New York State Psychiatric Institute, Unit 15, and Dept. Psychiatry, Columbia University, New York, New York 10032-2695
Address correspondence and requests for reprints to: Heino F. L. Meyer-Bahlburg, Department of Psychiatry, Columbia University College of Physicians and Surgeons, New York State Psychiatric Institute, Unit 15, 1051 Riverside Drive, New York, New York 10032.
THE current intense debate of the
management of patients with intersexuality and related conditions
focuses on three major issues: 1) the assignment of gender; 2) the
indication for genital surgery; and 3) the disclosure of medical
information to the patient. For instance, in recently published
guidelines for intersex management, Diamond and Sigmundson (1, 2)
emphasized the growing evidence of prenatal androgen effects on the
sexual differentiation of the brain and made strong recommendations for
assigning to the male gender all 46,XY patients of the following
categories: genital trauma (leading to loss of the penis), micropenis,
androgen insensitivity (AIS) of Quigley et al. (3) stages 2
and 3, hypospadias, 5-reductase deficiency (5
-RD), and
17ß-hydroxysteroid dehydrogenase (HSD) deficiency. The Intersex
Society of North America (4), a support group of mostly adult persons
with intersexuality founded in 1994, has published recommendations that
center on the avoidance of genital surgerybecause of the potential
consequences for sexual functioningwithout the patients informed
consent, "unless it is absolutely necessary for the physical health
and comfort of the intersexual child," and declare it "imperative
that intersexuals learn of their status in a properly supportive
emotional environment." Similarly, a growing number of medical
clinicians have begun to question prevailing practices of intersex
management (e.g., Refs. 5, 6, 7, 8). To evaluate such
recommendations, we have to place them in historical context and review
their empirical basis. The present comments are mostly confined to the
gender question. The discussion pertains to both classical intersex
patients born with ambiguous genitalia and to female-assigned 46,XY
individuals with such nonhormonal genital abnormalities as penile
agenesis, cloacal exstrophy of the bladder, and postnatal penile
ablatio or traumatic loss of the penis, which may occur, for instance,
after a circumcision accident.
Historically, the decision on gender assignment in cases of patients with ambiguous genitalia assumed a two-gender heterosexual society in which reproduction was essential for the survival of most individuals, and sex assignment was based on the predominant sexoriginally in terms of the appearance of the external genitalia, from the second half of the 19th century on in terms of the gonads and their histology, and from the 1950s on often also in terms of the sex chromosomes. Around the middle of the 20th century, reviews of the existing literature (9) and studies of intersex patients at Johns Hopkins Hospital (Baltimore, MD) by Money et al. (10) led to the conclusion that, in most cases, the gender assigned in infancy will be the one the patient stays with into adulthood, regardless of the status of the standard biological indicators of sex. As a consequence, the Hopkins group (11) replaced the then prevalent "true-sex policy" by an "optimal-gender policy" (12), which bases the assignment and reassignment of gender on the expected optimal outcome in terms of psychosexual, reproductive, and overall psychologic/psychiatric functioning. Early surgery for maximizing the gender-appropriate appearance of the external genitalia was recommended to facilitate gender-appropriate rearing. This policy was later extended to 46,XY cases with nonhormonal genital abnormalities (e.g., Ref. 13).
