How Common is Detransition? A Review of all the Evidence
ABSTRACT: To better constrain estimates of the prevalence of detransition and gender-affirming surgical regret, a wide range of studies (n=34+ original studies and 2 meta-analyses) were reviewed and sorted into high- and low-reliability subgroups. Studies whose methodology was broadly appropriate for determining detransition and regret rates (“high-reliability” studies) were separated into three topical categories and pooled to calculate composite rates within and across these categories. Studies that were not suitable for this purpose (“low-reliability” studies) were described in detail, but not factored into the rate calculations. The overall detransition rate was found to be 3.3% (n=141 of 4,221); the detransition rate for minors was 4.0% (n=102 of 2,528); and the surgical regret rate was 1.7–2.1% (n=70 of 3,279–4,029). Pooling all high-reliability studies, the combined regret/detransition rate was 2.5–2.7% (n=203 of 7,500–8,250). Potential sources of positive bias included failure to specify internal vs. external motivation and pervasive social pressure against transgender identity. Potential sources of negative bias included widespread loss to follow-up and overly-stringent definitions of regret.
1. Introduction
It is March 2023, and a fierce legislative battle is raging across the United States¹⁻². In the last three months, over 400 anti-trans bills have been proposed or passed, including bans on gender-affirming care (31 states), bans or restrictions on drag performance (18 states), and even bans on the legal recognition of transgender identity itself (6 states)³⁻⁴. Meanwhile, anti-trans hate has reached a new fever-pitch both online and in real life. Having successfully labeled us as groomers and pedophiles⁵⁻⁹, conservative figures are now calling for our total elimination from society¹⁰⁻¹³.
Through all this, one of the main justifications used for anti-trans measures has been the notion that most trans children eventually revert to identifying with their assigned sex at birth. From news articles¹⁴⁻¹⁶ to state government press releases¹⁷𝄒¹⁸ to courtroom hearings¹⁹𝄒²⁰, those opposed to so-called “gender ideology” are taking the stand to argue that, no, this is not the natural course of things. Children are not inherently trans; rather, they are led down a path of disfigurement and regret by those who have already given themselves over to delusion. As proof, they cite scientific papers which do, at first glance, seem to support their position. What wonderful news! All the struggle and strife of transition, unnecessary. Just wait a few years, and most of these children will learn to live in harmony with their bodies like God and Nature intended.
Not so fast, say trans allies. While it is true that some papers report high desistance rates, these papers have received widespread methodological criticism, and are at odds with an emerging body of work showing rates of less than 5%. Factor in the many studies proving that transition increases psychological well-being in trans youth and adults alike²¹𝄒²², and the numbers cited by anti-trans groups simply don’t add up. Yes, it is natural for parents to be worried for their child’s well-being. It is normal to feel surprised, and anxious, and confused. But in a time when we have extensive data showing positive outcomes, these concerns are being bolstered by misinformation from activists with an altogether more sinister purpose: to ban gender transition entirely.
With so many papers, articles, and anecdotes on each side, this debate can become a matter of choosing an ideological affiliation instead of uncovering the facts. Therefore, the goal of this review is to give a complete and unbiased summary of the research that might be cited in relation to the following question:
After the start of medical or social transition, what percentage of trans people experience regret or detransition?
2. High-reliability studies
Because different studies report different measures, it is useful to have a standard against which to judge them. For the purpose of this analysis, then, “detransition” will be defined as the intentional cessation of gender-affirming medical care or reversal of social transition. Where possible, this will be restricted to cases motivated by gender identity — which might be imperfectly reported as, for example, “gender dysphoria (GD) desisted”, “no longer wishes to pursue transition”, or “stopped identifying as gender variant”. In order to admit a range of papers, studies will be classified as “high-reliability” provided they can reasonably approximate detransition without any major errors. Of course, they will still have their shortcomings; but this at least eliminates those papers that are plainly unsuitable for the purpose, or that suffer from large and obvious sources of bias.
a. Detransition — in minors
[Detransition (minors)] Khatchadourian et al. (2014)²³ studied the medical records of 84 children (78 previously diagnosed with GD) treated at the British Columbia Children’s Hospital (BCCH) Transgender Program between January 1998 and December 2011. The median follow-up period was 2.0 years. Of the 52 patients that started on puberty suppression (27 gonadotropin-releasing hormone analogues (GnRHa; “puberty blockers”), 25 anti-androgens), 1 stopped due to emotional instability (but not unwillingness to pursue transition) and 1 stopped having decided not to pursue transition. Another stopped due to heavy smoking but was prescribed gender-affirming hormones (GAH) once the smoking resolved. In the GAH phase, “none of the 63 patients receiving cross-sex hormones discontinued his or her treatment permanently”. Thus, this study documented a 1.2% detransition rate (n=1 of 84). Note the short follow-up period of this study.
[Detransition (minors)] McCallion et al. (2021)²⁴ studied patient files of minors diagnosed with GD and referred to Paediatric Endocrinology at the Royal Hospital for Children in Glasgow between 2011 and 2019. The median follow-up period was 1.9 years. Of the 79 children who began GnRHa treatment, 6 discontinued — 3 because their GD had resolved and 3 from noncompliance. Subsequently, 1 patient who had started GAH desisted due to cessation of GD. This study reported discontinuation in both the GnRHa and GAH phases, but the GAH group (n=41) was much smaller than the GnRHa group (n=79) — likely because of the many new patients in 2017–2019 who had not yet become eligible for GAH by the end of the study period. Therefore, it is probably most accurate to multiply the two discontinuation rates together while eliminating the three patients with unknown outcomes, resulting in a 6.3% detransition rate (or 4.8 theoretical detransitioners out of n=76 with known outcomes). Note the short follow-up period of this study.
[Detransition (minors)] Butler et al. (2022)²⁵ studied minors referred by the NHS Gender Identity Development Service (GIDS) between 2008 and 2021. Of 1,151 eligible patients, 62 were lost to follow-up; 90 of the remaining 1,089 (8.3%) stopped identifying as gender variant. However, 32 of these did so before starting GnRHa treatment (after the first clinic appointment ) and with social transition status not specified. The remaining 58 had stopped either GnRHa or GAH treatment and ceased identifying as transgender — a 5.5% detransition rate (=58/(1,089–32)). The mean duration of treatment was not reported. It is unclear whether using discharge reports would have biased the results.
[Detransition (minors)] De Castro et al. (2022)ⁿ¹ studied minors referred between 1999 and 2016 to the Gender Identity Unit of Catalonia, in Barcelona, Spain. The mean follow-up period was 2.6 years. “Desisters” were those who met the following criteria at the end of follow-up: “(1) the subject’s gender identity did not differ from the assigned sex, (2) the subject experienced serious doubts about their gender identity leading to reverse the social transition, or (3) discontinuation of hormonal treatment.” Of 124 transgender minors, 3 desisted in some form — 1 after GAH, 1 after GAH and social transition, and another before any kind of transition. Therefore, this sample had a 2/124 = 1.6% detransition rate. Note the short follow-up period of this study.
