A deficiency of 21-hydroxylase activity is one of the most commonly inherited genetic disorders in man, with heterozygosity for CYP21 mutations affecting approximately 1 in 60 of the non-Jewish Caucasian population. We have hypothesized that heterozygosity for CYP21 mutations in women increases their risk of developing clinically evident hyperandrogenism, and that this risk is related to the severity of the mutation of CYP21 and/or the 17- hydroxyprogesterone (17-OHP) response to ACTH stimulation. To test these hypotheses, we studied 38 obligate carriers for 21 -hydroxylase deficiency (i.e. mothers of children with congenital adrenal hyperplasia or nonclassic congenital adreanl hyperplasia), comparing them to 27 weight-, parity-, and age-matched controls. Premenopausal carriers, not receiving hormonal treatment (n = 27), had higher mean total and free testosterone [T; 2.02 ± 0.55 vs. 1.56 ± 0.65 nmol/L (P < 0.007) and 0.018 ± 0.007 vs. 0.012 ± 0.006 nmol/L (P < 0.007), respectively] and lower mean sex hormone-binding globulin (214 ± 62 vs. 277 ± 129 nmol/L; P < 0.03) levels compared to controls. There was no difference in the mean basal levels of dehydroepiandrosterone sulfate, androstenedione (A4), or 17-OHP between carriers and controls. As expected, carriers exhibited higher stimulated and net increment 17-OHP levels than controls [21.1 ± 27.1 vs. 6.2 ± 3.1 nmol/L (P < 0.01) and 19.0 ± 26.5 vs. 4.4 ± 2.8 nmol/L (P < 0.009), respectively]. However, no difference was observed in the response of A4 to ACTH-(1-24) stimulation. Of the 27 carriers studied biochemically, 2(7.4%) had a stimulated 17-OHP value between 30.3-60.6 nmol/L, and 1 (3.7%) had a 17-OHP level above 60.6 nmol/L, suggestive of nonclassic adrenal hyperplasia. Of all carriers studied genetically (n = 36), 50.0% (18 of 36) had 1.33% (12 of 36) had 2, and 16.7% (6 of 36) had 3 or more mutations. In 27.8% (10 of 36) of carriers, the mutations were contiguous, consistent with a large gene conversion. All 38 carriers were examined for historical and physical features of hyperandrogenism. Hirsutism was defined as a Ferriman-Gallwey score of 6 or more, menstrual/ovulatory dysfunction as a history of menstrual cycles of more than 35-day, and hyperandrogenemia as total or free T, A4, and/or dehydroepiandrosterone sulfate levels above the upper 95th percentile of control values. Further, defining functional androgen excess (FAE) as the presence of at least 2 of the 3 hyperandrogenic features, 4 of 38 (10.5%) of carriers appeared to be affected (95% confidence interval, 2.9-24.8%). Assuming an expected prevalence rate of FAE in the general population of 5- 20%, the frequency of FAE among our carriers was not significantly higher than expected. In conclusion, heterozygosity for CYP21 mutations does not appear to increase the risk of clinically evident hyperandrogenism, although carrying the defect was associated with higher mean and free T levels. Finally, due to the low frequency of androgen excess in our heterozygote population, we were unable to correlate the severity of the CYP21 mutation and/or the 17-OHP response to ACTH stimulation with the presence of the phenotype.