In the original formulation of the optimal-gender policy, consideration of the sexual differentiation of the brain did not play a role because nothing was known about it at that time, and the intersex data then available were interpreted as showing that "psychologically, sexuality is undifferentiated at birth and that it becomes differentiated as masculine or feminine in the course of various experiences of growing up" (10), which is often referred to by others (e.g., Ref. 1) as the hypothesis of psychosexual neutrality at birth. In 1959, however, Phoenix et al. (14) demonstrated that perinatal sex hormones had a major organizing effect on the sexual differentiation of brain and behavior in guinea pigs, and these observations were soon replicated in other species. In the late1960s, Ehrhardt et al. (15, 16, 17) reported gender-role behavior shifts in the direction of masculinization in prenatally androgen-exposed 46,XX girls and women and demonstrated thereby that the pre-/perinatal sex hormone theory applied to the human species, at least to some extent. It is important to note, however, that also in the presence of markedly masculinized gender role behavior, gender identity usually remained female, even in the extreme situation of late-treated women with congenital adrenal hyperplasia who, in addition to prenatal masculinization, also showed massive postnatal virilization (16). Apparently, a given gender identity can accommodate wide variations in gender role behavior, a conclusion that should resolve at least some of the interpretive discrepancies between gender-role-behavior-based criticisms of the optimal-gender policy by biological determinists and gender-identity-focused statements by the policys defenders.
In 1974, Imperato-McGinley et al. (18) reported on a newly
identified syndrome of 46,XY patients with 5-RD; their external
genitalia looked female at birth and they were so assigned, but they
markedly virilized in adolescence. A large percentage of such patients
from an inbred kindred in a medically underserved country initiated
gender change from female to male in later adolescence and young
adulthood, without any medical intervention. This finding was
interpreted as showing that it was the combination of prenatal and
pubertal androgen effects on the brain that determined gender identity
(19). Similar observations were reported by Rösler and Kohn (20)
on the syndrome of 17ß-HSD deficiency. Critics (e.g., Ref.
21) of the conclusions by Imperato-McGinley et al.
questioned, however, whether direct sex-hormone effects on the brain
could be the sole or even the major mechanism underlying this late
gender change; it seemed likely that other psychosocial and
psychological mechanisms also contributed.
Thus, the key issues in the gender-assignment debate are the mechanisms involved in the development of gender identity. To what degree do prenatal hormones predispose an individual for a certain gender identity, and to what extent do postnatal social and psychological factors shape its development? Research on animals, especially lower mammals, has amassed an enormous body of evidence on the role of pre- and perinatal hormones in the sexual differentiation of brain and behavior, but even on the level of rodents social experiences seem to contribute. Also, there are wide differences between species in the specific hormonal mechanisms involved. Rodents, on which most of Diamonds (1) animal research was conducted, are not a good neuroendocrine model for nonhuman primates. Nonhuman primates, in turn, are not a good model for the development of human gender identity because we can not determine an analog of gender identity in primatesquite apart from the fact that primate species vary tremendously in the degree to which the sexes differ behaviorally and in the underlying developmental mechanisms. Human research has made it likely that prenatal sex hormones influence, but do not by themselves fully determine, the development of gender-role behavior (22), and there is a large body of evidence in support of marked effects of social and psychological factors on normal human gender development (23). It is the evidence from long-term follow-up of intersex patients themselves, however, that will be the final arbiter of the adequacy of a given management policy, and this evidence is extremely limited, especially in the case of male pseudohermaphroditism.
A few years ago, Diamond (1) began collecting scattered reports of cases of 46,XY patients who had been raised female and who later, on their own initiative, changed gender to male. Unfortunately, those cases were often not documented in sufficient psychological detail so that it is difficult to evaluate with at least some plausibility to what extent biological and/or social factors may have contributed to the gender change. It was the case known as "John/Joan" (24) that received the greatest public attention; it is also the case with the most data available. This child had been born a normal male and was so assigned ("John"). A circumcision accident at 7 months of age led to loss of the penis, a decision to reassign the child to the female gender ("Joan") at 17 months, and to perform feminizing surgery on her external genitalia at 21 months. Interview-based follow-up studies during childhood by Money and Ehrhardt (13, 25, 26) showed a reasonably well adjusted, although somewhat tomboyish, girl; but in adolescence, years after the family had stopped coming to Johns Hopkins Hospital, Joan had herself re-reassigned as John (who is reported to live now, in his mid-30s, as a married man with his wife and adopted children) (24).
