Content uploaded by Peggy T Cohen-Kettenis
Author content
All content in this area was uploaded by Peggy T Cohen-Kettenis
Content may be subject to copyright.
Sex reassignment: outcomes and predictors of
treatment for adolescent and adult transsexuals
YOLANDA L. S. SMITH*, STEPHANIE H. M. VAN GOOZEN, ABRAHAM J. KUIPER
AND PEGGY T. COHEN-KETTENIS
Department of Child and Adolescent Psychiatry, University Medical Centre Utrecht, The Netherlands ;
Department of Medical Psychology, VU University Medical Centre, Amsterdam, The Netherlands
ABSTRACT
Background. We prospectively studied outcomes of sex reassignment, potential differences between
subgroups of transsexuals, and predictors of treatment course and outcome.
Method. Altogether 325 consecutive adolescent and adult applicants for sex reassignment partici-
pated: 222 started hormone treatment, 103 did not; 188 completed and 34 dropped out of treat-
ment. Only data of the 162 adults were used to evaluate treatment. Results between subgroups were
compared to determine post-operative differences. Adults and adolescents were included to study
predictors of treatment course and outcome. Results were statistically analysed with logistic re-
gression and multiple linear regression analyses.
Results. After treatment the group was no longer gender dysphoric. The vast majority functioned
quite well psychologically, socially and sexually. Two non-homosexual male-to-female transsexuals
expressed regrets. Post-operatively, female-to-male and homosexual transsexuals functioned better
in many respects than male-to-female and non-homosexual transsexuals. Eligibility for treatment
was largely ba sed upon gender dysphoria, psychological stability, and physical appearance. Male-
to-female transsexuals with more psychopathology and cross-gender symptoms in childhood, yet
less gender dysphoria at application, were more likely to drop out prematurely. Non-homosexual
applicants with much psychopathology and body dissatis faction reported the worst post-operative
outcomes.
Conclusions. The results substantiate previous conclusions that sex reassignment is effective. Still,
clinicians need to be alert for non-homosexual male-to-females with unfavourable psychological
functioning and physical appearance and inconsistent gender dysphoria reports, as these are risk
factors for dropping out and poor post-operative results. If they are considered eligible, they may
require additional therapeutic guidance during or even after treatment.
INTRODUCTION
The phenomenon of transsexualism refers to
individuals who are born with the normal sexual
characteristics of one sex, but have the irrefut-
able conviction of belonging to the other.
Nowadays, many professionals who special-
ize in the treatment of transsexuals regard the
conviction of transsexuals as belonging to some-
one of the other sex as authentic and, conse-
quently, their wish for a sex change to be justified.
The recommended procedure of the Harr y
Benjamin International Gender Dysphoria
Association (Meyer et al. 2001), an inter-
national professional organization regardin g
transsexualism, is to approach the refer ral for
sex reassignmen t (SR) in two phases. In the first
phase, a DSM-IV diagnosis (APA, 1994) is
made. In addition, the eligibility of the patient
to move on to the second phase, the Real-life
* Address for correspondence: Prof. Dr P. T. Cohen-Kettenis,
VU University Medical Centre, Department of Medical Psychology,
P.O. Box 7057, 1007 MB Amsterdam, The Netherlands.
(Email: pt.cohen-kettenis@vumc.nl)
Psychological Medicine, 2005, 35, 89–99. f 2004 Cambridge University Press
DOI: 10.1017/S0033291704002776 Printed in the United Kingdom
89
Experience (RLE), is assessed. In this phase the
applicant’s ab ility to live in the desired role and
its consequences, and the strength of the desire
for SR are evaluated. If the social role change
during the RLE , which is usually supported by
hormonal therapy, results in a satisfactory out-
come, the applicant will be referred for surgery
(see Table 1).
Although SR is presently regarded as effective
in treating transsexualism, the most extreme end
of Gender Identity Disorder (GID) (Pfa
¨
fflin &
Junge, 1992, 1998; Eldh et al. 1997; Cohen-
Kettenis & Gooren, 1999 ; Rehman et al. 1999;
Meyer et al. 2001 ; Smith et al. 2001, 2002 ; Day,
2002; Lawrence, 2003), prospective studies are
needed to enhance knowledge about the benefits
and limitations of SR. In spite of strict prior
selection and counselling during treatment, an
estimated 1–2% of those treated express regrets
about SR (Pfa
¨
fflin, 1992; Pfa
¨
fflin & Junge,
1992, 1998 ; Kuiper & Cohen-Kettenis, 1998).
Considering the invasi ve and irreversible treat-
ment of SR, it is imperative to try and prevent
post-operative regret. This requires the identi -
fication of predictors of regret or poor post-
operative functioning. In some follow-up
studies different factors are proposed as influ-
encing the outcomes of SR negatively (e.g.
Wa
˚
linder et al. 1978; Spengler, 1980; Lothstein,
1982; Lundstro
¨
m et al. 1984; Blanchard, 1985;
Lindemalm et al. 1987; Blanchard et al. 1989;
Ross & Need, 1989; Pfa
¨
fflin, 1992 ; Pfa
¨
fflin &
Junge, 1992; Kuiper & Cohen-Kettenis, 1998 ;
Lande
´
n et al. 1998). These factors lie in the area
of psychological functioning, sexual orientation,
age at assessment, onset age of gender dys-
phoria, family history and support, professional
support during SR, and surgical outcomes.
However, the quality of the few existing follow-
up studies is rather poor given their mostly
retrospective nature. Sound prospective studies
are needed to identify predictors of post-
operative functioning more consistently and
reliably.
In the present large-scale prospective foll ow-
up study, we investigated two separate though
related topics : the outcomes of SR and the pre-
diction of favourable or poor outcomes. There-
fore, we first evaluated whether transsexuals
actually improve in important areas after SR,
and confirm some of the beneficial effects of SR
previously establ ished in the mostly retrospec-
tive follow-up studies (e.g. Mate-Kole et al.
1990; Day, 2002; Lawrence, 2003). Secondly,
we studied differences between sexes [male-to-
females (MFs) an d female-to-males (FMs)] and
subtypes (homosexuals and non-homosexuals)
in various areas of functioning (e.g. gender
dysphoria, body dissatisfaction, physical ap-
pearance, psychological functioning) after SR.
