Endocrine Abstracts

JOINT3758

Introduction: The cortisol concentration varies depending on the measurement method. Monoclonal immunoassays demonstrate greater specificity than their predecessor polyclonal immunoassays for the assessment of cortisol. This necessitates assay-specific cut-off limits for defining adrenal insufficiency (AI). A cortisol cut-off of 500 nmol/l has been recommended for the older cortisol immunoassay, while lower cut-off of 420 nmol/l for newer immunoassays. However these values are defined using the ACTH test and not specified for the insulin tolerance test (ITT).

Purpose: To conduct a direct comparison between the 1st and 2nd generation Roche Elecsys cortisol immunoassays and to establish test- and assay-specific cortisol cut-off limits for the ITT using a cohort of healthy participants.

Methods: Ninety healthy participants underwent an ITT after an overnight fast (60 men, median age 46 years (IQR 24.5–49.5)). Blood samples were collected before, 15, 30, 45, 60, 75 and 90 min after insulin injection. The samples were analyzed on the 1st and 2nd generation Roche Elecsys cortisol immunoassays. The cut-off limit for a normal cortisol response to the ITT was defined as the 2.5th percentile of the peak cortisol level for 1st and 2nd generation Roche Elecsys cortisol immunoassays, respectively. The percentage of participants failing the currently used cut-offs derived from the 250 μg ACTH test of 500 nmol/l (ElecsysCortisol I), and 420 nmol (ElecsysCortisol II) were calculated.

Results: ElecsysCortisol II measured peak cortisol concentrations on average 29 nmol/l (95% CI: −64 to 9 nmol/l) lower than ElecsysCortisol I. The median peak cortisol was 607 nmol/l (range 320–1020 nmol/l) with ElecsysCortisol I and 510 nmol/l (range 240–964 nmol/l) with ElecsysCortisol II. The 2.5th percentile cortisol cut-off limits for a normal cortisol response in the ITT were 379 nmol/l with ElecsysCortisol I and 312 nmol/l with ElecsysCortisol II. Sixteen (17.8%) of the healthy controls had a peak cortisol below 500 nmol/l when measured by ElecsysCortisol I, whereas 19 (21%) of the healthy controls had a peak cortisol below 420 nmol/l with ElecsysCortisol II.

Conclusion: The 2nd generation Roche Cortisol immunoassay systematically reported lower cortisol concentrations. Based on healthy controls, we defined a cut-off limit for the ITT of 312 nmol/l which is considerably lower than commonly applied cut-off limits to confirm or reject AI. Our data emphasize the importance of knowledge of the method used, especially in patients with low apriory likelihood of adrenal insufficiency where the diagnosis may be challenging and at risk of false positive results.