The John/Joan case has been so widely publicized that it has received an undue weight in the discussion of the thorny issue of gender assignment. In fact, this case is not particularly well suited to help the scientific side of the debate. It is only a single case of a rare condition, and the gender reassignment occurred so late that it is difficult to discount the contributions of social influences on the outcome. What we can conclude about this case is: 1) in all likelihood, this infant was a biologically normal male, with presumably normal effects of prenatal and early postnatal sex hormones on the sexual differentiation of the genitalia and, as far as it applies to humans, of the brain; 2) on the social side, the best estimate is that the parents at birth accepted this baby as a male and began raising him as a boy, along with his twin brother. Thus, for 7 months, the parents created for both twins a gender-typed social niche (in the ecological sense); 3) then, at 7 months of age, the penile trauma occurred. It took the family another 10 months, with numerous consultations, to come to a gender-reassignment decision. That is, until 17 months, John continued to live in his male role. After that, it took another 4 months until the feminizing surgery including castration was performed. Even if one were to accept the popular notion of the first 18 months of life as a safe window for gender-reassignment, this case straddles the border. However, as we now know from extensive child development research (23), the first 18 months of life are not free of social influences. During that time, gender-typing is pervasive in most families. Children begin to categorize people in their environment by gender at around 9 months of age (27), which is likely to be associated in some fashion with the development of a rudimentary sense of self. Thus, the concept of the 18-month window for a physician-imposed gender reassignment may be overoptimistic.
There were additional problems in the John/Joan case. Even after surgery, the position of the urethral meatus was so high that Joan, when urinating in the sitting position, needed to direct the urinary stream downward with her fingers; that is, urination in a boys standing position was a more reasonable procedure for Joan than for a girl with typical female genitalia. Note that John/Joan was operated after s/he had learned to walk. In addition, John/Joan had a twin to compare to and to compete with; John was even reported to have been the "dominant" twin all of which may have made it more difficult for Joan to accept a female role. Finally, we have to ask how typical the parents feminine sex-typing of Joan really was, after they had maintained a male niche for John over at least 17 months. The mother seems to have tried hard to adapt to the gender reassignment of her child, but she also needed psychotherapy to help her cope. We know much less about the fathers role, and some statements in the reports by Money (13, 25) makes one wonder to what extent the father was really part of the gender-reassignment effort.
Would John/Joan have developed differently if the accident or the gender-reassignment had happened earlier in his/her life? We do not know, of course. It is noteworthy, however, that a second case with the same condition, who was reassigned to female at 7 months of age (that is, much earlier in development than John/Joan) developed into a bisexual woman without gender dysphoria (28). Thus, it seems that for the only two North American cases with this condition on whom follow-up studies have been published, the outcome differed. That is remarkable, given that male infants with penile ablatio who experience normal prenatal plus early postnatal testosterone exposure before they are reassigned to female and undergo surgical castration, represent the extreme category of early androgenization within the intersex spectrum. Thus, even in this extreme condition, biology does not seem to dictate the outcome and we cannot make a prognosis with certainty. To move forward in this difficult area, we must carefully distinguish between conclusions for which we really have good evidence and statements based on interpolation. Recently, a third such 46,XY case of penile ablatio, followed by female reassignment and later patient-initiated change back to the male gender in adolescence was published (6), but the study lacks important details for full evaluation. Thus, at this stage of our knowledge, we can not state with any certainty which mechanisms underly the patient-initiated gender re-reassignment decisions in such cases. We even do not know whether the three reported cases are representative of what usually happens in female reassignment of infant males with ablatio penis. Nor can we conclude on the basis of just three cases in which gender the majority of such patients will experience a better psychological and sexual adjustment in the long run. One hopes, of course, that advances in techniques of microsurgery-supported penile reconstruction will make decisions on gender-reassignment in such cases obsolete.
