Dec-cjp

Adoptee Overrepresentation
Among Clinic-Referred Boys With
Gender Identity Disorder

Kenneth J Zucker, PhD¹, Susan J Bradley, MD²

Objectives: To test the hypothesis that adoptees are overrepresented among a sample of clinic-referred boys with gender identity problems (N = 238). To compare the adoptees and nonadoptees on demographic, behaviour problem, and gender-typed measures.

Method: The percentage of clinic-referred boys with gender identity problems adopted in the first 2 years of life (“early adoptees”) was compared to the base rate of boys adopted in Ontario. Parent-report and behavioural measures were used to compare the early adoptees with “late adoptees” (adopted after the second year of life) and nonadoptees.

Results: The percentage of boys with gender identity problems who were early adoptees (7.6%) was significantly higher than the base rate of males adopted in Ontario in the first 2 years of life (1.5%). Both the early and late adoptees were significantly less intelligent than the nonadoptees. The early adoptees also had significantly higher externalizing T scores on the Child Behavior Checklist than did the late adoptees and the nonadoptees. The 3 groups did not differ in the percentage who met the complete Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for gender identity disorder and on 4 other measures of gender-typed behaviour.

Conclusion: Adoptees are overrepresented among clinic-referred boys with gender identity problems. The reasons for this finding are not clear but may be accounted for by general risk factors that increase the likelihood of clinical referral or by psychosocial and biological factors associated with adoption.

(Can J Psychiatry 1998;43:1040–1043)

Key Words: gender identity disorder, adoption, boys

Adoptees are overrepresented among referrals to child mental health clinics (1), including in Canadian samples (2,3). This may reflect a genuine increase in behavioural difficulties, or it may be that adoptive parents make better use of clinical services (1,4).

Cursory inspection of demographic data on a clinic- referred sample of boys with gender identity problems suggested an overrepresentation of adoptees. The present study had 2 aims: 1) to verify that the percentage of adoptees was significantly elevated from the base rate of adoptees in the general population and 2) to examine some demographic and behavioural correlates of the child’s adoptive status.

Methods

Subjects

Between 1978 and 1996, 238 boys were assessed in a specialty clinic for children and adolescents with gender identity problems at the children’s department at a psychiatric research institute. Parent interview information was used to determine whether or not each boy met the complete Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for gender identity disorder (GID) of childhood (5).

Adoptive Status

Of the 238 boys, 18 (7.6%) were adopted or taken into permanent foster care prior to age 2 years: 16 boys were adopted between the time of birth and 1 year of age; the other 2 boys were adopted in their second year of life (median, 120 days; range, 1–540 days). In all cases, the adoption preceded the emergence of behavioural symptoms pertaining to GID. These boys were classified as early adoptees. Ten (4.2%) other boys were adopted or taken into long-term foster care after age 2 years (mean, 6.0 years; range, 2.8–11.7 years). In most cases, however, the symptoms of GID preceded adoption or foster care placement. These boys were classified as late adoptees. The remaining 210 boys had lived continuously since birth with at least 1 biological parent and were classified as nonadoptees.

Measures

Three sets of analyses were performed:

1. The percentage of early adoptees was compared with the percentage of boys adopted in Ontario in the first 2 years of life (6,7).

2. The 3 groups were compared on 3 demographic variables: age at assessment, parents’ social class, and parents’ marital status. For the early and late adoptees, social class and marital status were that of the adoptive parents (with the exception of the 4 late adoptees who were in long-term foster care).

3. The 3 groups were compared on a measure of general behaviour problems (the Child Behavior Checklist [CBCL]) (8), full-scale IQ, and 5 measures of gender-typed behaviour: 1) the percentage of boys who met the complete DSM criteria for GID; 2) the percentage of boys who drew a girl first on the Draw-a-Person test (9); 3) the percentage of cross-gender play on a free play task (10); 4) parent-report of feminine behaviour on 2 factors of the Revised Gender Behavior Inventory for Boys (11); and 5) parent-report of activity level (12).

