Different methods to estimate serum free cortisol: a comparison during cortisol tetracosactide testing.

BACKGROUND: Serum cortisol is routinely quantified by immunoassays. In intensive care units serum free cortisol (FC) determination has been described as a better indicator of survival than total cortisol (TC). To estimate FC different methods are available including saliva sampling. We compared five methods to estimate FC, before and after an ACTH stimulating test in patients suspected of adrenal insufficiency. METHOD: Serum and saliva was collected from 130 patients from the Endocrine Department of a university hospital before and after tetracosactide injection for TC determination. FC was estimated: after serum ultrafiltration, quadratic (Coolens') or cubic (Dorin's) equations, using TC/cortisol-binding globulin concentrations ratio or using cortisol concentration determination in saliva. RESULTS: FC concentrations obtained by different techniques were significantly correlated and Passing-Bablok regressions showed no deviation from linearity between salFC and filtFC or quadFC. Using the routine assumption that the patients were correctly diagnosed using a post-tetracosactide TC threshold of 550 nmol/L the FC methods generating the best ROC curves were salFC and filtFC or cubFC 30 min after tetracosactide injection. CONCLUSIONS: FC concentrations obtained by different techniques are significantly but not similarly correlated with TC. As, salFC and filtFC are more convenient to perform than methods involving CBG assays and are better correlated to TC during tetracosactide tests they may be preferred as FC surrogate assays.