What explains discrepant gender identity outcome in 46,XX individuals with 11-hydroxylase deficiency?

To The Editor: I read with great interest the report by Bin-Abbas et al1 of two 46,XX siblings with 11-hydroxylase deficiency (11-OHD) raised male with divergent gender identity (GI) outcome in adolescence. The authors conclude “that the extent of external genitalia virilization, particularly the degree of hypospadias, and not the duration or the level of prenatal and postnatal androgen exposure, is the main gender identity determinant” in 46,XX patients with CAH. I applaud the authors for having based their decisions for and against gender-reassignment in these two siblings on a careful psychiatric evaluation with extensive counseling, but I am concerned that their theoretical conclusions regarding the determinants of GI may lead some clinicians to problematic unifactorial decisions regarding gender assignment and re-assignment.

Clearly, the “extent of external genitalia virilization” determines gender assignment at birth in newborns with normal genital development, and to a considerable extent also in newborns with ambiguous genitalia, especially when sophisticated diagnostic techniques are not readily accessible, as it is common in resource-poor communities. However, the association of the degree of masculinization of the external genitalia with gender identity development is comparably weak, not only in 46,XY, but also in 46,XX individuals, and genital-status-based gender assignment in infancy does not firmly determine GI outcome later. Otherwise we would not see gender dysphoria and patient-desired gender re-assignment in 46,XX 21-OHD-CAH, which occurs–albeit with relatively low incidence–in both female- and male-assigned patients.2 In my reading of the literature and my clinical experience with 46,XX CAH patients raised female, patient-initiated gender change to male occurs more frequently in patients born with Prader stage 3–4 (clitorophallus with hypospadias) than in patients born with Prader stage 5 (completely masculinized external genitalia). The existence of non-intersex patients with female-to-male transsexualism who typically have entirely normal female genitalia also speaks against a determining role of the external genitalia for GI development.

So, which factors should one consider in explaining the different outcome in the two 11-OHD siblings of Bin-Abbas et al? First of all, the remarkable difference between the siblings in genital development may indicate behaviorally relevant differences in androgen exposure of the developing brain. Given that the two siblings are offspring from the same parents, it is highly likely (albeit not absolutely certain) that the molecular abnormality of the 17-OHD gene is identical in both. Nevertheless, the high local androgen concentration required at the early critical period of genital differentiation for the formation of a penile urethra3 was apparently not present in the older sibling, although later in development the androgen levels were high enough to promote substantial growth of a clitorophallus. We do not know how late during the fetal development of this patient the androgens increased to that level. The statement that the older sibling, in contrast to the younger one, “behaved as a female since early childhood”, raises the question whether the systemic androgen level influencing the sexual differentiation/masculinization of the brain during its hormone-sensitive period of prenatal development was lower in this sibling than in the younger one. Mammalian research has provided ample evidence that both dose and duration of androgen exposure during the homone-sensitive pre- or perinatal period of brain development affect later behavioral masculinization. This principle is likely to apply to humans as well. Research of my team in 46,XX individuals with 21-OHD CAH has clearly shown a dose-response relationship of syndrome severity (which is associated with the degree of androgen excess) with gender-related behavior (GR),4 and related research in both nonhuman primates and humans makes it likely that these GR effects are primarily prenatal rather than postnatal in origin.5,6 However, unlike the early critical period of genital differentiation, the hormone-sensitive period of sexual differentiation of the brain is much longer and extends far into the second half of pregnancy.

Our studies show considerable variability of GR in individuals with a given degree of CAH severity, as one also sees it in non-intersex males and females, so that prenatal androgen production is certainly not the only biological factor that explains interindividual differences in behavioral masculinization. In addition, one has to consider potential variations in systemic androgen metabolism, e.g., by aromatization,7 or in androgen utilization due to androgen receptor modulation by tri-nucleotide repeats in the androgen receptor gene,8 also the possibility of CNS-limited variations of endocrine processes such as production of neurosteroids and the action of steroid co-activators and co-inhibitors, 9,10 and last but not least, contributions from non-steroid related genetic effects on temperament.11

Although psychosocial factors contribute to GR,12 the markedly masculinized GR of 21-OHD CAH girls appears to be rather unresponsive to parental gender-normalizing pressure,13 and recognition of children’s gender by their peers is not primarily based on genital status.12 However, gender-related behaviors and interests that are atypical for the assigned gender are salient among the commonly given reasons for gender dysphoria and patient-desired gender-change by both intersex and transsexual patients in later adolescence and adulthood. On the other hand, core GI as male or female can accommodate considerable variations of GR, and my own team’s research has shown a lack of correlation of GI and GR in 21-OHD girls,14 which makes it unlikely that the processes underlying the development of GI and GR are identical. Many individuals with a marked degree of gender-atypical GR do not suffer from gender dysphoria, while others with a similar degree do, and we still cannot pinpoint the etiological difference. Overall, it currently appears more likely that it is the confluence of a variety of interacting biological, social, and psychological factors rather than one prominent factor among these that pushes a given child or adolescent towards gender dysphoria or actual gender-reassignment. Thus, when deciding for the newborn with a disorder of sex development on the gender assignment with optimal prognosis for stable GI development, we need to consider whether and to what extent the genital status at birth permits inferences on the androgenization of the brain, which varies considerably between syndromes, and to take into account all other known relevant factors including psychosocial and cultural ones. In cases of potential gender re-assignment after infancy, a comprehensive evaluation of the patient’s gender status (GR, GI, and any indications of gender dysphoria) must be added.