*Manuscript

Systematic review and meta-analysis of prevalence studies in transsexualism

Jon Arcelus, MD, PhD1,2, Walter Pierre Bouman MD1, Wim Van Den Noorgate PhD3,

Laurence Claes, PhD4, Gemma Witcomb PhD2, & Fernando Fernandez-Aranda

PhD5,6

1
Nottingham Centre for Gender Dysphoria, Nottingham, United Kingdom
2
School of Sport, Exercise, and Health Sciences, Loughborough University, United

Kingdom
3
Centre for Methodology of Educational Research,Katholieke Universiteit, Leuven,

Belgium

΀ Faculty of Psychology and Educational Sciences, KU Leuven, Leuven, Belgium

΁ Department of Psychiatry, University Hospital of Bellvitge-IDIBELL, Barcelona,

Spain

΂ CIBER Fisiopatología Obesidad y Nutrición (CIBERobn), ISCIII, Barcelona, Spain

Corresponding author:
Dr Walter Pierre Bouman
Nottingham Centre for Gender Dysphoria
Nottinghamshire Healthcare NHS Trust
3 Oxford Street
Nottingham, NG1 5BH
United Kingdom
Telephone: +44(0)115 876 0160
Fax: +44(0)115 947 5609

[email protected]
 2

Conflict of Interest: The authors have no relevant conflict of interest to declare.

Statement of Authorship
Category 1
(a) Conception and Design
Jon Arcelus, Laurence Claes; Walter Pierre Bouman; Gemma Witcomb; Wim

Van Den Noorgate

(b) Acquisition of Data

Jon Arcelus; Gemma Witcomb

(c) Analysis and Interpretation of Data

Jon Arcelus; Wim Van Den Noorgate; Laurence Claes

Category 2
(a) Drafting the Article
Jon Arcelus ; Walter Pierre Bouman; Laurence Claes ; Gemma Witcomb ;

Fernando Fernandez-Aranda

(b) Revising it for Intellectual Content

Jon Arcelus ; Walter Pierre Bouman; Laurence Claes ; Gemma Witcomb ;

Fernando Fernandez-Aranda

Category 3
(a) Final Approval of the Completed Article

Jon Arcelus; Walter Pierre Bouman; Wim Van Den Noorgate; Laurence

Claes; Gemma Witcomb; Fernando Fernandez-Aranda

 3

Abstract

Background: Over the last 50 years several studies have provided estimates of the

prevalence of transsexualism. The variation in reported prevalence is considerable and

may be explained by factors such as the methodology and diagnostic classification

used and the year and country in which the studies took place. Taking these into

consideration, this study aimed to critically and systematically review the available

literature measuring the prevalence of transsexualism as well as performing a meta-

analysis using the available data.

Methods: Databases were systematically searched and 1473 possible studies were

identified. After initial scrutiny of the article titles and removal of those not relevant,

250 studies were selected for further appraisal. Of these, 211 were excluded after

reading the abstracts and a further 18 after reading the full article. This resulted in 21

studies on which to perform a systematic review, with only 12 having sufficient data

for meta-analysis. The primary data of the epidemiological studies were extracted as

raw numbers. An aggregate effect size, weighted by sample size, was computed to

provide an overall effect size across the studies. Risk ratios and 95% confidence

intervals (CIs) were calculated. The relative weighted contribution of each study was

also assessed.

Results: The overall meta-analytical prevalence for transsexualism was 4.6 in

100,000 individuals; 6.8 for trans women and 2.6 for trans men. Time analysis found

an increase in reported prevalence over the last 50 years.

Conclusions: The overall prevalence of transsexualism reported in the literature is

increasing. However, it is still very low and is mainly based on individuals attending

clinical services and so does not provide an overall picture of prevalence in the

general population. However, this study should be considered as a starting point and
 4

the field would benefit from more rigorous epidemiological studies acknowledging

current changes in the classification system and including different locations

worldwide.

Keywords: Gender Dysphoria, transgender, population, systematic review, meta-

analysis
 5

1. Introduction

Transsexualism, as defined by the International Classification of Disease (ICD-10)(1)

and the previous editions of the Diagnostic and Statistical Manual of Mental

Disorders(2,3), describes individuals who experience discomfort or distress caused by

the discrepancy between their gender identity and the sex they were assigned at

birth. When this distress is sufficiently intense individuals wish to transition from one

point on a notional gender scale to another – most commonly from a man to a woman

(people known as trans women) or from a woman to a man (people known as trans

men(4,5). The diagnosis of transsexualism according to the ICD-10(1) is currently under

revision. It is proposed that the new edition of the ICD (ICD-11) will include a new

diagnostic term and will also include individuals who do not fit into the gender binary

category(6), as is the case in the DSM-5(7).

To estimate the prevalence of trans individuals is relevant for health service

development and policymaking, although this can be complex due to several factors.

Some of those factors relate to the complexity of undertaking general prevalence

studies, i.e. the fact that diagnoses change over the years or that results differ

depending on the period of time used to collect data. For example, point prevalence is

a measure of the proportion of people in a given population at an exact time point,

such as a particular date, which is in contrast to period prevalence that measures the

proportion of people in a given population over a specific time period, for example

several years(8).

Other factors that also add to the complexity of undertaking epidemiological studies

relate to the subject studied; in this case the number of transsexual individuals in the
 6

community. For example, the terminology and classification systems used have varied

over the years, and authors have used some of this terminology inconsistently, for

example, Gender Identity Disorder, Gender Dysphoria, or Transsexualism.

Nevertheless, it is reassuring (in relation to performing this meta-analysis) to find that

there are many epidemiological studies that have used the term “transsexualism” and

which have followed the ICD or DSM diagnostic criteria (9,1,2,3,10,11,7) or the definition

developed by Benjamin in 1966(12), which requires all of the following:

1) A sense of belonging to the opposite sex, of having been born into the wrong

sex, or being one of nature’s extant errors.

2) A sense of estrangement from one’s own body: all indications of sex

differentiation are considered as afflictions and repugnant.

3) A strong desire to resemble physically the opposite sex via therapy, including

surgery.

4) A desire to be accepted by the community as belonging to the opposite sex.

