Urinary glucocorticoid metabolites: biomarkers to classify adrenal incidentalomas?

Summary Objective Total urinary cortisol metabolites represent cortisol production and metabolism. We hypothesized that to assay metabolites could add some information to the one provided by a sole cortisol assay. Design and patients We set up an inexpensive multiplex mass spectrometry assay to quantify cortisol metabolites. We investigated 43 patients with benign secreting ( AT +) or silent ( AT −) adrenal tumours compared to 48 lean ( Nl ) or 143 obese (Ob) subjects, and to 26 patients with a Cushing's disease ( CD ). The initial investigation included immunoreactive quantification of urinary free cortisol ( UFC ). Results Cortisol metabolites were overexcreted in CD but not in Ob subjects. Nl and Ob were thus pooled in a control population (Ctl). Cortisol, tetrahydrocortisol ( THF ) and tetrahydrocortisone ( THE ) excretions were significantly increased in AT compared to Ctl subjects, whereas immunoreactive UFC was similar. A logistic regression retaining cortisol, THF , and α‐ and β‐cortolone as significant analytes allowed the construction of a receiver‐operating characteristics ( ROC ) curve significantly better than the curve generated by cortisol alone (area under the curve ( AUC ) 0·927 vs 0·729, respectively; P  < 0·0001). More importantly, although there was no significant difference between Ctl vs AT− subjects for cortisol metabolites, a logistic regression retaining cortisol, allo‐THF, and α‐ and β‐cortolone as significant analytes generated a ROC curve performing significantly better than cortisol alone ( AUC 0·910 vs 0·635, respectively; P  < 0·0001). Conclusion Cortisol metabolite excretion is modified in AT , including AT −, patients even without modification of UFC . Clinical usefulness of these biomarkers has to be investigated in prospective studies following up patients with AT .