Gender affirming hormone therapy for individuals with gender dysphoria aged <26 years: a systematic review and meta-analysis

Abstract

Objective

In this systematic review and meta-analysis, we assessed and summarised the certainty of the evidence about the effects of gender affirming hormone therapy (GAHT) in individuals with gender dysphoria (GD).

Methods

We searched Medline, Embase, PsychINFO, Social Sciences Abstracts, LGBTQ+ Source and Sociological Abstracts from inception to September 2023. We included studies comparing GAHT with no GAHT in individuals aged <26 years with GD. Outcomes of interest included psychological and physical effects. Pairs of reviewers independently screened articles, abstracted data and assessed the risk of bias in the included studies. We performed meta-analyses and assessed the certainty of the evidence using the grading of recommendations assessment, development and evaluation (GRADE) approach.

Results

We included 24 studies. Comparative observational studies (n=9) provided mostly very low certainty evidence regarding GD, global function and depression. One comparative observational study reported that the odds of depression may be lower (OR 0.73 (95% CI 0.61 to 0.88), n (number of studies)=1, low certainty) in individuals who received GAHT compared with those who did not. Before–after studies (n=13) provided very low certainty evidence about GD, global function, depression and bone mineral density. Case series studies (n=2) provided high certainty evidence that the proportion of individuals with cardiovascular events 7–109 months after receiving GAHT was 0.04 (95% CI 0.03 to 0.05, n=1, high certainty).

Conclusion

There is considerable uncertainty about the effects of GAHT and we cannot exclude the possibility of benefit or harm. Methodologically rigorous prospective studies are needed to produce higher certainty evidence.

Trial registration number

PROSPERO CRD42023452171.

WHAT IS ALREADY KNOWN ON THIS TOPIC.

WHAT THIS STUDY ADDS

• This study addressed the effects of GAHT in individuals with GD, while adhering to the highest methodological standards for conducting and reporting a systematic review and meta-analysis.

• The risk of bias in each included study and the certainty of the evidence for each outcome of interest were assessed.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE AND POLICY

• Evidence from this systematic review and meta-analysis can be used to inform individuals experiencing GD and considering GAHT, clinicians involved in their care as well as clinical practice guideline developers, policy makers and stakeholders who make decisions about treatment related to GD.

Introduction

Gender dysphoria (GD) refers to the intense distress caused by feelings of incongruence between one’s birth assigned sex and gender identity.¹ Individuals with persistent GD may seek hormonal and surgical interventions to align their physical bodies with their internal or expressed gender and alleviate this distress.²

Hormonal treatments for GD in youth include gonadotropin releasing hormone analogues (GnRHa) and gender affirming hormone therapy (GAHT). GnRHa (puberty blockers) may be administered as early as Tanner stage 2,³ followed by GAHT in adolescence to induce and maintain the desired secondary sex characteristics. Hormone therapies include the administration of testosterone for natal females (NFs) to create a masculinised appearance and oestrogen in conjunction with GnRHa for natal males (NMs) to produce a feminised appearance. Early interventions with puberty blockers followed by GAHT is believed to result in better physical outcomes aligned with the desired gender,^{4 5} although some individuals may receive only GAHT.

A high quality systematic review (SR) is needed to overcome methodological limitations in this field. This SR and meta-analysis aimed to summarise the effects of GAHT in individuals with GD aged <26 years.

Methods

We report this SR and meta-analysis following the guidance of the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) checklist (online supplemental appendix 1).

Eligibility criteria

Table 1 lists the eligibility criteria.

Table 1. Eligibility criteria.

