Gender Identity and Gender Role Orientation in Female AssignedPatients with Disorders of Sex Development

Aino K. Mattila, Riitta Fagerholm, Pekka Santtila, Päivi J. Miettinen and Seppo TaskinenFrom the School of Health Sciences, University of Tampere and Gender Identity Unit, Department of Adult Psychiatry, University Hospital of Tampere,Tampere (AKM), Department of Pediatric Surgery (RF, ST) and Department of Pediatrics (PJM), Children’s Hospital, University of Helsinki, Helsinki andÅbo Akademi University, Turku (PS), Finland

Purpose: Gender identity and gender role orientation were assessed in 24 femaleassigned patients with disorders of sex development.Materials and Methods: A total of 16 patients were prenatally exposed to an-drogens, of whom 15 had congenital adrenal hyperplasia and 1 was virilized dueto maternal tumor. Eight patients had 46,XY karyotype, of whom 5 had partialand 3 had complete androgen insensitivity syndrome. Gender identity was mea-sured by the 27-item Gender Identity/Gender Dysphoria Questionnaire for Ado-lescents and Adults with 167 female medical students as controls, and genderrole was assessed by the femininity and masculinity subscales of the 30-item BemSex Role Inventory with 104 female and 64 male medical students as controls.Results: No patient reached the cutoff for gender identity disorder on the GenderIdentity/Gender Dysphoria Questionnaire for Adolescents and Adults. However,patients with 46,XY karyotype demonstrated a somewhat more conflicted genderidentity, although the overall differences were relatively small. As to gender roleorientation, patients with complete androgen insensitivity syndrome had highscores on the femininity and masculinity scales of the Bem Sex Role Inventory,which made them the most androgynous group.Conclusions: Our findings, although clinically not clear cut, suggest that pa-tients with disorders of sex development are a heterogeneous group regardinggender identity and gender role outcomes, and that this issue should be discussedwith the family when treatment plans are made.

Key Words: adrenal hyperplasia, congenital; androgen-insensitivity

Abbreviations

and Acronyms

AIS � androgen insensitivitysyndrome

BSRI � Bem Sex Role Inventory

BSRI F � Bem Sex RoleInventory, Femininity Scale

BSRI M � Bem Sex RoleInventory, Masculinity Scale

CAH � congenital adrenalhyperplasia

CAHSL � congenital adrenalhyperplasia, salt losing

CAHSV � congenital adrenalhyperplasia, simple virilizing

CAIS � complete androgeninsensitivity syndrome

DSD � disorder of sexdevelopment

GIDYQ-AA � Gender Identity/Gender Dysphoria Questionnairefor Adolescents and Adults

syndrome; disorders of sex development; gender identity

obtained at https://www12.uta.fi/kirjasto/tup/pdf/Mattila_Aino_Supplementary_Table.pdf.

1930 www.jurology.com

GENDER identity development is con-sidered a complex process affected notonly by psychosocial factors, but alsoby biological factors such as prenatalexposure of the brain to testosteroneand the function of sex-linked genes.1

Although gender identity outcomecannot be predicted with certainty, ithas been common practice to treatchildren with disorders of sex devel-opment with genital surgery,2–4 oftenbefore the patients can express their

own opinion.

0022-5347/12/1885-1930/0THE JOURNAL OF UROLOGY®

© 2012 by AMERICAN UROLOGICAL ASSOCIATION EDUCATION AND RES

According to earlier studies, a ma-jority of patients with CAH and a46,XX karyotype who are raised asfemales exhibit female gender iden-tity, although among girls and womenwith CAH decreased heterosexual in-terest and increased identificationwith nonfemale gender are somewhatmore likely than among all women.1,2

It has also been found that girls withCAH may show masculinization ofgender related behavior but no gender

PAIS � partial androgeninsensitivity syndrome

Submitted for publication March 9, 2012.Study received institutional ethics committee

approval.Supplementary material for this article can be

confusion.5 Furthermore, gender iden-

http://dx.doi.org/10.1016/j.juro.2012.07.018Vol. 188, 1930-1934, November 2012

EARCH, INC. Printed in U.S.A.

