Endocrine Abstracts

Introduction: Gas chromatography tandem mass spectrometry (GC-MS/MS) quantitatively detects a large number of steroids at one time in a single serum sample. Most studies of subclinical hypercortisolism in adrenal incidentalomas have focused on detection of a few steroids like cortisol, DHEA/S, androstenedione, testosterone. We used GC-MS/MS to quantify 91 steroids in patients with adrenal incidentalomas to search for the most reliable marker of subclinical hypercortisolism.

Methods: Steroid profiles consisting of 91 steroids measured by GC-MS/MS from sera of 54 patients with unilateral (n=29) and bilateral (n=25) adrenal incidentalomas were compared. Suspected subclinical hypercortisolism was defined as cortisol >50 nmol/l (1.8 μg/dl) in 1 mg overnight dexamethasone (DXM) test. Sensitivity and specificity was calculated for urinary free cortisol >208 nmol/24 h, ACTH< 10 ng/l (2.2 pmol/l), midnight serum cortisol >149 nmol/l as sole criterion. Logistic regression and ROC curves were used to select steroids that best reflect non suppressibility in 1 mg dexamethasone test and ROC curves with sensitivity and specificity were calculated.

Results: Out of 54 patients 13 had suppressed cortisol in 1 mg DXM test under 50 nmol/l. Remaining 41 (21 unilat., 20 bilat.) had post DXM cortisol ranging from 51 to 381 nmol/l. UFC > 208 nmol/24 h showed 34% sensitivity and 76% specificity. ACTH < 10 ng/l had 17% sensitivity and 100% specificity. Midnight serum cortisol >149 nmol/l had 60% sensitivity and 84% specificity. DHEAS had sensitivity 71% and 85% specificity with cut off value =876 nmol/l, AUC=0,800 and Z-score=3.07. Androsterone sulphate showed sensitivity 76% and specificity 69% with cut off value=539 nmol/l, AUC=0.780 and z-score=3.33. Androstenediol had 78% sensitivity 77% specificity with cut off value=0.69 nmol/l, AUC 0,740, z-score=2.46. Other measured steroids had lower predictive values.

Conclusions: Non-suppressibility in dexamethasone test correlates with the suppression of DHEAS in most cases. Other steroids like androsterone sulphate, androstenediol and other show similar sensitivity and specificity but are difficult to measure in routine conditions. Diagnosis of subclinical hypercortisolism needs to be evaluated by a combination of parameters, which cannot be replaced by a sole criterion.