Plasma Steroid Profiles in Subclinical Compared With Overt Adrenal Cushing Syndrome

Jimmy Masjkur; Matthias Gruber; Mirko Peitzsch; Denise Kaden; Guido Di Dalmazi; Martin Bidlingmaier; Stephanie Zopp; Katharina Langton; Julia Fazel; Felix Beuschlein; Stefan Richard Bornstein; Martin Reincke; Graeme Eisenhofer


J Clin Endocrinol Metab. 2019;104(10):4331-4340. 

In This Article


Recruitment of Patients

Patients were enrolled in this bicentric, cross-sectional study from outpatients referred to the Department of Endocrinology at the Ludwig-Maximilians-Universität München (Munich, Switzerland) and the University Hospital Dresden (Dresden, Germany). Patients were tested for hypercortisolism because of adrenal masses detected by CT or MRI, as well as for investigation of secondary hypertension. No comparison was made between patients having bilateral adrenal masses with those having unilateral lesions. The test was also performed to screen for Cushing syndrome in subjects who were overweight and obese. Hormonal evaluations were also performed to rule out pheochromocytoma and primary hyperaldosteronism. The final study population was restricted to those with either confirmation or exclusion of disease according to current guidelines.[19] Patients with ACTH-dependent Cushing syndrome were excluded from the analysis.

After an overnight DST, cortisol values (fasting serum cortisol >1.8 μg/dL) of 24-hour UFC (upper cutoff, 75.4 μg/24 h), midnight SFC (>1.5 ng/mL), basal serum cortisol (upper cutoff, 240 μg/L), and ACTH (<10 pg/mL) concentrations were measured to confirm presence of AC or SC. An abnormal DST result as one initial screening test in combination with at least one other abnormal test result of the hypothalamic-pituitary-adrenal axis was used for diagnosis of AC and SC. Overt AC was thereby confirmed in 21 patients. Subjects with no relevant clinical features of Cushing syndrome were classified as having SC (n = 35) if they had biochemical evidence of hypercortisolism or as EX if they had no biochemical evidence of hypercortisolism (n = 152).

Reference Population

The reference control group was matched as closely as possible to combined patient groups according to age and sex. A total of 277 normotensive and hypertensive volunteers were thereby selected from 525 subjects (all in Dresden) recruited initially for establishing specific reference intervals for each of the steroids of the plasma panel.[18] The reference population was included for comparison with patients in whom hypercortisolism was excluded and to aid correction of impacts of differences in age and sex among the three different patient groups (Table 1). All subjects provided written informed consent under protocols approved by the local ethics committee at each center.

Steroid Profiling

The 15 adrenal-steroid panel, comprising aldosterone, cortisol, 11-deoxycortisol, 21-deoxycortisol, corticosterone, 11-deoxycorticosterone, 18-oxocortisol, 18-hydroxycortisol, cortisone, progesterone, 17-hydroxyprogesterone, pregnenolone, androstenedione, dehydroepiandrosterone (DHEA) and DHEA-sulfate (DHEA-S), was assayed by LC-MS/MS, as described elsewhere.[20] After DST, measurements of UFC, SFC, basal plasma cortisol, and ACTH were analyzed within the routine clinical laboratories at both centers.

Blood sampling was performed in the morning between 8:00 AM and 10:00 AM after an overnight fast. Samples were collected in blood tubes containing lithium heparin or EDTA and stored at −80°C until analyses of steroid profiles.

Statistical Analyses

Statistical analyses were carried out with the JMP statistical software package (SAS Institute, Cary, NC). The Wilcoxon and Steel Dwass all-pairs tests were used for nonparametric comparisons of demographic and routine biochemical data involving multiple groups. For parametric multivariate analyses, all data were logarithmically transformed before analyses. Corrections for age and sex in comparisons that involved the reference population involved multivariate calculations of least square means, using age and sex as covariates, with final display of data derived by the exponents of logarithmically transformed data to calculate geometric means and respective SEs. For least square means multivariate comparisons, significance of differences were assessed using the Tukey honest significant difference test. For other multivariate analyses (e.g., discriminant analyses), data were normalized according to age- and sex-specific upper cutoffs of reference intervals established elsewhere.[18]

For distinguishing patients with and without SC, receiver-operating characteristic (ROC) curves were constructed by logistic regression. Selection of a minimal panel of the most useful steroids for diagnosis was established using stepwise regression. Discriminant analysis with stepwise variable selection was used to assess the use of plasma steroid combinations for distinguishing patients with SC from AC and EX groups. Details of the discriminant platform of the JMP statistics software package are available online.[21] Results of steroid profiling were compared with those derived from routine measurements of UFC, SFC, plasma ACTH, and plasma cortisol before and after DST. Descriptions of underlying mathematical and statistical concepts, associated methodological details, and additional statistical analyses are available in an online repository.[22]