Please note that this summary was posted more than 5 years ago. More recent research findings may have been published.
This is a plain English summary of an original research article. The views expressed are those of the author(s) and reviewer(s) at the time of publication.
More use could be made of the STOP-Bang questionnaire as a self-completed ‘first check’ for detecting obstructive sleep apnoea or in selected people before surgery. The downside is that a high rate of false positives, identified by this review, means that many people in the community without the condition might test positive and receive unnecessary attention. This might not be a problem, as this common sleep-related breathing condition is under-diagnosed, but the consequences will need further evaluation. Sleep apnoea is associated with various health risks, including accidents due to daytime sleepiness, so once a firm diagnosis is made, people can benefit from the treatments available and anaesthetists can be prepared before surgery.
This review assessed the performance of the STOP-Bang questionnaire when used by people referred to sleep clinics, or before surgery. Compared with formal diagnosis in a sleep clinic, STOP-Bang demonstrated good sensitivity for detecting sleep apnoea in people with the condition (90%) but was not specific enough to be useful for making a diagnosis on its own (specificity 49%).
Questionnaires such as this could be a convenient first check for identifying an underdiagnosed condition, but is unlikely to take the place of a clinical assessment or reduce the use of sleep clinics for diagnosis. This study was conducted in hospital sleep clinics on people who are more likely to have sleep apnoea. The usefulness of the test in an unselected primary care population may be different and needs fuller assessment because false positives generated by use of the test could become a burden to the health system.
Other questionnaires exist. At present, UK guidance does not cover systematic screening for the condition, or advise which questionnaire is best for diagnosis or triage. Further research comparing the clinical and cost effectiveness and acceptability of different tools in UK settings and in primary care is needed before changing guidance or practice.
Why was this study needed?
Sleep apnoea is a condition where the throat relaxes during sleep and temporarily obstructs breathing, disturbing sleep. It is thought to affect up 1.5 million people in the UK, of whom 80% are undiagnosed. Research by the Office of Health Economics for the British Lung Foundation found that improving awareness, diagnosis and treatment of sleep apnoea could save the NHS up to £28 million and prevent up to 40,000 road traffic accidents a year.
Although sleep apnoea is common, its symptoms are not easy to detect yourself. A technique called polysomnography is used for reliable diagnosis, but is time-consuming, relies on sleep experts and is not available in all areas of the UK. It involves being monitored overnight at a sleep laboratory or using monitoring equipment at home, including breathing and blood oxygen sensors.
A simple, effective questionnaire could be a cheap and convenient way to identify people at risk of the condition, for example in primary care or before having surgery, so the reviewers here set out to see how accurate this one was. Other screening questionnaires exist, for example the Berlin scale, GASP (Graduated Apnoea Screening Protocol), and for sleepiness, the Epworth Sleepiness Scale. This review did not address comparisons between screening questionnaires, although previous reviews have shown mixed but generally favourable performance of STOP-Bang.
What did this study do?
This was a systematic review of 17 diagnostic test accuracy studies, including 9,206 participants. The findings of 11 studies of people referred to sleep laboratories and three studies of patients being screened before operations were pooled in meta-analysis. These two groups of people differed by risk of sleep apnoea (high for clinical referrals, compared with a lower risk for pre-operative patients), and in test availability (definitive overnight tests at sleep clinics, compared with need for rapid, accurate screening before surgery). However, the majority of the participants were middle aged men with high BMI.
The included studies examined STOP-Bang questionnaire performance by measuring the proportion detected compared with the most reliable diagnosis, directly measuring low oxygen breathing episodes during sleep, in a sleep laboratory (polysomnography).
The review was carried out to a high standard, using PRISMA guidelines. Applicability of the findings to UK care was limited by including only one, modestly sized UK study, which took place in a hospital.
What did it find?
- Pooled findings from the 11 sleep clinic studies (85% prevalence of sleep apnoea) found that the proportion of people correctly detected by the questionnaire, set at a three or more cut-off, was 90% (95% confidence interval [CI] 88% to 91%). The definition of sleep apnoea was made on a subsequent recording of apnoea-hypopnoea index [AHI] of 5 or more on polysomnography.
- In these sleep clinic studies the pooled specificity was 49% (95% CI 45% to 54%). This is the chance that someone without the condition is correctly detected as being free from sleep apnoea. The result suggests that more than half (51%) of people without sleep apnoea might be unnecessarily referred for further tests following the questionnaire.
