To establish the evidence base for the diagnosis of sleep apnea (SA) in adult patients, a systematic review of the literature from 1980 through November 1, 1997 was performed. Diagnostic studies were included if they reported results of any test to establish or support a diagnosis of SA, in comparison to a diagnosis from a full polysomnogram (PSG). Test results were meta-analyzed using fixed effects models and summary receiver operating characteristic curves (ROCs) to examine consistency of tests within and between diagnostics vs. the "gold standard" of PSG. From a total of 937 studies, 249 fit the broad eligibility criteria for inclusion in the clinical trial database and its data were extracted from these reports; useable data for statistical analyses were reported in 71 studies (7,572 patients). The sensitivity and specificity of partial channel and partial time PSGs appeared most promising as replacements for full PSG in patients suspected of obstructive SA. Clinical prediction rules (multivariate models) were also promising. Studies of portable sleep monitors, radiologic or morphologic features, and focused questionnaires were too heterogeneous to be meta-analyzed. In general, the diversity of study designs and objectives were very high and the methodological rigor of these studies as assessments of diagnostic tests was very low. Thus, we are still not in a position to recommend standardization of diagnostic methodology for sleep apnea. Instead, our recommendations for future research include standardization of terms and diagnostic criteria, and consistently reported statistics to enhance the utility of this literature.