Method-Specific Cortisol and Dexamethasone Thresholds Increase Clinical Specificity of the Dexamethasone Suppression Test for Cushing Syndrome

Abstract

Abstract Background The dexamethasone suppression test (DST) is the recommended first-tier test for suspected Cushing syndrome (CS). Missed dexamethasone intake or insufficient dexamethasone serum exposure may yield false positive results. Quantification of serum dexamethasone in DST samples may therefore improve test performance. Methods Simultaneous quantification of dexamethasone and cortisol by liquid chromatography-tandem mass spectrometry in 400 DST serum samples (100 overt CS, 200 excluded CS, 100 adrenal incidentalomas with (possible) autonomous cortisol secretion, AI-ACS) randomly selected within the indication groups. The 2.5th percentile of dexamethasone in patients with excluded CS was considered the lower limit of normal (LLN). Results Serum dexamethasone varied from undetectable to 20.2 ng/mL with a median of 4.8 ng/mL (95% CI 4.5-5.1 ng/mL). Dexamethasone was undetectable in only 16 patients (4%), suggesting non-compliance. The dexamethasone LLN was 1.8 ng/mL (4.6 nmol/L). Decreased glomerular filtration rate and diabetes mellitus were associated with higher serum dexamethasone concentration, while body mass index, sex, age, nicotine, and oral contraceptives had no significant effect. By excluding the 27 samples with dexamethasone <LLN and applying the method-specific cortisol cutoff of 2.4 µg/dL (66 nmol/L) to samples with suspected CS, the clinical specificity for CS increased from 67.5% to 92.4% while preserving 100% clinical sensitivity. Among 100 AI-ACS samples (defined by immunoassay), 4 samples had dexamethasone <1.8 ng/mL and 14 samples had cortisol <2.4 µg/dL, which excluded autonomous cortisol secretion. Conclusions Quantification of dexamethasone and method-specific cortisol cutoffs in DST samples may reduce the false positive rate and lower the proportion of patients requiring further workup.