SUN-LB061 Midnight Salivary Cortisol (MiSC): Diagnostic Concordance in a Normal Sample Population Using Different Immunoassays

Abstract

Salivary cortisol (SC) is a measure of free cortisol and follows diurnal rhythm of serum or plasma cortisol, so it is convenient for collection noninvasively on an outpatient basis. The Endocrine Society recommends measuring MiSC concentrations as a first-line test in screening for Cushing's syndrome. To define its utility as a diagnostic test, it is important to know whether assay techniques or thresholds affect the rate of false-positive results. Revision of literature showed that there are marked differences in sensitivity and specificity, and cutoff values of SC assays. The aim of this study was to compare MiSC results obtained by available immunoassays with different analytical sensitivity (AS) in an Argentinian population, and compare by pairs of methods their Diagnostic Concordance (DC) using individual reported cutoffs (RCO). Material and methods: Salivary samples obtained at 11 PM, from 133 adults subjects (18-74 years), with no corticoid therapy (last 6 months), psychiatric medication, depression, thyroid diseases, statins administration, or smoking. MiSC (nmol/L) were measured by ECLIA Cobas Roche (RCO: 4.9; AS: 1.5); RIA Beckman Coulter (RCO: 0.8; AS: 0.8), CLIA Liaison DiaSorin (RCO: 8.0; AS: 4.4), CLIA Access Beckman Coulter (RCO: 8.3; AS: 0.3) and CLIA Immulite 1000 Siemens (modified methodic) (RCO: 5.0 own data not reported; AS: 1.0). Bland & Altman biases were calculated between methods. Results: Calculated median, range and p95 (nmoL/L) for each method were: ECLIA Cobas (n=131): 1.6 (1.5-14.5) 6.4; RIA (n=90): 1.33 (0.8-7.7) 3.6, CLIA Liaison (n=126): 4.4 (8.4-22.1) 8.2, CLIA Access (n= 125): 4.1 (0.3-28) 15.6, and CLIA Immulite1000 (n: 99): 2.5 (1.4-20) 7.6. Calculated p95 was higher than RCO in all methods, being RIA the highest. ANOVA analysis for paired samples measured by all methods showed significant (p<0.005) differences except for Access and Liaison (p=1,000). Biases (nmol/L) obtained were: Cobas vs Immulite = -1; Access vs Cobas = 3.2; Cobas vs Liaison = -2.8; Cobas vs RIA = 0.9; Access vs Immulite = 2.4; Immulite vs Liaison = -1.8; Access vs Liaison = 0.3; Access vs RIA = 4.1; Liaison vs RIA = 3.6; Immulite vs RIA = 1.9. Lowest bias found was between Access and Liaison, result coincident with ANOVA. %DC between pairs of methods using RCOs was: Cobas vs Immulite = 87%; Access vs Cobas= 84%; Cobas vs Liaison = 95%; Cobas vs RIA = 64%; Access vs Immulite = 85%; Immulite vs Liaison = 90%; Access vs Liaison = 84%; Access vs RIA= 72%; Liaison vs RIA = 63%; Immulite vs RIA = 70%. Lowest %DC was found when RIA was involved in the comparison, which could be due to its low AS and RCO that are the same. Conclusions: Differences in %DC observed could be due to not only differences between methods but differences in composition of the population studied. Validating MiSC cutoff values in individual laboratories is important for good clinical use of results in the diagnosis and follow-up of Cushing’s syndrome. Unless otherwise noted, all abstracts...