Results

The percentage of boys adopted in Ontario prior to age 2 years between 1965 and 1989 (the birth-year range of the early adoptees) was 1.49% (24 613 of 1 651 266 male live births). This percentage was significantly lower than the percentage of early adoptees in our sample of clinic-referred, gender-disordered boys, c²[1] = 55.7, P < 0.000 01.

Table 1 shows the demographic data. At the time of assessment, both the early and late adoptees were significantly older than the nonadoptees (both Ps < 0.05). The 3 groups did not differ significantly with regard to their parents’ social class and marital status. It should, however, be noted that all but 1 of the adoptees were born to unwed mothers or to mothers who were not living with the biological father. With regard to biological parentage, therefore, it is quite likely that both the early and late adoptees were, on average, from a lower socioeconomic background than were the nonadoptees.

Table 1. Demographic variables as a function of group
	Nonadoptees			Early adoptees			Late adoptees
	Mean	SD	N	Mean	SD	N	Mean	SD	N
Mean age in years at assessment^a	6.9	2.3	210	8.7	2.8	18	9.8	2.1	10
Mean social class index^b	42.5	14.8	204	42.0	18.2	18	33.7	15.3	10
Mean full-scale IQ^c	106.8	15.7	201	95.5	17.9	18	95.6	17.2	10
Children living with:
Both parents			151			11			7
Mother only or reconstituted family			59			7			3

^aF[2,237] = 12.2, P < 0.0001. Post hoc comparisons based on Duncan’s multiple range test for unequal numbers (14).
^bSocial class derived from Hollingshead’s Four Factor Index of Social Status (absolute range 8–66) (13). Data on social class were unavailable for 6 nonadoptees.
^cF[2,228] = 6.2, P = 0.0023. IQ data based on either the Wechsler Preschool and Primary Scale of Intelligence (WPPSI or WPPSI-R) or the Wechsler Intelligence Scale for Children (WISC-R or WISC-III). IQ scores were not available for 9 nonadoptees.

Table 1 also shows the full-scale IQ data. Both the early and late adoptees had significantly lower IQs than the nonadoptees (both Ps < 0.05).

Table 2 shows the CBCL measures of behavioural psychopathology. Of the 5 measures, only 1 was significant: the early adoptees had a significantly higher externalizing T score than did the nonadoptees (P < 0.05) and the late adoptees (P < 0.10).

Table 2. Percentage of psychiatrists (N = 30) indicating the frequency of various behaviours
Behaviour	Always or almost always (%)	Sometimes (%)	Never or almost never (%)
Document in the chart that a competent patient was informed about risks and benefits of antipsychotic medication	13	60	27
Obtain written consent when prescribing antipsychotics to competent patients	0	0	100
Document in the chart that a relative of an incompetent patient was informed about risks and benefits of antipsychotic medication	44	28	28
Obtain written consent from relatives when prescribing antipsychotics to incompetent patients	28	8	64

^aCBCL data were not available for 10 nonadoptees and 1 early adoptee.
^bF[2,226] = 4.2, P = 0.0155.

Table 3 shows the data for the gender-typed measures as a function of group. Because the groups differed in age and because age was correlated with the behavioural and parent-report measures (in all cases, age was negatively related to the degree of feminine behaviour), age was partialled out or covaried in all of the analyses. There were no significant differences among the 3 groups on any of these measures.

Table 3. Measures of sex-typed behaviour as a function of group
	Nonadoptees		Early adoptees		Late adoptees
Measures	N	%	N	%	N	%
DSM
Full criteria Subthreshhold	127 82	60.8 39.2	9 9	50.0 50.0	4 6	40.0 60.0
Draw a person^a
Boy drawn first Girl drawn first	87 112	43.7 56.3	9 9	50.0 50.0	9 1	90.0 10.0
	Mean SD N		Mean SD N		Mean SD N
Free play task (same-sex play–cross-sex play)^b	–0.13 0.61 181		0.11 0.58 15		0.55 0.56 10
Activity level ^c	3.07 0.58 193		3.11 0.62 17		3.03 0.38 9
Revised Gender Behaviour Inventory for Boys^d
Factor 1	2.71 0.83 194		2.65 0.85 17		2.90 1.35 8
Factor 2	2.89 1.14 194		2.71 1.02 17		2.16 0.74 8

^aSome data were missing because a few parents did not wish their children to undergo psychological testing or because the family chose not to complete the assessment.