To our knowledge, subtype differences have not
yet been prospectively studied. These research
questions were, however, only studied in adult
transsexuals, because the adolescent results
have been published previously (Smith et al.
2001, 2002). Post-operatively, feelings of regret,
evaluation of treatment, satisfaction with surgi-
cal results, social and sexual functioning were
evaluated.
The other topic concerned predictors of the
course and outcomes of SR. We exami ned
which factors clinicians ba sed their referral on
for SR. The factors were age, sex, sexual orien-
tation, onset age of gender dysphoria, gender
dysphoria in childhood and at application,
social support, body dissatisfaction, physical ap-
pearance and psychological functioning. Finally,
we examined which factors predicted treatment
course (i.e. dropping out), post-operative func-
tioning, and treatment satisfaction.
Table 1. Progres sion in the SR procedure and matching sample sizes used in analyses
SR
procedure
Diagnostic
1st phase
a
(Applicants)
Not referred
for 2nd phase
b
(Non-starters)
Referred for,
started, continued
2nd phase
b
(Starters)
Referred for,
started, dropped out
2nd phase
b
(Drop-outs)
Referred for,
completed
SR surgery
(Completers)
1–4 years after
SR surgery
(Follow-up)
MFs 220 74 146 29 117 94
FMs 105 29 76 5 71 64
Included 325 103 222 34 188 158
a
In the first phase a DSM diagnosis is made and eligibility is assessed for starting hormone treatment and the Real-life Experience.
b
In the second phase hormone treatment and the Real-life Experience is started.
90 Y. Smith et al.
METHOD
Subjects
A total of 325 consecutive adolescents and
adults, who applied for SR at VU University
Medical Centre in Amsterdam (VUmc) or Uni-
versity Medical Centre Utrecht (UMCU), par-
ticipated. Of these, 222 (146 MFs, 76 FMs)
started hormone treatmen t : the ‘starter ’ group.
Twenty-nine MFs and 5 FMs stopped hormone
treatment: the ‘drop-out’ group. The group
who completed SR consisted of 188 patients
(117 MFs, 71 FMs) : the ‘completer’ group. The
group who never started hormone treatment
consisted of 103 patients (74 MFs, 29 FMs).
Pre-test data from this ‘no-starter’ group varied
from 89 to 103. At follow-up, some participants
had moved abroad, while others were untrace-
able, which resulted in 158 (94 MFs, 64 FMs)
participants who were interviewed. Follow-up
data ranged between 136 and 158 because not all
participants were willing or able to take part in
both an interview and a questionnaire session
(Table 1).
To examine the outcome issue we used data
of 162 ad ults. Pre-test data were obtained from
all ad ults (104 MFs, 58 FMs; 94 homosexuals,
68 non-homosexuals). Follow-up interview data
were gathered from 126 adults (i.e. 78 %;
77 MFs, 49 FMs; 71 homosexuals, 55 non-
homosexuals). Questionnaire data for different
measures fluctuated from 101 to 126. Since
SR patients do not undergo all possible oper-
ations, data on breast augmentation were
gathered from 52 MFs (21 homosexuals, 31
non-homosexuals), and on meta idoioplasty or
phalloplasty from 10 FMs (4 homosexuals, 6
non-homosexuals). Scores on the Appraisal of
Appearance Inventory were obtained from 57
adults.
Instruments
Biographical data
The Biographical Questionnaire for Trans-
sexuals, a semi-structured interview, contains
211 items on variables, such as cross-gender
feelings and behaviour, social and sexual con-
tacts (Verschoor & Poortinga, 1988; Doorn et al.
1994). The following items were used : sex (MF
or FM), onset age of cross-gender feelings,
childhood GID symptoms (11 items, see below),
age at application, and sexual orientation.
Concerning this last item, participants who
exclusively reported a homosexual preference
(MFs sexually attra cted to biological males ;
FMs to biological fema les) were included in
the homosexual group. Participants with an
asexual, heteros exual, and/or bisexual pref-
erence were included in the non-homosexual
group. Age at the start of hormone therapy and
surgery were obtained from medical files.
The GID in Childhood Scale was constructed
from the Biographical Questionnaire for
Transsexuals (Verschoor & Poortinga, 1988 ;
Doorn et al. 1994) to measure the self-reported
presence of GID symptoms in childhood. There
are 11 items (Cronbach’s a=0
.
81) concerning
the strong wish to be of the opposite sex in early
childhood, cross-gender appearance of the child,
cross-dressing, play and peer preference, and
cross-gender behaviour. Because of differences
in the numbers and types (i.e. quantitative
versus qualitative) of response categories, ans-
wers were dichotomized, resulting in a total score
ranging between 0 and 11, with the higher scores
indicating more childhood GID symptoms.
The Social Support Scale. This scale has 10
items enquiring about the patient’s eight closest
acquaintances (Van Tilburg, 1988). Scalability
coefficient H, calculated by means of a Mokken
analysis (Molenaar, 1982) and calculated for all
relationships together, was 0
.
38. Sumscores
range from 0 to 160, with the higher scores in-
dicating more support.
Gender dysphoria. This was measured with the
Utrecht Gender Dysphoria Scale, containing
12 items on which the subject rated agreement
on a 5-point scale. Scores range from 12 to 60.
Higher scores indicate more gender dysphoria
(Cohen-Kettenis & van Goozen, 1997).
Body dissatisfaction. A Body Image Scale
(Lindgren & Pauly, 1975), adapted for a Dutch
sample (Kuiper, 1991), was used. There are 30
items divided into three subscales: primary and
secondary sex characteristics, and neutral body
parts, with higher scores representing more dis-
satisfaction.
Physical appearance. On the Appraisal of Ap-
pearance Inventory three independent observers
Outcomes and predictors of sex reassignment 91
(the diagnostician, a nurse, the researcher) rated
the subject’s appearance on 14 five-point scales
of gender compatibility. The characteristics
were: hair, facial hair, larynx, voice, figure,
height, skin, hands/feet, muscularity, chin, nos e,
jaw, speech, and gestures/movement. Lower
scores reflect a better appearance in match-
ing the new gender. Intra-class correlation
coefficients between observers for each of the
14 items ranged from 0
.