[Detransition (minors)] Olson et al. (2022)²⁶ studied social transition in U.S.-based trans minors (no diagnosis required) who enrolled in Princeton University’s The Trans Youth Project²⁷ between July 2013 and December 2017. The mean follow-up period was 3.8 years. Of the 317 participants, 8 had resumed cisgender identification (as determined by pronouns) at an average of 5.4 years after the time of social transition — 2.5% detransition rate. 3.5% had come out as nonbinary, and a total of 7.3% had changed their gender identification at least once. All but one case of social detransition occurred by the age of 9 — well before any medical intervention would be considered. Although 26 of the subjects did not participate in the final 2 years of the study, the authors reported that many of the families had remained in contact with the research team and indicated that their non-participation was due to personal issues, not hesitancy to disclose detransition; thus, this was deemed unlikely to be a major source of error. The authors did, however, note the existence of known biases (high income, normal mental illness rates) in the sample, in addition to the fact that participants had socially transitioned an average of 1.6 years before joining the study.
— — — — Spotlight: paired GnRHa and GAH discontinuation studies — — —
This January, Van der Loos et al. (2023) published a study characterizing GnRHa treatment discontinuation rates in the Amsterdam Cohort of Gender Dysphoria (ACOG) — a population of 8,831 individuals who underwent diagnostic assessment and/or medical treatment for GD (and earlier, gender identity disorder, or GID) at the Vrije Universiteit Medical Center (VUmc) gender identity clinic (GIC) in Amsterdam between 1972 and 2018. Individually, this study would be unsuitable for measuring detransition rates, as it only covered the relatively short puberty suppression phase. But this was not the only discontinuation study performed on the full ACOG dataset. The previous year, Van er Loos et al. (2022) had studied GAH discontinuation in ACOG minors. Because these papers used the same dataset and studied almost precisely the same subset of patients, they can be combined into a total rate of discontinuation over the course of GnRHa and GAH treatment. Factoring in the GD remission data from the GnRHa phase, this becomes a passable measure of detransition.
[GnRHa discontinuation] Van der Loos et al. (2023)²⁸ studied 882 ACOG patients who visited the VUmc before age 18 and began GnRHa treatment; this would primarily have been in the 1987–2018 period, as treatment of minors was rare before 1987. Mean patient follow-up time was 4.6 years. Of these patients, 9 stopped without starting GAH due to remission of GD. Otherwise, 202 were not eligible for GAH for protocol reasons (too young, had not been on GnRHa for long enough), 5 for psychological reasons, 5 due to compliance issues, and another 5 because the diagnostic period had been extended. The remaining 656 began GAH. Median age at the start of GnRHa treatment was 14.0 for AMAB and 15.5 for AFAB. Median duration of GnRHa monotherapy was 1.6 years for AMAB and 0.7 years for AFAB. Median age at the start of GAH was 16.0 for AMAB and 16.7 for AFAB.
[GAH discontinuation] Van der Loos et al. (2022)²⁹ studied 720 ACOG patients who began GnRHa treatment before the age of 18, maintained this treatment for at least 3 months, and then started GAH. Of these patients, 16 had discontinued GAH as determined by a nationwide prescription registry coupled with VUmc records. An additional 2 patients had moved abroad, and so had unknown outcomes. Median age at the start of GAH was 16.0 for AMAB and 16.7 for AFAB; median age at the end of data collection was 20.2 for AMAB and 19.2 for AFAB. Median duration of GAH at the time of the study analysis was 3.5 years.
Combining these two studies, the overall detransition rate is
This calculation could be done either by multiplying the rates, as above, or by subtracting all detransitioners from the starting population, which would yield a 2.9% detransition rate. As with McCallion et al. (2021), the first method was chosen to account for the large fraction of young or new patients who were not included in the GAH pool. Note that those with unknown outcomes (n=10 for GnRHa discontinuation and n=2 for GAH discontinuation) were subtracted from the sample pool so as not to count them against the detransition rate. For category-mean rate calculations, the starting population will be 870: the original 882 from Van der Loos et al. (2023) minus the 12 with unknown outcomes. This results in n=28.2 theoretical detransitioners. This may be an overestimate, as reasons for GAH discontinuation were not specified.
The combined follow-up period after GnRHa initiation was ~4–5 years.
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
b. Detransition — combined or adults-only
[Detransition (adults)] Dhejne et al (2014)³⁰ studied applications for legal and surgical gender change by Swedish adults between 1960 and 2010 (n=767). Of the 681 who received a new legal gender and underwent gender-affirming surgery, 15 later submitted “regret applications” to apply for reversal — a 2.2% detransition rate. All regret applications were approved. The authors noted that the frequency of such applications decreased significantly with time. A unique strength of this study is that it catalogued long-term detransition; on the other hand, it lacked modern data and took place exclusively in a setting where surgery was a requirement for legal gender change³¹.
[Detransition (adults)] Pazos Guerra et al. (2020)³² studied a cohort of 796 patients who received treatment at the Identity Gender Unit of Doctor Peset University Hospital, Valencia, Spain, between January 2008 and December 2018. Based on follow-ups, the authors identified 8 individuals who requested some form of detransition, whether medical, social, or administrative — a 1.0% detransition rate. The data collection methodology was not described in detail. Depending on the methodology, this study may be vulnerable to loss to follow-up.
[Detransition (adults)] Hall et al. (2021)³³ studied the records of patients (age 17+) who had been discharged from the West of England GIC between September 2017 and August 2018. With detransition defined as reversion to the original gender role during or after the care period, they found that 12 of 175 eligible patients had detransitioned — a 6.9% detransition rate. Only 9 of these (5.1%) showed evidence of stopping hormones, and 4 were re-referred to the service after detransitioning. 5 patients (2.9%) discharged due to stopping transition, but the distinction between this and the higher detransition rate was not explained. Duration of treatment and reasons for detransition were not reported. It is unclear whether using discharge reports would bias the results.
[Detransition (adults)] Boyd et al. (2022)³⁴ studied patients diagnosed with GD (mean age 20 at time of presentation) within the UK GP practice population — a multisite training practice that provides primary care to over 20,000 individuals in southwestern England. Of the 41 patients who started GAH, 4 trans men stopped testosterone after a mean of 18 months and had records of comments expressing a desire to detransition — a 9.8% detransition rate. 4 additional patients voluntarily stopped GAH but continued to identify as transgender, with one of these having a history of regret. Mean treatment duration was not reported. Since this study has a very small sample size and an outlier result, it will skew the study-mean detransition rate.
c. Surgical regret
[Surgical regret] Pfafflin (1993)³⁵ surveyed 74 follow-up studies (n≥5) of post-operative transgender adults, comprising a total of 1,400 to 2,150 individuals. They found a regret rate of <1–1.5% (n~22: 25 minus 3 duplicates). Separately, they described their own sample of 295 patients, of whom 3 (1.0%) expressed regret. However, despite being regretful, these three patients still considered gender-affirming surgery to have been a “necessary step in their development.”
[Surgical regret] Smith et al. (2005)³⁶ studied a group of post-surgical patients at the VUmc GIC and University Medical Center Utrecht (158/188 = 84.0% follow-up). Follow-up occurred an average of 1.8 years after surgery and 3.5 years after GAH initiation. Of the 158 patients, 1.6% expressed regret about the surgery. There is some ambiguity in the presentation of this data; 2 surgical regret and 5 other or temporary in-treatment regrets were mentioned, neither of which makes sense with the provided percentage. For mean rate calculations, the value will be set at n=3 of 158, even though this is 1.9%, as it is the closest whole-number approximation. This paper also reported a “drop-out” group of 34 (of 222) people who stopped GAH earlier in the process; it was not specified whether these patients had detransitioned.