GENDER IDENTITY IN FEMALE ASSIGNED PATIENTS WITH DISORDERS OF SEX DEVELOPMENT 1931

tity development may depend on the severity of thedisorder. Meyer-Bahlburg et al found that womenwith the salt losing form of CAH are more likely tobe gender dysphoric than women with the simplevirilizing form of CAH.6

Patients with a 46,XY karyotype and PAIS areexposed to prenatal androgen effects to some extentand are commonly born with ambiguous genitalia. Ithas been estimated that signs of gender dysphoriadevelop in up to 25% of these patients.4 In the in-stance of CAIS patients are typically born with fe-male appearing external genitalia and, therefore,are assigned the female gender of rearing. It maynot be until adolescence that the condition is diag-nosed as the etiology of primary amenorrhea.7,8 Asto the gender identity of patients with CAIS, thecommon notion is that it is feminine and unambig-uous.2,8,9

We studied a clinical sample of patients with46,XX CAH and 46,XY AIS to determine whethertheir gender identity and gender role correspondedwith their assigned gender. We hypothesized thatpatients with CAHSV and CAIS would exhibit littlegender dysphoria and their gender role would befeminine, whereas patients with CAHSL and PAISwould exhibit more signs of conflicted gender iden-tity and a more masculinized gender role.

PATIENTS AND METHODS

We retrospectively reviewed the operative database ofChildren’s Hospital, University of Helsinki for feminizinggenitoplasty between 1980 and 2008. A total of 45 patients(30 with CAH, 2 with virilization due to maternal tumor,13 with AIS) older than 15 years were identified and weremailed a questionnaire up to 3 times. The ethics commit-tee of the hospital approved the study.

A total of 24 patients (53%) returned the questionnaire.The respondents did not differ statistically significantlyfrom the nonrespondents in terms of age or major diag-nostic groups. A total of 16 respondents had prenatalandrogen exposure with 46,XX karyotype. Of these pa-tients 3 had CAHSV and 12 had CAHSL. One patient withvirilization due to a maternal virilizing tumor was placedin the CAHSV group. Eight patients had androgen insen-sitivity syndrome with 46,XY karyotype (3 with CAIS and5 with PAIS). All patients had undergone a first feminiz-ing genitoplasty and/or gonadectomy at a median age of4.3 years (range 0.1 to 14.9). The patients were mostlyfollowed by a surgeon and an endocrinologist throughoutadolescence but no systematic psychological or psychiatriccounseling on gender identity issues was provided forpatients or their parents.

Patient attitude toward timing and necessity of genitalsurgery was evaluated, with response categories of “tooyoung,” “proper age,” “too old” and “should not have beendone.” Data on patient attitudes regarding surgery havebeen reported in more detail elsewhere.10 To investigate if

the attitude toward surgery was associated with gender

identity, 2 groups of the 23 patients who gave an answerwere formed, ie those who reported that surgery was doneat the appropriate age or that they were too old at surgery(surgery acceptable group, 20 patients) and those whowere against surgery or did not express an opinion (sur-gery controversial group, 3 patients).

As a control group we had 104 or 167 (depending on thequestionnaire) female medical students and 64 male med-ical students. The medical students filled out gender iden-tity and gender role questionnaires voluntarily in semi-nars in 3 different years. Permission for recruiting thevolunteer student sample was given by the PlanningGroup of Undergraduate Medical Education at the Uni-versity of Tampere Medical School.

Patient age ranged from 15.5 to 36.7 years (mean � SD23.7 � 5.8, median 25). Corresponding figures were 23 to47 years (mean � SD 25.8 � 3.4, median 25) for the 167female controls, 23 to 43 years (mean � SD 25.8 � 3.1,median 25) for the subgroup of 104 female controls and 23to 47 years (mean � SD 26.4 � 4.1, median 25) for themale controls. Age differences between patients and con-trol groups were not statistically significant.