68 to 0
.
79.
Psychological functioning. The Dutch Short
MMPI (Luteyn et al. 1980) contains 83 items
measuring Negativism, Somatization, Shyness,
Psychopathology, and Extroversion. Higher
scores indicate more dysfunction on the first
four scales but less on Extroversion.
The Dutch version of the Symptom Check
List (Derogati s et al. 1973 ; Arrindell & Ettema,
1986) has 90 items enquiring about recent
complaints. Subscales are: Agoraphobia,
Anxiety, Depression, Somatization, Obsession/
Compulsion, Suspicion, Hostility, and Sleeping
problems. The total score for Psychoneuroticism
ranges from 90 to 450. Higher scores indicate
more psychological instability.
Treatment evaluation and post-treatment func-
tioning. To evaluate post-operative functioning
13 items measured post-operative functioning
and (dis)satisfaction (e.g. with questions about
treatment, regret, social and sexual functioning,
and social experiences) (Doorn et al. 1996).
Patients also completed a questionnaire about
the functioning of vagina/penis and breasts (aug-
mentation or removal), and surgical satisfaction
(Cohen-Kettenis & van Goozen, 1997).
Post-operative Functioning Scale. Twenty-one
items (Cronbach’s a=0
.
87) measured post-
operative functioning and satisfaction with SR
(Doorn et al. 1996) and resulted in a single score
with higher scores reflecting worse functioning
and more dissatisfaction. (See Jo urnal’s website
for Appendix with specific items.)
Procedure
The GID in Childhood Scale and the Social
Support Scale were used at pre-test. Gender
dysphoria, body dissatisfaction, physical ap-
pearance and psychological functioning were
assessed before and after SR to measure change.
The remaining instruments were administered at
follow-up.
Except for the GID in Childhood Scale, the
Social Support Scale, and the Post-operative
Functioning Scale, all instruments were used to
examine the issue of outcome. All instruments
administered at pre-test were used to investigate
predictors of eligibility for and drop-outs of SR.
Follow-up data of the ‘completer’ group were
used to develop the Post-operative Funct ioning
Scale, investigating predictors of outcomes
of SR.
Pre-test data were gathered during the first
diagnostic procedure after the first interview.
Follow-up data were gathered at least 1 year
after surgery. Sessions took between 2 and
3 hours. The Ethics Committees of the UMCU
and VUmc approved the study.
Statistical analyses
Changes over time in treated adults were
analysed with univariate paired-sample t tests,
applying the Bonferroni correction [dividing
the number of tests (19) by 5%] and using a
significance level of 0
.
003 for these results
(Table 2). Post-operative Sex (MF v. FM) and
Subtype (homosexual v. non-homosexual) dif-
ferences were studied with univariate or multi-
variate analyses of variance [(M)ANOV As]
(Table 3). Nominal or ordinal data were ana-
lysed per item with the x
2
test or Mann–Whitney
U test respectively.
To identify which factors predicted eligibility
for hormone treatment and premature drop-
out, logistic regression analyses were performed
with group membership as the criterion variable
(no-starter and starter, completer and drop-out
respectively). Since we had no a priori hypoth-
eses about group prediction, the first (stepwise)
analysis included all 17 factors : age, sex, sexual
orientation, onset age of gender dysphoria, GID
symptoms in childhood, gender dysphoria at
assessment, social support, body dissatisfaction
(3 scales), physical appearance, and psycho-
logical functioning (2 tests: 1 and 5 scales).
Next, we conducted a (simultaneous) logistic
regression analysis using the significant pre-
dictors. In case of unequal sample sizes cut
values were reset to achieve the highest sensi-
tivity and specificity.
92 Y. Smith et al.
Predictors of post-operative functioning were
identified with a multiple linear regression analy-
sis with the Post-operative Functioning Scale as
the dependent variable. Seven of the 17 factors
were relatively independent and included as
predictors in the first (stepwise) analysis: sex,
sexual orientation, physical appearance, sec-
ondary sex characteristic s, extroversion, psycho-
pathology, and psychoneuroticism. Patients
with missing values were deleted listwise. Sig-
nificant predictors were analysed in a second
(simultaneous) multiple linear regression.
RESULTS
Outcomes of the adult transsexuals
Biographical data
The mean age of the transsexuals who com-
pleted SR was 30
.
9 years (range 17
.
7–68
.
1 years)
at application and 35
.
2 years (range 21
.
3–71
.
9
years) at follow-up. Cross-sex hormone treat-
ment started at the mean age of 31
.
6 years
(range 17
.
9–68
.
3 years). The average duration
between starting hormone treat ment and surgery
was 20
.
4 months (range 12–73 months). The
Table 2. Pre-t est and post-test scores of the adult follow-up sample
Pre-test Post-test
Paired Two-tailed
Mean
S.D. Mean S.D. tp
Gender dysphoria 54
.
37
.
114
.
83
.
049
.
5 <0
.
001
Physical appearance 44
.
79
.
633
.
810
.
210
.
9 <0
.
001
Body dissatisfaction
Primary sex characteristics 18
.
12
.
76
.
63
.
225
.
5 <0
.
001
Secondary sex characteristics 34
.
86
.
925
.
26
.
813
.
7 <0
.
001
Neutral body characteristics 46
.
89
.
636
.
58
.
011
.
3 <0
.
001
Psychological functioning
Negativism 22
.
67
.
717
.
17
.
86
.
8 <0
.
001
Somatization 9
.
17
.
66
.
65
.
33
.
10
.
003
Shyness 14
.
79
.
310
.
07
.
35
.
8 <0
.
001
Psychopathology 3
.
23
.
02
.
42
.
62
.
80
.
006
Extraversion 13
.
86
.
515
.
55
.
62
.
90
.
005
Psychoneuroticism 143
.
040
.
7 120
.
331
.
45
.
5 <0
.
001
Anxiety 15
.
25
.
313
.
04
.
54
.
0 <0
.
001
Agoraphobia 9
.
43
.
68
.
63
.
22
.
10
.
040
Depression 29
.
311
.
322
.
58
.
45
.
3 <0
.
001
Somatization 18
.
27
.
016
.
74
.
42
.
30
.
024
Inadequacy 15
.