The definition of transsexualism is different in the DSM-IIIR(3). This edition of the

diagnostic criteria did not include those individuals who were not interested in

undergoing sex reassignment surgery and who received the alternative diagnosis of

Gender Identity Disorder of Adolescents and Adulthood, Nontranssexual Type

(GIDAANT). This great variation in how transsexualism is defined in different

studies (based on the diagnostic criteria used at the time that the study took place) will

clearly affect the reported prevalence.

 7

A second complexity in developing true epidemiological studies of transsexualism

relates to the methods used to identify this population. For example, the country

where the study takes place can influence the prevalence of individuals found, as trans

people tend to live in larger cities and especially in areas, or countries, which are

defined as “trans friendly”(13). Therefore studies from specific countries may describe

a high prevalence of transsexual individuals, which may not be generalizable to other

countries. Thirdly, the timing of the study may also affect the findings. The fact that

in some countries tolerance to trans individuals has improved over the years has

allowed trans people to “come out” more easily in order to access clinical services (14,
15)
. This may be reflected by the fact that older studies(16) report lower prevalence

than more recent ones(17). Finally, the recruitment process used to collect

epidemiological data will also influence the findings. Many studies are based on

clinical populations of individuals which, by definition, only include those who have

the capacity and motivation to ask for help, but, importantly, can also access clinical

services(18). This is reflected in the large number of studies from the Netherlands(18,19)

where trans services have been available for many years and where society is

generally tolerant (14, 15).

Therefore, as highlighted by the WPATH Standards of Care(20) and other authors(21)

efforts to formally calculate the prevalence of transsexualism present with enormous

difficulties, due to the differences in cultural manifestation of gender behavior. It is

because of this that researchers who have studied prevalence rates have focused on

the most easily counted subgroup of gender-nonconforming people - those who

present for gender-transition-related care at specialist gender identity clinics(21). As a

consequence many prevalence studies published in this field have their origins in the
 8

Western world(22). There is only one prevalence study from the Eastern world(23),

which is surprising given that many countries like Thailand, India and Pakistan are

known to have an apparently tolerant culture towards trans* identities, although there

are many ethnographic related studies(24,25,26).

The large number of studies investigating prevalence of transsexual individuals, and

reviewed in this study, provide the best available insight into the rates of

transsexualism, and vary from 0.45(22) to 23.6(23) per 100,000 people. Although

prevalence studies are welcome, such a great variation in findings leaves the reader

confused. Therefore, the aim of this study was to respond to the reported variation in

prevalence by critically and systematically reviewing prevalence studies in

transsexualism. Where data were available, a meta-analysis of the studies was carried

out which took population, diagnosis, the time period studied and gender into

consideration.

2. Methods

2.1. Search strategy

This meta-analytic review adheres to the guidelines detailed in the PRISMA

Statement(27). A systematic literature search, appraisal and meta-analysis was

conducted using a broad range of subject headings in order to identify relevant

prevalence studies in the field. The following data bases were used: Web of Sciences,

Medline/Pubmed, Biosis, Science Direct, and Scielo. For each database, combinations

of the following search components were used: Transsexual, Transgender, Gender

Dysphoria, Gender Identity Disorder, Gender non conforming, Gender Variant,

epidemiology, incidence, and prevalence. Studies published between 1945 and June
 9

2014 were selected. Two researchers independently selected the studies, extracted the

data, cross-checked them and resolved disagreements. Case studies or studies

describing small populations of individuals which could not be proved to reflect the

prevalence of a given area, region or country were excluded. Only studies describing

adolescent or adult populations were included. Reference lists of relevant articles

were screened for further potential studies and citation searches were conducted. Only

studies describing transsexualism as per Benjamin(12), ICD(1,9) or DSM(2,3,7,10,11)

definitions were selected. Table 1 details the criteria for search used for this review.

Table 1 insert around here

When the study did not describe some of the above information, whenever possible,

this was calculated by the authors. For example, in some cases the mean population of

the studied area was calculated. Studies were excluded where there was ambiguity in

the number of individuals with a clear diagnosis or studies that primarily included

individuals who were self-diagnosed or had not been diagnosed by a professional(19,

28)
.

2.2. Procedure

Studies meeting the inclusion criteria were examined. The study collected the

available information for the following outcomes of interest: 1) prevalence of

transsexual individuals in general per 100,000 individuals; 2) prevalence of trans

women per 100,000 individuals; 3) prevalence of trans men per 100,000 individuals;

4) sex ratio between males and females; 5) country or region where the study took

place; 6) number of years during which information was collected; 7) information

 10

regarding methods used to identify trans individuals; and 8) the change in the trans

men/trans women ratio as a function of time (descriptive information).

2.3. Included and excluded studies

Studies were screened in three phases, namely title, abstract, and full text. In the first

instance the titles were screened (n= 1724). The number of studies found using the

above search terms were: Transgender (TG) plus epidemiology (259), TG plus

incidence (28) and TG plus prevalence (257). Transsexualism (TS) plus epidemiology

(143), TS plus incidence (51) and TS plus prevalence (130); Gender dysphoria (GD)

plus epidemiology (24), GD plus incidence (4), GD plus prevalence (11); Gender

identity disorder (GID) plus epidemiology (390), GID plus incidence (41) and GID

plus prevalence (384). Gender non conforming (GNC) plus epidemiology (1), GNC

plus incidence (0) and GNC plus prevalence (1). Duplicates were removed (n=251),

and the two independent reviewers (JA and GW) independently screened and coded

the remaining titles (n=1473). Based on the titles, 1223 papers were excluded. The

main reason for exclusion was that the studies did not describe prevalence of

Transsexualism, Gender Identity Disorder or Gender Dysphoria, but prevalence of

other disorders such as mental health problems or HIV in trans individuals. Out of the

250 studies selected to be screened in detail, 211 were excluded after reading the

abstracts. The reasons for the exclusions were: inadequate sample size (case studies);

2) no prevalence data; and 3) no specific area, region or country covered by the study.