Gender identity was measured using the Finnish ver-sion of the 27-item GIDYQ-AA, which is designed to assessthe level of gender dysphoria and how well respondentsidentify with their assigned gender.11,12 This tool hasadolescent and adult versions. In the present study theadult version for women was used. There are validationstudies of the GIDYQ-AA in transgendered/transsexualpatients but not in patients with DSD.11,12 The GIDYQ-AAwas chosen because we were unaware of any other genderidentity/dysphoria questionnaire in Finnish.

The items on the GIDYQ-AA are statements on a5-point Likert scale ranging from 1 (always) to 5 (never).Three items are reverse coded, and the item concerningdreams can be left blank. The GIDYQ-AA index is formedby dividing the composite score of the scale by the numberof responses and theoretically ranges from 1 to 5. Lowerscores indicate more problems with gender identity andscores lower than 3 indicate gender identity disorder.11,12

The GIDYQ-AA was filled out by 167 female medicalstudents in 3 different years. Apart from the allowablenonresponse dream item, 2 patients had 1 missing item, 1patient had 3 missing items and 4 students had 1 missingitem each. In these cases the GIDYQ-AA index was formedby dividing the sum score by the actual number of re-sponses. The reverse coded item in the middle of theGIDYQ-AA was apparently misresponded to, probablydue to inattention, by 13 female students and 3 patients,and the reverse coded item at the end of the questionnaireby 1 patient. These responses were corrected.

Gender role orientation was assessed using the Finnish30-item version of the BSRI.13–15 This tool comprises 2gender role subscales, ie masculinity (BSRI M, 10 items)and femininity (BSRI F, 10 items), plus 10 neutral filleritems that were not used in this study. The items onBSRI M describe culturally desirable male traits, namelyassertiveness, leadership ability, dominance, strong per-sonality, forcefulness, aggressiveness, willingness to takea stand, independence, defense of own beliefs and willing-ness to take risks. The items on BSRI F describe culturally

desirable female traits, namely capacity for understand-

GENDER IDENTITY IN FEMALE ASSIGNED PATIENTS WITH DISORDERS OF SEX DEVELOPMENT1932

ing and sympathy, eagerness to soothe hurt feelings, sen-sitivity to the needs of others, compassion, love of children,affection, gentleness, warmth and tenderness. The itemsare statements on a 7-point Likert scale ranging from 1(never or almost never) to 7 (always or almost always).

BSRI M and BSRI F indices were formed by summingthe items on the respective scales and then dividing thecomposite scores by 10. BSRI was filled out by medicalstudents in 2 different years (104 females, 64 males). Onestudent had 1 missing response on the BSRI M scale, and5 students and 1 patient had 1 missing response each onthe BSRI F scale. The missing values were replaced by themean of the other responses on the respective scale.

To our knowledge the BSRI has not previously beenused in patients with DSD. This tool was chosen to assesspossible relative differences in feminine and masculinegender role orientation between groups.

Differences between groups were analyzed using theMann-Whitney 2-sample statistic. Computations werecarried out with Stata® statistical software, version 9.2.

RESULTS

The GIDYQ-AA index for all patients with 46,XYkaryotype and separately for patients with PAISwas statistically significantly lower than that forfemale students (see table). The GIDYQ-AA indexwas statistically significantly lower in the surgerycontroversial group (mean 4.02, range 3.42 to 4.58,median 4.07, IQR 1.15) than in the surgery accept-able group (mean 4.86, range 4.56 to 5.00, median4.91, IQR 0.24, p � 0.010).