Types of studies	We included randomised controlled trials, comparative observational studies and before–after studies addressing the intervention and comparison of interest. We also included case series addressing the intervention of interest in special instances. We did not find any randomised controlled trials and included all eligible comparative observational and before–after studies. As for case series, if an outcome of interest was not reported in the eligible comparative observational or before–after studies, we included all eligible case series studies addressing that outcome. We included studies published in full and in the English language.
Population	We included individuals aged <26 years who were diagnosed, experienced, self-identified or were identified by a parent as having GD, gender identity disorder or gender incongruence, or who identified as transgender or non-binary. To be as inclusive as possible, we included all studies where the mean age of participants was <26 years. We decided to include individuals aged <26 years because the definition of youth, the target population of this review, is commonly defined as extending into the mid-twenties.41 42
Intervention	We included studies assessing the effects of GAHT. We defined GAHT as stated by the authors or as the use of feminising hormones in an individual assigned male at birth or as the use of masculinising hormones in an individual assigned female at birth.
Comparator	The comparator of interest was no GAHT (eg, psychological therapy, no treatment). In case series studies, a comparator group was not necessary.
Outcomes	We included studies reporting on any of the following outcomes if follow-up was short term (≤6 months) or long term (≥1 year): GD, completed suicides, global function, depression, sexual dysfunction from physiological perspective (ie, lack of erection, dyspareunia, problems related to dry and degenerated mucosal tissue, and anorgasmia), bone mineral density and cardiovascular events.

GAHT, gender affirming hormone therapy; GD, gender dysphoria.

Information sources

With the assistance of an information specialist (RC), we searched Medline, Embase, PsycINFO, Social Sciences Abstracts, Contemporary Women’s Issues, LGBTQ+Source, Sociological Abstracts, Studies on Women, Gender Abstracts and Google Scholar from inception to September 2023. This search was part of an umbrella search for a related SR.⁶ All search strategies are detailed in online supplemental appendix 2.

Study selection

Two reviewers (SI, YMR), using Covidence software (https://www.covidence.org/) and after training and calibration exercises, independently screened titles and abstracts and full texts of potentially eligible studies. A third reviewer (AM) resolved any conflicts. Study selection for this SR was completed in tandem with another related SR at the abstract and full text stages.⁶

Data collection

Online supplemental appendix 3 has details on data collection.

Risk of bias in individual studies

We assessed the risk of bias using a modified version of the Cochrane risk of bias tool for non- randomised studies of interventions (ROBINS-I) for each study design (onlinesupplemental appendices 4 5).

Data synthesis

Although the authors of the included studies used various observational study designs, we classified studies according to how the data were analysed for this review (online supplemental appendix 6). For dichotomous outcomes, we summarised the effect of interventions using odds ratios (ORs) in comparative observational and before–after studies, and proportions (ie, number of events per number of participants in the study group) in case series studies. For continuous outcomes, we summarised the effects of interventions using mean difference in comparative observational studies (ie, difference in scores between the study groups), mean change in before–after studies (ie, difference in scores before and after intervention) and mean in case series. Because the study authors did not provide correlation coefficients, we imputed a moderate correlation coefficient (r=0.5) when calculating mean change. We calculated 95% CIs around all estimates.

We conducted a meta-analysis using a random effects model when appropriate, as determined by subject area experts (CK-M, SM), for studies addressing the same outcome and with no clinical heterogeneity between them (ie, study design, population, intervention/comparator and outcome definition). When studies reported outcomes using different scales, we calculated the standardised mean change for before–after studies. If we could not perform a meta-analysis, we summarised the evidence across studies. We used the meta and metafor packages in R Studio V.4.2 for the analyses.

Certainty of the evidence

We assessed the certainty of the evidence following the grading of recommendations assessment, development and evaluation (GRADE) approach⁷ (online supplemental appendix 7). We assessed the certainty in the causal effects of GAHT on the outcomes of interest rather than the association between GAHT as an exposure. We followed GRADE guidance and principles to address questions about interventions using observational studies. This process involves clarifying the question (target of certainty), defining the intent of the question (causality) and assessing the certainty of the evidence under those parameters.

When assessing risk of bias for each outcome, we rated down the certainty of evidence from observational studies by up to three levels due to prognostic imbalance. In case series, outcomes requiring a comparison group (eg, GD, completed suicides, global function, depression, sexual dysfunction and bone mineral density (BMD)) were rated down three levels due to the absence of such a group. However, outcomes not requiring a comparison group (eg, cardiovascular events linked to GAHT) were not rated down, as these events were specific to intervention recipients.

To minimise value judgments, we used a null effect threshold (1 for relative measures and 0 for absolute measures, mean differences or mean changes) to rate the certainty of any benefit or harm (of any magnitude) from receiving GAHT over not receiving GAHT. We did not define a minimally important difference to determine whether an effect was clinically meaningful or important.

Subgroup and sensitivity analyses

Online supplemental appendix 8 describes the subgroup and sensitivity analyses.