Out of the 39 papers retrieved for more detailed evaluation, five were excluded as

there was no exact epidemiological information or they were not covering a specific,

identifiable area (these were not excluded in the previous stage as this information

was not identifiable by reading the abstract). Four more studies were excluded as they
 11

only described the prevalence of psychiatric co-morbidity, and a further five studies

were excluded as the population was self-identified or it was unclear as to whether the

participants fulfilled diagnostic criteria. In addition, a case study, a general review on

the etiology of transsexualism (with no new data), and two studies that described only

children were also excluded. In total 21 studies were systematically reviewed.

Insert figure 1 around here

2.4.Assessment of quality

The study used a checklist for cross sectional studies based on the NICE checklists (29)

developed by Gilbert(30) and used in previous studies(31). The NICE rating system rates

the studies from good quality (when all or most of the criteria have been fulfilled;

[++]), reasonable quality (when some of the criteria have been fulfilled; [+]), to poor

quality (when few or no criteria are fulfilled; [-]). The review and scoring was based

on the quality of the study reporting prevalence of transsexualism. Some of the

studies were qualitatively excellent but reported prevalence of sex reassignment

surgery instead of transsexualism. In these cases, studies were scored as +* to indicate

this.

2. 5. Statistical analysis

To calculate the meta-analytical prevalence, only studies that reported new data were

included. When several studies used the same data but at different times (i.e., data

from clinical databases across years), only the newest point prevalence data was

selected. In contrast, to calculate time series analysis, information from studies

describing period prevalence were included and the mean year of the studied period
 12

was used as a moderator for the analysis. For studies that spanned several years, the

total number of cases (trans men, trans women) were divided by the number of years

covered by the study in order to calculate the prevalence per year. Any missing

information was estimated.

The primary data of the epidemiological studies were extracted as raw numbers. An

aggregate effect size, weighted by sample size, was computed to provide an overall

effect size across the studies. Homogeneity among studies was computed using the Q

statistic and the I2 statistic. A significant Q statistic suggests that the distribution of

effect size around the mean is greater than would be predicted from sampling error

alone, whereas I2 provides an estimate of the proportion of the variance in the

aggregate effect size that is attributable to between-studies heterogeneity(32). Random-

effects models were fitted if there was heterogeneity. Risk ratios and 95% confidence

intervals (CIs) were calculated. The relative weighted contribution of each study was

also assessed. Meta-analysis was performed using the Comprehensive Meta-Analysis

software programme-2 for Windows(33) according to the Cochrane reviewers’

handbook(34). Significance was set at p < 0.05.

3. Results

3.1. Study characteristics

Most of the studies published were from Europe (18; 85.7%): five from Sweden (17, 22,
35, 36, 37)
, three from the Netherlands(18, 38, 39), three from the United Kingdom(40, 41, 42),

two from Germany(13, 43) and one each from Spain(44), Belgium(45), Serbia(46),

Ireland(47), and Poland(48). There was only one study from the United States of

America (USA), which was also the oldest(16), only one from Singapore(23), and only
 13

one from Australia(49). Most of the studies from the same country, such as the Dutch

and the Swedish studies, used the same data but at different periods therefore

covering different years. Other studies from the same country focused on different

areas, regions or counties within the country, such as East and West Germany, and

Northern Ireland, Scotland or England (three from the United Kingdom). The study

period varied from three years(22) to an impressive 50 years(17).

Employing the NICE rating system described above, there were four studies that were

scored as (-). This was because they were particularly old studies reporting

epidemiological data calculated from approximations(16), or from information

collected from third parties (such as psychiatrists via questionnaires(50), or from

primary care(42)). They were also rated as (-) if the information regarding how

individuals could access the gender clinic and whether the clinic covered a specific

geographical area was unclear(48). Six studies scored (+) as although methodologically

strong, the information was gathered from third parties (courts or government(28, 43),

endocrinologists(47) or surgeons(45)), or the geographical area covered was unclear(40,

41)
. Two studies that focused specifically on sex reassignment surgery were scored as

(+*)(17, 46). The rest of the studies scored (++) as they were methodologically strong

and provided clear epidemiological information regarding the prevalence of

transsexualism in the specific country, based on individuals attending a gender

identity clinic. Most of the studies were from the same countries, such as the

Netherlands and Sweden.

3.2. Population studied

 14

All of the 21 studies selected for the systematic review defined the population studied

as transsexual individuals. The majority of the studies used data gathered

retrospectively from gender identity clinics. The clinics covered the whole of their

countries, such as the Netherlands(18), or a specific geographical area within the

country, such as Catalonya in Spain(44). The studies used the ICD(1), DSM(2,3) or

Benjamin’s(12) criteria to define the population.

The reviewed studies provided reliable data and identified a population of transsexual

individuals in the following way: seven studies described individuals referred to a

gender identity clinic (who fulfilled diagnostic criteria), four studies described

individuals who were treated with cross-sex hormones who fulfilled the diagnosis,

five studies described individuals who had or were referred for sex reassignment

surgery (SRS) and five studies collected information regarding the possible number of

transsexual individuals by accessing Governmental organizations, i.e. the bureau of

records or virtual statistics. In spite of the complication of reaching a clear diagnosis,

most studies were able to provide the estimate of the number of transsexual

individuals, trans men and trans women from the age of 15 years.

3.3. Prevalence of transsexualism

Out of the 21 studies selected, 20 provided information on the number of transsexual

individuals in their study and gave information about the population that the clinic or

studied area covered. The American study(16) provided only an approximation of trans

individuals. Although it is historically important, it does not provide enough

information for analysis. As explained above, studies using the same database, but at

different time periods were removed for this analysis and only the most recent ones
 15

selected. Therefore out of the 20 studies, 12 provided enough new data on prevalence.

Those studies provided point prevalence as they reported the number of transsexual

individuals at the point when the study took place by reporting the number of

individuals since the clinic opened or records began. The prevalence of trans

individuals was studied in 12 countries with a total population of 95,141,541

individuals. The largest population studied was in West Germany with nearly 35

million individuals included(43). There was a similar group of natal males and females

within the overall general population with 30,651,864 males and 31,689,246 females.

The 12 studies identified a total of 4355 trans individuals. The largest number of trans

individuals were identified in the German study that looked at gender identity clinics

in specific areas of Germany, with a total of 1773 trans individuals(13). However, this

study did not provide the highest prevalence of transsexual individuals per 100,000

people. Rather, this was reported in the Singaporean study (23.60 Singapore-born

trans individuals per 100,000 people(23). This study may underestimate the total

number of trans individuals in Singapore, as it only includes those who go for sex-

reassignment surgery, which is only a proportion of all trans people. Not

unexpectedly, more recent studies found higher prevalence rates than older studies.