85
.
813
.
54
.
54
.
1 <0
.
001
Sensitivity 28
.
29
.
124
.
46
.
54
.
4 <0
.
001
Hostility 7
.
82
.
47
.
42
.
11
.
50
.
147
Sleeping problems 5
.
42
.
94
.
62
.
22
.
30
.
024
Table 3. Differences between the adult sexes and subtypes at follow-up
MFs
[mean (
S.D.)]
FMs
[mean (S.D.)]
HOs
[mean (S.D.)]
NHs
[mean (S.D.)]
Sex
F(p)
Subtype
F(p)
Age 38
.
6 (12
.
3) 29
.
6(8
.
3) 31
.
7 (10
.
7) 39
.
6 (11
.
7) 16
.
0(<0
.
001) 6
.
4(0
.
01)
Gender dysphor ia 15
.
4(3
.
1) 13
.
9(2
.
8) 6
.
6(0
.
01) 1
.
3(0
.
27)
Physical appearance 38
.
2(9
.
3) 26
.
0(6
.
9) 28
.
3(<0
.
001) 2
.
0(0
.
16)
Body dissatisfaction 3
.
1(0
.
03) 1
.
2(0
.
33)
Primary sex characteristics 6
.
0(2
.
2) 7
.
6(4
.
2) 7
.
0(0
.
01)
Dutch Short MMPI 2
.
1(0
.
07) 2
.
2(0
.
06)
Somatization 5
.
6(4
.
8) 7
.
9(5
.
5) 4
.
0(0
.
047)
Extraversion 13
.
8(5
.
4) 18
.
0(5
.
0) 17
.
1(5
.
3) 13
.
6(5
.
4) 9
.
2(0
.
003) 6
.
9(0
.
01)
Symptom Check List 2
.
4(0
.
02) 2
.
6(0
.
009)
Depression 24
.
6(9
.
8) 19
.
7(4
.
7) 6
.
5(0
.
01)
Somatization 15
.
2(3
.
7) 18
.
2(4
.
7) 11
.
0(0
.
001)
Sleeping problems 4
.
3(2
.
0) 5
.
0(2
.
3) 5
.
2(0
.
02)
MFs, male-to-female transsexuals; FMs, female-to-male transsexuals; HOs, homosexual transsexuals; HNs, non-homosexual transsexuals.
Outcomes and predictors of sex reassignment 93
average duration between surgery and follow-up
was 21
.
3 months (range 12–47 months).
At follow-up, main effects for Sex and Sub-
type were found for age. FMs and homosexuals
were younger than MFs and non-homosexuals
respectively (Table 3).
At follow-up 5 subjects (4
.
9%) were student s,
38 (36
.
9%) had jobs, 3 (1
.
9%) had retired, and
58 (56
.
3%) were unemployed. The majority
(n=59) lived independently (56
.
2%), 27 sub-
jects (25
.
7%) each lived together with another
adult with or without children, 9 (8
.
6%) were
living with (one of) their parents, 2 (1
.
9%) were
head of an incomplete family, and the remaining
8(7
.
6%) lived in a guest house or boarding
house.
Gender dysphoria. At follow-up there was less
gender dysphoria; the low post-test scores rep-
resent a virtual absence of gender dysphoria
after SR (Table 2). A main effect of Sex was
found with FMs feeling less gender dysphoric.
No Subtype difference in post-operative gender
dysphoria was found (Table 3).
Body dissatisfaction. The majority (n=98,
91
.
6%) were (very) satisfied with their overall
appearance; 9 (8
.
4%) were neutral ; no one was
dissatisfied. Satisfaction with primary sex, sec-
ondary sex, and neutral characteristics had
increased after SR (Table 2). A MANOVA
showed that FMs were more dissatisfied with
their primary sex characteristics at post-test
than MFs. No Subtype differences were found
(Table 3).
Physical appearance. The group scored lower
on the Appraisal of Appearance Inventory at
post-test (Table 2), indicating that their ap-
pearance better matched the new gender. The
physical appearance of FMs was more compat-
ible than that of the MFs, but there was no
Subtype difference (Table 3).
Psychological functioning. At follow-up the
group functioned psycho logically better. Scores
on Negativism and Shyness had improved.
Scores on Somatization, Psychopathology, and
Extroversion showed a tendency towards im-
provement (pf0
.
006). In general, follow-up
scores indicated fewer psychological problems
(Table 2). Comparing pre- and post-test g roup
means with Dutch normative data, most scores
remained within the average range at follow-up,
although Extroversion scores were below aver-
age. Somatization scores were high at pre-test.
The mean Psych oneuroticism score was lower
after SR [see Table 2 for lower scores on four of
the eight subscales (p<0
.
001)]. These scores can
only be compared with Dutch normat ive data
for males and females separately. Both the MF
(p=0
.
001) and FM (p<0
.
001) group showed
improvement in mean scores. The MFs went
from above average at pre-test (mean=143,
S.D.=38
.
0) to average at post-test (mean=123,
S.D.=36
.
0); the FMs went from high (mean=
143,
S.D.=44
.
8) to above average at follow-up
(mean=116,
S.D.=22
.
8).
The Dutch Short MMPI showed a marginally
significant Sex effect, with FMs being more
extrovert. The Sex effect on the Symptom Check
List showed MFs as being more depressed than
FMs (Table 3). There was also a marginally
significant Subt ype effect on the Dutch Short
MMPI, with homosexuals scoring more favour-
ably on Somatization and Extroversion. The
Symptom Check List showed a Subtype effect
with homosexuals scoring lower on Sleeping
Problems and Somatization (Table 3).
Thus, although the groupas a whol e functioned
psychologically rather well at application, their
psychological stability had improved after SR.
In addition, and post-operatively, FMs and
homosexuals functioned psychologically better
than MFs and non-homosexuals respectively.
Treatment evaluation and post-treatment func-
tioning. The vast majority (98
.
4%) expressed
no regrets about SR. One non- homosexual
MF had experienced such strong regrets
during and after treatment that she would not
elect for SR again, if given a second oppor-
tunity. In contrast, a second non-homosexual
MF, who expressed some regrets, reported she
would choose SR again. Five non-homosexuals
(4 MFs, 1 FM) reported some regrets during
treatment only, but expressed no desire or in-
tention to resume their original gender role. No
differences were found between the sexes in re-
ported regret during (Z=x1
.