Individual GIDYQ-AA items were analyzed for fe-male students and patient groups according to karyo-type. Comparisons between patients with 46,XXkaryotype and female students yielded 2 statisticallysignificant differences. Altogether 13 of 27 items dif-fered statistically significantly between patients with46,XY karyotype and female students.

There were no statistically significant differencesfor any of the patient groups on the BSRI F index

GIDYQ-AA Index and Bem Sex Role Inventory scores for stude

GIDYQ-AA Index

Mean (range) Media

Female students 4.87 (3.96–5.00) 4.93Pts with 46,XX karyotype: 4.85 (4.56–5.00) 4.91

CAHSV 4.86 (4.56–5.00) 4.94CAHSL 4.84 (4.56–5.00) 4.87

Pts with 46,XY karyotype: 4.53 (3.42–5.00) 4.70PAIS 4.64 (4.07–4.96) 4.78CAIS 4.35 (3.42–5.00) 4.63

Male students — —

GIDYQ-AA scores ranged from 1 to 5, with scores below 3 indicating gender dy* p � 0.019 compared to female students.† p � 0.034 compared to female students.‡ p � 0.009 compared to female students, p � 0.040 compared to male studen§ p �0.001 compared to female students.

� p � 0.002 compared to female students.

(see table). The BSRI M index for patients withCAIS was statistically significantly higher than itwas for female students, male students or patientswith PAIS.

DISCUSSION

The main finding of the present study was that interms of gender dysphoria no patient reached thecutoff for gender identity disorder on the GIDYQ-AA, although as a group patients with 46,XY karyo-type demonstrated a somewhat more conflicted gen-der identity than those with a 46,XX karyotype.Regarding our hypothesis based on the literature,the findings concerning patients with CAHSV wereclose to those expected, as this group did not essen-tially differ from female students. The same findingalso applied to patients with CAHSL, although theywere expected to show more signs of gender dyspho-ria. In line with our hypothesis the GIDYQ-AA indexfor patients with PAIS was significantly lower com-pared to female students, meaning that these pa-tients reported slightly more signs of conflicted gen-der identity than did the students. Unexpectedly,the patients with CAIS had the lowest GIDYQ-AAindex, although (possibly because of the small sam-ple size) they did not differ statistically significantlyfrom the other groups.

The overall differences on the GIDYQ-AA wererelatively small, and their clinical significance isunclear. The average GIDYQ-AA index is reportedly4.87 for heterosexual women (the same as for thefemale students and essentially the same as for thepatients with 46,XX karyotype in our study), 4.38 fornonheterosexual women, 2.56 for male gender iden-tity patients and 2.20 for female gender identitypatients.11 In all of our patients with 46,XY karyo-type the average index (4.53) was somewhat higherthan what has been reported for nonheterosexual

patients

Mean � SDBSRI F Index (range)

Mean � SDBSRI M Index (range)

5.50 � 0.68 (3.80–6.80) 4.67 � 0.61 (3.10–6.10)5.40 � 0.87 (3.90–6.60) 4.86 � 0.71 (3.00–5.90)5.64 � 0.74 (4.80–6.60) 5.03 � 0.47 (4.70–5.70)5.33 � 0.92 (3.90–6.50) 4.80 � 0.79 (3.00–5.90)5.26 � 0.75 (4.10–6.20) 5.00 � 0.65 (3.90–5.80)5.12 � 0.77 (4.10–6.20) 4.64 � 0.53 (3.90–5.30)5.50 � 0.82 (4.60–6.20) 5.60 � 0.20 (5.40–5.80)‡4.99 � 0.65 (3.50–6.30)§ 4.98 � 0.56 (3.90–6.50)�

BSRI scores ranged from 1 to 7.