Management of conflicts of interest

Infomation on management of conflicts of interest is presented in online supplemental appendix 9. Other SRs under the described agreement include SRs about the effects of social gender transition, mastectomy,⁸ chest binding and genital tucking, and puberty blockers (all submitted for publication).

Results

After screening 6736 titles and abstracts for this SR and another related SR,⁶ we included 24 studies. Figure 1 shows the study search and selection process. We present the reasons for exclusion at the full text screening stage (n=311), with references, in online supplemental appendix 10.

Figure 1. Preferred Reporting Items for Systematic Reviews and Meta-analyses 2020 flow diagram for new systematic reviews which included searches of databases and register only. ^ΏThis was an umbrella search completed for two related systematic reviews and meta-analyses. 24 studies were included in this systematic review. The studies that were in another review are part of the studies excluded for wrong intervention. *Ten of 27 studies excluded for wrong intervention were included in another review. Source: Page et al.⁴³.

Characteristics of included studies

Of 24 included studies, 9 were comparative observational,9,17 13 were before–after18,30 and 2 were case series^{31 32} (figure 1). Thirteen studies included NMs and NFs, and 11 included NFs only. Mean age of participants at the time of GAHT ranged from 15.1 (SD 1.8) to 25.1 (SD 4.8). Characteristics of the included studies are presented in online supplemental appendix 11. Online supplemental appendix 12 describes the outcome measurement instruments used in the studies and their interpretability.

Risk of bias in included studies

Across comparative observational studies, the domains commonly judged as serious or critical risk of bias were confounding, missing data and deviation from intended intervention (ie, administration of co-interventions). Before–after studies were at serious or critical risk of bias due to missing data and deviation from intended intervention (ie, administration of co-interventions). In addition to lacking a comparison group, case series were at critical risk of bias due to measurement of the outcome (online supplemental appendix 13).

Effects of interventions

We described the effects of the interventions for each study design. Tables24 provide summaries of our findings, and online supplemental appendix 14 displays forest plots of the meta-analysis. If sex specific data were available, we included separate data points for NMs and NFs in each meta-analysis (online supplemental appendix 14). When studies reported data for both groups and no important heterogeneity was found, we presented a single combined effect estimate.

Table 2. Gender affirming hormone therapy versus no gender affirming hormone therapy: evidence from comparative observational studies.

Outcomes	Anticipated absolute effects* (95% CI)		Relative effect(95% CI)	No of participants(studies)	Certainty of evidence(GRADE)	Comments
	Risk with no GAHT	Risk with GAHT
GD, current (no follow-up) assessed with: participant reported Gender Distress Scale, higher scores indicate higher gender distress, scale 1–5	Mean GD, current (no follow-up) was 4.17	MD 0.4 lower (0.24 lower to 0.16 higher)	–	146 (1 non-randomised study)13	⨁◯◯◯ Very low†‡	The evidence is very uncertain about the effect of GAHT on GD (no follow-up) in natal males and natal females
Global function, long term follow-up assessed with: participant reported various scales (SCL-90-R, Global Severity Index, CGAS), higher scores indicate better global function, follow-up 12–24 months§	–	SMD 0.87, SD higher (0.25 lower to 2 higher)	–	125 (2 non-randomised studies)10 12	⨁◯◯◯ Very low¶**††	The evidence is very uncertain about the effect of GAHT on global function at long term follow-up in natal males and natal females
Depression, long term follow-up assessed with: participant reported various scales (SCL-90-R, Depression Domain, Children's Depression Inventory), higher scores indicate worse depression, follow-up mean 12 months§	–	SMD 0.3, SD lower (0.85 lower to 0.25 higher)	–	154 (2 non-randomised studies)12 14	⨁◯◯◯ Very low††‡‡§§	The evidence is very uncertain about the effect of GAHT on depression at long term follow-up in natal males and natal females
Other outcomes, not measured¶¶	–	–	–	–	–	–

Grading of recommendations assessment, development and evaluation (GRADE) Working Group grades of evidence. High certainty=we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty=we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty=our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty=we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^*The risk in the intervention group (with 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (with 95% CI).

^†Rated down three levels due to critical risk of bias because of lack of adjustment for important confounders (ie, psychiatric interventions, mental health comorbidities, socioeconomic status and family support) and missing data (ie, 41.71% provided outcome data).

