MON-172 Comparison of a Late-night Salivary Cortisol Immunoassay and a Cortisol and Cortisone LCMS Assay in Patients with Cushing’s Disease

Late-night salivary cortisol (LNSC) measured by enzyme immunoassay (EIA) is established as a reliable screening test and recommended as a first-line test for Cushing’s syndrome.(1,2) However, liquid chromatography-tandem mass spectrometry (LCMS), validated to measure salivary cortisol (F) and cortisone (E), has been proposed to be superior diagnostically as well as helpful in detecting saliva contaminated with topical hydrocortisone (i.e. cortisol).(3,4) We measured EIA-F, LCMS-F, and LCMS-E in 913 consecutive late-night saliva samples from patients suspected of Cushing’s syndrome. For the current report, we focus only on the patients with Cushing’s disease (CD) who had at least one elevated salivary result using cutoffs established previously [EIA-F ≥3.3 nmol/L, LCMS-F ≥2.8 nmol/L, and LCMS-E ≥8.7 nmol/L].(5) We identified 27 patients who had newly diagnosed CD (n-CD) or had persistent/recurrent CD after pituitary surgery (r-CD). Data are shown as mean (SD). There were 19F/8M patients aged 50 (17) y; their BMI was 35 (11) kg/m(2). 8 patients had n-CD and 19 had r-CD. In all 27 patients, EIA-F was 9.1 (6.0), LCMS-F was 5.1 (4.0), and LCMS-E was 19.6 (12.3) nmol/L. In n-CD, EIA-F was 14.3 (7.7), LCMS-F was 7.7 (4.6), and LCMS-E was 29.4 (12.8) nmol/L. In r-CD, EIA-F was 6.9 (3.6), LCMS-F was 4.0 (3.2), and LCMS-E was 15.5 (10.0) nmol/L. Mean EIA-F was greater than mean LCMS-F in all patients, and mean LNSC in n-CD patients was greater than mean LNSC in r-CD patients. The LCMS-F/E ratio was <0.8 in all subjects indicating that no saliva samples were contaminated with topical hydrocortisone.(3) Furthermore, there was no difference in the LCMS-F/E ratio between n-CD [0.3 (0.1)] and r-CD [0.3 (0.2)] patients. Using previously identified cutoffs, 9 of the 27 CD patients had normal LCMS-F whereas only 1 patient had a normal EIA-F (P=0.011). The EIA-F results of patients with normal LCMS-F were significantly lower than patients with increased LCMS-F demonstrating that LCMS-F was more likely to provide false negative results with milder hypercortisolism. Most of these false-negative LCMS-F results occurred in the r-CD patients (8 out of 19) who had milder hypercortisolism compared to the n-CD patients (P=0.006). No differences in diagnostic performance were found between EIA-F and LCMS-E. In summary, EIA-F appears to be superior to LCMS-F in identifying CD patients with milder hypercortisolism using established diagnostic cutoffs.(5) Neither LCMS-E nor F/E ratio improved the diagnostic sensitivity in newly diagnosed CD or persistent/recurrent CD after pituitary surgery. We suggest that late-night salivary cortisol measured by EIA provides the best sensitivity for CD diagnosis. (1)J Clin Endocrinol Metab 2008;93:1526–1540. (2)Clin Chem 2003;49:203–204. (3)Clin Chem 2012;58:947–948. (4)J Clin Endocrinol Metab 2010;95:4951–4958. (5)J Endocr Soc 2019;3:1631–1640.