Concordance of the late night salivary cortisol in patients with Cushing's syndrome and elevated urine-free cortisol.

The concordance of the late night salivary cortisol (LNSC) results with the 24-h urine-free cortisol (UFC) results in the biochemical screening for Cushing's syndrome is unknown. We investigated this in a population of Cushing's syndrome subjects. We used meta-analytic methods to pool proportions of LNSC-positive subjects from diagnostic evaluations of Cushing's syndrome subjects where both tests were performed and the UFC was elevated (any level). Cushing's syndrome was confirmed in all subjects by two out of three conventional tests. LNSC was collected between 22:00 to 24:00 h and measured around the same time period as the UFC. Minimum cutoffs of ≥4 and ≥10 nmol/L were used to determine concordance with the UFC and studies were limited to those that used radioimmunoassays or electrochemiluminiscence immunoassays for LNSC. The concordance of LNSC ≥4 nmol/L was 97 % (95 % CI 95-99 %) and studies were homogeneous. With LNSC ≥10 nmol/L, there was heterogeneity and two groups were discernible with a pooled concordance of 69 % (95 % CI 60-77 %) and 95 % (95 % CI 92-97 %). Within these sub-groups, studies were homogeneous and there was no difference between them in collection methods, assays used, geographic location, year of publication, or the quality of the underlying studies. The LNSC at a very specific cutoff detects at best 95 % of cases and at worst 69 % of cases of Cushing's syndrome that are UFC positive. The two tests become equivalent at the more sensitive cutoff (>4 nmol/L). We conclude that, given its many benefits and the currently documented equivalence to the UFC, the LNSC should replace the conventional 24-h UFC as the frontline test when screening for Cushing's syndrome.