A high prevalence of increased DHEAS (dehydroepiandrosterone sulfate) levels (about a third of cases) has been reported in women with polycystic ovary syndrome (PCOS). This excess of adrenal androgens remains a mystery in this ovarian pathology. It is well known that DHEAS production correlates negatively with age, and study populations of women with PCOS are generally young. To avoid this bias, a study was carried out on a large population of women with PCOS and control women, using normal DHEAS values for each age group, to better assess prevalence and better understand the link between PCOS and DHEAS.

A retrospective cross-sectional study was conducted at the Lille University Hospital. A total of 1223 patients with PCOS according to the Rotterdam criteria and 517 control women were included. DHEAS elevation was diagnosed according to the standards of the Lille University Hospital Institute of Biochemistry and Molecular Biology, based on patient age. The prevalence of increased serum DHEAS levels was calculated in each population and according to PCOS phenotype. Correlations were assessed between serum DHEAS levels and clinical, hormonal, and metabolic markers, with adjustment for age.

Prevalence of increased DHEAS was significantly higher in the PCOS group than in the control group (8.1 vs. 4.3%; OR=1.98 (95%CI: 1.23–3.19), P=0.005, and OR=1.07 (95%CI: 1.05–1.09), P=0.014 without and with adjustment for BMI respectively), and in phenotypes A and C than in controls (OR=2.88 (95%CI: 1.76 to 4.72), P<0.001 and OR=2.81 (95%CI: 1.39 to 5.67), P=0.004 respectively), but not in phenotype D. A correlation was found between DHEAS level and total testosteronemia (r=0.34, P<0.001), androstenedione (r=0.24, P<0.001), 17 hydroxyprogesteronemia (r=0.22, P<0.001) and age (r=0.25, P<0.001). No correlations were found with AMH, LH or FSH, and a very weak positive correlation was found with BMI (r=0.15; P<0.001).

Using age-dependent norms, DHEAS elevation was found in only 8.1% of women with PCOS (11% in the case of phenotypes A and C) versus 4.3% in controls and women with phenotype D. DHEAS levels correlated only with other androgens, and not (or only minimally) with other ovarian, pituitary or metabolic markers. DHEAS assay therefore appears to be of no interest for positive diagnosis or understanding of the pathophysiology of PCOS, except in case of very high testosterone levels.