^‡Rated down one level for imprecision as the CIs crossed the threshold of no effect (ie, MD=0), suggesting both a possibility of a benefit or harm in the outcome.

^§Long term follow-up: outcome measured at ≥12 months of follow-up.

^¶Rated down three levels due to critical risk of bias because of lack of adjustment for important confounders in the two included studies (ie, psychiatric interventions, mental health comorbidities, socioeconomic status and family support) and missing data in one of the two included studies (ie, 37% provided outcome data).

^**Statistically, there was considerable heterogeneity with I2=88% and p<0.01. However, we did not rate down for inconsistency as this heterogeneity could be explained by the fact that one of the two included studies measured the outcome only in natal female participants, while the other study measured the outcome in natal female and male participants.

^††Rated down one level for imprecision as the CIs crossed the threshold of no effect (ie, SMD=0), suggesting both a possibility of a benefit or harm in the outcome.

^‡‡Rated down three levels due to critical risk of bias because of lack of adjustment for important confounders in the two included studies (ie, psychiatric interventions, mental health comorbidities, socioeconomic status and family support) and serious risk of bias due to deviation of intended intervention in one of the included studies (28.26% of the participants in the no GAHT group were receiving puberty blockers or spironolactone as monotherapy).

^§§Statistically, there was moderate heterogeneity with I2=63% and p=0.03. However, we did not rate down for inconsistency as this can be explained because one of the two included studies measured the outcome only in natal female participants, while the other study measured in natal female and male participants.

^¶¶Outcomes not measured: death by suicide, sexual dysfunction from a physiological perspective (ie, lack of erection, dyspareunia, problems related to dry and degenerated mucosal tissue and anorgasmia), bone density and cardiovascular events.

CGAS, Children’s Global Assessment Scale; GAHT, gender affirming hormone therapy; GD, gender dysphoria; MD, mean difference; SCL-90-R, Symptom Checklist-90 Revised; SMD, standardised mean difference.

Table 4. Gender affirming hormone therapy versus no gender affirming hormone therapy: evidence from case series.

Outcomes	Anticipated absolute effects* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Certainty of evidence (GRADE)	Comments
	Risk with no GAHT	Risk with GAHT
Death by suicide, long term follow-up assessed with: medical records, follow-up mean 24 months†	No comparison group available	6 per 1000 (1 to 18)	Proportion 0.006 (0.001 to 0.018)	315 (1 non-randomised study)20	⨁◯◯◯ Very low‡	The evidence is very uncertain about the effect of GAHT on death by suicide at long term follow-up in natal males and females
Cardiovascular events, long term follow-up assessed with: medical records, number of events, follow-up 7–109 months†§	No comparison group available	40 per 1000 (30 to 50)	Proportion 0.04 (0.03 to 0.05)	3875 (1 non-randomised study)31	⨁⨁⨁ ⨁ High¶	The proportion of natal females experiencing cardiovascular events at long term follow-up is 40 per 1000
Cardiovascular events, long term follow-up assessed with: medical records, number of participants with an event, follow-up mean 26 months†**	No comparison group available	0 per 1000 (0 to 10)	Proportion 0.00 (0.00 to 0.01)	1893 (1 non-randomised study)32	⨁⨁⨁◯ Moderate††¶	The proportion of natal females experiencing cardiovascular events at long term follow-up is 1 per 1000
Other outcomes, not measured‡‡	–	–	–	–	–	–

Grading of recommendations assessment, development and evaluation (GRADE) Working Group grades of evidence. High certainty=we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty=we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty=our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty=we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^*The risk in the intervention group (with 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (with 95% CI).

^†Long term follow-up: outcome measured at ≥12 months of follow-up.

^‡Rated down three levels for risk of bias due to lack of a comparison group.

^§Cardiovascular events included stroke, myocardial infarction and venous thromboembolism.

^¶We did not rate down for risk of bias because this outcome does not need a comparison group, as the study participants can only experience this outcome if they have received the intervention.

^**Cardiovascular events included thromboembolism.

^††Rated down one level for indirectness because this study included natal males only.