� 0.025 compared to patients with PAIS.

nts and

n (IQR)

(0.19)(0.28)(0.24)(0.28)(0.56)*(0.23)†(1.58)

sphoria.

ts and p

GENDER IDENTITY IN FEMALE ASSIGNED PATIENTS WITH DISORDERS OF SEX DEVELOPMENT 1933

women,11 while in our patients with CAIS it wasapproximately the same (4.35). We have earlierreported that nonheterosexual orientation issomewhat more prevalent in the sexually activepatients in our sample, especially in patients withAIS, compared to the general female population.10

This finding may be a factor reflected in the presentfindings.

According to individual GIDYQ-AA items, pa-tients with AIS as a group were inclined to identifythemselves not entirely as women but as “trans” or“intersexual.” Moreover, there were 3 patients whoeither were against genital surgery or could not givetheir opinion concerning it, all of whom belonged tothe 46,XY group (1 patient with CAIS and 2 withPAIS). These patients had statistically significantlymore signs of gender dysphoria than those whofound the surgery necessary. In a study on a diag-nostically heterogeneous sample of adults with46,XY DSD it was found that a third had at somepoint in their lives been unsure about their gen-der, although 85% nevertheless reported beingmainly satisfied with it, while a minority alsoendorsed a third gender option.16 In the presentstudy the patient who had the lowest GIDYQ-AAindex, and who also deplored the genital surgerythat had been performed in adolescence, provided awritten explanation on the GIDYQ-AA form: “I amnot a man or a woman; I am something else and itsuits me.”

As to gender role orientation, the BSRI femininityscale index differed statistically significantly onlybetween female and male students. When the pa-tients were divided into 4 diagnostic subgroups,those with CAIS had high scores on the BSRI femi-ninity and masculinity scales, and the differences onBSRI M between patients with CAIS and femalestudents, CAIS and male students, and CAIS andPAIS were statistically significant. These resultsmake patients with CAIS the most androgynousgroup of all,13 which may not be pathological, as ithas been suggested that androgyny may be associ-ated with better mental health.17 According to ear-lier studies, and the assumed prenatal androgenexposure of the central nervous system, one wouldexpect individuals with CAHSL and PAIS to be themost masculine of the patient groups.1,2,6 However,surprisingly this was not the case.

Our study has several limitations. The number ofeligible patients was small and only 53% partici-pated, which makes generalizing the results prob-

lematic. Moreover, because of the small size of the

diagnostic subgroups, there possibly was not enoughstatistical power to detect such significant differ-ences that may have been evident had the groupsbeen larger. Therefore, our study should primarilybe regarded as descriptive.

Another important limitation is the lack of a clin-ical, age matched control group. Our control groupwas a convenience sample of medical studentswhose age distribution was different from that of thepatients, although the age differences did not reachstatistical significance. Furthermore, the design ofour naturalistic followup study was cross-sectionalregarding gender identity and gender role orienta-tion, and we had no data concerning the early devel-opmental years of the patients or the age at onset ofgender identity problems when these were implied.In regard to gender identity formation, 6 patients(25% of sample) were younger than 20 years andthus adolescents.18 Moreover, with our methods itwas impossible to establish whether the signs ofgender conflict were caused by psychobiological orpsychosocial factors, or both.

We used self-report measures that, comparedwith interviews, usually give more suggestive ratherthan definitive estimates of psychological phenom-ena. The measures we used should also be consid-ered critically. The GIDYQ-AA has been designedto screen for gender identity disorders.11,12 Addi-tionally, findings from the BSRI may vary acrosscultures and samples, making the results possiblysample dependent.19 Furthermore, neither theGIDYQ-AA nor the BSRI has been validated in pa-tients with DSD. A strength of our study was thatour original database was quite representative ofthe population, as most DSD surgery in Finland hasbeen performed at Helsinki University Central Hos-pital.

CONCLUSIONS

Our findings suggest that patients with DSD are aheterogeneous group regarding gender identity andgender role outcome, which should be taken intoaccount when treatment plans are made. In recentyears a shift toward multidisciplinary evaluationand treatment of DSD with active involvement ofthe patient and family has been taking place.3,4 Anindividual treatment plan with gender identitycounseling for the patient and family, and a widerrange of treatment options based on up-to-date psy-chobiological research findings should be provided

for all patients.

