Late-night salivary cortisol versus urinary free cortisol in the monitorization of cushing's syndrome patients treated with steroidogenesis inhibitors

A. Vega-Beyhart*^a (Dr), D. Díaz-Catalán^a (Dr), D. Salas-Bravo^a (Mr), L. Boswell^a (Dr), M. Mora^a (Dr), I. Halperin^a (Dr), G. Casals^a (Dr), FA. Hanzu^a (Dr)

^a Hospital Clinic Barcelona, Barcelona, SPAIN

* arturovega32@gmail.com

Introduction

Twenty-four-hour urinary free cortisol (24h-UFC) is the main cortisol biomarker employed in the follow-up of patients with Cushing Syndrome (CS) during treatment with steroidogenesis inhibitors (SEI). However, 24h-UFC has pitfalls as the intraindividual variability and cross-reactive interferences. Use of late-night salivary cortisol (LNSC) to monitor SEI is in debate because its unclear whether only targeting diurnal secretion resembles the disease activity.

Objective

To determine whether LNSC is a helpful biomarker in the monitorization of patients with CS treated with Ketoconazole (KTZ) or Metyrapone (MTP), comparing it with 24h-UFC.

Methods

Between 2014 and 2020, 34 patients were included and followed during the initiation, dose titration and maintenance phase (>3 months) of the SEI (KTZ = 24, MTP = 10). Hormonal and cardiometabolic outcomes were assessed in each clinical visit. LNSC was measured by immunoassay (IA) whereas 24h-UFC by IA and gas-chromatography/mass-spectrometry (GC-MS).

Results

Before treatment, median LNSC was 3.59 (2.3 - 7.2) µg/L (upper limit of normal [ULN*LNSC = 2.8 ± 1.5]). 24h-UFC*IA was 335.6 µg (151.9 - 603.1, ULN*IA = 3.5 ± 2.6) whereas 24h-UFC*MS was 204.8 µg (97.2 - 344.2, ULN*MS = 4.3 ± 2.3). LNSC held equal linear agreement with 24h-UFC*IA (R2 = 0.73, p = 0.000) than with 24h-UFC*MS (R2= 0.74, p =0.000). During treatment with KTZ (>400mg/day), the agreement between LNSC and UFC*IA as well as their ULN, declined but persisted (R2 = 0.46, p = 0.22) however the association was lost with the UFC*MS (R2 = 0.33, p = 0.63). In patients with MTP (>500mg/day), there was also a decrease in the method agreement of LNSC and UFC*IA (R2 = 0.46, p = 0.03) but interestingly, LNSC and UFC*MS held strong linear association (R2 = 0.73, p = 0.000). Hormonal and cardiometabolic associations with the extent of LNSC and 24h-UFC are under analysis.

Conclusion

LNSC and its ULN were comparable to UFC*IA and UFC*MS before treatment and in patients taking MTP, however in those with KTZ, method agreement was lost which could lead to discrepancies in the disease activity classification.

The author has declared no conflict of interest.