This retrospective study was carried out to investigate, in a large group of hyperandrogenized women with polycystic ovary syndrome (PCOS) and nonhyperandrogenized control women, the interrelationships between sex steroids and indices of body fat distribution. Moreover, we investigated the relationships between these parameters and insulin blood levels, since obese women with abdominal pattern of fat distribution (A-BFD), as well as those with PCOS (either obese and nonobese) are characterized by moderate to severe hyperinsulinemia. A sample of 100 women with PCOS and that of 138 women without clinical signs of hyperandrogenism, who served as a control group, were investigated. The waist to hip circumference ratio (WHR) which was used to define different patterns of fat topography was significantly (p < 0.05) higher in PCOS (0.84 ± 0.10) than in control women (0.81 ± 0.08). In both groups, women with WHR values lower than or equal to 0.85 were considered as having a peripheral pattern of body fat distribution (P-BFD) whereas those having WHR values higher than 0.85 had A-BFD. Compared to controls, women with PCOS had higher LH, androgen and estrogen concentrations. In both PCOS and controls there were no differences in sex hormone levels between women with different patterns of fat distribution, except androstenedione, which levels were significantly higher in women with A-BFD than with P-BFD. Women with PCOS showed significantly higher insulin levels than controls. Moreover, in both groups fasting and stimulated insulin were significantly higher in women with A-BFD than in those with P-BFD. Significant simple correlation coefficients were found between WHR and androstenedione in PCOS only, and between WHR and insulin in both PCOS and controls. In multiple regression analysis, in which age and body mass index were included as independent variables, a weak but significant correlation between WHR and androstenedione values was found both in PCOS and controls separately, as well as in the whole sample of women considered together. On the contrary, correlation coefficients between WHR and insulin were found in PCOS and controls considered separately and in the whole sample of women, even when the independent effects of androgens had been taken into account in the analysis. These results are consistent with the hypothesis that the degree of androgenization is only weakly correlated with A-BFD in premenopausal women. Hyperinsulinemia, on the contrary, appears to be more consistently correlated with A-BFD, regardless of the presence of PCOS. Whether these associations may have a causal relationship with different patterns of fat distribution remains to be further elucidated.