This study examined the prevalence of both basal and glucose-stimulated hyperinsulinemia and acanthosis nigricans (AN) as well as the relationship between insulin and androgen levels in hyperandrogenic women. Sixty-two women who had an elevation of 1 or more plasma androgen levels were studied. The results in these women, grouped for analysis on the basis of obesity and ovulatory status, were compared to those in 36 control women of similar ages and weights. The anovulatory hyperandrogenic women had the clinical and biochemical features of the polycystic ovary syndrome (PCO). Oral glucose tolerance tests were performed with measurement of glucose, insulin, sex hormone-binding globulin (SHBG), and total and non-SHBG-bound sex steroid levels. AN was present in 29% of the hyperandrogenic women, the majority of them obese. Fifty percent of obese PCO women had AN, but they did not otherwise differ from PCO women lacking this dermatological change. Only women with PCO had significant hyperinsulinemia independent of obesity, and obese PCO women with AN had the highest serum insulin levels. Plasma glucose values during the oral glucose tolerance test were significantly increased in obese PCO women independent of the presence of AN, and 20% of these women had frank impairment of glucose tolerance. Ovulatory hyperandrogenic women had normal insulin levels and glucose tolerance. Obese and nonobese women had different relationships between sex steroid and insulin levels; obese women had significant correlations between insulin and non-SHBG testosterone levels (r = 0.30; P less than 0.05), whereas nonobese women had significant correlations between insulin and FSH (r = 0.40; P less than 0.01), dehydroepiandrosterone sulfate (r = 0.33; P less than 0.05), and SHBG (r = 0.37; P less than 0.05) levels, suggesting that the mechanisms underlying the association between sex steroid and insulin levels are complex. These findings suggest that 1) only women with PCO have hyperinsulinemia independent of obesity; hyperinsulinemia is not a feature of hyperandrogenic states in general; 2) AN is a common finding in obese hyperandrogenic women, particularly those with PCO; 3) only obese PCO women are at risk for impairment of glucose tolerance, independent of the presence of AN, suggesting that the negative impact of PCO and obesity on insulin action is additive; and 4) PCO women with AN can be considered as a subgroup of PCO and do not appear to have a distinct endocrine disorder.

We randomized 32 cycling female Sprague-Dawley rats (82 days old) into experimental and control groups (16 animals/group). Hyperinsulinemia was induced and maintained for 22 days in the experimental group with NPH human insulin (Novolin, Squibb-Novo, Princeton, NJ) as previously described. Controls received an identical volume of vehicle. Fifteen minutes before death, each rat received a sc injection of 100 ng synthetic GnRH (Factrel, Ayerst Laboratories, New York, NY). The mean serum insulin level was significantly higher in the insulin-treated group than in the control group (165 +/- 57 vs. 49 +/- 9 microU/ml; P less than 0.05). The mean final weight also was significantly higher in the insulin-treated group (283 +/- 4 vs. 242 +/- 7 g; P less than 0.001). There were no significant differences in mean final serum levels of testosterone, estradiol, estrone, or androstenedione or in GnRH-stimulated serum levels of LH or FSH. The androstenedione to estrone ratio, however, was significantly lower in the insulin-treated group (2.5 +/- 0.3 vs. 3.4 +/- 0.2; P less than 0.01), suggesting that aromatase activity increased with hyperinsulinemia. Specific [125I]insulin binding to ovarian tissue homogenates was lower in the insulin-treated group (1.7 +/- 0.1% vs. 2.6 +/- 0.6%/0.2 mg protein; P greater than 0.05), suggesting that ovarian insulin receptors tended to down-regulate with hyperinsulinemia. Specific [125I]insulin-like growth factor I [( 125I]IGF-I) binding to ovarian tissue homogenates, in contrast, was significantly higher in the insulin-treated group (13.3 +/- 1.4% vs. 7.2 +/- 0.6%/0.2 mg protein; P less than 0.05), suggesting that ovarian IGF receptors up-regulated with hyperinsulinemia. The affinity of neither [125I]insulin binding nor that of [125I]IGF-I binding changed significantly, with the 50% inhibition point remaining between 2.0 and 5.0 ng/ml for each peptide in both groups. We conclude that hyperinsulinemia increases ovarian [125I]IGF-I binding and stimulates aromatase activity in the rat. These phenomena, if also true in women, could be important factors contributing to the ovarian hyperstimulation observed in various hyperinsulinemic states.

No abstract