^‡‡Other outcomes not measured: gender dysphoria, global function, depression, sexual dysfunction from physiological perspective (ie, lack of erection, dyspareunia, problems related to dry and degenerated mucosal tissue, and anorgasmia) and bone mineral density.

GAHT, gender affirming hormone therapy.

Comparative observational studies

Table 2 provides a summary of the findings.

Gender dysphoria

Current GD, using the Gender Distress Scale ranging from 1 to 5, may be lower (mean difference 0.4 lower (95% CI 0.24 lower to 0.16 higher), number of studies (n)=1, very low certainty) in NMs and NFs who received GAHT compared with those who did not. However, we are very uncertain about the causal effect of the intervention on GD.¹³

Global function

A meta-analysis suggested that global function, measured within the past 12–24 months, may be higher (standardised mean difference (SMD) 0.87 higher (95% CI 0.25 lower to 2 higher), n=2, very low certainty) in NMs and NFs who received GAHT compared with those who did not. However, we are very uncertain about the causal effect of the intervention on global function.^{10 12}

Depression

Eight studies reported this outcome using seven different measurement instruments. Due to variability in instruments, time points and reporting, we could not include all of the studies in a single meta-analysis. A meta-analysis suggested that depression, measured within the past 12 months, may be lower (SMD 0.3 (95% CI 0.85 lower to 0.25 higher), n=2, very low certainty) in NMs and NFs who received GAHT compared with those who did not. However, we are very uncertain about the causal effect of the intervention on depression^{12 14} (onlinesupplemental appendices 15 16 show low to very low certainty evidence about depression from studies not pooled with this evidence).

Before–after studies

Table 3 provides a summary of findings table.

Table 3. Gender affirming hormone therapy versus no gender affirming hormone therapy: evidence from before–after studies.

Outcomes	Anticipated absolute effects* (95% CI)		No of participants (studies)	Certainty of evidence (GRADE)	What happens
	Risk with no GAHT	Risk with GAHT
GD, short term follow-up assessed with: participant reported Gender Preoccupation and Stability Questionnaire, higher scores indicate higher levels of GD, scale 14–70, follow-up mean 6 months†	–	Standardised mean change 0.26 lower (1.64 lower to 1.13 higher)	36 (1 non-randomised study)19 21	⨁◯◯◯ Very low‡§	The evidence is very uncertain about the effect of GAHT on GD at short term follow-up in natal females
Global function, short term follow-up assessed with: participant reported various scales (RAND Short Form-36 Health Survey, Symptom Checklist-90 Revised, Global Severity Index), higher scores indicate better global function, scale 0 –100, follow-up mean 6 months†	–	Standardised mean change 0.25 higher (0.09 higher to 0.4 higher)	73 (2 non-randomised studies)21 25	⨁◯◯◯ Very low§¶**††	The evidence is very uncertain about the effect of GAHT on global function at short term follow-up in natal females
Depression, long term follow-up assessed with: participant reported various scales (Beck Depression Inventory, Hospital Anxiety and Depression Scale), higher scores indicate worse depression, follow-up 18–24 months‡‡	–	Standardised mean change 0.41 lower (0.65 lower to 0.17 lower)	389 (2 non-randomised studies)18 20	⨁◯◯◯ Very low§§¶¶	The evidence is very uncertain about the effect of GAHT on depression at long term follow-up in natal males and females
Sexual dysfunction (ie, vaginal dryness or itch), long term follow-up assessed with: participant report of symptoms, follow-up mean 12 months‡‡	In 193 participants, a linear regression analysis showed that there was no change from baseline in symptoms of vaginal dryness or itch after receiving GAHT (b=0.053, 95% CI − 0.03 to 0.13)***		193 (1 non-randomised study)26	⨁◯◯◯ Very low†††	The evidence is very uncertain about the effect of GAHT on sexual dysfunction (ie, vaginal dryness or itch) at long term follow-up in natal females
Sexual dysfunction (ie, vaginal dryness or itch), short term follow-up assessed with: participant report of symptoms, follow-up mean 6 months†	In 193 participants (ie, natal females), a linear regression analysis showed that there was no change from baseline in symptoms of vaginal dryness or itch after receiving GAHT (b=− 0.01, 95% CI −0.09 to 0.8)***		193 (1 non-randomised study)26	⨁◯◯◯ Very low†††	The evidence is very uncertain about the effect of GAHT on sexual dysfunction (ie, vaginal dryness or itch) at short term follow-up in natal females
Bone mineral density, femoral neck, long term follow-up assessed with: DXA, z-scores, scale − 3 to 3, follow-up mean 12 months‡‡	Mean bone mineral density, femoral neck, long term follow-up was 0.84	Mean change 0 (0.01 lower to 0)	199 (1 non-randomised study)30	⨁◯◯◯ Very low‡‡‡	The evidence is very uncertain about the effect of GAHT on bone mineral density, femoral neck, at long term follow-up in natal females
Bone mineral density, hip, long term follow-up assessed with: DXA g/cm2, follow-up 12–36 months‡‡	Mean bone mineral density, hip, long term follow-up was 0.95	Mean change 0.01 higher (0.01 higher to 0.01 higher)	199 (1 non-randomised study)30	⨁◯◯◯ Very low‡‡‡	The evidence is very uncertain about the effect of GAHT on bone mineral density, hip, at long term follow-up in natal females
Bone mineral density, lumbar spine, long term follow-up assessed with: DXA g/cm2, follow-up 12–36 months‡‡	Mean bone mineral density, lumbar spine, long term follow-up was 1.04	Mean change 0.01 higher (0 to 0.01 higher)	234 (2 non-randomised studies)27 30	⨁◯◯◯ Very low‡‡‡	The evidence is very uncertain about the effect of GAHT on bone mineral density, lumbar spine, at long term follow-up in natal females
Other outcomes, not measured§§§	–	–	–	–	–