[Surgical regret] De Vries et al. (2011, 2014)³⁷𝄒³⁸ tracked a cohort of 55 trans youth who were referred to the VUmc GIC between 2000 and 2008 and went on to receive puberty blockers, hormone replacement therapy, and gender-affirming surgery. Mean age was 14.8 at the start of puberty suppression, 16.7 at the start of GAH, and 19.2 at the time of surgery. No patients reported regret associated with any of these stages, including surgery (a 0% regret rate), although at least 6 eligible patients failed to follow up. The cohort reported statistically significant decreases in gender dysphoria and behavioral and emotional problems.
[Surgical regret] Bustos et al. (2021)³⁹ conducted a meta-analysis of studies (n≥10) on post-op regret for gender-affirming surgery. The 27 included studies contained no duplicates from Pfafflin (1993), although the Pfafflin (1993) original cohort was included. Out of 7,928 patients, 77 (1.0%) regretted their procedure(s). An error identified⁴⁰𝄒⁴¹ in the original analysis — the inflation of the Wiepjes et al. (2018) surgical cohort from 2,627 to 4,863 due to miscalculation from the confusingly-presented Table 1 of that paper — was corrected in a 2022 erratum⁴². However, given this error, a detailed reanalysis of the Bustos et al. (2021) studies was performed here (see Appendix A).
The results were as follows: Wiepjes et al. (2018)⁴³ and Landén et al. (1998)⁴⁴ (which studied a subset of the Dhejne et al. 2014 sample) were eliminated because their definitions of surgical regret entailed hormonal detransition and application for reversal, respectively. Kuiper et al. (1998)⁴⁵ — a detransitioner study that made no claims of quantifying the overall rate of regret — was also eliminated. Finally, the Jiang et al. (2018)⁴⁶ sample size was reduced from 80 to 14, as there was no evidence of regret having been queried outside of the vulvoplasty subgroup. After these omissions, the 24 remaining studies had a 45/1666 = 2.7% regret rate. Of these 24 studies, 21 utilized questionnaires/surveys and/or clinician interviews; and of these 21, 12 reported at least 70% follow-up, and 7 reported at least 80% follow-up. The 70% follow-up group had a pooled 20/601 = 3.33% regret rate, while the 80% follow-up group had a pooled 15/398 = 3.77% regret rate. There was also considerable heterogeneity in the types of regret reported. 26 of 45 regretful patients experienced “minor” or “occasional” regret, regret “sometimes”, “partial regret”, or, in one case, externally-motivated regret due to difficulty finding employment. An additional 3 were regretful but still felt that the surgery had been necessary at the time, and another 4 were reported under a definition of regret that included ambivalence. If these 33 patients were removed from the count, the consistent and/or major regret rate would be (at most) 0.72%. Evidently, some amount of uncertainty remains within this body of work.
[Surgical regret] Tang et al. (2022)⁴⁷ analyzed the records of 209 teenage patients who underwent top surgery (gender-affirming mastectomy) between January 2013 and July 2020 (median age 16 at referral), finding a 1.0% regret rate (n=2). Neither of the two patients who expressed regret elected to undergo reversal surgery. However, this figure is currently disputed, as Reuters has claimed to have found two more detransitioners within the same patient pool⁴⁸. Accordingly, this result will not be factored into the mean rate calculation, but is still listed here, due to the media correction being unverified.
3. Low-reliability studies
Low-reliability studies are those that fail to satisfactorily answer the posed question (without factoring in paper quality or reliability for other measures).
[Detransition/regret] Deutsch (2012)⁴⁹ reported extremely low regret (0.87%) and detransition (0.15%) rates among 1,944 patients accessing gender-affirming care at 12 U.S.-based informed consent clinics. Patients had been treated for an average of 3.13 years. While this study is useful as a survey of informed consent practices, the methodology is unsuitable for quantifying regret and detransition rates; reliance on clinic survey responses of patient self-reports makes these statistics especially vulnerable to loss to follow-up.
[GnRHa discontinuation, surgical regret] Wiepjes et al. (2018)⁴³ studied ACOG patient files from 1972 to 2015. Of the ~333 minors receiving GnRHa treatment, 1.9% had stopped the medication without starting GAH. However, this study should not be counted, as Van der Loos et al. (2023) assessed the same measure within a more updated version of ACOG. The study also reported a 0.5% regret rate for the 2,627 patients who had undergone gonadectomy (any gender-affirming surgery involving the removal of the testes or ovaries). This was actually a reversal rate: “transwomen who started testosterone treatment after vaginoplasty or transmen who started estrogen treatment after oophorectomy and expressed regret.” Therefore, it cannot be used as a measure of surgical regret.
[Detransition] Davies et al. (2019)⁵⁰ searched for keywords related to regret and detransition in patient files created between August 2016 and August 2017 at a national UK GIC, finding a 0.47% (n=16 of 3,398) regret/detransition rate. Unfortunately, this research appears only as a one-page poster session summary, and so fails to provide detailed methodology, results, or limitations analysis. Using automated keyword search of active patient files seems to be a weakness, but it is not possible to accurately assess the reliability of this research without the full text.
[Detransition] Richards and Doyle (2019)⁵¹ examined the records of a random sample of 303 current patients at the NHS-run Nottingham Center for Transgender Health. With detransition defined as in Hall et al. (2021), they found 3 cases: 1 while in active care and 2 while not in active care. All cases of detransition were associated with a lack of social support and led to subsequent re-transition. Thus, the overall detransition rate (as defined here) was 0%. However, restricting the sample pool to current GIC patients is likely to introduce a large bias in favor of low detransition rates. Duration of treatment was not reported, and all historical instances of detransition among current patients were counted.
[Surgical regret] Narayan et al. (2021)⁵² reported data from the responses of 46 surgeons to a survey about post-surgical regret and reversal surgery among trans patients following breast- and genital-related procedures. These surgeons had collectively treated between 18,125 and 27,325 such patients, with only 62 having expressed regret and/or requested reversal surgery (a 0.2–0.3% regret rate). Since the study relied on surgeon survey responses of (presumably) patient self-reports, it cannot be justified as anything other than a lower bound on the prevalence of regret. This result seems also to have been reported in Danker et al. (2018)⁵³.
[Surgical regret] Jedrzejewski et al. (2023)⁵⁴ collected data from the Oregon Health & Science University (OHSU) Transgender Health Program’s (THP) “Regret and Request for Reversal” work group, a multidisciplinary work group of medical and psychological professionals that met over the course of 14 months. Their results included 1,989 individuals who underwent gender-affirming surgery between 2016 and 2021, with a 0.3% “regret rate” (n=6). However, as this study only documented requests for reversal, it is of low reliability for assessing surgical regret.
[GnRHa discontinuation] Brik et al. (2020)⁵⁵ studied a population of 143 minors who began puberty blockers at the Curium-Leiden University Medical Center GIC between November 2010 and January 2018. Nine patients (6.2%) stopped GnRHa treatment overall and 5 (3.5%) of these did not desire further gender-affirming treatment. Like Wiepjes et al. (2018) and other studies of this kind, Brik et al. (2020) only examined GnRHa discontinuation, and so cannot provide a reliable approximation of detransition rates.
