Berlin Questionnaire for Sleep Apnea Risk Assessment

The Berlin Questionnaire is a self-administered screening tool designed to identify people at high risk for obstructive sleep apnea (OSA). Developed in 1996 at a conference of sleep and primary care researchers in Berlin, Germany, it groups questions into three categories covering snoring, daytime sleepiness, and body weight plus blood pressure. A person flagged as high risk in at least two of those three categories is considered likely to have sleep apnea and a candidate for formal testing. The questionnaire has been translated into dozens of languages and validated in settings ranging from primary care offices to pre-surgical clinics, though its accuracy varies more than many clinicians expect.

What the Questionnaire Asks

The Berlin Questionnaire contains about ten items, organized into three symptom categories. The first category asks about snoring: whether you snore, how loud it is, how often it happens, and whether it bothers other people. The second category focuses on daytime fatigue and sleepiness, including whether you feel tired after sleeping, whether you experience daytime drowsiness, and whether you have ever nodded off while driving. The third category records whether you have high blood pressure and captures your body mass index (BMI) based on self-reported height and weight.

Scoring is straightforward. Each category gets marked as positive or negative based on the frequency and severity of the symptoms reported. If two or more categories come back positive, the questionnaire classifies you as high risk for sleep apnea. If only one or none is positive, you land in the low-risk group. The original validation study found strong internal consistency across the questions and reported that the high-risk classification predicted a respiratory disturbance index above 5 with a sensitivity of 86% and specificity of 77%.¹

How Accurate Is It in Practice

The numbers from that original 1999 study painted an encouraging picture, but subsequent research in broader populations tells a more nuanced story. In a general population study, the questionnaire’s sensitivity for moderate-to-severe sleep apnea (defined as 15 or more breathing disruptions per hour) was about 59%, with a specificity around 78%.² When the threshold was raised to severe sleep apnea (30 or more events per hour), sensitivity climbed to roughly 77%, though the questionnaire’s ability to correctly identify people without the condition dropped slightly. The pattern is consistent: the Berlin Questionnaire tends to catch more severe cases while letting milder ones slip through.

One study in patients with metabolic syndrome found a similar trend. For any level of OSA, sensitivity sat at about 65% with specificity around 38%. But for severe OSA, sensitivity jumped to 91%.³ The questionnaire is better at flagging the people who most urgently need treatment than at catching everyone with some degree of sleep-disordered breathing.

Context matters, too. In a sleep clinic population where most patients already had symptoms prompting a referral, the Berlin Questionnaire performed poorly, identifying only 62% of true positives while misclassifying many patients in both directions.⁴ That makes intuitive sense. In a population that already skews heavily toward sleep problems, a screening tool built for general populations loses its ability to separate signal from noise. The questionnaire works best when it is doing what it was originally designed to do: flagging unsuspected risk in people who have not yet been referred to a sleep specialist.

Berlin Versus STOP-Bang and Other Screening Tools

The Berlin Questionnaire is far from the only game in town. The STOP-Bang questionnaire, developed for surgical patients, asks about snoring, tiredness, observed apnea, blood pressure, BMI, age, neck circumference, and gender. The Epworth Sleepiness Scale (ESS) takes a different angle entirely, measuring subjective daytime sleepiness by asking how likely you are to doze off in various situations.

Head-to-head comparisons show trade-offs rather than a clear winner. In one study comparing all three tools, the Berlin Questionnaire had the highest sensitivity at about 86%, meaning it caught the most true cases, but STOP-Bang had markedly better specificity at roughly 82%, meaning it produced fewer false alarms.⁵ The ESS consistently lagged behind both on overall diagnostic accuracy. A separate study in people with type 2 diabetes found that the Berlin and STOP-Bang questionnaires performed similarly for moderate-to-severe OSA, with sensitivities near 65–69% and specificities around 49–50%.⁶

The Epworth Sleepiness Scale’s weakness as a standalone OSA screener comes down to what it actually measures. It captures how sleepy you feel, which depends on individual perception, sleep habits, and even cultural attitudes toward napping. Plenty of people with significant airway obstruction during sleep do not report feeling unusually drowsy during the day. Because the Berlin Questionnaire incorporates snoring patterns, witnessed breathing pauses, and physical risk factors alongside sleepiness, it casts a wider net.⁷

In patients with type 2 diabetes specifically, one study found that for severe OSA, both the Berlin Questionnaire and STOP-Bang reached 100% sensitivity, while the ESS offered the highest specificity at about 53%.⁸ In practice, some clinicians use multiple questionnaires together, hoping that the combination catches what any single tool might miss.