4, p=0
.
2) or after
SR (Z=x1
.
1, p=0
.
3). During treatment more
non-homosexuals reported feelings of regret
(Z=x3
.
1, p=0
.
002).
94 Y. Smith et al.
Social life and social contacts. The majority
(n=90, 89
.
1%) felt accepted by most people, 8
(7
.
9%) by some, 3 (3%) by no one. Altogether
84 individuals (83
.
2%) felt supported in their
new gender role by (almost) everyone they
knew, whereas 11 (10
.
9%) felt supported by
some people. Despite the fact that 6 subjects
(5
.
9%) did not feel supported, they were able
to rely on some individuals during difficult
times. Four subjects (3
.
9%) had no one to
turn to when times got hard. Still, the vast
majority (99, 96
.
1%) could rely on at least
some others during difficult times. In total,
18 individuals (17
.
3%) sometimes felt they
were being laughed at, 2 (1
.
9%) had experi-
enced being ridiculed by strangers; 84 (80
.
8%)
had never experienced any such adverse reac-
tions. Over 98% (n=102) felt they were com-
pletely taken seriously by most people. Two
(1
.
9%) only felt taken seriously by a few close
friends. No one reported not being taken
seriously by anyo ne.
MFs and FMs felt equally accepted
(Z=x 0
.
8, p=0
.
4). However, FMs had more
support in the new gender role (Z=x 2
.
5,
p=0
.
01) and were more able to rely on signifi-
cant others during difficult times (Z=x 2
.
2,
p=0
.
03). Although MFs were more often
laughed at or ridiculed (Z=x 3
.
5, p<0
.
001),
they reported feeling taken equally seriously by
(almost) all people (Z=x 1
.
7, p=0
.
08). Homo-
sexuals felt more supported (Z=x 2
.
0, p=0
.
04)
and taken more seriously than non-homo-
sexuals (Z=x 2
.
5, p=0
.
01).
Relationships and sexuality. The majority (n=
46, 88
.
5%) of the 50 subjects who had a steady
sexual partner were satisfied with their sex life, 3
(5
.
8%) expressed a neutral view, and 3 (5
.
8%)
were dissa tisfied. Of the 84 subjects (82
.
4% of
the follow-up sample) who were sexually active,
the majority (53, 63
.
1%) achieved orgasm al-
ways or regularly, 16 (19%) sometimes, and 15
(17
.
9%) never.
A larger percentage (x
2
=4
.
2, p=0
.
04) re-
ported a homosexual (94, 58
.
0%) than a non-
homosexual orientation (68, 42
.
0%). Within the
FMs a greater proportion (x
2
=5
.
9, p=0
.
015)
had a homosexual orientation (70
.
7%) than the
MFs (51
.
0%). More of the sexually active FMs
(81
.
6%) than of the MFs (42
.
1%) achieved
orgasm always or regularly (Z=x 2
.
4, p=0
.
01).
Yet, both sexes reported equal satisfaction with
their sex life (Z=x0
.
6, p=0
.
5). No Subtype
differences were found.
Satisfaction with surgery. For FMs breast
removal is emotionally the most important sur-
gery. They are advised to postpone metaidoio-
plasty (transformation of the hypertrophic
clitoris into a micropenis) or phalloplasty in
view of the fact that surgical techniques are
steadily improving. Eleven FMs (28
.
9%) were
satisfied with their breast removal, 5 (13
.
2%)
were dissatisfied due to the visibility of the scars,
and 22 (57
.
9%) were not completely satisfied.
Four FMs were satisfied with their metaidoio-
plasty or phalloplasty. One FM was dissatisfied
because of urinary problems, while four were
not completely satisfied.
For the MFs vaginoplasty is the most im-
portant surgical intervention. The majority of
MFs (47, 70
.
1%) were satisfied; 15 (22
.
4%)
were not completely satisfied, mostly because
they considered their vagina not deep or femi-
nine enough. Five MFs (7
.
5%) were dissatisfied,
because they were unable to achieve sexual
arousal or orgasm, or because corrective surgery
was needed. The majority (34, 65
.
4%) were
satisfied with their breast augment ation ; 15
(28
.
8%) were not co mpletely satisfied, and three
felt uneasy about their breasts being too far
apart.
Predictors of the course and outcomes of
adolescent and adult transsexuals
Prediction of elig ibility criteria
Eligibility for SR was largely based upon the
factors gender dysphoria, psychoneuroticism,
and physical appearance. For the precise weight
of each predictor variable and the constant in
this equation model, see Table 4. Stronger
gender dysphoria (higher scores), more psycho-
logical stability (lower scores on Psycho-
neuroticism), and a feminine look for MFs and
a masculine look for FMs (lower scores on
physical appearance), increased the probability
of the clinician referring the applicant to start
hormone treatment. With these three predic-
tor variables 78% of the applicants were cor-
rectly assigned to the ‘no-starter’ (52%) or the
‘starter ’ (88 %) group (cut value=0
.
63).
Outcomes and predictors of sex reassignment 95
Prediction of the course of treatment
The probability that a transsexual discontinued
hormone treatment depended on sex, psycho-
pathology, childhood GID symptoms, and gen -
der dysphoria (Table 4). A negative coefficient
means that a factor contributes negatively to the
probability of being a drop-out. The relatively
high beta value of the factor sex reflects being a
FM. Thus, the combinat ion of being a MF with
more psychopathology and childhood GID
symptoms, yet less gender dysphoria at assess-
ment, increased the likelihood of premature
drop-out. Together these four predictors cor-
rectly assigned 68% of the transsexuals to the
‘completer’ (68%) or the ‘ dr op-out ’ (69 %)
group (cut value=0
.
15).
Prediction of post-operative functioni ng
The level of post-operative functioning could be
predicted by the patient’s sexual orientation,
psychological stability, and dissatisfaction with
secondary sex characteristic s at assessment. The
beta weights (see Beta column in Table 5) show
the relative importance of the variables con-
tributing to the predictability of the quality of
post-operative functioning (R
2
=0
.