The situation is not any better for patients with penile agenesis where only one long-term follow-up case has been published (29): assigned as female at birth, reassigned as male at 4 months, re-reassigned as female and castrated during the second half of the 4th year of life, self-initiated gender change to male in late adolescence. An ongoing follow-up study of female-raised 46,XY patients with cloacal exstrophy of the bladder has identified a few young patients who are opting for gender change to male, whereas others live as females (Ref. 30 and unpublished data).
The gender change accounts on 5-RD (19, 31) and 17ß-HSD (20) are
mostly limited to underdeveloped countries, where the dramatic
virilizing changes of puberty in the absence of any medical
intervention make life as a female very difficult, and where the male
role confers a clear social advantage. But, we urgently need
information about the usual life course of such patients in
industrialized societies. Long-term follow-up of 46,XY patients with
micropenis have been described in several studies (e.g.,
Refs. 5, 32, 33, 34), both of male-raised and female-raised patients. It
is certainly remarkable how good some of the male-raised patients have
managed their sex life, but there are contrary reports as well, and in
the absence of larger samples and a resolution to the question of
selection bias a definitive answer on the management policy is still
outstanding. Patients with complete AIS are usually assigned female
(35, 36), and no such case has been reported who initiated a gender
change to male. Long-term outcome data on patients with partial AIS are
scanty and often poor in psychological detail (3), except for a patient
series from Money and Ogunro (37). If raised male, some patients seem
to be behaviorally undermasculinized. If raised female, some seem
behaviorally overmasculinized, and at least one such case has been
reported to initiate gender change to male (38). The majority seem to
stay in their assigned gender.
In his later follow-up studies of the long-term outcome of 46,XY pseudohermaphrodites at Johns Hopkins Hospital, Money et al. (39, 40) confirmed that the majority of such intersex patients remained with their assigned gender, although somewhat more 46,XY patients raised female changed from female to male than 46,XY patients raised male changed from male to female. The only predictor of "gender transposition," denoting gender-identity change and/or the development of a homosexual orientation, that could be identified was protracted ambiguity of the external genitalia. The implicit finding of a gender-assignment-confirming function of genital surgery, to date tested only in this one study, contrasts with the recommendation of The Intersex Society of North America of avoiding surgery for other than medical reasons. Finally, a new follow-up study (41) of 58 adult male pseudohermaphrodites from the Johns Hopkins clinic by way of a written semistructured self-report again showed that the majority of both women and men were satisfied with the gender to which they had been assigned and also with their sexual functioning.
In summary, the evidence available to date permits only tentative policy-relevant conclusions: 1) the organizational effects of prenatal androgens are more noticeable in gender-role behavior than in gender identity; 2) gender identity can develop as female or male over wide variations of gender role behavior; 3) the majority of 46,XY intersex patients seem to develop an identity commensurate with the assigned gender and do not change their gender later; 4) childhood gender identity will in most cases continue into adolescence and adulthood, but patient-initiated gender change in intersex patients seems to happen more often in adolescence and adulthood than in childhood. Thus, long-term follow-up into mid-adulthood is essential if one wants to arrive at definitive conclusions; 5) more female-assigned 46,XY patients initiate gender change to male than male-assigned 46,XY patients initiate gender change to female, and there is suggestive, but not conclusive, evidence that this is more frequent in patients with a history of fully male-typical prenatal androgenization; 6) there is, at this stage of research, no unambiguous evidence for or against female gender assignment of 46, XY patients, even in the prenatally most androgenized conditions; 7) The number of well documented cases is uncomfortably small for definitive conclusions (especially regarding prevalence rates of gender change), and psychological details and assessment measures often leave much to be desired; 8) the only way of obtaining a sufficient empirical basis for an intersex-management policy is to conduct sophisticated comprehensive psychological follow-up studies with reasonable sample sizes that require collaboration across clinics.
Received July 9, 1999.
Revised August 6, 1999.
Accepted August 6, 1999.
References