METABOLIC FEATURES OF POLYCYSTIC-OVARY-SYNDROME ARE FOUND IN ADOLESCENT GIRLS WITH HYPERANDROGENISM

Recently, we reported that hyperandrogenism in adolescent girls is accompanied by augmented LH pulsatility and elevated LH/FSH ratio with increased ovarian volume. Together with higher concentrations of 17-hydroxyprogesterone, androstenedione, testosterone, and estrone that are ovarian in origin, these neuroendocrine features are identical to those seen in adult women with polycystic ovary syndrome. In the present study, we report the metabolic characteristics of these hyperandrogenic adolescent girls. The GR insulin-like growth factor I(IGF-I)-binding protein (BP)-3 axis, insulin sensitivity, and insulin-IGFBP-1/insulin sex hormone binding globulin axes were evaluated in 13 adolescent girls (ages 11-18 yr) with mild to moderate signs of hyperandrogenism (HA) and 28 age-matched normal girls. Insulin sensitivity was assessed by a frequent-sample iv glucose tolerance test (ivGTT, 0.3 g/kg). Twenty-four hour blood samples were obtained at 10-min intervals and were used to determine GH pulsatility (20-min samples), IGFBP-3 levels (0800-0900 h), and fluctuations of insulin, IGFBP-1, and IGF-I (hourly samples) during feeding and fasting phases of the day. In addition, GH responses to GHRH stimulation (1 mu g/kg) were assessed. Pasting insulin concentrations, but not plasma glucose levels, were significantly elevated in the KA group compared with those in the normal group (256 +/- 35 vs. 103 +/- 24 pmol/L, P = 0.0008), as were insulin responses to ivGTT and meals (P < 0.01) and 24-h mean insulin concentrations (P < 0.01). Thus, hyperinsulinemia with normal fasting glucose levels in HA girls may reflect insulin resistance, as suggested by the increased ratio of insulin and glucose (P < 0.001). All measures of insulin were correwith body mass index (BMI); however, insulin remained significantly higher in the HA group after correcting for BMI, suggesting that decreased insulin sensitivity was related to other factors in addition to BMI. Twenty-four hour IGFBP-1 concentrations showed a diurnal pattern with an inverse relationship to insulin, and 24-h mean concentrations were lower in the HA group (0.35 +/- 0.13 vs. 0.76 +/- 0.09 mu g/L, P = 0.02). Reduced sex hormone binding globulin levels were also inversely related to insulin levels (P = 0.0007). In contrast, GH pulsatile characteristics and IGF-I/IGFBP-3 levels, as well as GH responses to GHRH, were similar between the groups. The activity of the GH-IGF-I axis in HA girls exhibited age- and BMI-dependent declines identical to those seen in,normal adolescent girls. Ovarian volume was larger in the HA group (P < 0.0001) and correlated independently with both 24-h mean concentrations of LH (P = 0.0002) and fasting insulin (P = 0.002). Basal serum concentrations of 17-hydroxyprogesterone, androstenedione, testosterone, and estrone were positively correlated with LH levels; however, an independent influence of insulin on steroid levels was unmasked during suppression of LH by the GnRH antagonist Nal-Glu. In conclusion, intrinsic abnormalities of the GH-IGF-IGFBP-3 axis were not found in adolscent girls with hyperandrogenism. The salient metabolic feature identified is hyperinsulinemia, which in concert with elevated LH levels and reduced IGFBP-1 and sex hormone binding globulin may contribute to the development of manifestations of polycystic ovary syndrome, especially increased ovarian volume, and ovarian hyperandrogenism. These observations reaffirm the peripubertal onset of this syndrome.