Specificity of late-night salivary cortisol measured by automated electrochemiluminescence immunoassay for Cushing's disease in an obese population.

Abstract

Data about the specificity of late-night salivary cortisol (LNSC) in obese subjects are still conflicting. Therefore, with this study, we aimed to evaluate the specificity of LNSC measurement in an obese cohort with or without type 2 diabetes mellitus (T2DM) using an automated electrochemiluminescence immunoassay (ECLIA). A total number of 157 patients involving 40 healthy subjects (HS) with BMI < 25kg/m2, 83 obese subjects (OS) with BMI ≥ 35kg/m2, and 34 histopathologically proven Cushing's disease (CD) were included. All patients underwent LNSC testing. Salivary cortisol was measured at 11p.m. for all groups using an ECLIA. Reference range was established using values of LNSCs of HS and ROC curves were used to determine diagnostic cutoffs. In the HS group, mean LNSC was 4.7nmol/l (SD ± 3.1), while the OS group had a mean value of 10.9nmol/l (SD ± 7.5) and the CD group of 19.9nmol/l (SD ± 15.4). All groups differed significantly (p < 0.001). The ROC analysis of CD against HS alone showed a sensitivity of 85.3% and a specificity of 87.5% with a cut-off value of 8.3nmol/l. The ROC analysis between OS and CD showed a maximum sensitivity of 67.6% and specificity of 78.3% for a cut-off value of 12.3nmol/l. Taken both (HS and OS) groups together against the CD group, ROC analysis showed a maximum sensitivity of 67.6% and specificity of 85.4% for a cut-off value of 12.3nmol/l. No correlation was found between BMI, T2DM, and LNSC for all groups. In our obese cohort, we found that LNSC assayed by ECLIA had a low specificity in the diagnosis of CD.