Performance of LC–MS/MS and immunoassay based 24-h urine free cortisol in the diagnosis of Cushing's syndrome

• Published in: The Journal of steroid biochemistry and molecular biology Vol. 190; pp. 193 - 197
• Main Authors: Oßwald, Andrea, Wang, Rong, Beuschlein, Felix, Hartmann, Michaela F., Wudy, Stefan A., Bidlingmaier, Martin, Zopp, Stephanie, Reincke, Martin, Ritzel, Katrin
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.06.2019; Elsevier BV
• Subjects: Adrenocorticotropic hormone; Chemiluminescence; Cortisol; Cross-reactivity; Cushing syndrome; Cushing’s syndrome; Glucocorticoids; Hormones; Immunoassay; LC–MS/MS; Liquid chromatography; Mass spectroscopy; Metabolites; Nervous system diseases; Patients; Pituitary; Statistical analysis; Urine; Urine free cortisol; Immunoassay; Cushing’s syndrome; LC–MS/MS; Urine free cortisol
• ISSN: 0960-0760; 1879-1220
• DOI: 10.1016/j.jsbmb.2019.04.004

•Immunoassays overestimate urinary free cortisol (UFC) due to interfering steroid metabolites.•UFC measurement with established immunoassays correlate with LC–MS/MS.•Measured by ROC AUC, diagnostic accuracy of immunoassays is comparable to LC–MS/MS. 24-h urine free cortisol (UFC) is an indicator of integrated cortisol secretion and established screening tool for Cushing’s syndrome (CS). Doubts have been raised regarding specificity of immunoassays, and mass spectrometric techniques have been proposed as an alternative. In the present study we compared diagnostic accuracy of UFC measured with LC–MS/MS vs. immunoassay in patient with CS and patients where CS has been excluded. We examined 24-h urine samples from patients with surgically confirmed CS (n = 77; Cushing’s disease (n = 44), ectopic CS (n = 5), adrenal CS (n = 28)) and patients in whom Cushing’s syndrome was excluded (n = 97) by long-term follow up. UFC was first measured by automated chemiluminescence immunoassays (ADVIA Centaur, Siemens; LIAISON, DiaSorin). Aliquots of all samples were also analyzed by liquid chromatography-tandem mass spectrometry (LC–MS/MS). Statistics: Passing-Bablok Regression, Receiver operating characteristic (ROC) analysis with Youden’s index calculation. UFC of CS patients were higher with both immunoassays compared to LC–MS/MS (913 +/- 235 vs. 303 +/- 155 μg/24 h (ADVIA) and 898 +/-216 vs. 399 +/- 196 μg/24 h (LIAISON)). Similarly, UFC were higher with immunoassays than with LC–MS/MS in the control group (223 +/- 10 vs. 23 +/- 2 μg/24 h (ADVIA) and 105 +/- 6 vs. 27 +/- 4 ug/24 h for (LIAISON)). Passing-Bablok regression showed good correlation between LC–MS/MS and ADVIA as well as between LCMS/MS and LIAISON (r = 0.96 and r = 0.99, p < 000.1) but less correlation in controls (r = 0.83 and r = 0.74, respectively, p < 000.1). ROC calculation revealed the highest ROC AUC (0.89) for the LIAISON immunoassay, followed by LC–MS/MS (0.82) and the ADVIA (0.80). In direct comparison, AUCs from LC–MS/MS and immunoassays in the same patient were not statistically different (p < 0,001). Best cut-off concentration to identify patients with CS was 234 μg/24 h (LIAISON), 51 μg/24 h for LC–MS/MS and 330 μg/24 h (ADVIA Centaur). In summary, UFC values were measured substantially higher by both immunoassays compared to LC–MS/MS. This is most likely due to cross-reactivity from interfering glucocorticoid metabolites. Nevertheless, all three methods correlated well. ROC analysis revealed the highest AUC for one of the immunoassays, although differences between the three methods were not significant. Direct comparison with LC–MS/MS indicates that high diagnostic accuracy can be obtained with suitable immunoassays.