- Pooled findings from the three pre-operative studies (a lower prevalence of 68% compared to sleep clinics) found similar results: sensitivity 84% and specificity 43% for sleep apnoea (two studies).
- In one general population study (with the lowest prevalence of 12.7% for moderate-to-severe sleep apnoea) the sensitivity was 88% to detect moderate-to-severe OSA, the specificity remained at 30% but the positive predictive value fell to 16%. This means that for those testing positive on the questionnaire only 16% actually will turn out to have sleep apnoea if tested in a sleep lab later.
- In general, the higher the STOP-Bang score cut-off, the higher the probability of having moderate to severe sleep apnoea, but this varied by study setting and underlying prevalence of the condition. For example, in pooled results from the sleep clinic studies, a STOP-Bang score of 3 had a 25% likelihood of having severe sleep apnoea, whereas the pooled results from pre-operative studies showed that people with a score of 3 had a15% likelihood of having severe sleep apnoea.
What does current guidance say on this issue?
NICE guidance from 2008 states that diagnosis of sleep apnoea should be carried out by a specialist service. It does not cover population screening for the condition, which is not as yet a policy assessed or recommended by the UK National Screening Committee.
US guidance from 2014 on the management of people undergoing surgery, notes that the performance of sleep apnoea screening questionnaires varied widely between studies. It recommends drawing diagnostic information from sources such as clinical examination, medical history and sleep studies if sleep apnoea is suspected.
What are the implications?
A simple, effective questionnaire such as STOP-Bang could help identify people at risk of this under-diagnosed condition, but over half of people who test positive using the tool in any setting do not have sleep apnoea. If the questionnaire is used in primary care, without a clinical suspicion of sleep apnoea, many people many people may be referred onwards to scarce specialist sleep clinic resources.
So far there has been limited UK-based research. Most studies assessing the performance of sleep apnoea screening questionnaires have been carried out in North America. Before implementing in the NHS, further research is needed to investigate how the tool performs in UK settings, such as in primary care and in healthy populations before planned surgery. Comparison of STOP-Bang’s performance against other scales is also needed to assess their relative clinical and cost effectiveness and acceptability in the UK.
Citation and Funding
Nagappa M, Liao P, Wong J, et al. Validation of the STOP-Bang questionnaire as a screening tool for obstructive sleep apnea among different populations: a systematic review and meta-analysis. PLoS One. 2015;10(12):e0143697.
This work was supported by University Health Network Foundation and Department of Anesthesia, Toronto Western Hospital, University Health Network, University of Toronto.
No funding bodies had any role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Bibliography
Abrishami A, Khajehdehi A, Chung F. A systematic review of screening questionnaires for obstructive sleep apnea. Can J Anesth. 2010;57:423–38.
BLF. Obstructive sleep apnoea (OSA) [internet]. London: British Lung Foundation; 2014.
Cowan DC, Allardice G, MacFarlane D, et al. Predicting sleep disordered breathing in outpatients with suspected OSA. BMJ open. 2014;4(4):e004519.
Gross JB, Apfelbaum JL, Caplan RA, et al. Practice guidelines for the perioperative management of patients with obstructive sleep apnea. An updated report by the American Society of Anesthesiologists Task Force on perioperative management of patients with obstructive sleep apnea. Anesthesiology. 2014;120:268–86.
Johns MW. A new method for measuring daytime sleepiness: The Epworth Sleepiness Scale. Sleep 1991; 14:540-5.
Netzer NC, Stoohs RA, Netzer CM, et al. Using the Berlin Questionnaire to identify patients at risk for the sleep apnoea syndrome. Ann Intern Med. 1999;131(7):485-91.
NHS Choices. Obstructive sleep apnoea [internet]. Leeds: NHS Choices; 2014.
NICE. Continuous positive airway pressure for the treatment of obstructive sleep apnoea/hypopnoea syndrome. TA139. London: National Institute for Health and Care Excellence; 2008.
Rejón-Parrilla JC, Garau M, Jon Sussex J. Obstructive sleep apnoea health economics report. London: British Lung Foundation, Office for Health Economics; 2014.
Produced by the University of Southampton and Bazian on behalf of NIHR through the NIHR Dissemination Centre
