisms (such as arousal) that help to overcome hypoxia are suppressed during anesthesia. Accordingly, life-threatening respiratory complications have been reported in the peri-surgical and post-surgical period in patients with unrecognized severe sleep apnea. In the anesthetic setting, screening for sleep apnea using the Berlin Questionnaire would be of immense value to help prevent peri- and post-surgical complications. At present, anesthesia researchers at the Toronto Western Hospital are interested in developing an abbreviated form of the Berlin Questionnaire as a quick screening tool for all surgical patients. Information on which items of the Berlin Questionnaire are more predictive of a greater risk of sleep apnea would be desirable for these purposes.

All the aforementioned issues must be explored before the Berlin Questionnaire scale can be relied upon to identify those at high risk for having sleep apnea, not only in the sleep clinic, but in other clinical settings. The aim of this study is to compare the findings on the Berlin Questionnaire to that of overnight polysomnography in patients who have had two nights of sleep studies.

As the healthcare system has indicated that only a single night of diagnostic sleep studies is preferred, it is necessary to investigate whether the Berlin Questionnaire is more predictive of sleep apnea based on the first, second or both nights of sleep apnea screening. A goal of this study is to take into account whether subjective measures of sleepiness, insomnia and fatigue are a useful adjunct in improving the predictive power of the Berlin Questionnaire. An additional goal is to establish the feasibility of abbreviating the Berlin Questionnaire by determining if certain specific items on the scale are more predictive of sleep apnea with a higher sensitivity and specificity. Lastly, as it has been suggested that the presentation of sleep apnea is different in men and women, gender differences in the sensitivity and specificity of the Berlin Questionnaire need to be explored.


Primary: Overall, the Berlin Questionnaire will detect sleep apnea in a sleep clinic patient population with a high sensitivity and specificity. The second night of the diagnostic study will demonstrate a stronger correlation to findings on the questionnaire.

Secondary: A shorter version of the Berlin Questionnaire can be developed to include only those items that are more predictive of sleep apnea.

Tertiary: The Berlin Questionnaire will have reduced sensitivity and specificity in women. Use of the Athens Insomnia Scale and the Fatigue Severity Scale, but not the Epworth Sleepiness Scale, will improve the predictive power of the Berlin Questionnaire in women.

Study design:

This is a retrospective, chart review study of sleep clinic patients who have completed the Berlin Questionnaire for sleep apnea in addition to undergoing two overnight sleep study. All clinic charts from patients who had 2 overnight sleep studies as well as having completed the Berlin Questionnaire in a one year period were included.

Frequently Asked Questions

Please check this section regularly for updates. (Updated May 19, 2006.)

Research Question: 

The objectives of this study are as follows:

  1. To determine the sensitivity and specificity of the Berlin Questionnaire compared to findings of respiratory disturbance based on one versus two nights of polysomnographic recording in a sleep clinic population.
  2. To ascertain which items of the Berlin Questionnaire are more predictive of sleep apnea towards developing an abbreviated version of the questionnaire.
  3. To explore if gender differences alter the specificity and sensitivity of the Berlin Questionnaire.
  4. To determine if a questionnaire battery comprised of the Epworth Sleepiness Scale and the Fatigue Severity Scale in addition to the Berlin Questionnaire improves the sensitivity and specificity of correctly identifying patients with sleep apnea.



Data are available for 133 subjects. Categorical variables are recorded with 1 under the relevant category and blank under other categories.

Subject number 
Alcohol - alcohol usage (Yes or No; amount not specified) 
Caffeine - caffeine usage (Yes or No; coffee, colas, etc) 
AIS - Athens Insomnia Scale, range 0-24, over 6 indicative of insomnia 
Berlin Scale - for sleep apnea, Scoring details and Berlin Questionnaire are here (in pdf) (1? in the data file indicates that the “Don’t know” option was selected for question 1) 
ESS - Epworth Sleepiness Scale, range 0-24, over 10 indicative of excessive daytime sleepiness 
FSS - Fatigue Severity Scale, Scoring information is here 
CES-D - Depression scale, range 0-42, over 16 indicative of depression 
BMI - body mass index (units: kg/m2), height and weight were measured at clinic (Berlin Questionnaire value of BMI was calculated based on height and weight reported by subjects) 
Neck Size - in cm 
BP - Blood Pressure (systolic/diastolic)

First Night test: 

FS - Fatigue scale, range 1-7, subject chooses which response (from 1 to 7) best suits them, higher scores indicate greater fatigue 
SSS - Sleepiness scale, range 1-7, subject chooses which response (from 1 to 7) best suits them, higher scores indicate greater sleepiness 
Sleep Measures: 
SOL - sleep onset latency (minutes taken to fall asleep) 
TST - total sleep time (in minutes) 
SE - sleep efficiency (percentage of time in bed actually spent asleep) 
REM latency - latency (in minutes) to start of REM sleep 
S1 - percentage of stage 1 sleep 
S2 - percentage of stage 2 sleep 
S3 - percentage of stage 3 sleep 
S4 - percentage of stage 4 sleep 
REM - percentage of REM sleep 
Wakefulness - percentage of wakefulness 
RDI - respiratory disturbance index (number of sleep breathing disorders per hour) 
Min O2 sat - minimum oxygen saturation during sleep 
Snoring (self-explanatory) 
PLMI - periodic limb movement index (number of limb movements during sleep per hour of sleep) 
AI - arousal index (number of arousals from sleep per hour of sleep)

MSLT - Mean Sleep Latency Test: day testing adminisitered after the first night in the sleep lab to measure sleepiness; 4 sessions per day, each 30 minutes long during which patients are encouraged to fall asleep. s= fell asleep, w = awake throughout sessions. 

Mean sleep latency - average time it took to fall asleep throughout the 4 sessions. If patient was awake during one or more sessions, 30 minutes sleep latency is used nominally for those sessions to calculate the mean sleep latency. 

REM latency - if the patient had REM sleep during the daytime tests, and if so, how long it took them to enter REM sleep.

Second Night test: 

The above first night measures are repeated here.

MWT - Maintenance of Wakefulness Test: day testing adminsistered which measures the ability to stay awake. Again 4 sessions took place during which the patient was told to try to stay awake and it was recorded whether the patient remained awake (w) or fell asleep (s). Again, if patient was awake for one or more of the 4 sessions, those sessions are designated as a 30 minute sleep latency.



(The following are available from Alison Gibbs.)

  • Chung, SA, Jairam, S, Hussain, MRG, Shapiro, CM. How, what, and why of sleep apnea: Perspectives for primary care physicians. Canadian Family Physician. 2002; 48: 1073-80.
  • Gami AS, Pressman G, Caples SM, Kanagala R, Gard JJ, Davison DE, Malouf JF, Ammash NM, Friedman PA, Somers VK. Association of atrial fibrillation and obstructive sleep apnea. Circulation. 2004; 110(4):364-7.
  • Netzer et. al. Using the Berlin Questionnaire To identify Patients at risk for the sleep apnea syndrome. Annals of internal medicine. 1999; 131: 485-491