As an example, the recent Swedish study found a prevalence of 16.67 per 100,000(17).

Meta-analysis found an overall prevalence of transsexualism using random effects

meta-analysis of 0.000046 (95% CI= 0.000028-0.000077). Heterogeneity was high (I²

= 99.6%; Q value=3314.7) therefore the random effect was selected. This indicates a

prevalence of transsexualism of 4.6 per 100,000 or 1 in every 21,739 individuals.

 16

Insert table 2 around here

3.4. Prevalence of trans women

All studies identified a higher prevalence of trans women compared to trans men,

with the exception of the study by Godlewski(48) with a score of (-). Prevalence rates

of trans women varied between 0.44(22) and 35.2 per 100,000(23).

As one of the studies did not report the number of female or male individuals in the

population studied, this study was removed from subsequent analysis(43). Therefore,

eleven studies were included in the meta-analysis. An overall prevalence rate of

0.000068 (95% CI= 0.00004 – 0.00010) of trans women was found. Heterogeneity

was high (I² = 99.0%; Q value=1070.35) therefore random effect was selected. This

indicates a prevalence of trans women of 6.8 per 100,000 or 1 in every 14,705

individuals.

Insert Table 3 around here

3.5. Prevalence of trans men

Most studies reported a smaller number of trans men when compared to trans women.

Prevalence rates varied from 0.25(16) to 6.64 per 100,000(17). The trans women to trans

men ratios varied from 6.1 trans women to every one trans men (6:1)(49) to 1 to 1

(1:1)(46).

Eleven studies were included in the meta-analysis. An overall prevalence rate of

0.000026 (95% CI= 0.000017 – 0.00004) was found. Heterogeneity was also high (I²
 17

= 97.7%; Q value=435.1). This indicates a prevalence of trans men of 2.6 per 100,000

or 1 in every 38,461 individuals. Using meta-analytical prevalence rates, the ratio of

trans women to trans men was found to be 2.6 trans women for every trans man

(2.6:1).

Insert table 4 around here

Insert table 5 around here

3.6. Change of prevalence rates over the years

For the time series analysis 17 studies that provided period prevalence data at

different time points were included. Most of the studies provided data from a period

of time between 4 and 10 years. In order to undertake time series analysis a regression

analysis was undertaken with date as a moderator variable. The date selected was

calculated as the mean of the numbers of years of the study period. Time series

analysis showed a statistically significant effect for general transsexualism (z=23.45,

p < 0.001), trans women (z = 22.13, p < 0.001), and trans men (z = 8.00, p < 0.001)

suggesting a higher prevalence of transsexual individuals, trans women and trans men

in more recent studies.

Insert figure 2 around here

4. Discussion

This is the first study that has aimed to critically review and meta-analytically

investigate the prevalence of transsexualism. Twenty-one studies were identified as

 18

having enough data and using similar diagnostic criteria suitable for a critical review,

with 12 studies having enough data to perform a meta-analysis. Most epidemiological

studies in this area have investigated the prevalence rates of HIV in this

population(51,52,53) and very few have investigated the number of trans people in the

general population, which may be due to the complex methodology required to

undertake such studies.

This study showed that there has been a clear increase in the prevalence of individuals

diagnosed with transsexualism over time, with newer studies reporting statistically

significant higher rates than older studies. Countries using the same database over the

years have reported an increased prevalence of individuals attending clinical services,

receiving prescribed cross-sex hormones, or applying for sex reassignment surgery.

The overall meta-analytical prevalence of transsexualism was found to be 4.6 per

100,000, with the meta-analytical prevalence of trans women being higher (6.8 in

100,000) than that of trans men (2.6 in 100,000). Although the sex ratio has moved

closer to 1:1, using the meta-analytical prevalence, the trans women to trans men ratio

was calculated to be 2.6 to 1.

The increase in prevalence over the years is likely due to several factors: the increased

visibility of trans people in the media, which likely contributes to at least a partial

destigmatization of being trans(54); the wide availability of information on the Internet

about transsexualism or Gender Dysphoria, which also likely contributes to

destigmatization(54); the increased awareness of the availability of biomedical

treatment(4,5, 20); and the development of societal tolerance towards trans

individuals(15). With regards to the latter, most of the studies investigating social

attitudes to gender and sexual diversity have primarily explored peoples’ attitudes
 19

towards gay, lesbian, and bisexual (LGB) individuals. The 2012 edition of the

Eurobarometer however, included questions aiming at measuring the public attitudes

to transgender people in European countries for the first time. The most tolerant

countries appear to be Denmark, Luxembourg, Ireland, Sweden and the United

Kingdom. Interestingly, the proportion of individuals who feel uncomfortable with

trans people was larger than for LGB people in all European countries. In six

European countries the majority of people report to feel comfortable with a trans

person as a government leader: Denmark, Sweden, Luxembourg, Spain, the United

Kingdom and the Netherlands(15). This report is the first to specifically explore

societal tolerance towards trans people.

Recent reports indicate that the number of individuals with gender dysphoria who

attend clinical services for an assessment has increased substantially over the years in

many European countries(54,55). There is also a significant increase in people who

self-diagnose as suffering from gender dysphoria(19). Although this review did not

include studies of people who self-diagnose, such studies are important as they may

indicate the level of future demand for clinical services. For example Kuyper and

Wijsen(19) found that 4.6% of natal men and 3.2% of natal women in their Dutch

population sample reported an ambivalent gender identity (equal identification with

other sex as with sex assigned at birth) and 1.1% of the natal men and 0.8% of the

natal women reported an incongruent gender identity (stronger identification with

other sex as with sex assigned at birth). It remains unknown how many of their

sample will seek assessment and treatment via a gender identity clinic service.

Many trans individuals require clinical services as they wish to have cross-sex

hormone treatment and surgery to alleviate their gender dysphoria. However, some
 20

feel that only one of these treatment modalities is necessary for them, whilst others

may decide to not take cross-sex hormones or undergo any operations(56, 57).