17). As high-
er scores on the Post-operative Functioning
Scale reflect more dysfunction and dissatis-
faction, the predicted score of an applicant on
this scale increased with a non-homosexual
orientation, more psychopatholog y and dissat-
isfaction with secondary sex characteristics at
assessment.
DISCUSSION
One aim of this prospective study was to inves-
tigate which areas of functioning improve as a
consequence of SR. The main symptom for
which the patients had requested treatment,
gender dysphoria, had decreased to such a de-
gree that it had disappeared. Satisfaction of the
patients with their sex characteristics had im-
proved to the point of content, confirming pre-
vious results (Green & Fleming, 1990 ; Pfa
¨
fflin &
Junge, 1998), and according to observers, their
appearance better matched the new gender.
Psychological functioning had also improved
(see Mate-Kole et al. 1990). Thus, it seems safe
to conclude that the transsexuals had improved
in important areas of functioning and that 1–4
years after surgery, SR appeared therapeutic
and beneficial. Furthermore, the vast majority
expressed no regrets about their SR.
Post-operative evaluation showed that the
majority functioned quite well socially. A small
minority, however, lacked support and accept-
ance, and were ridicu led. Surprisingly, 98 % felt
taken seriously. This somewhat rose-coloured
view may be explained by the fact that the social
support received and the relief about the new
situation may have put adverse reactions into
perspective, whereas disappointing experiences
may have been played down to reduce co gnitive
dissonance after undergoing such invasive and
irreversible interventions.
At follow-up, the majority were content with
their sex life, and those who were sexually
active, reported achieving orgasm. This has
been reported previously (e.g. Rakic et al. 1996;
Rehman et al. 1999), but in MFs the capacity
Table 4. B coefficients and constants of the
factors predicting group membership
Predictor variables
Starter group Drop-out group
Bpvalue Bpvalue
Sex x1
.
82 0
.
006
Sexual orientation
Onset age of gender dysphoria
Age at application
GID symptoms in childhood 0
.
18 0
.
026
Gender d ysphoria 0
.
08 <0
.
001 x0
.
05 0
.
030
Social support
Primary sex characteristics
Secondary sex characteristics
Neutral sex characteristics
Physical appearance x0
.
05 0
.
003
Psychoneuroticism x0
.
01 <0
.
001
Negativism
Somatization
Shyness
Psychopathology 0
.
12 0
.
024
Extroversion
Constant 1
.
00 0
.
442 x0
.
04 0
.
972
Table 5. Factors predicting post-operative
functioning
Model B Beta p value
Sexual orientation x3
.
70 x0
.
24 0
.
002
Psychopathology 0
.
43 0
.
17 0
.
028
Dissatisfaction secondary
sex characteristics
0
.
31 0
.
28 <0
.
001
Constant 16
.
80 <0
.
001
96 Y. Smith et al.
for orgasm has been reported to decrease post-
surgically (Lief & Hubschman, 1993).
The findings supp ort the conclusion that after
SR most transsexuals functioned socially and
sexually well. One MF expressed deep regrets.
She indicated that professional guidance re-
garding adverse consequences (i.e. intolerance
of society, family and her own children), would
have made the transition more endurable. This
stresses the need for good aftercare.
Comparing the sexes, the FMs showed better
results, supporting the results of earlier studies
(see Introduction). This might be due to their
more convinci ng gender role behaviour and
looks and their ‘ type ’ of transsexualism, imply-
ing an earlier age at application. More FMs
than MFs were capable of achieving orgasm.
This can be attributed to hormonal effects (van
Goozen et al. 1995) or to the fact that most FMs
lived with their enlarged clitoris. Then again, it
may also portray different meanings of sexuality
in males and females, since both sexes reported
equal satisfaction with their sex life. Contrasting
most of the more favourable FM findings are
the greater reported satisfaction of the MFs
with surgical results. This is unde rstandable
given that most FMs did not (yet) have a penis.
For the FMs the ability to live in the new gender
and sexual role clearly awaits the advancement
of surgical techniques.
With respect to subtype differences, homo-
sexuals were younger and functioned psycho-
logically better than non-homosexuals. No dif-
ferences were found in gender dysphoria, body
dissatisfaction, or physical appearance. Only
non-homosexuals reported some regrets during
treatment, and two during and after SR, which
they all related to a lack of acceptance and
support from others. The better functioning
of homosexual s may also be explained by their
sexual orientation. Subtype differences could
reflect different aetiological backgrounds. Be-
cause the onset age and age at application have
been found to be earlier in homosexuals, it is
likely that non-homosexuals encounter more
problems in life before applying for SR. Also,
post-surgically, ‘ homosexuals’ will have op-
posite gender partners, thus forming hetero-
sexual couples. This still is socially more
acceptable.
The less favourable outcomes of the non-
homosexuals carry significant implications for
clinical practice. If considered eligible for SR,
non-homosexuals should be able to receive
additional guidance in coping with adverse
consequences, such as a more troubled psycho-
logical functioning, or a more critical environ-
ment.
In conclusion, our data substantiate findings
from mostly retrospective follow-up studies that
SR is effective. Some individuals probably need
a more thorough diagnostic procedure and
more therapeutic support, sometimes even after
treatment, than is currently the case. For most
transsexuals in this study, the strict eligibility
criteria and profes sional guidance as currently
provided appears to be sufficient, as reflected by
the overall favourable outcomes of SR. How-
ever, alleviation of the gender problem is not
equivalent with an easy life. Apparently, clin-
icians need to be alert for signs that a trans -
sexual applicant will not be able to cope with
adversities during treatment.
Another goal was to identify predictors of
the course and outcomes of SR. We found that
clinicians assessed applicants to be eligible for
hormone treatment when they were more gen-
der dysphoric, psychologically more stable, and
when the physical appearance better matched
the new gender role. Given the nature of the
problem, it is not surprising that strong gender
dysphoria was one of the main predictors. Since
an unfavourable physical appearance could be a
risk factor for post-operative regret (Wa
˚
linder
et al. 1978), it is interesting to observe that the
clinicians also took this factor into account
when deciding upon referral. Furthermore,
clinicians greatly valued the applicant’s psycho-
logical functioning (see also Kuiper & Cohen-
Kettenis, 1998). These factors predicted 88 % of
the starter group. Clearly, clinicians must have
had other reasons for referring the remaining
12%, the most likely factor being the diagnosis.
