Late-night salivary cortisol for the diagnosis of Cushing syndrome: a meta-analysis.

OBJECTIVE To report a meta-analysis of late-night salivary cortisol testing for the diagnosis of Cushing syndrome. METHODS MEDLINE and EMBASE computer databases were searched to identify relevant articles published between January 1950 and December 2007. The search strategy used the following medical subject headings and keywords: cortisol, Cushing or Cushing's, saliva, salivary, late-night, nocturnal, and nighttime. The results were limited to studies in humans older than 18 years. Titles and abstracts of all articles, as well as full text of relevant articles, were reviewed. Sensitivity, specificity, likelihood ratio positive, likelihood ratio negative, and diagnostic odds ratio were extracted by 2 authors. Discrepancies were resolved by mediation and discussion with a third author. RESULTS Seven articles contained sufficient information to be included in the analysis. A total of 947 patients (339 with Cushing syndrome) were identified. Pooled data from the 7 studies revealed a sensitivity of 92% (95% confidence interval [CI], 88%-94%), specificity of 96% (95% CI, 94%-97%), and diagnostic odds ratio of 311 (95% CI, 92-1059). Likelihood ratio positive was 21 (95% CI, 10-43), with a likelihood ratio negative of 0.08 (95% CI, 0.02-0.32). Inconsistencies for each of these results measured by the I2 statistic ranged from moderate to high. CONCLUSION This analysis demonstrates that late night salivary cortisol has excellent diagnostic characteristics and as such, is a robust, convenient test for screening and diagnosis of Cushing syndrome.