Psychotherapy can be helpful for some individuals to integrate their transgender

feelings into their gender role assigned at birth and do not feel the need to feminize or

masculinize their body; for those changes in social gender role and expression can be

sufficient to alleviate gender dysphoria(20,57). Moreover, many trans individuals may

have socially transitioned with or without partial treatment through self-prescribed

cross sex hormones or private means(58,59).

The complex care pathway of trans individuals, makes the organization of service

delivery difficult to plan ahead. The prevalence of people requiring assessment and

treatment for gender dysphoria may be more accurately estimated by looking at

community studies. Studies reporting prevalence rates of SRS are likely to

underestimate the true prevalence of transsexualism in the community. Since one of

the largest studies in this meta-analysis identified their participants via applications

for SRS(21) , the overall prevalence rate is likely to be higher than the one reported in

this meta-analysis.

The main strength of this paper is the fact that this is the first study that summarizes

and critically assesses all the available data in the subject of trans epidemiology

taking into consideration the weight of the study in the analysis, in order to avoid

biased results based on large studies. While studies have become methodologically

stronger over the years, this review and meta-analysis is limited by the available data.

The majority of the studies were conducted in Western countries, particularly in

Europe, and the results may be related to the level of tolerance of society, the
 21

healthcare system available, legislation regarding the rights of trans people and the

academic interest in the area of trans healthcare. This is clearly reflected by the large

body of research originating from Europe, particularly in Sweden and the

Netherlands. There are a limited number of reports regarding trans individuals from

developing countries(60, 61, 62, 63). This may simply indicate that while trans individuals

do indeed have access to services, no epidemiological studies have been undertaken.

Alternatively, for some countries it may indicate that there are no clinical services

available and that trans individuals in these areas need to suppress their real self, the

consequence of which may be the development of mental health problems(64, 65, 66).

The study is limited by the high heterogeneity of the included studies, which is not

surprising as there are clear differences between the methodology of the studies

included in the review. This is reflected in the great variation of prevalence data from

the different studies. We use a random effects model in an effort to incorporate

heterogeneity between the studies in our analysis but recognize that this does not

eliminate the fact that heterogeneity was present. In spite of the limitations of this

meta-analytical study, and the majority of the prevalence studies reviewed, the

existing data should be considered as a starting point. The field would benefit from

more rigorous epidemiological studies acknowledging current changes in

classification system and including different locations worldwide.

 22

References

1. World Health Organization. International Classification of Diseases 10 (ICD-10).

Geneva: World Health Organization.1992.

2. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders (DSM3), 3rd ed. Washington, DC: American Psychiatric Press.
1980.

3. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders, (DSM-III-R), Revised 3rd ed. Washington, DC: American
Psychiatric Press. 1987.
4. Ahmad S, Barrett J, Beaini AY, et al. Gender dysphoria services: a guide for
general practitioners and other healthcare staff. Sex Relat Ther.
2013;28(3):173–186. doi:10.1080/14681994.2013.808884

5. Wylie K, Barrett J, Besser M, et al. Good Practice Guidelines for the Assessment
and Treatment of Adults with Gender Dysphoria. Sex Relat Ther.
2014;29(2):154–214. doi:10.1080/14681994.2014.883353

6. Drescher J, Cohen-Kettenis P, Winter S. Minding the body: situating gender

identity diagnoses in the ICD-11. Int Rev Psychiatr. 2012;24:568–77.
doi:10.3109/09540261.2012.741575

7. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders (DSM5), 5th ed. Washington, DC: American Psychiatric Press.
2013.

8. Gerstman B. Epidemiology Kept Simple: An Introduction to Traditional and

Modern Epidemiology, 2nd ed. Hoboken, NJ: Wiley-Liss. 2003. Retrieved
from http://eu.wiley.com/WileyCDA/WileyTitle/productCd-0471400289.html
on 11th January 2015.

9. World Health Organization. International Classification of Diseases 9 (ICD-9).

Geneva: World Health Organization. 1978.

10. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders (DSM4), 4th ed. Washington, DC: American Psychiatric Press.
1994.
 23

11. American Psychiatric Association. Diagnostic and Statistical Manual of Mental

Disorders (DSM4R), Revised 4th ed. Washington, DC: American Psychiatric
Press. 2000.

12. Benjamin H. The transsexual phenomenon. New York: The Julian Press.1966.

13. Garrels L, Kockott G, Michael N, et al. Sex ratio of transsexuals in Germany: the
development over three decades. Acta Psychiat Scand. 2000;102(6):445–448.

14. FRA. Being trans in the European Union: Comparative analysis of the EU LBGT
survey data. Vienna. FRA – European Union Agency for Fundamental Rights.
2014.

15. Keuzenkamp S, Kuyper L. Acceptance of LGBT individuals in the Netherlands

2013. The Hague. The Netherlands Institute of Social Research (ED). 2013.

16. Pauly IB. The current status of the change of sex operation. J Nerv Ment Dis.
1968;147(5):460–471.

17. Dhejne C, Oberg K, Arver S, Landén M. An Analysis of All Applications for Sex
Reassignment Surgery in Sweden, 1960-2010: Prevalence, Incidence, and
Regrets. Arch Sex Behav. 2014:43(8):1535 - 1545. doi:10.1007/s10508-014-
0300-8

18. Eklund P, Gooren L, Bezemer P. The prevalence of transsexualism in The

Netherlands. Brit J Psychiat. 1988;152:638–640. doi:10.1192/bjp.152.5.638

19. Kuyper L, Wijsen C. Gender identities and gender dysphoria in the Netherlands.
Arch Sex Behav. 2014:43(2):377–385. doi:10.1007/s10508-013-0140-y

20. Coleman E, Bockting W, Botzer M, et al. Standards of Care for the Health of

Transsexual, Transgender, and Gender-Nonconforming People, Version 7. Int

J Transgen. 2012;13(4):165–232. doi:10.1080/15532739.2011.700873

21. Zucker KJ, Lawrence AA. Epidemiology of Gender Identity Disorder:

Recommendations for the Standards of Care of the World Professional
Association for Transgender Health. Int J Transgen. 2009:11:8–18.
doi:10.1080/15532730902799946

22. Wålinder J. Incidence and sex ratio of transsexualism in Sweden. Brit J Psychiat.
1971;119:195–196. doi:10.1192/bjp.119.549.195
 24

23. Tsoi W. The prevalence of transsexualism in Singapore. Acta Psychiat Scand.

1988;78:501–504.

