Influence of various clinical variables and storage conditions on urinary cortisol levels: gas chromatography–mass spectrometry (GC–MS) versus immunoassay
Timo Deutschbein1,2, Martina Broecker-Preuss2, Michaela Hartmann3, Stefan Wudy3, Ricarda Althoff2, Klaus Mann2 & Stephan Petersenn2,4
Introduction: Measurement of urinary cortisol is often used to assess disease activity in patients with suspected or proven hypercortisolism. However, specific reference ranges are lacking for some of the newer assays. This study analyzed upper limits of normal (ULN, mean+2S.D.) for two analytical procedures (GCMS, ECLIA) in relation to various independent variables. Besides, the influence of different storage conditions was investigated (by ECLIA).
Methods: Each ten healthy subjects were grouped by age (1830; 3150; >50), BMI (<25; >25), and sex (60 males, 60 females; age 39.3±1.3; BMI 25.9±0.4). Subjects collected 24 h urines on two (n=120) to three (n=11) separate days. Total urinary cortisol was measured after enzymatic hydrolysis by GCMS and urinary free cortisol (UFC) was measured by ECLIA (Roche). ULN were calculated for each procedure and then applied to 12 patients with histologically confirmed hypercortisolism (four males, eight females; age 53.1±3.1; BMI 29.1±1.8). In order to determine degradation, samples were stored at 4 °C (without light) or 22 °C (with and without light) for 0, 24, and 72 h.
Results: Results of both procedures were significantly correlated (r=0.77, P<0.001). For each procedure, multiple stepwise regression analysis identified sex as the only significant predictor, resulting in sex-dependent ULN (each males versus females): 900 vs 679 nmol/d for GCMS, 344 vs 255 nmol/d for ECLIA. These ULN classified samples from patients as hypercortisolemic in 100% (GCMS), and 96% (ECLIA). Mean coefficients of variation (CV) for three collection periods were 42% (2470%) for ECLIA. Different storage conditions over 72 h did not alter UFC levels significantly.
Conclusion: Results of the two analytical procedures for urinary cortisol were well correlated. ECLIA as well as GCMS allowed excellent identification of hypercortisolic states if assay- and sex-specific ULN were used. Although patients were intensively educated to provide adequate 24 h urines, we observed relevant variation between different collection periods. UFC is stable over 72 h irrespective of the storage conditions applied when measured by ECLIA.