^bSome data were missing because the task was not administered to children at the upper end of the age range (mainly 11- and 12-year olds), on the grounds that it was age-inappropriate. Absolute range –1.0 to 1.0.

^cSome data were missing because the parents could not read English or because the parents chose not to complete the assessment process. Absolute range 1.0–5.0.

^dAbsolute range was 1.00–5.80 (Factor 1) and 1.00–6.84 (Factor 2).

Discussion

Compared with the base rate of boys adopted in Ontario prior to age 2 years, early adoptees were overrepresented among a clinic-referred sample of boys with gender identity problems; however, our percentage of early adoptees is comparable to the percentage of adoptees reported in other Canadian child samples referred for more common clinical problems (2,3). This latter comparison, however, should be viewed with some caution, since the age at adoption is often not specified in such studies (1). Perhaps the percentage of early adoptees, at least as this status was defined in the present study, would be higher than it would be for boys adopted at a similar age but with other types of clinical problems. This should be explored in a future study. Nevertheless, from the present data, one can ask why early adoptees are overrepresented among clinic-referred boys with gender identity problems.

One explanation is that the early adoptees had characteristics other than the gender identity problem that increased the likelihood of a clinical referral. In the present study, 2 potential variables were identified: lower intelligence and a higher externalizing T score on the CBCL. Thus, the overrepresentation of adoptees among boys with gender identity problems may be a function of other risk factors that lead to clinic referrals.

Several psychosocial and biological factors have been hypothesized as playing either predisposing or causal roles in the development of GID (12). At least 2 of these factors might be relevant in understanding the overrepresentation of adoptees among boys with GID.

The first explanation implicates the role of early disruptions in the mother–son attachment relationship, which predisposes the son to develop gender identity conflict.

Adoption, particularly when it occurs under adverse conditions, may predispose the child to an insecure attachment with the adoptive mother (15). It has been argued that insecure attachments and concomitant separation anxiety disorder precede the emergence of boyhood GID (16). In such instances, cross-gender behaviour (for example, cross-dressing) has been interpreted as a defensive manoeuvre in which the boy confuses “being Mommy” with “having Mommy” (16, p 708). Unfortunately, this account has several interpretive problems, including the fact that most young boys with insecure attachments or separation anxiety do not develop GID. Thus, the proposed causal sequence lacks specificity.

A second explanation pertains to the role of prenatal maternal stress. Experimental induction of stress during a critical period in rat pregnancy shows that the sex-dimorphic behaviour of male offspring is both demasculinized and feminized (17,18). Several retrospective human studies have examined whether or not the mothers of homosexual men were more likely than those of heterosexual men to experience stress during their pregnancies (19–22). In these studies, sexual attraction to men was viewed as a form of behavioural feminization, since such sexual attraction to men is more typical of women than men. The results of these studies, however, have been equivocal, and the methodologically strongest study (22) found no effect.

GID in boys can also be viewed as a form of feminization, since their sex-dimorphic behaviour is more typical of girls than of boys. It is possible that the mothers of the adopted boys were under severe stress during their pregnancies—some were troubled teenagers, others were financially unable to keep their sons, and some were substance abusers. If prenatal maternal stress actually increases the likelihood of postnatal behavioural feminization, this might account for the overrepresentation of adopted boys with gender identity problems. Unfortunately, the validity of this sex-dimorphic psychobiologic explanation is uncertain.

This scenario also does not account for at least 2 facts: most boys who are adopted do not develop gender identity problems, and most boys with gender identity problems are not adopted. Nonetheless, if the stress hypothesis is accurate, it could be accounted for by at least 2 factors: the pregnancies of these mothers were particularly stressful, or the male fetuses had a specific predisposition to respond to the maternal stress with feminization of the central nervous system.