Grading of recommendations assessment, development and evaluation (GRADE) Working Group grades of evidence. High certainty=we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty=we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty=our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty=we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^*The risk in the intervention group (with 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (with 95% CI).

^†Short term follow-up: outcome measured at ≤6 months of follow-up.

^‡Rated down three levels due to risk of bias stemming from prognostic imbalance associated with the observational study design and critical risk of bias due to missing data (ie, 46.75% provided outcome data).

^§Rated down one level for imprecision because the optimal information size (200) was not met. Low sample size importantly increases the risk of random error.

^¶Rated down two levels due to risk of bias stemming from prognostic imbalance associated with the observational study design and critical risk of bias due to missing data in one of the two included studies (ie, 46.75% provided outcome data).

^**Statistically, there was considerable heterogeneity with I2=94% and p<0.01. However, we did not rate down for inconsistency as the overall effect estimate was not importantly affected by the studies contributing to statistical heterogeneity.

^††Rated down one level for indirectness because one of the two included studies reported the outcome only for natal females.

^‡‡Long term follow-up: outcome measured at ≥12 months of follow-up.

^§§Rated down three levels due to risk of bias stemming from prognostic imbalance associated with the observational study design, as well as critical and serious risk of bias due to missing data in the two included studies (ie, 20% and 69% of participants, respectively, provided outcome data).

^¶¶Statistically, there was considerable heterogeneity with I2=100% and p<0.01. However, we did not rate down for inconsistency as the overall effect estimate was not importantly affected by the studies contributing to statistical heterogeneity.

^***In the linear mixed model, time was an added categorical variable to detect changes in symptom scores between 0 and 3 months, 0 and 6 months and 0 and 12 months of GAHT. Differences in changes in symptom scores between different administration forms were corrected for baseline differences to avoid regression to the mean. An increase or decrease in symptom scores of 0.2 was considered clinically relevant.

^†††Rated down three levels due to risk of bias stemming from prognostic imbalance associated with the observational study design and critical risk of bias due to concerns with measurement of the outcome (ie, subjective and self-reported outcome).

^‡‡‡Rated down three levels due to risk of bias stemming from prognostic imbalance associated with the observational study design and critical risk of bias due to missing data (ie, 48% of participants provided outcome data).

^§§§Other outcomes: gender dysphoria, sexual dysfunction from physiological perspective (ie, lack of erection, dyspareunia and anorgasmia) and cardiovascular events.

DXA, dual energy x-ray absorptiometry; GAHT, gender affirming hormone therapy; GD, gender dysphoria.