They might also have appraised certain risk
factors as relatively harmless in view of existing
protective factors (e.g. strong social support,
adequate coping skills).
We found transsexuals to be more at risk
for dropping out of treatment when they were
MFs, showed more psychopathology, more
GID symptoms in childhood, yet less gender
dysphoria at application. The greater vulner-
ability of MFs to drop out is understandable
given that FMs fare better post-operatively.
Outcomes and predictors of sex reassignment 97
Unfortunately, our data do not permit us to
distinguish during treatment between the impac t
of psychopathology , on the one hand, and of
interactive effects of psychopathology with ex-
ternal forces, on the other. We cannot rule out
the possibility that it is not psychopathology
per se that increases the probability to drop out,
but rather a combination of psychological vul-
nerability and personal circum stances, such as
abandonment by a partner. One should also
bear in mind that the drop-outs stopped hor-
mone treatment during our data collection; it
is possible, however, that they will reapply later
in life.
Our finding of an association between more
childhood GID symptoms an d greater drop
out seems puzzling. It is in contrast with the
literature on SR risk factors and clinically
counter-intuitive. Early gender dysphoria has
been associated with early-onset transsexualism
and favourable SR outcome (see Lothstein,
1982; Blanchard, 1985; Lindemalm et al. 1987 ;
Blanchard et al. 1989; Pfa
¨
fflin, 1992). Here, it is
the combination of factors that is crucial. Still,
the contradicting presence of more gender
dysphoria in childhood but less at application
should alert the clinician when assessing eligi-
bility. This inconsistency may reflect confusion
about development, an (unconscious) exagger-
ation of history if current feelings are not
clear-cut, or a conscious effort to mislead the
clinician.
Finally, we investigated which assessment
factors predicted post-operative functioning. It
is important to bear in mind that we applied a
continuous scale from good to bad, as opposed
to the dichotomy ‘no regret ’–‘ regret ’, because
hardly any transsexuals reported regret. A non-
homosexual orientation, with more psycho-
pathology and dissatisfaction with secondary
sex characteristics predicted unfavourable post-
operative functioning. The finding that non-
homosexuals and those with more psychological
instability are at risk for unfavourable func-
tioning and more dissatisfaction after SR fits
with earlier studies (see Introduc tion ; Blanchard
et al. 1989; Lande
´
n et al. 1998). We found that
two non-homosexual s expressed regret about
SR. Finally, dissatisfaction with appearance
predicted poor post-o perative functioning,
either because it directly and adversely affected
psychological stabili ty or mood, or it indirectly
affected the way they were socially treated (or a
combination of both).
Taking all the findings into account, our
‘sample’ of clinicians appropriately assessed
some risk factors that predict the course and
outcomes of SR, yet they underestimated
others. They particularly recognized the impact
of the applicant’s psychological functioning and
physical appearance on post-operative func-
tioning. However, clinicians might want to take
special notice of MFs who report inconsistencies
in past and present gender dysphoria, in the
presence of psychopathology, and of non-
homosexuals with strong dissatisfaction about
their appearance and clear psychopathology.
They may benefit from additional guidance after
SR, while adjusting to their new lives and coping
with unexpected or adverse consequences.
The results of this study subscribe to the sig-
nificance of some of the risk factors described in
the literature with more conclusive data. Fur-
thermore, factors were found that could assist
clinicians in identifying individuals who might
be at risk for poor outcome.
ACKNOWLEDGEMENTS
The authors thank Mrs W. Harmsen and Mr
Jos Megens in particular for their invaluable
help in the data collection process and appreci-
ate the contribution of Dr Anton M. Verschoor
in the setting up of this research project. This
work was financially supported by the Stichting
Fondsenwervingsacties, the Ziekenfondsraad,
and the Stichting Nederlands Gender Centrum.
DECLARATION OF INTEREST
None.
NOTE
An Appendix accompanies this paper on the
Journal’s website (http//journals.cambridge.
org).
REFERENCES
APA (1994). Diagnostic and Statistical Manual of Mental Disorders
(4th edn). American Psychiatric Association: Washington, DC.
Arrindell, W. A. & Ettema, J. H. M. (1986). SCL-90 : Handleiding bij
een multidimensionele psychopathologie-indicator [SCL-90, Manual
98 Y. Smith et al.
of a multidimensional psychopathology-indicator]. Swets en Zei-
tlinger: Lisse, The Netherlands.
Blanchard, R. (1985). Typology of male-to-female transsexualism.
Archives of Sexual Behavior 14, 247–261.
Blanchard, R., Steiner, B. W., Clemmensen, L. H. & Dickey, R.
(1989). Prediction of regrets in postoperative transsexuals.
Canadian Journal of Psychiatry 34, 43–45.
Cohen-Kettenis, P. T. & Gooren, L. J. G. (1999). Transsexualism : a
review of etiology, diagnosis and treatment. Journal of Psycho-
somatic Research 46, 315–333.
Cohen-Kettenis, P. T. & van Goozen, S. H. M. (1997). Sex reassign-
ment of adolescent transsexuals: a fo llow-up study. Journal of the
American Academy of Child and Adolescent Psychiatry 36,
263–271.
Day, P. (2002). Tech Brief Series : trans-gender reassignment surgery.
New Zealand Health Technology Assessment Report 1, 1–38.
Derogatis, L. R., Lipman, R. S. & Covi, L. (1973). SCL-90: an
outpatient psychiatric rating scale – preliminary report. Psycho-
pharmacology Bulletin 9, 13–27.
Doorn, C. D., Kuiper, A. J., Verschoor, A. M. & Cohen-Kettenis,
P. T. (1996). Het verloop van de geslachtsaanpassing: Een 5-jarige
prospectieve studie [The course of sex reassignment : A 5-year pro-
spective study]. Report for the Dutch National Health Council.
Doorn, C. D., Poortinga, J. & Verschoor, A. M. (1994). Cross-gender
identity in transvestites and male transsexuals. Archives of Sexual
Behavior 23, 185–201.