GENDER IDENTITY IN FEMALE ASSIGNED PATIENTS WITH DISORDERS OF SEX DEVELOPMENT1934

REFERENCES

1. Hines M: Gender development and the humanbrain. Annu Rev Neurosci 2011; 34: 69.

2. Yang JH, Baskin LS and DiSandro M: Genderidentity in disorders of sex development: reviewarticle. Urology 2010; 75: 153.

3. Wiesemann C, Ude-Koeller S, Sinnecker GH et al:Ethical principles and recommendations for themedical management of differences of sex de-velopment (DSD)/intersex in children and adoles-cents. Eur J Pediatr 2010; 169: 671.

4. Barthold JS: Disorders of sex differentiation: apediatric urologist’s perspective of new terminol-ogy and recommendations. J Urol 2011; 185: 393.

5. Meyer-Bahlburg HF, Dolezal C, Baker SW et al:Prenatal androgenization affects gender-relatedbehavior but not gender identity in 5–12-year-oldgirls with congenital adrenal hyperplasia. ArchSex Behav 2004; 33: 97.

6. Meyer-Bahlburg HF, Dolezal C, Baker SW et al:Gender development in women with congenitaladrenal hyperplasia as a function of disorderseverity. Arch Sex Behav 2006; 35: 667.

7. Allen L: Opinion one: a case for delayed gonad-

ectomy. J Pediatr Adolesc Gynecol 2009; 22: 381.

8. Kiddo DA: Opinion two: a case for early gonad-ectomy. J Pediatr Adolesc Gynecol 2009; 22: 384.

9. Wisniewski AB, Migeon CJ, Meyer-Bahlburg HFet al: Complete androgen insensitivity syndrome:long-term medical, surgical, and psychosexualoutcome. J Clin Endocrinol Metab 2000; 85:2664.

10. Fagerholm R, Santtila P, Miettinen P et al: Sexualfunction and attitudes toward surgery after fem-inizing genitoplasty. J Urol 2011; 185: 1900.

11. Deogracias JJ, Johnson LL, Meyer-Bahlburg HFet al: The Gender Identity/Gender DysphoriaQuestionnaire for Adolescents and Adults. J SexRes 2007; 44: 370.

12. Singh D, Deogracias JJ, Johnson LL et al: TheGender Identity/Gender Dysphoria Questionnairefor Adolescents and Adults: further validity evi-dence. J Sex Res 2010; 47: 49.

13. Bem SL: The measurement of psychological an-drogyny. J Consult Clin Psychol 1974; 42: 155.

14. Choi N, Fuqua DR and Newman JL: Exploratoryand confirmatory studies of the structure of the

Bem Sex Role Inventory Short Form with two

divergent samples. Educ Psychol Meas 2009; 69:696.

15. Colley A, Mulhern G, Maltby J et al: The shortform BSRI: instrumentality, expressiveness andgender associations among a United Kingdomsample. Pers Individ Dif 2009; 46: 384.

16. Meyer-Bahlburg HF, Migeon CJ, Berkovitz GD etal: Attitudes of adult 46,XY intersex persons toclinical management policies. J Urol 2004; 171:1615.

17. Lefkowitz ES and Zeldow PB: Masculinity andfemininity predict optimal mental health: a be-lated test of the androgyny hypothesis. J PersAssess 2006; 87: 95.

18. UNICEF: The state of the world’s children 2011.Adolescence: an age of opportunity. Available athttp://www.unicef.org/sowc2011. Accessed March 9,2012.

19. Juujärvi S: The Ethic of Care and its Develop-ment. Helsinki: University of Helsinki, Faculty ofSocial Sciences, Department of Social Psychology2003. Available at http://urn.fi/URN:ISBN:952-10-

1436-9. Accessed March 9, 2012.