It is also unclear whether the mothers of the nonadoptees had greater prenatal stress than those of control boys. If one could demonstrate that the prenatal stress hypothesis has validity, it could then be deemed a general risk factor for the development of GID in boys. Unfortunately, such data are not available. If prenatal maternal stress is not a general risk factor, it remains a possibility that it is operative for only a subgroup of boys with GID, including those who are adopted early in life. Since the early adoptees were no more or less feminine than the nonadoptees, at least as judged by the behavioural measures used in this study, this would suggest that the behavioral “end state” (that is, GID) is the product of multiple casual pathways, a conclusion that is believed by some to best account for the development of psychopathology in general (23).

Clinical Implication

Early adoption may constitute a “small” risk factor for the development of gender identity problems in boys.

Limitations

A clinical control group of early adoptees was not ascertained.
It is not clear whether or not the percentage of early adoptees among clinic-referred boys with gender identity problems would be higher or lower than or comparable to the percentage of early adoptees referred for other clinical problems.

References

1. Haugaard JJ. Is adoption a risk factor for the development of adjustment problems? Clin Psychol Rev 1998;18:47–69.

2. Jerome L. Overrepresentation of adopted children attending a children’s mental health centre. Can J Psychiatry 1986;31:526–31.

3. Kotsopoulos S, Cote A, Joseph K, Pentland N, Stavrakaki C, Sheahan P, and others. Psychiatric disorders in adopted children: a controlled study. Am J Orthopsychiatry 1988;58:608–12.

4. Cohen NJ, Coyne J, Duvall J. Adopted and biological children in the clinic: family, parental and child characteristics. J Child Psychol Psychiatry 1993;34:545–62.

5. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 3rd ed. Revised. Washington (DC): American Psychiatric Association; 1987.

6. Information Systems Branch, Ministry of Community and Social Services, North York, Ontario.

7. Registration Division, Office of the Registrar General, Ministry of Consumer and Commercial Relations, Thunder Bay, Ontario.

8. Achenbach TM, Edelbrock C. Manual for the Child Behavior Checklist and Revised Child Behavior Profile. Burlington (VT): University of Vermont, Department of Psychiatry; 1983.

9. Zucker KJ, Finegan JK, Doering RW, Bradley SJ. Human figure drawings of gender-problem children: a comparison to sibling, psychiatric, and normal controls. J Abnorm Child Psychol 1983;11:287–98.

10. Zucker KJ, Doering RW, Bradley SJ, Finegan JK. Sex-typed play in gender-disturbed children: a comparison to sibling and psychiatric controls. Arch Sex Behav 1992;11:309–21.

11. Zucker KJ, Bradley SJ, Corter CM, Doering RW, Finegan JK. Cross-gender behaviour in very young boys: a normative study. In: Sampson J, editor. Childhood and sexuality. Montreal: Editions Etudes Vivantes; 1980. p 599–622.

12. Zucker KJ, Bradley SJ. Gender identity disorder and psychosexual problems in children and adolescents. New York: Guilford Press; 1995.

13. Hollingshead AB. Four factor index of social status. Unpublished manuscript. Department of Sociology, Yale University, New Haven (CT); 1975.

14. Kramer CY. Extensions of multiple range tests to group means with unequal numbers of replications. Biometrics 1956;12:307–10.

15. Goldberg S. Attachment and childhood behavior problems in normal, at-risk, and clinical samples. In: Atkinson L, Zucker KJ, editors. Attachment and psychopathology. New York: Guilford Press; 1997. p 171–95.

16. Coates S, Person ES. Extreme boyhood femininity: isolated behavior or pervasive disorder? J Am Acad Child Psychiatry 1985;24:702–9.

17. Ward IL. The prenatal stress syndrome: current status. Psychoneuroendocrinology 1984;9:3–11.

18. Ward IL. Sexual behavior: the product of perinatal hormonal and prepubertal social factors. In: Gerall AA, Moltz H, Ward IL, editors. Handbook of behavioral neurobiology. Volume 11. Sexual differentiation. New York: Plenum Press; 1992. p 157–80.

