Androgen and psychosexual development: core gender identity, sexual orientation, and recalled childhood gender role behavior in women and men with Congenital Adrenal Hyperplasia

Journal of Sex Research, Feb, 2004 by Melissa Hines, Charles Brook, Gerard S. Conway

Human psychosexual development involves three primary components: core gender identity, the sense of self as male or female; sexual orientation, erotic interest in individuals of the same or the other sex; and gender role behaviors, the myriad characteristics that are associated with being male or female or that differ on average for males and females (Green, 1974; Hines, 2004). Determinants of these components are thought to be both biological (genetic and hormonal) and social-cognitive (involving reinforcement and modelling of gender typical behavior; Hines, 2004). This paper focuses on the role of androgens, the primary hormones produced by the testes, in psychosexual development.

The hypothesis that androgens influence human psychosexual development is based on evidence that androgens influence sex-related development in other mammals. During critical periods of early life, high levels of androgen promote male-typical neural and behavioral characteristics, whereas low levels produce female-typical characteristics (see De Vries & Simerly, 2002; Goy & McEwen, 1980, for reviews). These hormonal effects have been demonstrated most clearly for reproductive behaviors, such as the mounting of female animals by males and the receptive lordosis posture adopted by females in response to these mounts. For example, female rodents and nonhuman primates who are treated with testosterone or other androgens during early development show increased male-typical behavior (mounting) and decreased female-typical behavior (lordosis) in adulthood. Similarly, male animals denied testosterone during early life show decreased mounting and increased lordosis as adults. Early manipulations of androgens also influence other behaviors that show sex differences. For instance, treating female rats or rhesus monkeys with testosterone during early development increases subsequent rough-and-tumble play, a behavior that normally is more common in males than in females (Goy & McEwen, 1980; Meaney & Stewart, 1981).

Hormonal influences on human development have been harder to document than those on other species, because true experiments in which participants are randomly assigned to be treated with hormones or placebo are generally not possible. However, some information has come from studies of individuals who developed in unusual hormone environments for other reasons--for example, because of genetic disorders. Information also has come from studies relating normal variability in the early hormone environment to subsequent variability in behavior.

The strongest evidence of hormonal influences on human behavioral development has come from studies of childhood play. Girls with the genetic disorder congenital adrenal hyperplasia (CAH) are exposed to high levels of androgen prenatally (Pang et al., 1980; Wudy, Dorr, Solleder, Djalali, & Homoki, 1999). This androgen exposure causes partial to complete masculinization of the external genitalia in utero, and the masculinized genitalia in the newborn female typically lead to diagnosis within days of birth. Postnatally, androgen levels are regulated by hormone treatment, and the external genitalia are surgically feminized, usually in infancy. Despite postnatal treatment, girls with CAH show altered play behavior (see Hines, 2002, 2004, for reviews). They are more likely than other girls to prefer toys that are normally preferred by boys (e.g., cars) and less likely to prefer toys that are normally preferred by girls (e.g., dolls). They also show increased preferences for boys as playmates and for boy-typical activities. These differences in play behavior are seen on questionnaires, in interviews, and in direct observation of toy choices. They also are seen when girls with CAH are compared to unaffected female relatives, as well as to controls matched for background factors like age and parental socioeconomic status. Similar outcomes have been seen for girls exposed to high levels of androgenic hormones prenatally because their mothers were prescribed hormones during pregnancy (Ehrhardt & Money, 1967). In addition, normal variability in maternal testosterone levels during pregnancy has been found to relate positively to male-typical play behavior in female offspring at the age of 3 1/2 years (Hines et al., 2002a).

Androgen levels prenatally also may influence sexual orientation and core gender identity, although fewer studies have been conducted in this area. Women with CAH have been found to be more likely to say they are bisexual or homosexual than female controls with other endocrine disorders (Money, Schwartz, & Lewis, 1984), to score higher on a measure of homosexual interest and lower on a measure of heterosexual interest than their unaffected sisters (Dittmann, Kappes, & Kappes, 1992), and to show less heterosexual interest as well as less sexual interest in general than unaffected female relatives (Zucker et al., 1996). Two studies have not reported reduced heterosexuality in women with CAH (Kuhnle & Bullinger, 1997; Lev-Ran, 1974). However, the author of one of the studies (Lev-Ran, 1974) suggested that social prohibitions in the then Soviet Union, where the study was conducted, may have inhibited participants from discussing their sexuality frankly. The methodology of the second study (Kuhnle & Bullinger, 1997) may have limited its power to detect relationships between androgen and sexual orientation. It included a substantial number of women with late-onset CAH, whose androgen elevation would have begun after the presumed critical prenatal period for hormonal influences, and it classified women as either heterosexual or homosexual based solely on whether they were currently living with a female partner.