Gender dysphoria

Meta-analysis suggested that GD, measured within the past 6 months with the Gender Preoccupation and Stability Questionnaire ranging from 14 to 70, may be lower (SMD 0.26 lower (95% CI 1.64 lower to 1.13 higher), n=2, very low certainty) in NFs after receiving GAHT compared with before, although we are very uncertain about the causal effect of the intervention on GD.^{19 21}

Global function

Three studies reported global function using three different measures at two different time points. Global function, measured within the past 6 months, may be higher (SMD 0.25 higher (95% CI 0.09 higher to 0.4 higher), n=2, very low certainty) in NFs after receiving GAHT compared with before, although we are very uncertain about the causal effect of the intervention on global function^{21 25} (onlinesupplemental appendices 15 16 show very low certainty evidence about global function from studies not pooled with this evidence).

Depression

Four studies reported this outcome using four different scales. Due to variability in measurement instruments, time points and reporting, we could not include all of the studies in a single meta-analysis. A meta-analysis suggested that depression, measured within 18–24 months, may be lower (SMD 0.41 lower (95% CI 0.65 lower to 0.17 lower), n=2, very low certainty) in NMs and NFs after receiving GAHT compared with before, although we are very uncertain about the causal effect of the intervention on depression^{18 20} (onlinesupplemental appendices 15 16 show very low certainty evidence about depression from studies not pooled with this evidence).

Sexual dysfunction

A study reported a linear regression analysis with no statistically significant change in sexual dysfunction (ie, vagina dryness or itch) reported by NFs after 6 months of receiving GAHT (b=− 0.01, 95% CI −0.09 to 0.8) and after 12 months of receiving GAHT (b=0.053, 95% CI − 0.03 to 0.13) compared with before the intervention. This evidence was rated as very low certainty, and therefore we are very uncertain about the causal effect of the intervention on sexual dysfunction.²⁶

Bone mineral density

Six studies reported lumbar spine BMD, three studies reported femoral neck BMD and three studies reported hip BMD using z scores and g/cm². Lumbar spine BMD, measured within the past 12–36 months with g/cm² may be higher (0.01 higher (95% CI 0 higher to 0.01 higher), n=2, very low certainty) in NFs receiving GAHT compared with before, although we are very uncertain about the causal effect of the intervention on lumbar spine BMD.^{27 30}

Femoral neck BMD, measured within the past 12 months assessed with the dual energy x-ray absorptiometry, z scores ranging from − 3 to 3, may not change (mean change 0 (95% CI 0.01 lower to 0), n=1, very low certainty) in NFs after receiving GAHT compared with before, although we are very uncertain about the causal effect of the intervention on femoral neck BMD.³⁰

Hip BMD, measured within the past 12–36 months with g/cm², was higher (0.01 higher (95% CI 0.01 higher to 0.01 higher), n=1, very low certainty) in NFs receiving GAHT compared with before, although we are very uncertain about the causal effect of the intervention on hip BMD. ³⁰ Onlinesupplemental appendices 15 16 have details of very low certainty evidence about BMD from studies not pooled with this evidence.

Case series

Table 4 summarises our findings. One of the before–after studies reported data on death by suicide only after the intervention, and we classified it as a case series for that outcome.²⁰

Death by suicide

As retrieved from medical records, death by suicide within 24 months of receiving GAHT occurred in 2 of 315 NMs and NFs (0.6%) (proportion 0.006 (95% CI 0.001 to 0.018), n=1, very low certainty). We are very uncertain about the effects of GAHT on death by suicide.²⁰

Cardiovascular events

As retrieved from medical records, cardiovascular events within 7–109 months of receiving GAHT occurred in 151 of 3875 NFs (3.9%) (proportion 0.04 (95% CI 0.03 to 0.05), n=1, high certainty).³¹

Cardiovascular events

As retrieved from medical records, cardiovascular events within 26 months of receiving GAHT occurred in 3 of 1893 NFs (0.2%) (proportion 0.00 (95% CI 0.00 to 0.01), n=1, moderate certainty).³²

Discussion

This SR and meta-analysis synthesised the available evidence regarding the effects of GAHT in young individuals with GD. Comparative observational studies provided mostly very low certainty evidence for GD, global function and depression. One study provided low certainty evidence that depression may be lower in NMs and NFs who received GAHT compared with those who did not. Before–after studies provided very low certainty evidence. Case series provided very low certainty evidence on death by suicide and high to moderate certainty evidence for cardiovascular events.