[GnRHa discontinuation] Carmichael et al. (2021)⁵⁶ studied a cohort of 44 minors who commenced GnRHa treatment at the UK Gender Identity Development Service (GIDS) in London between June 2011 and April 2015. Medical and psychological testing was performed at 12-month intervals thereafter until the end of the treatment pathway at age 16. At this point, one patient had elected to stop GnRHa treatment without starting GAH — a 2.3% discontinuation rate.
[GnRHa, GAH discontinuation] Masic et al. (2022)⁵⁷ analyzed the treatment pathways of transgender adolescents who had been referred to UK endocrine clinics under the GIDS between January 2017 and December 2018. Of the 431 patients who consented to start GnRHa treatment, 30 (7.0%) did not commence or stopped treatment. Of these, only 3 listed “feeling unsure” as the cause; others transferred to private care, paused for fertility preservation, or had unknown reasons. In all, 17 participants (3.9%) did not access or discontinued GnRHa treatment and left the service, while 13 stayed with the service. No one who started GAH stopped it during the study period. Because of uncertainty in trajectories after cessation and large loss to follow-up, this study cannot accurately determine a detransition rate.
[GnRHa discontinuation] Nos et al. (2022)⁵⁸ tracked prescriptions of trans minors using the TRICARE health plan benefit of the U.S. Military Healthcare System (MHS) between October 2009 and April 2018. Of the 70 trans youth who were started on GnRHa treatment, 7 discontinued treatment and did not start GAH — a 10% discontinuation rate. This study only tracked data within a single insurance plan; those who continued to access the military health system but got their prescriptions or started GAH elsewhere would have been erroneously classified as discontinuing. Thus, it is not a reliable measure of GnRHa discontinuation.
[GAH discontinuation] Roberts et al. (2022)⁵⁹ — seemingly a companion study to Nos et al. (2022) — tracked prescriptions of trans adults and minors using the TRICARE health plan benefit of the MHS during the same 2009–2018 period. Active-duty service members were excluded due to their special requirements for accessing gender-affirming care. Of the 952 qualifying individuals, 29.8% had stopped receiving refills by the end of a 4-year period. This study suffers from the same limitation as its counterpart; in this case, the authors themselves note that the result is likely to be an overestimate. It is unclear how the adult 60.9% of the population having a 35.6% discontinuation rate and the minor 39.1% of the population having a 25.6% discontinuation rate is compatible with the overall 29.8% discontinuation rate.
[Desistance] Korte et al. (2008)⁶⁰ reviewed evidence on gender identity disorder (GID) in childhood, stating that “only 2.5% to 20% of all cases of GID in childhood and adolescence are the initial manifestation of irreversible transsexualism.” As proof, they cited two previous studies and one other review. One of the studies — Green et al. (1987)⁶¹ — concerned the “sexual fantasies of feminine boys” and included no explicit analysis of gender identity. The other study — Drummond et al. (2008)⁶² — used a sample of 25 girls who had been referred to a GIC, only 15 of whom were diagnosed with GID. The pattern is much the same within the cited review (Zucker 2005)⁶³: study samples generally consisted of children who had been referred to a GIC but not necessarily diagnosed with GID. Even beyond this, the paper lacked self-consistency: in the text, the original assertion is amended to “gender-atypical behavior in childhood” leading to persistent GID in 2.5% to 20% of cases.
[Desistance] Wallien and Cohen-Kettenis (2008)⁶⁴ conducted follow-up assessments of adolescents who had been referred to the VUmc GIC as young children between 1989 and 2005. The mean follow-up period was 10.3 years. 19 of the 77 patients had never met the criteria for a (DSM-III-R) GID diagnosis; yet the percentage of persistence (27%) was calculated over this whole group, including the non-responders and sub-threshold cases. This and other desistance studies are not relevant to detransition rates, as they mostly study children who never transitioned. Interestingly, but not surprisingly, this study found a strong relationship between the intensity of gender dysphoria and the likelihood of persistence.
[Desistance] Steensma et al. (2011)⁶⁵ conducted follow-up interviews of a cohort of minors who had received a DSM-IV or DSM-IV-TR GID diagnosis between January 2000 and January 2007 at the VUmc GIC. The mean follow-up period was 6.7 years. Of 53 participants, 24 (45.3%) had desisted. This paper shares the methodological shortcomings of Wallien and Cohen-Kettenis (2008). All 24 children who did not reapply for treatment during adolescence were assumed to have desisted, even though only 11 of them were interviewed. Moreover, all early-childhood referrals (mean age 9.4) were included regardless of whether they later returned to the clinic or began medical treatment. By definition, none of the desisters ever started medical transition in the first place.
[Desistance] Steensma et al. (2013)⁶⁶ picked up where the 2011 study left off, assessing 127 adolescents who were referred for treatment before the age of 12 and identifying 80 (63.0%) as desisters. The mean follow-up period was 7.0 years. Only 80 of the 127 patients met DSM-IV-TR GID diagnostic criteria in childhood. Like Wallien and Cohen-Kettenis (2008), this study showed that the probability of persistence was well-predicted by the intensity of gender dysphoria, as determined by the Gender Identity Interview for Children (GIIC) and pre-existing social role transition. But like other desistance studies, it did not study detransition among those who actually transitioned (n=4 social transitioners at T0 was too small to include here).
Importantly, concerns have been raised about the DSM-IV GID criteria employed by these desistance studies. According to Dr. Kelly Winters⁶⁷, “…the DSM-IV Subcommittee on Gender Identity Disorders deliberately chose to allow diagnosis of GIDC without any ‘explicit wish to be of the opposite sex’ (Bradley, et al., 1991⁶⁸) — a loophole that sidestepped gender dysphoria and was corrected in the DSM-5.” Review of these documents confirms this to be true: in both the DSM-IV and the DSM-IV-TR, “desire to be, or insistence that he or she is, the other sex” appeared as one of five criteria in subsection A, only four of which were required for the diagnosis⁶⁹𝄒⁷⁰. The other criteria were either (a) indicators of gender nonconformity and associated distress or (b) not mandatory for the diagnosis. Meanwhile, in the DSM-5, “a strong desire to be of the other gender or an insistence that one is the other gender” is a requirement for gender dysphoria in children⁷¹. This helps explain why Steensma et al. (2011, 2013) and others consistently found high desistance rates even within the GID subgroups of their studies: because GID could be diagnosed solely based on behavior as opposed to identity. Note that, while Steensma et al. (2011, 2013) did mention children wishing to be or identifying as the opposite sex, this was reported qualitatively or as a part of the GIIC score, and not as a precondition for GID diagnosis.
[Desistance] Ristori and Steensma (2016)⁷² reviewed previous follow-up studies of children referred to GICs. These papers include Green et al. (1987), Drummond et al. (2008), and Wallien and Cohen-Kettenis (2008), in addition to six papers⁷³⁻⁷⁸ from the 1960s, 1970s, and 1980s, and a doctoral dissertation which later became Singh et al. (2021; see below). The first three papers have low reliability for quantifying detransition rates (see above). The six older papers collectively reported only a single AFAB subject; these were studies about “feminine boys” who exhibited many of the signs of what today we might call gender dysphoria. These studies cannot characterize detransition rates because they used (a) extremely outdated GID diagnosis criteria (b) applied retrospectively if at all, (c) with no evidence of social transition and no option for medical transition. Furthermore, they often had extremely small sample sizes (n~10).