24. Nanda, S. The Hijras of India: cultural and individual dimensions of an

institutionalized third gender role. J Homosex. 1985;11:35-54.

25. Nanda, S. Neither Man nor Woman. The Hijras of India. Second edition.
Belmont, USA: Wadsworth. 1998.

26. Reddy, G. With respect to sex. Chicago, USA: University of Chicago. 2005.

27. Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group. Preferred
reporting items for systematic reviews and meta-analyses: The PRISMA

Statement. PLoS Med. 2009;6, e1000097. DOI:

10.1371/journal.pmed.1000097

28. Veale JF. Prevalence of transsexualism among New Zealand passport holders.
Aust NZ J Psychiat. 2008;42(10):887–889. doi:10.1080/00048670802345490

29. National Institute of Health and Clinical Excellence. The Guidelines Manual.
London: National Institute for Clinical Excellence. 2007. Retrieved online
from www.nice.org.uk on 11th January 2015.

30. Gilbert N. Disclosure of eating disorders and subsequent help seeking.

Unpublished thesis. 2009. Birmingham University, UK. Retrieved online from
http://etheses.bham.ac.uk/428/ on 11th January 2015.

31. Arcelus J, Haslam M, Farrow C, Meyer C. The role of interpersonal functioning

in the maintenance of eating psychopathology: A systematic review and
testable model. Clin Psychol Rev. 2013;33(1):156-167. doi:
10.1016/j.cpr.2012.10.009

32. Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. Stat

Med. 2002;21(11):1539–58. doi:10.1002/sim.1186

33. Borenstein M, Hedges L, Higgins J, Rothstein H. Comprehensive meta-analysis.

Version 2. Englewood NJ: Biostat. 2005.

34. Higgins JP, Green S. Cochrane handbook for systematic reviews of interventions
Version 5.1.0. 2011. Retrieved online from www.cochrane-handbook.org on
11th January 2015.
 25

35. Landén M, Wålinder J, Hambert G, Lundström B. Factors predictive of regret in

sex reassignment. Acta Psychiat Scand. 1998;97(4):284–289.

36. Olsson SE, Moller AR. On the Incidence and Sex Ratio of Transsexualism in
Sweden, 1972-2002. Arch Sex Beh. 2003;32(4):381–386. doi:
10.1023/A:1024051201160

37. Wålinder J. Transsexualism: A study of fourty-three cases. Goteborg:

Scandinavian University Books.1967.

38. Bakker A, Van Kesteren PJM, Gooren L, Bezemer P. The prevalence of

transsexualism in the Netherlands. Acta Psychiat Scand. 1993;87:237–238.

39. Van Kesteren PJ, Gooren LJ, Megens JA.An epidemiological and demographic
study of transsexuals in the Netherlands. Acta Psychiat Scand. 1996;
87(4):237–238.

40. Hoenig J, Kenna JC. The prevalence of transsexualism in England and Wales. Brit
J Psychiat. 1974;124(579):181–190. doi:10.1192/bjp.124.2.181

41. O’Gorman EC. A preliminary report on transsexualism in Northern Ireland. Ulster

Med J. 1981;50(1):46–49.

42. Wilson P, Sharp C, Carr S. The prevalence of gender dysphoria in Scotland: a

primary care study. Brit J Gen Pract. 1999;49(449):991–992.

43. Weitze C, Osburg S. Transsexualism in Germany: empirical data on epidemiology

and application of the German Transsexuals’ Act during its first ten years.
Arch Sex Behav. 1996;25(4):409–425.

44. Gómez-Gil E, Trilla Garcia A, Godas Sieso T, et al. Estimacion de la prevalencia,

incidencia y razon de sexos del transexualismo en catalunya segun la demanda
asistencial. Acta Espan Psychiat. 2006;34(5):295–302.

45. De Cuypere G, Van Hemelrijck M, Michel A, et al. Prevalence and demography

of transsexualism in Belgium. Eur Psychiat. 2007;22(3):137–141.
doi:10.1016/j.eurpsy.2006.10.002

46. Vujovic S, Popovic S, Sbutega-Milosevic G, Djordjevic M, Gooren LJ.

Transsexualism in Serbia: A twenty-year follow-up study. J Sex Med.
2009;6(4):1018–1023. doi: 10.1111/j.1743-6109.2008.00799.x
 26

47. Judge C, O’Donovan C, Callaghan G, Gaoatswe G, O’Shea D. Gender dysphoria -

prevalence and co-morbidities in an Irish adult population. Front Endocrinol.
2014;5:87. doi:10.3389/fendo.2014.00087

48. Godlewski J. Transsexualism and anatomic sex ratio reversal in Poland. Arch Sex
Behav. 1988;17(6):547–548.

49. Ross M, Wålinder J, Lundstrom B, Thuwe I. Cross-cultural approaches to

transsexualism. Acta Psychiat Scand. 1981;63:75–83.

50. Wålinder J. Transsexualism: definition, prevalence and sex distribution. Acta

Psychiat Scand. 1968;203:255-257.

51. Operario D, Soma T, Underhill K. Sex work and HIV status among transgender
women: Systematic review and meta-analysis. JAIDS 2008; 48:97-103.

52. Nuttbrock L, Bockting W, Rosenblum A, et al. Gender abuse, depressive

symptoms, and HIV and other sexually transmitted infections among male-to-
female transgender persons: a three-year prospective study. Am J Public
Health 2013; 103:300-307.

