Karlijn Koops^a,b,c, Peter H. Bisschop^b,d, Annemieke C. Heijboer^a,b,c,e, Jacquelien J. Hillebrand^b,c

• a: Amsterdam UMC location Vrije Universiteit Amsterdam, Department of Laboratory Medicine, Endocrine Laboratory, Boelelaan 1117, Amsterdam, The Netherlands
• b: Amsterdam Gastroenterology, Endocrinology & Metabolism, Amsterdam, The Netherlands
• c: Amsterdam UMC location University of Amsterdam, Department of Laboratory Medicine, Endocrine Laboratory, Meibergdreef 9, Amsterdam, The Netherlands
• d: Amsterdam UMC, University of Amsterdam, Department of Endocrinology and Metabolism, Meibergdreef 9, Amsterdam, The Netherlands
• e: Amsterdam Reproduction & Development Research Institute, Amsterdam, The Netherlands.

Objective:

One of the recommended initial tests for diagnosing endogenous hypercortisolism is the measurement of late-night salivary cortisol (cortisol_LNSa). Published cut-off values of cortisol_LNSa as well as late-night salivary cortisone (cortisone_LNSa) vary widely between studies. This study aims to establish cut-off values for cortisol_LNSa and cortisone_LNSa using liquid chromatography-tandem mass spectrometry (LC-MS/MS) to distinguish between people with and without endogenous hypercortisolism.

Design:

A retrospective study was conducted on subjects who had late-night saliva samples taken for evaluation of endogenous hypercortisolism at Amsterdam UMC between December 2015 and February 2022.

Methods:

25 subjects with and 430 subjects without endogenous hypercortisolism were included. Salivary samples were collected between 22:00-00:00 on two consecutive days. Cortisol_LNSa and cortisone_LNSa concentrations were measured using LC-MS/MS. The diagnosis of endogenous hypercortisolism was confirmed by histology and postoperative clinical features, and excluded based on a follow-up of ≥12 months without clinical evidence for progressive signs of hypercortisolism.

Results:

The median cortisol_LNSa was 7.8 nmol/L for subjects with and 0.8 nmol/L for subjects without endogenous hypercortisolism (p<0.001). For cortisone_LNSa, this was 35.5 nmol/L and 5.3 nmol/L, respectively (p<0.001). Optimal diagnostic accuracy was established at 2.25 nmol/L for cortisol_LNSa (sensitivity: 92%, specificity: 89%, positive predictive value (PPV): 32.9%, negative predictive value (NPV): 99.5%) and 15.5 nmol/L for cortisone_LNSa(sensitivity: 93.8%, specificity: 94.5%, PPV: 46.9%, NPV: 99.7%). The sum of cortisol_LNSa+cortisone_LNSa provided a slightly higher diagnostic accuracy with a cut-off of 18.5 nmol/L (sensitivity: 93.8%, specificity: 95.2%, PPV: 50%, NPV: 99.7%).

Conclusion:

We established cut-off values for cortisol_LNSa and cortisone_LNSa for the diagnosis of endogenous hypercortisolism. Cortisone_LNSa and the sum cortisol_LNSa+cortisone_LNSa provided the highest diagnostic accuracy.