[Desistance] Singh et al. (2021)⁷⁹ assessed 139 boys referred to the Gender Identity Service, Child, Youth, and Family Program at the Centre for Addiction and Mental Health (CAMH) in Toronto between 1975 and 2009, reporting a whopping 88% desistance rate. The mean follow-up period was 12.9 years. Only 63.3% of participants initially met the diagnostic criteria for GID (DSM-III, DSM-III-R, or DSM-IV, depending on the time of referral), with the authors stating that the remainder would have met the criteria “at some point in their lives.” Throughout the paper, “gender identity at follow-up” was conflated with “persistent gender dysphoria”, which in turn was assessed by means of gender dysphoria questionnaires, clinical interviews, medical chart reviews, and even parental input.
A far more serious threat to the integrity of these results emerged in 2015, when CAMH was closed under questionable circumstances⁸⁰. An external review had alleged that Dr. Kenneth Zucker, CAMH’s chief psychologist, pushed a form of “reparative” therapy that involved encouraging dysphoric children to repress their transness, questioning children as young as 9 about their sexual fantasies, and hiding affirming community and medical resources from patients⁸¹. CAMH later issued a formal apology and paid Zucker a sum of $586,000, claiming that some of the statements in the review had been inaccurate⁸². Nevertheless, in his own words, Zucker promotes “active psychosocial treatment to reduce gender dysphoria so that the child’s eventual gender identity is more congruent with his or her biological sex.”⁸³ He depicts early social transition as causing persistent transgender identity⁸³, and, as described by J. Michael Bailey in The Man Who Would Be Queen⁸⁴ — itself criticized for transphobic bias⁸⁵⁻⁸⁷ — “thinks an important goal of treatment is to help children accept their birth sex and to avoid becoming transsexual.” This is a reparative therapy approach⁸⁸, and indeed, patients outside of the review have reported Zucker practicing this type of therapy⁸⁹𝄒⁹⁰. Between this and the 42% of children admitted to CAMH that never met the criteria for GD in the first place⁸⁰, it is unsurprising that studies undertaken at this clinic — including Singh et al. (2021) — reported anomalously high desistance rates⁹¹.
A common argument arising from these desistance studies is that transition actually causes transgender identity — that it is an instance of iatrogenic (treatment-induced) harm. This theory is not supported by the evidence. Wallien and Cohen-Kettenis (2008) and Steensma et al. (2013) established a strong relationship between GD intensity and probability of subsequent transition. Rae et al. (2019)⁹² went further, showing that children who would go on to socially transition already had similar gender identity and preferences to already-transitioned children and to cisgender children of the gender to which they would transition. It’s a classic case of correlation vs. causation: transitioning doesn’t make one trans, but being trans does make one more likely to transition and persist.
[Off-topic] Cohen et al. (2023)⁹³ studied a sample of 68 trans youth requesting GAH and/or surgery at a U.S. gender services program between 2010 and 2021, finding a 29% rate of “shifts in gender-related medical requests.” However, the sample was pre-selected to identify such shifts; those with short follow-up periods (<2 years; n=58) were excluded only if they did not experience a shift. Shifts were defined as “requesting gender-affirming hormones and/or surgery and later withdrawing that request one or more times” — a vague category that could include anything from temporarily canceling surgery to full detransition. Furthermore, the percentage of autistic patients in this study was substantially inflated (47%) due to the presence of an autism and gender diversity-specialized clinic.
4. Conclusions
The aim of this review was to investigate the body of work that is commonly cited in response to the question, “After the start of medical or social transition, what percentage of trans people experience regret or detransition?” During the analysis, studies were sorted into two categories based on their reliability for this purpose. “High-reliability” studies were those which answered the question explicitly with no glaring methodological errors. “Low-reliability” studies were those which either failed to answer the question or suffered from significant and pervasive methodological errors, or, in some cases, both. Composite rates of detransition and regret were then calculated using the high-reliability studies.
The high-reliability study group was as follows: 7 studies investigated detransition in 2,528 trans minors, 4 studies investigated detransition 1,693 trans adults (and possibly minors), and 4 publications [100 follow-up studies] investigated surgical regret in 3,279–4,029 trans adults and adolescents. The overall detransition rate was 3.3%; the detransition rate for trans youth was 4.0%; and the surgical regret rate was 1.7–2.1%. Averaging study results instead of pooling samples, the mean detransition rates were 4.0% overall and 3.4% for minors.
Low-reliability studies generally fell into one of three subgroups. First, there were 6 studies which found extremely low rates (~0.0–0.5%) but used questionable definitions or were especially vulnerable to loss to follow-up. Wiepjes et al. (2018) and Jedrzejewski et al. (2023) looked at hormonal detransition and requests for reversal, respectively, and so could not accurately describe surgical regret. Deutsch (2012), Davies et al. (2019), and Narayan et al. (2021) used automated search or survey methods that were highly susceptible to loss to follow-up. Richards and Doyle (2019) searched for detransition only among current GIC patients.
Another 7 studies quantified GnRHa or GAH treatment discontinuation specifically, usually over short timeframes. Brik et al. (2020), Carmichael et al. (2021), and Masic et al. (2021) all reported GnRHa discontinuation rates of between 2% and 7%, although many of these were for reasons unrelated to gender identity (pausing for fertility preservation, switching to a different provider, etc.). Wiepjes et al. (2018) reported a 1.8% GnRHa discontinuation rate, but Van der Loos et al. (2023) recently duplicated this statistic on a newer version of the same ACOG dataset. Smith et al. (2005) reported a 15% “drop-out group” of patients who stopped GAH; this figure is hard to evaluate, since it was presented only briefly (the paper itself was about post-surgical outcomes). Nos et al. (2022) and Roberts et al. (2022) reported rates of 10% and 30%, respectively, but were not reliable measures of discontinuation because they used data from a single insurance plan within the U.S. Military Health System.
The final and largest low-reliability subgroup consisted of studies that were simply not relevant to this topic, as they primarily studied young children who never transitioned. These studies often included children without a diagnosis of GID or GD as well as those who received a diagnosis at a very young age but never sought social or medical intervention. An additional source of error was present in those that used DSM-IV or DSM-IV-TR GID diagnostic criteria, which did not require explicit identification with or desire to be the other sex. Some of these studies produced useful results anyway, as in the case of Wallien and Cohen-Kettenis (2008) and Steensma (2013), which showed a strong relationship between GD intensity and likelihood of transition. Others, unfortunately, are tainted by credible allegations that children were pressured to desist in therapy.
