Assessment of breast and prostate cancer risk in trans and gender diverse individuals receiving gender-affirming hormone therapy

Abstract

Introduction

Gender-affirming hormone therapy (GAHT)—most commonly involving estradiol combined with either cyproterone or spironolactone—plays a central role in gender affirmation. However, its expanding use has raised concerns regarding potential health implications, particularly in relation to breast and prostate cancer. This study examines the association between GAHT and the incidence of these cancers in individuals with gender incongruence.

Methods

We conducted a retrospective analysis using clinical data from TriNetX, a federated health research network. Individuals diagnosed with gender incongruence (ICD-10-CM: F64) were divided into two groups: the GAHT group, who received estradiol with or without cyproterone or spironolactone, and the non-GAHT group, who received none of these agents. Propensity score matching was performed based on age and sex (as recorded in electronic health records) to balance the groups. Data on confounding lifestyle risk factors such as smoking and alcohol use were not available. Breast and prostate cancer incidence was then compared using risk differences (RD) and risk ratios (RRs).

Results

After matching, 18,332 trans individuals (mean age: 30.0 ± 12.5 years) were included in the breast cancer analysis. Breast cancer was diagnosed in 1.0% of the GAHT group and 0.3% of the non-GAHT group, corresponding to an RD of 0.7% (95% CI: 0.6–0.9%) and an RR of 3.57 (95% CI: 2.64–4.84). For prostate cancer, a separate matched cohort of approximately 22,338 individuals was analyzed. Incidence was 0.13% in the GAHT group and 0.06% in the non-GAHT group, with an RD of 0.06% (95% CI: 0.01–0.12%) and an RR of 2.00 (95% CI: 1.05–3.80); however, this difference did not reach statistical significance.

Conclusions

While the absolute incidence rates of both breast and prostate cancer were very low in this cohort, we observed modestly higher rates among trans individuals receiving GAHT compared to those who did not. However, in the absence of data on key confounders such as smoking and alcohol use, these associations should be interpreted with caution, as residual confounding cannot be excluded. The results highlight the need for ongoing cancer surveillance, individualized risk-table assessment, and the development of inclusive clinical guidelines to ensure safe and equitable care for gender-diverse populations.

Introduction

Research on breast and prostate cancer in the context of gender incongruence remains sparse, with most existing studies based on small study populations (de Blok et al., 2019; Dijkman et al., 2023; Nyame et al., 2022). As gender-affirming hormone therapy (GAHT)—a core component of gender-affirming care—becomes more widespread, there is a growing need to assess its long-term health implications, particularly regarding cancer risk in this population.

Globally, breast cancer is the most commonly diagnosed malignancy and ranks fifth in overall cancer-related mortality (Sung et al., 2021). In the United States, data from the American Cancer Society indicate that in 2023, approximately 12,800 individuals assigned male at birth were diagnosed with breast cancer, resulting in 530 deaths—a mortality rate of 19%. By contrast, the mortality rate for individuals assigned female at birth with breast cancer was slightly lower, at 15% (American Cancer Society, 2023). Prostate cancer remains the most frequently diagnosed cancer among individuals assigned male at birth in the United States and is the third leading cause of cancer-related deaths (Haas et al., 2008).

GAHT is a non-surgical form of gender affirmation that involves the administration of exogenous hormones to induce physical changes consistent with an individual’s gender identity. Treatment regimens typically include estrogen with or without anti-androgens for transfeminine individuals, or testosterone with or without progestogens for transmasculine individuals, depending on the treatment goals. Such interventions are generally recommended to be undertaken under the guidance of qualified healthcare professionals following thorough consultation (T’Sjoen et al., 2019).

Estrogen formulations constitute a principal component of transfeminine GAHT, wherein they are administered to achieve circulating estradiol concentrations within the physiologic range observed in cisgender females, thereby promoting feminization and inducing proliferative changes in breast tissue. Estrogens and the risk of breast cancer (Al-Shami et al., 2023). A nested case–control study indicated that use of systemic estrogen therapy for more than five years increases the risk of breast cancer, particularly among women also receiving progestogen (Vinogradova et al., 2020). In addition, the combination of estrogen with cyproterone acetate or spironolactone has been shown to produce a greater reduction in testosterone levels (Angus et al., 2019). For another hormone-sensitive malignancy—prostate cancer, epidemiological studies have reported that transfeminine individuals receiving androgen deprivation and estrogens have a substantially lower risk compared with cisgender men. Given that many transgenders and gender-diverse (TGD) individuals receive GAHT over extended periods, evaluating its long-term oncological effects remains an important area of research (de Nie et al., 2020). Given that many TGD individuals receive GAHT over extended periods, evaluating the oncological effects remains an important area of research (Getahun et al., 2018; Hembree et al., 2017).

The objective of this study is to retrospectively assess whether GAHT use among TGD individuals is associated with an increased incidence of breast and prostate cancer. Using data from TriNetX, a large federated health research network, this study leveraged comprehensive electronic health record data to clarify the relationship between gender-affirming hormone therapy and cancer risk in this population. The platform enables access to a large, diverse sample, facilitating robust epidemiological insights.

