OBJECTIVE To investigate the basal levels of gonadotrophins and sex steroids, with special reference to the effects of obesity and body fat distribution, In premenopausal women, both those with polycystic ovary syndrome (PCOS) and those with normal ovaries and regular menstrual cycles.
DESIGN Cross-sectional study. The separate effects of obesity (and body fat distribution and fasting Insulin levels) and PCOS on endocrine variables were evaluated by means of analysis of covariance.
PATIENTS Sixty-seven women with anovulatory menstrual cycles and polycystic ovaries according to ultrasonography and 59 women with normal ovaries and regular cycles, both groups covering a wide range of body mass index (BMI, PCOS, 17·6-37·4, mean 25·7 kg/m2; controls, 18·8-40·9, mean 25·1 kg/m2).
MEASUREMENTS Serum levels of gonadotrophins, sex steroid hormones, prolactin and GH obtained in the early follicular phase in the controls, fasting insulin levels, anthropometric measures (BMI, skinfolds, waist hip ratio).
RESULTS Mean serum concentrations of LH, andro-stenedione, testosterone, the free androgen index (FAI; all P < 0·0001) and DHEAS (P < 0·01) were higher, and serum FSH (P < 0·01) and serum SHBG levels lower (P < 0·0001), in the PCOS group than in the controls. Women with PCOS had a more pronounced upper body fat distribution and higher fasting insulin levels than the controls. Independent of PCOS, BMI was positlvely associated with serum levels of FSH (P < 0·001) and negatively with levels of LH (P < 0·05), LH/FSH ratio (P < 0·0001), SHBG (P < 0·0001) and androstenedione (P < 0·01), whereas for levels of testosterone, FAI and DHEAS the impact of obesity differed significantly between the groups. Thus, in the PCOS group, testosterone levels (P < 0·05) and the FAI (P < 0·001) were positively associated with BMI, whereas they were constant throughout the entire range of BMI in the controls. DHEAS levels were positively associated with BMI in the PCOS group (P < 0·05) and negatively in the controls (P < 0·01). Measures of upper body fat were related to testosterone and FAI levels, independent of BMI.
CONCLUSIONS Lower FSH levels were found in women with PCOS than during the early follicular phase of normally ovulating women, suggesting a role in anovulation in PCOS. Obesity itself exerted effects on endocrine variables, with the net result of a reduced LHIFSH ratio and lower serum levels of androstenedione and SHBG in both groups; obesity was associated with increased levels of DHEAS, testosterone and FAI exclusively in the women with PCOS. The results underline the endocrine impact of obesity and body fat distribution and the necessity of applying reference values of BMI matched subjects when establishing the endocrine profile of women with PCOS.