53. Poteat T, Reisner SL, Radix A. HIV epidemics among transgender women. Curr
Opin HIV AIDS 2014; 9: 168-173.

54. Aitken M, Steensma TD, Blanchard R, VanderLaan DP, Wood H, Fuentes A,

Spegg C, Wasserman L, Ames M, Lindsay Fitzsimmons C, Leef JH, Lishak
V, Reim E, Takagi A, Vinik J, Wreford J, Cohen-Kettenis PT, de Vries ALC,
Kreukels BPC, Zucker KJ. Evidence for an altered sex ratio in clinic-referred
adolescents with gender dysphoria. J Sex Med. 2015;12: 756-763.
Doi:10.1111/jsm.12817
55. de Vries ALC, Kreukels BPC, T’Sjoen G, Ålgars M, Mattila A. Increase of
referrals to gender identity clinics: A European trend? In: Transgender
Healthcare in Europe. Book of Abstract. pp10. Ghent, Belgium: European
Professional Association of Transgender Health (EPATH). 2015.
Retrieved online from http://epath.eu/wp-content/uploads/2014/07/EPATH-
2015-Book-of-Abstracts.pdf on 8th April 2015.

56. Bockting W, Coleman E, De Cuypere G. Care of transsexual persons. New Engl J

Med. 2011;364:2259-2560. doi:10.1056/NEJMc1104884
 27

57. Bockting WO. Psychotherapy and the real-life experience: From gender
dichotomy to gender diversity. Sexologies. 2008;17:211–224.
doi:10.1016/j.sexol.2008.08.001

58. Davies A, Bouman WP, Richards C, et al. Patient satisfaction with gender identity
clinic services in the United Kingdom. Sex Relat Ther. 2013;28(4):400–418.
doi:10.1080/14681994.2013.834321

59. Mepham NL, Bouman WP, Arcelus J, Wylie KR, Hayter M. People with Gender
Dysphoria who self prescribe cross sex hormones: Prevalence, sources and
side effects knowledge. J Sex Med. 2014;11: 2995-3001. doi:
10.1111/jsm.1269

60. Franz RC. Sigmoid colon vaginoplasty: A modified method. Brit J Obstet Gynaec.
1996;103:1148-1155.

61. Ruan FF, Bullough VL, Tsai YM. Male transsexualism in mainland China. Arch
Sex Behav. 1989;18:517-522.

62. Zhaoji X, Chuanmin W, Bullough VL. Transsexualism in China in 1995. In B.

Bullough, V. L. Bullough, & J. Elias (Eds.), Gender blending (pp. 377-382).
Amherst, NY: Prometheus Books. 1997.

63. Hahn T. Jin Xing, a dancer and leader of the Xing Dance Theatre in Shanghai,
China, and a transsexual. Ballettanz. 2004;3:22–25.

64. Claes L, Bouman WP, Witcomb G, Thurston M, Fernandez-Aranda F, Arcelus J.

Non-Suicidal Self-Injury in Trans People: Associations with Psychological
Symptoms, Victimization, Interpersonal Functioning and Perceived Social
Support. J Sex Med. 2014. doi: 10.1111/JSM.12711

65. Davey A, Bouman WP, Arcelus J, Meyer C. Social support and psychological
wellbeing: A comparison of patients with gender dysphoria and matched
controls. J Sex Med. 2014;11(12):2976-2985. doi: 10.1111/jsm.12681

66. Heylens G, Elaut E, Kreukels BPC, et al. Psychiatric characteristics in transsexual

individuals: multicentre study in four European countries. Brit J Psychiat: J
Ment Sci. 2013;204(2):151–156. doi:10.1192/bjp.bp.112.121954
28
Figure 1. Literature review method used
Click here to download high resolution image
Figure 2: Time-series analysis
Click here to download high resolution image
Table 1

Table 1: Criteria for searches on prevalence of Transsexualism

Category Criteria

Study population All races, ethnicities, and cultural groups

Adults (>18 years) or adolescents (18-15
years); no children (<15 years)

Study settings and geography All nations

Time period Published from 1954 through June 2014

Publication criteria Included:

- All languages
- Articles in print or online
Excluded:
- Articles in grey literature or non-peer-
reviewed journals or unobtainable during
the review period

Admissible evidence Transsexualism, Gender Identity Disorder

(Study design and or Gender Dysphoria must be diagnosed
other criteria) according to DSM III, DSM III-R, DSM
IV, DSM-IV-TR, DSM-V, ICD-9, ICD-10,
or Benjamin criteria.
Prevalence must be available or be able to
be extracted from the data.

Country, number of years used to extract

the data, number of trans individuals, natal
or preferred gender, mean age, general
population of the studied area (divided by
natal males and female), to be recorded
when possible.

Eligible study designs include:

Observational studies including descriptive
studies, evidence that the population
described belongs to a defined area, region
or country.
Table 2

Table 3: Meta-analysis on the prevalence of Transsexualism

study Statistics for each study Event rate and 95% CI

Event Lower Upper
rate limit limit Z-Value p-Value
Hoenig & Kenna 1974 0.000019 0.000015 0.000024 -88.374312 0.000000
Ross et al 1981 0.000024 0.000021 0.000027 -175.566362 0.000000
O'Gorman 1982 0.000019 0.000013 0.000028 -57.492379 0.000000
Tsoi et al 1988 0.000237 0.000216 0.000259 -178.635953 0.000000
Van Kesteren et al 1996 0.000081 0.000076 0.000087 -297.482874 0.000000
Weitze and Osburg 1996 0.000021 0.000020 0.000023 -291.609808 0.000000
Wilson et al 1999 0.000048 0.000041 0.000056 -125.794153 0.000000
Gomez Gil et al 2006 0.000039 0.000034 0.000045 -137.007418 0.000000
De Cuypere et al 2007 0.000053 0.000048 0.000058 -199.916997 0.000000
Vujovic et al 2008 0.000023 0.000019 0.000027 -129.715639 0.000000
Dhejne et al 2014 0.000167 0.000155 0.000180 -226.985486 0.000000
Judge et al 2014 0.000068 0.000059 0.000077 -141.727714 0.000000
-0.25 -0.13 0.00 0.13 0.25
Table 3

