Endocrine Abstracts

JOINT2830

Background: Measurement of 24-hour urinary free cortisol (UFC) is recommended for biochemical diagnosis of Cushing’s syndrome (CS). However, measured cortisol concentrations differ depending on the test method. Results of chemiluminescence immunoassays (CLIA), in contrast to liquid chromatography–mass spectrometry (LC–MS/MS), may be impacted by the presence of cortisol metabolites in urine. Poor agreement across CLIA methods is primarily attributable to cross-reactivity and to whether or not an extraction procedure is performed. Accordingly, the Endocrine Society recommends using the upper limit of the method-specific reference interval (RI) as a cut-off in diagnosis of CS. We assessed the performances of a new automated CLIA that does not require extraction, and compared clinical performance to two other CLIAs and a LC–MS/MS method.

Methods: Analytical performance of the IDS Urinary Cortisol CLIA (Immunodiagnostic Systems Ltd, UK) was verified. Clinical performance was studied in 24-hour urine samples from 24 CS patients and 50 patients in which the diagnosis of CS was clinically suspected, but finally excluded. Results obtained by the IDS assay were compared to those from Beckman Access Cortisol CLIA (Beckman Coulter Inc, USA) without extraction, DiaSorin Liaison Cortisol CLIA (DiaSorin Spa, Italy) after extraction with dichloromethane and an in-house LC–MS/MS (CHU of Liege, Belgium). Analytical agreement across methods was estimated using Passing–Bablok regression and diagnostic performances were compared using clinical sensitivity, specificity, and accuracy.

Results: UFC concentrations measured with IDS were lower compared to Beckman (slope: 0.81, intercept: 0.31 μg/dl; R²: 0.96), but higher than DiaSorin (slope: 1.45, intercept: 1.68 μg/dl; R²: 0.75) and LC–MS/MS (slope: 4.09, intercept: 1.44 μg/dl; R²: 0.73). At the RI-based cutoff (μg/24 h: IDS 324, Beckman 403, DiaSorin 83, LC–MS/MS 45), the IDS CLIA results exhibited 92% sensitivity and 100% specificity (Beckman: 83% sensitivity, 98% specificity; DiaSorin: 100% sensitivity, 56% specificity; LC–MS: 96% sensitivity, 82% specificity). At an optimized, method-specific cut-off providing 100% sensitivity (μg/24 h: IDS 306, Beckman 266, DiaSorin 171, LC–MS 44), specificity was 98% with IDS and DiaSorin CLIAs, 84% with Beckman CLIA and 80% with LC–MS/MS.

Conclusions: The new IDS CLIA is a clinically accurate automated method for measuring UFC concentration. As expected, due to the omission of an extraction procedure, the assay measures higher concentrations compared to LC–MS/MS and other CLIA methods. Comparison of 3 immunoassays and an LC–MS/MS method demonstrated acceptable correlation but significant differences in absolute concentrations. Using method-specific cut-offs are used, the new CLIA provides similar or better clinical sensitivity and specificity.