19. Dörner G, Geier T, Ahrens L, Krell L, Münx G, Sieler, H, and others. Prenatal stress as possible aetiogenic factor of homosexuality in human males. Endokrinologie 1980;75:365–8.

20. Dörner G, Schenk B, Schmiedel B, Ahrens L. Stressful events in prenatal life of bi- and homosexual men. Experimental and Clinical Endocrinology 1983;81:83–7.

21. Ellis L, Ames MA, Peckham W, Burke D. Sexual orientation of human offspring may be altered by severe maternal stress during pregnancy. Journal of Sex Research 1988;25:152–7.

22. Bailey JM, Willerman L, Parks C. A test of the maternal stress theory of human male homosexuality. Arch Sex Behav 1991;20:277–93.

23. Cicchetti D, Rogosch FA. Equifinality and multifinality in developmental psychopathology. Dev Psychopathol 1996;8:597–600.

Résumé

Objectifs : Vérifier l’hypothèse selon laquelle les enfants adoptés sont sur- représentés au sein d’un échantillonnage de garçons présentant des problèmes d’identité sexuelle qui ont été adressés à une clinique (N = 238). Comparer les enfants adoptés et non adoptés selon les mesures démographiques, de problèmes de comportement et de catégorisation selon le sexe.

Méthode : On a comparé le pourcentage des garçons présentant des problèmes d’identité sexuelle qui ont été adressés à une clinique et qui ont été adoptés dans les deux premières années de leur vie (« adoptés tôt ») au taux de base des garçons adoptés en Ontario. On a utilisé des mesures de déclaration des parents et de comportement pour comparer les enfants adoptés tôt aux enfants adoptés plus tard (après leur deuxième année) et aux enfants non adoptés.

Résultats : Le pourcentage des garçons présentant des problèmes d’identité sexuelle qui ont été adoptés en bas âge (7,6 %) était significativement plus élevé que le taux de base des garçons adoptés dans leurs deux premières années de vie en Ontario (1,5 %). Tant les enfants adoptés en bas âge que plus tard étaient significativement moins intelligents que les enfants non adoptés. Les enfants adoptés en bas âge affichaient également des notes externalizing T (externalisation T) significativement plus élevées à la Child Behavior Checklist (liste de vérification du comportement des enfants) que les enfants adoptés plus tard et que les enfants non adoptés. Les trois groupes ne présentaient pas de différence en ce qui concerne le pourcentage de ceux qui satisfont aux critères complets du DSM des troubles d’identité sexuelle, et les quatre autres mesures du comportement pour la catégorisation selon le sexe.

Conclusion : Les enfants adoptés sont sur-représentés au sein des garçons présentant des problèmes d’identité sexuelle qui ont été adressés à une clinique. Les raisons de ce résultat ne sont pas évidentes mais peuvent s’expliquer par les facteurs de risque généraux qui accroissent la probabilité de l’aiguillage à une clinique, ou par les facteurs psychosociaux et biologiques associés à l’adoption.

A version of this paper was presented at the meeting of the International Academy of Sex Research, Baton Rouge, Louisiana, July 1997.

Manuscript received January 1998, revised, and accepted May 1998.

¹Senior Psychologist and Head, Child and Adolescent Gender Identity Clinic, Child and Family Studies Centre, Clarke Institute of Psychiatry, Toronto, Ontario.

²Consultant Psychiatrist, Child and Adolescent Gender Identity Clinic, Child and Family Studies Centre, Clarke Institute of Psychiatry, Toronto, Ontario.

Address for correspondence: Dr KJ Zucker, Child and Adolescent Gender Identity Clinic, Child and Family Studies Centre, Clarke Institute of Psychiatry, 250 College St, Toronto, ON M5T 1R8

email: ZUCKERK@cs.clarke-inst.on.ca

Can J Psychiatry, Vol 43, December 1998