Diagnostic characteristics of late-night salivary cortisol using liquid chromatography-tandem mass spectrometry

Objective: The objective of the study was to describe the diagnostic performance of a commercially available late-night salivary cortisol (NSC) assay using liquid chromatography tandem mass spectrometry. Methods: We retrospectively identified 90 patients who had one or more NSC determinations: 52 patients in whom Cushing syndrome (CS) was excluded or could not be confirmed [group 1 (G1)] and 38 patients in whom CS was confirmed [group 2 (G2)]. Eighteen healthy volunteers served as controls. Results: Baseline demographics in all groups were similar with regards to age, ethnicity, gender, and body mass index. NSC levels [median (range)] were higher in G2, 381 (64-13,500) ng/dl [10.51 (1.77-372.46) nmol/liter], compared with controls, 19.3 (2.1-416) ng/dl [0.53 (0.06 -11.48) nmol/liter], and G1, 26 (4-176) ng/dl [0.72 (0.11-4.86) nmol/liter, P < 0.001]. The highest combined sensitivity (92%) and specificity (92%) was achieved at a cut point of 107 ng/dl (2.95 nmol/liter). Two or more NSCs were done in 32 of 52 G1 and 31 of 38 G2 patients. In G1 eight of 32 (25%) had at least one elevated [>100 ng/dl (2.76 nmol/liter)] NSC including two in whom both NSCs were elevated. In contrast, four of 31 (13%) in G2 had at least one normal NSC including one with four of five normal NSC values. None of the patients with CS had a NSC less than 60 ng/dl (<1.66 nmol/liter). Comparing G1 and G2, obtaining more than one saliva sample did not improve the diagnostic accuracy of NSC measurement (P = 0.64). Conclusion: The liquid chromatography tandem mass spectrometry assay to measure NSC is a simple and reliable test to screen patients suspected to have CS. Clinicians should be aware of appropriate cutoff values for proper interpretation of NSC and use additional tests when necessary.