Although some may view our modification of the ROBINS-I tool as a limitation, we strongly believe it produced similar conclusions than if we had used the original tool or alternatives, such as the Newcastle–Ottawa scale.³³ Given the widespread methodological limitations in this field, any risk of bias tool would give similar conclusions. Comparative observational and before–after studies were at serious or critical risk of bias due to missing data and deviation from intended intervention (ie, administration of co-interventions). Case series, which lack a comparison group, were at critical risk of bias due to measurement of the outcome. These studies should only be used to generate hypotheses for more rigorous study designs, such as prospective cohorts.

The target question of this SR, and of the decision makers considering these interventions, is what are the effects of GAHT? In the absence of randomised controlled trials or comparative observational studies, case series and before–after studies provide the best available evidence to answer this question. While these study designs answer single group questions (eg, what is the functional status among people who received GAHT), they are limited in assessing intervention effects (eg, whether functional status is better in people who received GAHT than those who did not). We accounted for these limitations and assessed the certainty of the available evidence following current methodological standards.³⁴

We rated down the certainty of the evidence mostly because of risk of bias and imprecision, often resulting from an insufficient sample size, for most outcomes and study designs. We did not find evidence about sexual dysfunction in NMs. The overarching theme from this and other SRs on GAHT is the lack of high quality evidence for individuals with GD. Unlike this SR, other reviews did not assess the certainty of evidence for each outcome.

Taylor et al focused on individuals aged ≤18 years, rating most studies as low to moderate quality using the Newcastle–Ottawa Scale. They found limited evidence on GD, body satisfaction, psychological and cognitive outcomes, and infertility.³⁵ Doyle et al reported on psychosocial functioning changes after GAHT among transgender individuals of all ages. They concluded that risk of bias, assessed with the Newcastle–Ottawa Scale, varied among studies. Small sample sizes and undadjusted confounders limited the ability to draw causal inferences.³⁶

Van Leerdam et al concluded that GAHT may reduce GD, body dissatisfaction and uneasiness, subsequently improving psychological well being and quality of life in transgender individuals of all ages.³⁷ They rated the evidence as low to moderate in quality, based on longitudinal cohort and cross sectional studies, without clarifying their rating methods. Chew et al suggested that GAHT helps adolescents achieve intended physical effects, with limited evidence on its psychosocial and cognitive impact. ³⁸ Furthermore, a SR by Connelly et al concluded that current data are insufficient to determine the impact of GAHT on blood pressure in transgender individuals.³⁹ Across all of these SRs, the findings highlight methodological limitations, low quality evidence and significant gaps.

Evidence about the effects of GAHT in individuals aged <26 years with GD is predominantly of very low certainty, with lack of moderate and high certainty evidence about the effects of this intervention. This information is crucial for patients, caregivers, clinicians, guideline developers and policy makers involved in treatment decisions. Beyond evidence certainty, decision making should consider other factors, including the magnitude and consequences of potential benefits and harms, patients ’ and caregivers’ values and preferences, resource use, feasibility, acceptability and equity.⁴⁰ Guideline developers and policy makers must transparently state which and whose values they prioritise when developing treatment recommendations and policies.

Strengths and limitations of the review process

This SR and meta-analysis had multiple strengths. We rigorously followed the highest methodological standards, we assessed the risk of bias for each study using the ROBINS-I tool and evaluated the certainty of the evidence for each outcome using the GRADE approach. A limitation of our review is the inclusion of only English language studies, although we do not expect this to fundamentally alter our conclusions. Due to feasibility considerations, we prioritised specific outcomes and could not address others that may be important to readers, such as regret, anxiety, pelvic pain or cancers (eg, breast, gynaecological, prostate and colon cancer).

Conclusions

The best available evidence reporting on the effects of GAHT in individuals with GD ranged from moderate to high certainty for cardiovascular events and low to very low certainty for the outcomes of GD, global function, depression, sexual dysfunction, BMD and death by suicide. We did not find evidence of sexual dysfunction in NMs. The evidence in this SR and meta-analysis does not exclude the possibility of benefit or harm upon receipt of GAHT. Prospective studies giving higher certainty evidence are needed to understand the short and long term effects of GAHT.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.