Table 4: Meta-analysis on the prevalence of trans women

Model study Statistics for each study Event rate and 95% CI
Event Lower Upper
rate limit limit Z-Value p-Value
Hoenig & Kenna 1974 0.000030 0.000022 0.000039 -72.978449 0.000000
Ross et al 1981 0.000042 0.000036 0.000048 -145.805211 0.000000
O'Gorman 1982 0.000029 0.000019 0.000044 -47.947147 0.000000
Tsoi et al 1988 0.000350 0.000315 0.000389 -147.330866 0.000000
Van Kesteren et al 1996 0.000121 0.000113 0.000130 -243.491123 0.000000
Wilson et al 1999 0.000078 0.000066 0.000093 -106.547824 0.000000
Gomez Gil et al 2006 0.000055 0.000046 0.000065 -112.230595 0.000000
De Cuypere et al 2007 0.000078 0.000069 0.000087 -161.694523 0.000000
Vujovic et al 2008 0.000022 0.000018 0.000028 -90.222401 0.000000
Dhejne et al 2014 0.000116 0.000105 0.000127 -187.715946 0.000000
Judge et al 2014 0.000098 0.000084 0.000115 -116.323709 0.000000
-0.25 -0.13 0.00 0.13 0.25
Table 4

Table 5: Meta-analysis on the prevalence of trans men

study Statistics for each study Event rate and 95% CI

Event Lower Upper
rate limit limit Z-Value p-Value
Hoenig & Kenna 1974 0.000009 0.000006 0.000015 -47.810027 0.000000
Ross et al 1981 0.000007 0.000005 0.000009 -69.495292 0.000000
O'Gorman 1982 0.000010 0.000005 0.000021 -30.460159 0.000000
Tsoi et al 1988 0.000120 0.000100 0.000145 -96.769109 0.000000
Van Kesteren et al 1996 0.000043 0.000038 0.000048 -164.890182 0.000000
Wilson et al 1999 0.000019 0.000014 0.000027 -62.373363 0.000000
Gomez Gil et al 2006 0.000022 0.000017 0.000029 -76.557473 0.000000
De Cuypere et al 2007 0.000030 0.000025 0.000035 -114.211995 0.000000
Vujovic et al 2008 0.000023 0.000018 0.000028 -93.198522 0.000000
Dhejne et al 2014 0.000066 0.000059 0.000075 -152.689491 0.000000
Judge et al 2014 0.000036 0.000028 0.000047 -78.561306 0.000000
-0.25 -0.13 0.00 0.13 0.25
Table 5

Table 6: Studies reviewed for the systematic review on prevalence of Transsexualism

Dates Prevalence Sex ratio M/F

Name Date studied Country Age Method Total for 100,000 TW 100,000 TM 100,000 (biological)

Pauly(14) 1968 n/a USA n/a Approximation n/a 0.40 1 0.25 4 to 1

Ask all psychiatrists to inform

them of all the Transsexuals
Wålinder(31)** 1967 n/a Sweden >15 known to them. 3/4 replied 110 1.42 2.7 0.97 2.8 to 1

Wålinder(20) ** 1971 1967-1970 Sweden >15 Referrals to gender clinic 27 0.45 0.44 0.44 1 to 1

Hoenig & Kenna(34)*(**) 1974 1958-68 England and wales >15 Referrals to gender clinic 66 1.92 2.94 0.92 3.2 to 1

Questionnaire to all
psychiatrist in Australia via
(43)
Ross et al * (**) 1981 1979-1981 Australia >15 the journal 272 2.40 4.16 0.66 6.1 to 1

O'Gorman(35)* (**) 1982 1968-1982 N-Ireland >15 Referrals to Gender clinic 28 1.92 2.85 1 2.8 to 1

Treated at the gender clinic

1976-1980 The Netherlands >15 (with hormones) 120 1.22 2.22 0.5 3 to 1

Eklund et al (16) ** 1988 Treated at the gender clinic

1976-1983 The Netherlands >15 (with hormones) 233 1.58 3.84 1 3 to 1

Treated at the gender clinic

1976-1986 The Netherlands >15 (with hormones) 538 2.77 5.55 1.85 3 to 1
Godlewski et al(42) 1988 1974-78 Poland >14 Referrals to sexology clinic 26 n/a n/a n/a 1 to 5.5

alive in
Tsoi et al(21)* (**) 1988 1986 Singapore >15 Sex reassignment surgery 458 23.60 35.2 12 3 to 1

Treated at the gender clinic

Bakker et al(32)** 1993 1976-1990 The Netherlands >15 (with hormones) 713 4.42 8.4 3.28 2.5 to 1

Van Kesteren et al(33)* Treated at the gender clinic

(**) 1996 1975-1992 The Netherlands >15 (with hormones) 998 8.05 12.11 4.3 3 to 1

Questionnaires to courts that

Weitze & Osburg(37)* handles legal registration of
(**) 1996 1981-1990 West Germany n/a sex change 733 2.10 2.38 0.96 2.3 to 1

Use of care and sex

Landén et al(29)** 1996 1972-1996 Sweden >15 assignment 233 3.42 4.03 2.83 1.4 to 1

Patients register in GP
practice, via questionnaires, I
selected those taking
Wilson et al(36)* (**) 1999 1998 Scotland >15 hormones or post operative 160 4.79 7.82 1.92 4 to 1

Referrals to different Gender 1.2 to 1 last 4

Garrels et al(13) 2000 1964-1998 Germany >14 services 1773 n/a n/a n/a years.

Application for sex re-

Olsson & Möller(30) ** 2003 1972-2002 Sweden >15 assignment 402 5.91 7.34 4.54 1.61 to 1
 Gómez-Gil et al(38)* (**) 2006 1996-2004 Spain (Cataluña) >15 Referral to gender Clinic 182 3.88 4.75 2.07 2.6 to 1

All TS who underwent SR

De Cuypere et al(39)* (**) 2007 1985-2006 Belgium >15 from plastic surgeons 412 4.28 7.75 2.95 2.43 to 1

Vujovic et al(40)* (**) 2008 1987-2006 Serbia >18 Referral to gender Clinic 147 2.25 2.23 2.27 1 to 1

Applications for legal and

Dhejne et al(15)* (**) 2014 1960-2010 Sweden >17 surgical sex re-assignment 681 16.67 11.57 6.64 1.6 to 1

Referral to endocrine clinic

Judge et al(41)* (**) 2014 2005-2014 Ireland >15 for hormones 218 6.77 9.84 3.61 2.7 to 1

TS: Transsexuals; TW: Transwomen; TM: Transmen

*Studies for meta-analysis. Studies describing point prevalence
** Studies for time line. Studies providing information on period prevalence