Method-specific serum cortisol responses to the adrenocorticotrophin test: comparison of gas chromatography-mass spectrometry and five automated immunoassays

Objective: The serum cortisol response to the ACTH test is known to vary significantly by assay but lower reference limits (LRL) for this response have not been established by the reference gas chromatography-mass spectrometry (GC-MS) method or modern immunoassays. We aimed to compare the normal cortisol response to ACTH stimulation using GC-MS with five widely used immunoassays. Design, Patients and Measurements: An ACTH test (250 micrograms iv ACTH1-24) was undertaken in 165 healthy volunteers (age 20-66yrs; 105 female, 24 of whom were taking an oestrogencontaining oral contraceptive pill [OCP]). Serum cortisol was measured using GC-MS, Advia Centaur (Siemens), Architect (Abbott), Modular Analytics E170 (Roche), Immulite 2000 (Siemens) and Access (Beckman) automated immunoassays. The estimated LRL for the 30min cortisol response to ACTH was derived from the 2.5th percentile of log-transformed concentrations. Results: The GC-MS-measured cortisol response was normally distributed in males but not females, with no significant gender difference in baseline or post-ACTH cortisol concentration. Immunoassays were positively biased relative to GC-MS, except in samples from women on the OCP, who showed a consistent negative bias. The LRL for cortisol was method-specific (GC-MS: 420 nmol/l; Architect: 430 nmol/l; Centaur: 446 nmol/l; Access 459 nmol/l; Immulite (2000) 474 nmol/l) and, for the E170, also gender-specific (Female: 524 nmol/l; Male 574 nmol/l). A separate LRL is necessary for women on the OCP. Conclusions: Normal cortisol responses to the ACTH test are influenced significantly by assay and oestrogen treatment. We recommend the use of separate reference limits in premenopausal women on the OCP, and warn users that cortisol measurements in this subgroup are subject to assay interference