Pre-Surgical Screening

Undiagnosed sleep apnea in surgical patients is a genuine safety concern. Anesthesia can relax airway muscles and interact with the breathing disruptions that OSA causes, raising the risk of complications during and after an operation. The Berlin Questionnaire has been studied as a way to catch these patients before they reach the operating room.

In a validation study of surgical patients who underwent full polysomnography, the Berlin Questionnaire’s sensitivity ranged from about 69% to 87% depending on the severity threshold used, performing comparably to both the ASA checklist and the STOP questionnaire.⁹ A separate study of elective surgical patients found that the questionnaire classified about 24% of patients as high risk, and it correctly identified every patient who had a prior diagnosis of OSA.¹⁰ The implication is practical: for busy pre-operative clinics where polysomnography is not feasible for every patient, a quick paper questionnaire can at least flag the people who need closer monitoring.

Using a Bed Partner’s Answers

Sleep apnea has an odd feature: the person who suffers from it is often the last to know. Snoring, gasping, and breathing pauses happen during sleep, and the bed partner is frequently more aware of these symptoms than the sleeper. Researchers have tested whether having a bed partner fill out the Berlin Questionnaire improves its accuracy.

In one study, the bed-partner version predicted moderate-to-severe OSA with a sensitivity of 82% compared to 76% when the patient filled it out themselves. For severe OSA, the bed-partner version reached 79% sensitivity versus 71% for self-report, with specificity remaining similar in both cases.¹¹ The improvement is modest but consistent. If you sleep alone, your answers are still useful, but a partner’s observations about your snoring and breathing can sharpen the result.

Pregnancy Complicates Things

Pregnant women are at increased risk for sleep-disordered breathing, particularly later in pregnancy when weight gain, fluid retention, and hormonal changes narrow the upper airway. Screening would be valuable, but the Berlin Questionnaire has a complicated relationship with pregnancy.

One study tracking pregnant women across trimesters found that the questionnaire’s predictive value was essentially useless in the first trimester but improved substantially in the second and third trimesters, with area-under-the-curve values reaching 0.84 and 0.81 respectively.¹² The trouble is that by later pregnancy, nearly everyone gains weight and many develop elevated blood pressure, which are two of the three Berlin categories. A separate analysis found that the questionnaire flagged 72% of pregnant women as high risk, with sensitivity of 83% but specificity of only 37%.¹³ When nearly three-quarters of your population screens positive, the questionnaire is not really screening anymore.

Researchers have pointed out that in pregnancy, the Berlin Questionnaire’s risk classification is driven predominantly by the obesity category, which picks up normal gestational weight changes alongside genuinely elevated risk.¹⁴ Pregnancy-specific screening tools are under development, but for now, clinicians working with pregnant patients should interpret a high-risk Berlin result with extra caution.

Stroke Patients Present a Different Challenge

Sleep apnea is extremely common after stroke, with estimates suggesting it affects a majority of acute stroke patients. Screening this population matters because untreated OSA can impair recovery and increase the risk of recurrent stroke. But the Berlin Questionnaire faces a unique obstacle here: many stroke patients cannot reliably self-report due to communication difficulties, cognitive impairment, or reduced awareness of their pre-stroke symptoms.

Studies in stroke populations have found that the questionnaire underperforms. In one evaluation, the Berlin Questionnaire achieved only about 46% sensitivity and 73% specificity for moderate-to-severe sleep apnea in acute ischemic stroke patients. A modified version that isolated BMI as a separate risk category improved sensitivity modestly to about 54%.¹⁵ When the questionnaire was administered to the informants living with stroke patients rather than the patients themselves, results were only slightly better, with sensitivity around 67% and specificity near 56%.¹⁶ Another study of ischemic stroke patients found the STOP-Bang questionnaire outperformed the Berlin Questionnaire on both sensitivity and specificity in this group.¹⁷

The evidence is fairly clear that in neurological populations, the Berlin Questionnaire should not be the primary screening method. Direct sleep testing is more appropriate when stroke is involved.

The Cardiovascular Connection

Beyond its role as a screening tool, the Berlin Questionnaire has been used in research to explore the relationship between sleep apnea risk and cardiovascular disease. The association between OSA and heart disease is well established, but the questionnaire has offered a practical way to study it in large groups where polysomnography would be impractical.