The uncertainty remaining within this body of work is largely a function of opposing biases. On the one hand, widespread loss to follow-up suggests that rates may be underestimated, since those who detransition might disengage from their health care provider. This goes doubly for surgical regret studies, few of which have more than 85% follow-up. There is also a distinct lack of long-term (>5 year) follow-up studies which could capture delayed de- and re-transition. On the other hand, pervasive social pressure against transgender identity suggests that rates may be overestimated, with many cases of detransition being the result of outside circumstances. This issue is highlighted by the results of the 2015 U.S. Transgender Survey, in which 13.1% of 17,151 post-transition respondents reported a history of detransition but only 0.7% listed “transition was not right for them” as a motivating factor⁹⁴𝄒⁹⁵. Far more common were external factors such as pressure from a parent or other family member, difficulty finding a job or pressure from an employer, harassment and discrimination, and transition simply being too daunting to continue.
Efforts were made to mitigate these biases throughout the high-reliability study selections. Detransition and surgical regret studies that were especially vulnerable to loss to follow-up were eliminated. The Bustos et al. (2021) meta-analysis was re-analyzed, resulting in the exclusion of three large surgical regret studies (n ~4,000 total) and raising its rate from 1.0% to 2.7%. And finally, patient motivation was factored in where possible, with remission of GD, resumed cisgender identity, or explicit mention of detransition being preferred. In cases where inadequate information was provided, all cases of discontinuation were retained.
It is often said that there is no reliable evidence about detransition rates in minors, but that is no longer true; multiple large and rigorous studies on the topic (Olson et al. 2022, Butler et al. 2022, Van der Loos et al. 2022, 2023) have been published in the last year. With this, childhood detransition frequency is arguably better characterized than adult detransition frequency — a domain in which even the high-reliability papers have obvious caveats. Pazos Guerra et al. (2020) simply alluded to “follow-ups”. Boyd et al. (2022) had a particularly small sample size, and Hall et al. (2021) mentioned that only 9 of their 12 detransitioners had evidence of GAH discontinuation and 4 of the 12 were re-referred to the GIC; their subsequent trajectories were not specified.
Regardless of the minutiae, the rates are reassuringly low: ~1.7–2.1% surgical regret and ~3.3% detransition over the course of several years. These figures were derived from what I hope was a truly neutral evaluation of the research. Unreliable studies showing high rates were purged, but so were those showing low rates. Certain considerations subtracted detransitioners, others added them. Results were recalculated. Errors were explained. In a debate drawn along sharply political lines, this was not a partisan analysis. Yet, at the end of the day, the data are clear: those who transition report high rates of persistence.
Appendix A. Analysis of the Bustos et al. (2021) studies
Three of the studies counted in Bustos et al. (2021) — Kuiper et al. (1998), Landén et al. (1998), and Wiepjes et al. (2018) — were deemed to be of low reliability for quantifying surgical regret. The characteristics of the remaining 24 studies are described below.
End Statement
I do want to clarify something. This review eliminated studies that have any potentially significant bias; however, there are levels within this. Deutsch (2012), Davies et al. (2019), Richards and Doyle (2019), Narayan et al. (2021), and Jedrzejewski et al. (2023) are clearly more reliable than the desistance studies, which are not only flawed, but also irrelevant for this topic. In fact, studies which show extremely low (<1%) surgical regret rates could well be closer to the reality than the rest of the surgical regret literature, much of which comes from 30–50 years ago when trans medical care, surgical practice, and cultural acceptance were nothing like they are today. This review should NOT be taken as an indictment of these papers. If Narayan et al. (2021) is included in the surgical regret statistic, it becomes 0.43% — 0.62%.
I would also like to note that there is uncertainty around the inclusion of Pazos Guerra et al. (2020), which did not describe their data collection methodology in detail but simply stated that they searched for cases of medical, social, or administrative detransition via follow-ups. If this paper were eliminated, the overall detransition rate would be 3.9% and the combined regret/detransition rate would be 2.7% — 3.0%.
Disclaimer
I am not a medical or psychiatric researcher. If mistakes were made, let it be known that it was due to incompetence.
Edit Log
3/24/23: “End statement” added
3/25/23: De Castro et al. (2022) added
3/25/23: “End statement” expanded
References
n1. De Castro, C., and Coauthors, 2022: High persistence in Spanish transgender minors: 18 years of experience of the Gender Identity Unit of Catalonia. Revista de Psiquiatría y Salud Mental, ISSN 1888–9891, https://doi.org/10.1016/j.rpsm.2022.02.001.
1. https://www.mapresearch.org/2023-press-release-under-fire-report
3. https://www.aclu.org/legislative-attacks-on-lgbtq-rights
4. https://www.tracktranslegislation.com/
7. https://twitter.com/RepMTG/status/1595113301176111104
8. https://twitter.com/tedcruz/status/1584588208238575616
11. https://www.mediamatters.org/daily-wire/daily-wire-host-you-have-ban-transgenderism-entirely
15. https://breakpoint.org/when-mental-illness-goes-viral-social-contagions-are-destroying-our-girls/
17. https://www.floridahealth.gov/newsroom/2022/04/20220420-gender-dysphoria-press-release.pr.html
19. https://twitter.com/ErinInTheMorn/status/1628429657451143173
23. Khatchadourian, K., S. Amed, and D. L. Metzger, 2014: Clinical Management of Youth with Gender Dysphoria in Vancouver. The Journal of Pediatrics, 164, 906–911, https://doi.org/10.1016/j.jpeds.2013.10.068.
24. McCallion, S., S. Smith, H. Kyle, M. G. Shaikh, G. Wilkinson, and A. Kyriakou, 2021: An appraisal of current service delivery and future models of care for young people with gender dysphoria. Eur. J. Pediatr., 180, 2969–2976, https://doi.org/10.1007/s00431-021-04075-2.
25. Butler, and Coauthors, 2022: Discharge outcome analysis of 1089 transgender young people referred to paediatric endocrine clinics in England 2008–2021. Archives of Disease in Childhood, 107, 1018–1022, http://dx.doi.org/10.1136/archdischild-2022-324302.
26. Olson, K. R., L. Durwood, R. Horton, N. M. Gallagher, and A. Devor, 2022: Gender Identity 5 Years After Social Transition. Pediatrics, 150, e2021056082, https://doi.org/10.1542/peds.2021-056082.
27. https://hudl.princeton.edu/research-0
28. Loos, M. A. T. C. van der, and Coauthors, 2023: Children and adolescents in the Amsterdam Cohort of Gender Dysphoria: trends in diagnostic- and treatment trajectories during the first 20 years of the Dutch Protocol. The Journal of Sexual Medicine, 20, 298–409, https://doi.org/10.1093/jsxmed/qdac029.
29. Loos, M. A. T. C. van der, S. E. Hannema, D. T. Klink, M. den Heijer, and C. M. Wiepjes, 2022: Continuation of gender-affirming hormones in transgender people starting puberty suppression in adolescence: a cohort study in the Netherlands. The Lancet: Child and Adolescent Health, 6, 869–875, https://doi.org/10.1016/S2352-4642(22)00254-1.
30. Dhejne, C., K. Öberg, S. Arver, and M. Landén, 2014: An Analysis of All Applications for Sex Reassignment Surgery in Sweden, 1960–2010: Prevalence, Incidence, and Regrets. Arch. Sex. Behav. 43, 1535–1545, https://doi.org/10.1007/s10508-014-0300-8.
