This review assessed the diagnostic value of questionnaires, clinical scales or prediction equations compared with standard overnight polysomnography for diagnosis of obstructive sleep apnea. The authors concluded that most of the clinical screening tests would miss a significant proportion of such patients. A number of issues with the review process made the reliability of this conclusion difficult to determine.

To compare clinical screening tests for sleep apnoea and establish an evidence base for their preoperative use.

PubMed and Cochrane Central Register of Controlled Trials were searched from 1966 to May 2008. Search terms were reported. References of identified studies were checked. Nine relevant journals were handsearched. Only studies published in English were eligible for inclusion.

Studies that assessed the diagnostic value of questionnaires, clinical scales or prediction equations compared with standard overnight polysomnography in a hospital or laboratory facility for diagnosis of obstructive sleep apnoea were eligible for inclusion. Studies were required to report pretest probabilities and data to enable construction of a 2x2 table, sensitivity and specificity or positive and negative likelihood ratios.

A number of questionnaires and clinical models were assessed. These included Berlin questionnaire, Sleep Disorders Questionnaire, checklist of the American Society of Anesthesiologists, Snoring Questionnaire, Kushida Index and Epworth Sleepiness Scale. A range of apnoea-hypopnoea index thresholds were used (between 5 and 10 for apnoea and 25 or 30 for severe apnoea).The prevalence of obstructive sleep apnoea ranged from 9% to 85%. The proportion of men in the included studies ranged from 24% to 100%.

Two reviewers independently assessed the studies for inclusion in the review; disagreements were resolved through discussion and consultation with a third reviewer.

The studies were assessed for validity using the QUADAS framework.

The authors did not state how many reviewers performed validity assessment.

Data on true and false positives and negatives were extracted or calculated; these were used to obtain sensitivity, specificity, likelihood ratios and diagnostic odds ratios (DORs), with 95% confidence intervals (CI). The following categories of test were defined: excellent (DOR of 81 or greater); good (DOR between 10 and 80); average (DOR between 5 and 10); poor (DOR between 2 and 5); and no value (DOR below 2).

It appeared that one reviewer performed data extraction, which was checked by a second reviewer.

Studies were combined using a DerSimonian and Laird random-effects meta-analysis. Separate estimates were calculated for obstructive apnoea and severe obstructive apnea, and for questionnaires and clinical models. Statistical heterogeneity was assessed using Cochran's Q and the I² statistic. It was planned to use a Mantel-Haenszel fixed-effect model in the absence of significant heterogeneity. Subgroup analyses were performed for each test that was assessed by more than one study. A meta-regression was used to identify sources of heterogeneity and elements that were important to diagnostic performance. This assessed study characteristics including the impact of the apnoea-hypopnoea index threshold used and patient characteristics.

Twenty-six studies with a total of 6,794 patients with suspected obstructive sleep apnoea were included in the review. Median sample size was 123 (range 33 to 1,409). QUADAS scores ranged from 6 to 13 .

Statistical heterogeneity between studies was high (I² ranged from 74.4% to 90.6%). No single questionnaire or clinical model satisfied criteria for the ideal preoperative screening test. Only the Kushida Index had a reproducibly excellent DOR, but this was based on only two studies. Clinical models had marginally better pooled DOR than questionnaire models for both diagnosis of obstructive sleep apnoea (10.49 versus 5.02) and severity (17.24 versus 10.12). Rates of false negatives varied widely.

Due to substantial false negative rates, it is likely that most of the clinical screening tests would miss a significant proportion of patients with obstructive sleep apnea.

The review question was clear. The authors searched two relevant databases and additional sources. However, the restriction to published studies reported in English increased the risk of language and publication biases, and the chance that relevant studies were omitted. An appropriate tool was used to appraise study validity, but calculation of a summary score based on the criteria evaluated was not informative. Methods to reduce reviewer bias and error were used in study selection and data extraction, but were not reported for validity assessment. The decision to use meta-analysis may not have been advisable given high levels of clinical and methodological heterogeneity between studies. DOR is not the most informative measure of a test's performance. The limitations of the search and concerns about the synthesis made it difficult to determine the reliability of the conclusions.

Practice: The authors did not state any implications for practice.

Research: The authors stated that further validation of Berlin questionnaire and Sleep Disorders Questionnaire in a typical surgical population were required.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.