Effects on Gender Identity of Prenatal Androgens and Genital Appearance: Evidence From Girls With Congenital Adrenal Hyperplasia

Abstract

Females with congenital adrenal hyperplasia (CAH) resulting from 21-hydroxylase deficiency are relatively common among those with intersex conditions. Prenatal androgen exposure is independent of rearing gender, and it is possible to assess variations in gender identity in relation to genital appearance and genital surgery. For these reasons, the authors determined gender identity, using a 9-item interview, in 43 girls with classic CAH who ranged in age from 3-18 years. Seven girls identified as tomboys and 29 control girls also were evaluated. The latter included 25 unaffected sisters and cousins of girls and boys with CAH as well as 4 sisters of tomboys. Each of the interview items was scored for sex-typicality. Girls with CAH, on average, had gender scores intermediate between those of tomboys and control girls. Only 5 of the 43, however, had scores outside the control range. The difference in scores between girls with CAH and control subjects was statistically significant and could not be ascribed to outliers. Tomboys also had significantly higher scores than control girls; 3 of 7 were outside the control range. There was significantly more variation in gender identity in tomboys than in girls with CAH or control subjects. Gender identity did not correlate significantly with age. The CAH and control groups were similar in some items of gender identify such as discomfort with being a girl and the wish to be a boy. In the girls with CAH, there was no evident connection between gender identity and either the degree of genital virilization or the age at which genital reconstructive surgery was carried out. Moderate androgen excess early in development, in the form of CAH, clearly increases the risk of atypical gender identity, but this risk could not be predicted from genital virilization in the present study. These females are sex-atypical in some respects but not in others. It does not seem likely that gender role behaviors can predict gender identity or sexual orientation in this population. In addition, behavior cannot be predicted from any single parameter such as the degree of androgen exposure, the appearance of the genitals, or the child-rearing environment.