Abstract
INTRODUCTION: Cortisol concentration is often estimated by competitive chemiluminescent immunoassays, which are prone to interference due to cross-reactivity. The extent of interference is inconsistent across different adrenal disorders and among different analyzers. Although liquid chromatography-mass spectrometry (LC-MS/MS) possesses better analytical specificity, it is not widely available. We aimed to compare cortisol values measured across four different analyzers across different adrenal disorders. METHODS: Cortisol concentrations were measured in the serum of patients with congenital adrenal hyperplasia (CAH) (n = 12), primary adrenal insufficiency (PAI) (n = 11), endogenous Cushing's syndrome (CS) (n = 6), adrenal incidentaloma (AI) (n = 8), and healthy subjects (HS) (n = 10) in the following analyzers: Advia Centaur XP (Siemens) (SACXP), Immulite1000 (Siemens) (SI1000), Cobas e411 (Roche) (RCe411), and Architect (Abbott)(AA). RESULTS: In CAH patients, a poor agreement was observed between SAC XP [median (IQR) 14.6 (4.7) μg/dL] and RC e411 [median (IQR) 4.6 (3.9) μg/dL] [ICC: -0.016, (-0.55, 0.55)]. The correlation was also poor between SAC XP and SI 1000 [ICC: 0.00, (0.558, 0.551)] as well as between SAC XP and AA [ICC: 0.089, (-0.488, 0.612)]. The agreement was good between RCe411, SI 1000 [median (IQR) 9.6 (3) μg/dL], and AA [median (IQR) 5.1 (3.4) μg/dL] platforms in the same group. Measured cortisol values correlated well across all analyzers in PAI, CS, AI, and HS. CONCLUSIONS: Cortisol concentration demonstrated variable agreement among different analyzers in different adrenal disorders. In CAH, cortisol values measured by SAC XP poorly correlate with RC e411, SI 1000, and AA platforms. Reassessment in another analyzer therefore would be prudent to avoid devastating consequences of unrecognized hypocortisolism in this subset of patients. A good concordance was observed among platforms in other conditions.