One study of patients undergoing coronary angiography found that those classified as high risk by the Berlin Questionnaire had significantly more severe coronary artery disease, with higher scores on a standard measure of arterial blockage.¹⁸ Another study reported that high-risk status on the questionnaire was associated with roughly 4.5 times the odds of having coronary artery disease after adjusting for age, sex, and BMI.¹⁹ In patients who had already experienced an acute coronary event, a high-risk Berlin result predicted roughly 3.7 times the odds of composite cardiovascular events during follow-up.²⁰

These associations do not mean the questionnaire is diagnosing heart disease. What they suggest is that the cluster of symptoms it captures, including obesity, hypertension, and disordered breathing during sleep, overlaps substantially with cardiovascular risk. For cardiologists, a positive Berlin result on a heart patient may be a practical reason to pursue formal sleep testing.

Commercial Drivers and Workplace Screening

Few occupations make the consequences of untreated sleep apnea as concrete as long-haul trucking. Drowsy driving is a major safety hazard, and OSA is one of its most treatable causes. The U.S. Federal Motor Carrier Safety Administration’s Medical Review Board has recommended that commercial drivers undergo formal OSA testing if they screen positive on the Berlin Questionnaire or have a BMI of 30 or above.²¹

When researchers deployed an anonymous online Berlin-based screening tool for commercial drivers, roughly 56% of the 595 respondents screened positive. About 70% were obese, and one in five reported having fallen asleep at a stoplight. A study of truck drivers in Brazil found that about 26% were classified as high risk by the Berlin Questionnaire.²² A Serbian validation study in commercial drivers found that the questionnaire showed reasonable measurement properties, with sensitivity ranging from about 51% for mild OSA up to 75% for severe cases, and specificity between 70% and 86%.²³

The appeal of the Berlin Questionnaire in occupational settings is its simplicity. It does not require medical equipment, can be completed in a few minutes, and can be administered at scale. The risk is that its moderate specificity generates a large number of false positives, potentially triggering expensive follow-up testing for drivers who do not actually have OSA. The debate over the right balance between catching dangerous cases and minimizing unnecessary testing burdens is ongoing.

What Clinical Guidelines Actually Say

Despite the Berlin Questionnaire’s wide use, the American Academy of Sleep Medicine issued a strong recommendation that clinical tools, questionnaires, and prediction algorithms should not be used to diagnose OSA in adults without polysomnography or home sleep apnea testing.²⁴ This is the single most important thing to understand about the questionnaire: it is a screening tool, not a diagnostic one. Screening identifies people who need further testing. Diagnosis requires measuring what actually happens during sleep.

In patients with type 2 diabetes, the Berlin Questionnaire shows high sensitivity around 80% but low specificity near 48%, meaning it flags many diabetic patients who do not actually have clinically significant OSA.²⁵ In patients who had recently suffered a heart attack, overall diagnostic accuracy was only 63%, with a specificity of 34%.²⁶ In a population where nearly everyone is already at elevated cardiovascular risk, the questionnaire’s risk factors overlap too heavily with the conditions these patients already have.

The guideline’s message is not that the Berlin Questionnaire is worthless. It is that no paper questionnaire replaces a sleep study, and clinicians should not skip testing based on a low-risk Berlin result if clinical suspicion remains high.

Translations and Global Use

One of the Berlin Questionnaire’s practical strengths is its adaptability across languages and cultures. It has been validated in Greek for primary care use, where researchers confirmed its utility for flagging OSA risk in that setting.²⁷ A Modern Standard Arabic adaptation demonstrated strong validity and reliability, with near-perfect agreement on test-retest measures and Cronbach’s alpha values indicating adequate internal consistency across the three categories.²⁸

These translations matter because OSA is a global problem, and sleep labs are expensive, scarce, and unevenly distributed. In healthcare systems where formal sleep testing has long wait times or limited availability, a validated questionnaire in the local language can at least help clinicians prioritize who gets tested first. The questionnaire’s simplicity is its greatest asset in resource-limited settings, even though that same simplicity limits its precision.

Tracking Changes After Weight-Loss Surgery

The Berlin Questionnaire has also been used longitudinally to monitor whether OSA risk changes after major interventions. Bariatric surgery offers a compelling test case, since significant weight loss often reduces or resolves sleep apnea. In a prospective study of bariatric surgery patients, the proportion classified as high risk by the Berlin Questionnaire dropped from about 93% before surgery to roughly 30% afterward.²⁹ STOP-Bang scores and Epworth Sleepiness Scale scores dropped significantly as well. This kind of before-and-after use gives clinicians a quick read on whether a patient’s risk profile has shifted enough to warrant repeat sleep testing, without requiring a new polysomnography for every follow-up visit.

The questionnaire was never intended to replace objective measurement, but it fills a real gap in the many situations where objective measurement is not immediately available. Understanding where it works well, where it breaks down, and what it cannot do on its own is the difference between using it wisely and leaning on it too hard.