Novel threshold value of midnight serum cortisol for diagnosis of hypercortisolism using the Roche Cortisol II assay

Background: The diagnosis of hypercortisolism requires multiple biochemical investigations, due to variations in cortisol production during the 24-hour circadian cycle. Midnight serum cortisol is difficult to interpret since the threshold value is dependent on the analytical method used and is often not provided by the manufacturer. Second-generation assays are more specific than first-generation assays and may have lower threshold values. Objectives: The aim of this study was to determine a novel threshold value of midnight serum cortisol for the biochemical diagnosis of hypercortisolism, using the Roche Cobas Cortisol (R) second-generation assay. Methods: This study was performed in adult patients hospitalized in the endocrinology unit of a university hospital. Patients had a complete assessment of their 24-hour cortisol cycle, i.e., a serum cortisol test every four hours and at least two first-line tests: late night salivary cortisol, dexamethasone suppression test and/or 24-hour urinary free cortisol. First-line tests were used to identify patients with hypercortisolism. Serum samples were analyzed by second-generation electrochemiluminescence immunoassays (ECLIA) from Roche Cobas Cortisol (R). Results: Midnight serum cortisol samples were obtained from 175 hospitalized patients. The novel threshold value obtained was 157 nmol/L with a sensitivity of 82.9% (68.6 to 94.3%) and a specificity of 90.0% (85.0 to 95.0%). Conclusion: Our study confirms that the threshold value of midnight serum cortisol is not comparable between first- and second-generation Roche Cobas Cortisol (R) assays and that the threshold value is lower with the secondgeneration assay.

Automated summary

This study aimed to determine a novel threshold value of midnight serum cortisol for the biochemical diagnosis of hypercortisolism, using the Roche Cobas Cortisol(R) second-generation assay. The results showed a novel threshold value of 157 nmol/L, with a sensitivity of 82.9% and a specificity of 90.0%, indicating that the threshold value of midnight serum cortisol differs between first- and second-generation assays.