Eldh, J., Berg, A. & Gustafsson, M. (1997). Long-term follow up after
sex reassignment surgery. Scandinavian Journal of Plastic and
Reconstructive Surgery and Hand Surgery 31, 39–45.
Green, R. & Fleming, D. (1990). Transsexual surgery follow-up :
status in the 1990s. Annual Review of Sex Research 1, 163–174.
Kuiper, A. J. (1991). Transseksualiteit : Evaluatie van de geslachts-
aanpassende behandeling [Transsexualism: An evaluation of sex
reassignment]. Elinkwijk: Utrecht, The Netherlands.
Kuiper, A. J. & Cohen-Kettenis, P. T. (1998). Gender role reversal
among postoperative transsexuals. International Journal of Trans-
genderism 2 (http://www.symposion.com/ijt/ijtc0502.htm).
Lande
´
n, M., Wa
˚
linder, J., Hambert, G. & Lundstro
¨
m, B. (1998).
Factors predictive of regret in sex reassignment. Acta Psychiatrica
Scandinavica 97, 284–289.
Lawrence, A. A. (2003). Factors associated with satisfaction or regret
following male-to-female sex reassignment surgery. Archives of
Sexual Behavior 32, 299–315.
Lief, H. & Hubschman, L. (1993). Orgasm in the postoperative
transsexual. Archives of Sexual Behavior 22, 145–155.
Lindemalm, G., Ko
¨
rlin, D. & Uddenberg, N. (1987). Prognostic fac-
tors vs. outcome in male-to-female transsexualism: a follow up of
13 cases. Acta Psychiatrica Scandinavica 74, 268–274.
Lindgren, T. & Pauly, I. (1975). A body image scale for evaluating
transsexuals. Archives of Sexual Behavior 4, 639–656.
Lothstein, L. M. (1982). Sex reassignment surgery : historical,
bioethical, clinical and theoretical issues. American Journal of
Psychiatry 139, 417–426.
Lundstro
¨
m, B., Pauly, I. & Wa
˚
linder, J. (1984). Outcome of sex
reassignment surgery. Acta Psychiatrica Scandinavica 70, 289–294.
Luteyn, F., Kok, A. R. & van der Ploeg, F. A. E. (1980). NVM,
Nederlandse verkorte MMPI, Handleiding [Dutch short version of
the Minnesota Multiphasic Personality Inventory, Manual]. Swets
en Zeitlinger : Lisse, The Netherlands.
Mate-Kole, C., Freschi, M. & Robin, A. (1990). A controlled study of
psychological and social change after surgical gender reassignment
in selected male transsexuals. British Journal of Psychiatry 157,
261–264.
Meyer III, W., Bockting, W. O., Cohen-Kettenis, P. T., Coleman, E.,
DiCeglie, D., Devor, H., Gooren, L. J. G., Hage, J. J., Kirk, S.,
Kuiper, A. J., Laub, D., Lawrence, A., Menard, Y., Patton, J.,
Schaefer, L., Webb, A. & Wheeler, C. C. (2001). The standards of
care for gender identity disorders (6th Version). International
Journal of Transgenderism 5 (http://www.symposion.com/ijt/soc/
index.htm).
Molenaar, I. W. (1982). Mokken scaling revisited. Kwantitatieve
methoden 3, 145–164.
Pfa
¨
fflin, F. (1992). Regrets after sex reassignment surgery. Journal of
Psychology & Human Sexuality 5, 69–85.
Pfa
¨
fflin, F. & Junge, A. (1992). Geschlechtsumwandlung: Abhandlun-
gen zur Transsexualita
¨
t. Schattauer : Stuttgart, Germany.
Pfa
¨
fflin, F. & Junge, A. (1998). Sex Reassignment : Thirty years of
international follow-up studies SRS : A Comprehensive Review,
1961–1991 (English edn). Symposion Publishing : Du
¨
sseldorf,
Germany (http://www.symposion.com/ijt/pfaefflin/1000.htm).
Rakic, Z., Starcevic, V., Maric, J. & Kelin, K. (1996). The outcome of
sex reassignment surgery in Belgrade : 32 patients of both sexes.
Archives of Sexual Behavior 25, 515–525.
Rehman, J., Lazer, S., Benet, A., Schaefer, L. & Melman, A. (1999).
The reported sex and surgery satisfactions of 28 postoperative
male-to-female transsexual patients. Archives of Sexual Behavior
28, 71–89.
Ross, M. W. & Need, J. A. (1989). Effects of adequacy of gender
reassignment surgery on psychological adjustment : a follow-up of
fourteen male-to-female patients. Archives of Sexual Behavior 18,
145–153.
Spengler, A. (1980). Kompromisse statt stigma und unsicherheit.
Transsexuelle nach der operation [Compromises instead of
stigma and doubts. Transsexuals after surgery]. Sexualmedizin 9,
98–103.
Smith, Y. L. S., Cohen, L. & Cohen-Kettenis, P. T. (2002). Post-
operative psychological functioning of adolescent trans-
sexuals: a Rorschach study. Archives of Sexual Behavior 31,
255–261.
Smith, Y. L. S., van Goozen, S. H. M. & Cohen-Kettenis, P. T. (2001).
Adolescents with gender identity disorder who were accepted or
rejected for sex reassignment surgery: a prospective follow-up
study. Journal of the American Academy of Child and Adolescent
Psychiatry 40, 472–481.
van Goozen, S. H. M., Cohen-Kettenis, P. T., Gooren, L. J. G.,
Frijda, N. H. & van de Poll, N. E. (1995). Gender differences in
behavior: activating effects of cross sex hormones. Psychoneuro-
endocrinology 20, 343–363.
Van Tilburg, T. (1988). Verkregen en gewenste ondersteuning in het
licht van eenzaamheidservaringen [Received and desired support
related to the experience of loneliness]. Elinkwijk : Utrecht, The
Netherlands.
Verschoor, A. M. & Poortinga, J. (1988). Psychosocial differences
between Dutch male and female transsexuals. Archives of Sexual
Behavior 17, 173–178.
Wa
˚
linder, J., Lundstro
¨
m, B. & Thuwe, I. (1978). Prognostic factors in
the assessment of male transsexuals for sex reassignment. British
Journal of Psychiatry 132, 16–20.
Outcomes and predictors of sex reassignment 99