31. https://reproductiverights.org/sweden-drops-law-forcing-sterilization-of-trans-people/
32. Pazos Guerra, M., M. G. Balaguer, M. G. Poras, F. H. Murillo, E. S. Izquierdo, and C. M. Ariño, 2020: Transsexuality: Transitions, detransitions, and regrets in Spain. Endocrinol. Diabetes Nutr. (Engl. Ed), 67, 562–567, https://doi.org/10.1016/j.endinu.2020.03.008.
33. Hall, R., L. Mitchell, and J. Sachdeva, 2021: Access to care and frequency of detransition among a cohort discharged by a UK national adult gender identity clinic: Retrospective case-note review. BJPsych Open, 7, E184, https://doi.org/10.1192/bjo.2021.1022.
34. Boyd, I., T. Hackett, and S. Bewley. 2022: Care of Transgender Patients: A General Practice Quality Improvement Approach. Healthcare, 10, 121, https://doi.org/10.3390/healthcare10010121.
35. Pfafflin, F., 1993: Regrets After Sex Reassignment Surgery. Journal of Psychology & Human Sexuality, 5, 69–85, https://doi.org/10.1300/j056v05n04_05.
36. Smith, Y. L. S., S. H. M. van Goozen, A. J. Kuiper, and P. T. COHEN-KETTENIS, 2005: Sex reassignment: outcomes and predictors of treatment for adolescent and adult transsexuals. Psychological Medicine, 35, 89–99, https://doi.org/10.1017/s0033291704002776.
37. Vries, A. L. C. de, T. D. Steensma, T. A. H. Doreleijers, and P. T. Cohen‐Kettenis, 2011: Puberty Suppression in Adolescents With Gender Identity Disorder: A Prospective Follow‐Up Study. The Journal of Sexual Medicine, 8, 2276–2283, https://doi.org/10.1111/j.1743-6109.2010.01943.x.
38. Vries, A. L. C. de, J. K. McGuire, T. D. Steensma, E. C. F. Wagenaar, T. A. H. Doreleijers, and P. T. Cohen-Kettenis, 2014: Young Adult Psychological Outcome After Puberty Suppression and Gender Reassignment. PEDIATRICS, 134, 696–704, https://doi.org/10.1542/peds.2013-2958.
39. Bustos, V. P., and Coauthors, 2021: Regret after Gender-affirmation Surgery: A Systematic Review and Meta-analysis of Prevalence. Plast. Reconstr. Surg. Glob. Open, 9, e3477, https://doi.org/10.1097/GOX.0000000000003477.
40. Expósito-Campos, P., and R. D’Angelo, 2021: Letter to the Editor: Regret after Gender-affirmation Surgery: A Systematic Review and Meta-analysis of Prevalence. Plastic and Reconstructive Surgery — Global Open, 9, p e3951, https://doi.org/10.1097/GOX.0000000000003951.
41. Escandón, J. M., and Coauthors, 2021: Reply: Regret after Gender-affirmation Surgery: A Systematic Review and Meta-analysis of Prevalence. Plastic and Reconstructive Surgery — Global Open, 9, p e3954, https://doi/org/10.1097/GOX.0000000000003954.
42. Bustos, V. P., and Coauthors, 2022: Regret after Gender-affirmation Surgery: A Systematic Review and Meta-analysis of Prevalence — Erratum. Plastic and Reconstructive Surgery — Global Open, 10, p e4340, https://doi.org/10.1097/GOX.0000000000004340.
43. Wiepjes, C. M., and Coauthors, 2018: The Amsterdam Cohort of Gender Dysphoria Study (1972–2015): Trends in Prevalence, Treatment, and Regrets. J. Sex. Med., 15, 582–90, https://doi.org/10.1016/j.jsxm.2018.01.016.
44. Landén, M., J. Wålinder, G. Hambert, and B. Lundström, 1998: Factors predictive of regret in sex reassignment. Acta. Psychiatr. Scand., 97, 284–289, https://doi.org/10.1111/j.1600-0447.1998.tb10001.x.
45. Kuiper, A. J., and P. T. Cohen-Kettenis, 1998: Gender role reversal among postoperative transsexuals. International Journal of Transgenderism, 2, ISSN 1434–4599.
46. Jiang, D., J. Witten, J. Berli, and D. Dugi, 2018: Does Depth Matter? Factors Affecting Choice of Vulvoplasty Over Vaginoplasty as Gender-Affirming Genital Surgery for Transgender Women. The Journal of Sexual Medicine, 15, 902–906, https://doi.org/10.1016/j.jsxm.2018.03.085.
47. Tang, A., and Coauthors, 2022: Gender-Affirming Mastectomy Trends and Surgical Outcomes in Adolescents. Annals of Plastic Surgery, 88, 325–331, PMCID: PMC9555285 (available on 2023–05–01).
48. https://www.reuters.com/investigates/special-report/usa-transyouth-outcomes/
49. Deutsch, M. B., 2012: Use of the Informed Consent Model in the Provision of Cross-Sex Hormone Therapy: A Survey of the Practices of Selected Clinics. International Journal of Transgenderism, 13, 140–146, https://doi.org/10.1080/15532739.2011.675233.
50. Davies, S., S. McIntyre, and C. Rypma, 2019: Detransition rates in a national UK Gender Identity Clinic. Poster session, 3rd Biennial EPATH Conference Inside Matters: On Law, Ethics, and Religion. https://epath.eu/wp-content/uploads/2019/04/Boof-of-abstracts-EPATH2019.pdf
51. Richards, C., and J. Doyle, 2019: Detransition rates in a large national gender identity clinic in the UK. Counselling Psychology Review, 34, https://doi.org/10.53841/bpscpr.2019.34.1.60.
52. Narayan, S. K., and Coauthors, 2021: Guiding the conversation-types of regret after gender-affirming surgery and their associated etiologies. Ann. Transl. Med., 9, 605, https://doi.org/10.21037/atm-20-6204.
53. Danker, S., S. K. Narayan, R. Bluebond-Langner, L. S. Schechter, and J. U. Berli, 2018: Abstract: A Survey Study of Surgeons’ Experience with Regret and/or Reversal of Gender-Confirmation Surgeries. Plastic and Reconstructive Surgery — Global Open, 6, 189, https://doi.org/10.1097/01.GOX.0000547077.23299.00.
54. Jedrzejewski, B. Y., M. C. Marsiglio, J. Guerriero, A. Penkin, OHSU Transgender Health Program “Regret and Request for Reversal” workgroup, and J. U. Berli, 2023: Regret after Gender Affirming Surgery — A Multidisciplinary Approach to a Multifaceted Patient Experience. Plastic and Reconstructive Surgery, published online ahead of print, https://doi.org/10.1097/PRS.0000000000010243.
55. Brik, T., L. J. J. J. Vrouenraets, M. C. de Vries, and S. E. Hannema, 2020: Trajectories of Adolescents Treated with Gonadotropin-Releasing Hormone Analogues for Gender Dysphoria. Arch. Sex. Behav., 49, 2611–2618, https://doi.org/10.1007/s10508-020-01660-8.
56. Carmichael, P., and Coauthors, 2021: Short-term outcomes of pubertal suppression in a selected cohort of 12 to 15 year old young people with persistent gender dysphoria in the UK. PLoS ONE, 16, e0243894, https://doi.org/10.1371/journal.pone.0243894.
57. Masic, U., G. Butler, P. Carruthers, and P. Carmichael, 2022: Trajectories of transgender adolescents referred for endocrine intervention in England. Archives of Disease in Childhood, 107, 1012–1017, http://dx.doi.org/10.1136/archdischild-2022-324283.
58. Nos, A. L., and Coauthors, 2022: Association of Gonadotropin-Releasing Hormone Analogue Use With Subsequent Use of Gender-Affirming Hormones Among Transgender Adolescents. JAMA Network Open, 5, e2239758, https://doi.org/10.1001/jamanetworkopen.2022.39758.
59. Roberts, C. M., D. A Klein, T. A. Adirim, N. A. Schvey, and E. Hisle-Gorman, 2022: Continuation of Gender-affirming Hormones Among Transgender Adolescents and Adults. The Journal of Clinical Endocrinology & Metabolism, 107, e3937–e3943, https://doi.org/10.1210/clinem/dgac251.
60. Korte, A., D. Goecker, H. Krude, U. Lehmkuhl, A. Grüters-Kieslich, and K. M. Beier, 2008: Gender Identity Disorders in Childhood and Adolescence: Currently Debated Concepts and Treatment Strategies. Dtsch Arztebl International, 105, 834–841, https://doi.org/10.3238/arztebl.2008.0834.
61. Green, R., C. W. Roberts, K. Williams, M. Goodman, and A. Mixon, 1987: Specific cross-gender behaviour in boyhood and later homosexual orientation. Br. J. Psychiatry, 151, 84–88, https://doi.org/10.1192/bjp.151.1.84.
62. Drummond K. D., S. J. Bradley, M. Peterson-Badali, and K. J. Zucker, 2008: A follow-up study of girls with gender identity disorder. Developmental Psychology, 44, 34–45, https://doi.org/10.1037/0012-1649.44.1.34.
63. Zucker, K. J., 2005: Gender identity disorders in children and adolescents. Annu. Rev. Clin. Psychol., 1, 467–492, https://doi.org/10.1146/annurev.clinpsy.1.102803.144050.
64. Wallien, M. S., and P. T. Cohen-Kettenis, 2008: Psychosexual outcome of gender-dysphoric children. J. Am. Acad. Child Adolesc. Psychiatry, 47, 1413–1423, https://doi.org/10.1097/CHI.0b013e31818956b9.
65. Steensma, T. D., R. Biemond, F. de Boer, and P. T. Cohen-Kettenis, 2011: Desisting and persisting gender dysphoria after childhood: a qualitative follow-up study. Clin. Child Psychol. Psychiatry, 16, 499–516, https://doi.org/10.1177/1359104510378303.
66. Steensma, T. D., J. K. McGuire, B. P. Kreukels, A. J. Beekman, and P. T. Cohen-Kettenis, 2013: Factors associated with desistence and persistence of childhood gender dysphoria: a quantitative follow-up study. J. Am. Acad. Child Adolesc. Psychiatry, 52, 582–590, https://doi.org/10.1016/j.jaac.2013.03.016.
68. S. Bradley, and Coauthors, 1991: Interim Report of the DSM-IV Subcommittee on Gender Identity Disorder. Arch. Sex. Behav., 20, 333–343, https://doi.org/10.1007/BF01542614.
69. American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.).
70. American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental Disorders (4th ed., text rev.).
71. American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th ed.).
72. Ristori, J., and T. D. Steensma, 2016: Gender dysphoria in childhood. International Review of Psychiatry, 28, 13–20, https://doi.org/10.3109/09540261.2015.1115754.
73. Bakwin, H., 1968: Deviant gender-role behavior in children: Relation to homosexuality. Pediatrics, 41, 620–629, https://doi.org/10.1542/peds.41.3.620.
74. Lebovitz, P. S., 1972: Feminine Behavior in Boys: Aspects of Its Outcome. American Journal of Psychiatry, 128, 1283–1289, https://doi.org/10.1176/ajp.128.10.1283.
75. Zuger, B., 1984: Early Effeminate Behavior in Boys. The Journal of Nervous and Mental Disease, 172, 90–97, https://doi.org/10.1097/00005053-198402000-00005.
76. Money, J., and A. J. Russo, 1979: Homosexual Outcome of Discordant Gender Identity/Role in Childhood: Longitudinal Follow-Up. Journal of Pediatric Psychology, 4, 29–41, https://doi.org/10.1093/jpepsy/4.1.29.
77. Davenport, C. W., 1986: A follow-up study of 10 feminine boys. Archives of Sexual Behavior, 15, 511–517, https://doi.org/10.1007/bf01542316.
78. Kosky, R. J., 1987: Gender‐disordered children: does inpatient treatment help? Medical Journal of Australia, 146, 565–569, https://doi.org/10.5694/j.1326-5377.1987.tb120415.x.
79. Singh, D., S. J. Bradley, and K. J. Zucker, 2021: A Follow-Up Study of Boys With Gender Identity Disorder. Frontiers in Psychiatry, 12, https://doi.org/10.3389/fpsyt.2021.632784.
80. https://www.huffpost.com/entry/the-end-of-the-desistance_b_8903690
81. https://www.transadvocate.com/wp-content/uploads/GIC-Review-26Nov2015-TA1.pdf
82. https://www.cbc.ca/news/canada/toronto/camh-settlement-former-head-gender-identity-clinic-1.4854015
83. Zucker, K. J., 2019: Debate: Different strokes for different folks. Child Adolesc. Ment. Health, 25, 36–37, https://doi.org/10.1111/camh.12330.
84. https://faculty.wcas.northwestern.edu/JMichael-Bailey/TMWWBQ.pdf
85. Dreger, A. D., 2008: “The Controversy Surrounding The Man Who Would Be Queen: A Case History of the Politics of Science, Identity, and Sex in the Internet Age”. Archives of Sexual Behavior, 37, 366–421, https://doi.org/10.1007/s10508-007-9301-1.
86. https://web.archive.org/web/20040614034224/http://www.seedmagazine.com/?p=article&n=above&id=130
87. https://reason.com/2003/11/01/queer-science-2/
88. http://www.florenceashley.com/uploads/1/2/4/4/124439164/ashley_reparative_therapy.pdf
89. https://www.npr.org/2008/05/07/90247842/two-families-grapple-with-sons-gender-preferences
92. Rae, J. R., and Coauthors, 2019: Predicting early-childhood gender transitions. Psychological Science, 30, 669–681, https://doi.org/10.1177/0956797619830649.
93. Cohen, A., and Coauthors, 2023: Shifts in Gender-Related Medical Requests by Transgender and Gender-Diverse Adolescents. J. Adolesc. Health, 72, 428–436, https://doi.org/10.1016/j.jadohealth.2022.10.020.
94. James, S. E., J. L. Herman, S. Rankin, M. Keisling, L. Mottet, and M. Anafi, 2016: The Report of the 2015 U.S. Transgender Survey. Washington, DC: National Center for Transgender Equality, https://www.ustranssurvey.org/reports.
95. Turban, J. L., S. S. Loo, A. N. Almazan, and A. S. Keuroghlian, 2021: Factors Leading to “Detransition” Among Transgender and Gender Diverse People in the United States: A Mixed-Methods Analysis. LGBT Health, 8, 273–380, https://doi.org/10.1089/lgbt.2020.0437.