Materials and methods

Data source and study design

This retrospective observational study utilized data from the TriNetX Global Health Research Network, a federated platform that aggregates de-identified electronic health records (EHRs) from 87 healthcare organizations (HCOs) across 21 countries and regions, including the Americas, Europe, the Middle East, Africa, and the Asia-Pacific. These HCOs represent a diverse mix of academic medical centers, community hospitals, and outpatient clinics, ranging from small specialty practices to large multi-hospital systems. While institutional inclusion is determined by TriNetX and may vary, there are no reported exclusions based on size or geographic location.

Data within TriNetX are collected passively during routine clinical care and integrated into the network with monthly updates. Participant information is not prospectively recruited but retrospectively identified based on standardized coding systems, including ICD-10-CM for diagnoses and RxNorm for medications (Palchuk et al., 2023). This approach has been widely utilized in previous studies leveraging real-world data from electronic health records to investigate clinical outcomes and treatment patterns (Heym et al., 2022; Wüster et al., 2023).

The primary outcomes assessed were the incidence of breast and prostate cancer. Statistical analyses included the calculation of risk ratios (RR), odds ratios (OR), and risk differences (RD), each with corresponding 95% confidence intervals (CI). Survival analysis was performed using the log-rank test, with statistical significance set at p < 0.05.

Cohort definition, data acquisition, and matching

To define the treatment and control cohorts, we searched the TriNetX database for individuals recorded with gender incongruence (ICD-10-CM: F64) between November 14, 2003, and November 14, 2023. The treatment group (GAHT group) included individuals who met all of the following criteria:

1. A diagnosis of gender incongruence,

2. Documented exposure to estradiol (RxNorm: 4083), and

3. Exposure to either cyproterone (RxNorm: 3014) or spironolactone (RxNorm: 9997).

The control group (non-GAHT group) included individuals with a diagnosis of gender incongruence but no record of exposure to any of the three medications listed above. Exclusion was implemented using a logical OR operator: any record of estradiol, cyproterone, or spironolactone led to exclusion from the control group. All included individuals had at least one documented healthcare encounter. The detailed inclusion and exclusion criteria used to define the GAHT and non-GAHT groups are shown in Table 1.

Table 1.

Search criteria used to define treatment (GAHT group) and control (non-GAHT group) groups in the TriNetX database.

Cohort(s)	Condition type	Logical operator	Term operator	Code system	Code	Description
GAHT and non-GAHT group	Diagnosis	must have	and	ICD-10	F64	Gender incongruence
	Visit	must have	–	TNX	Visit	Healthcare encounter
GAHT group	Medication	must have	–	RxNorm	4083	Estradiol
	Medication	must have	and any of	RxNorm	3014	Cyproterone
	Medication	must have		RxNorm	9997	Spironolactone
non-GAHT group	Medication	cannot have	–	RxNorm	4083	Estradiol
	Medication	cannot have	or	RxNorm	3014	Cyproterone
	Medication	cannot have	or	RxNorm	9997	Spironolactone

In this study, gender information was extracted as recorded in the EHR, which may reflect either self-identified gender or administrative sex markers, depending on documentation practices across providers. Due to limitations in the dataset, sex assigned at birth was not available. Therefore, gender classification reflects recorded gender, which may not fully represent individuals’ gender identity or biological characteristics.

To reduce confounding and improve group comparability, one-to-one propensity score matching was performed based on age and recorded gender. After matching, the GAHT and non-GAHT cohorts each included 18,332 individuals (Figure 1). A separate matching procedure was conducted for the prostate cancer analysis, yielding approximately 22,338 matched individuals per group.

Figure 1.

Modified CONSORT flowchart.

Due to database constraints, detailed information on GAHT dosage, formulation, duration, and treatment adherence was not consistently available. Therefore, GAHT exposure was treated as a binary variable (exposed vs. unexposed). This limitation is discussed further in the study.

Cancer outcomes were identified using ICD-10 diagnostic codes recorded in the EHRs of participating HCOs. No linkage to national or population-based cancer registries was available, which may limit the completeness of outcome capture. Additionally, the TriNetX platform does not provide data on the frequency or total number of healthcare visits per individual. Consequently, outcome ascertainment may be influenced by individual healthcare utilization patterns, introducing potential detection bias if individuals with more frequent contact with the healthcare system were more likely to receive a cancer diagnosis.

Ethics statement

In line with the Health Insurance Portability and Accountability Act (HIPAA), TriNetX adheres to the law’s regulations, focusing on the privacy and security of healthcare data. This study, being retrospective and involving de-identified data, is exempt from ethical review or clearance in accordance with institutional and local laws. The research follows TriNetX Global Health Research Network’s access policies and publication criteria. Consequently, no conflicts of interest or ethical concerns were identified, allowing the study to proceed without formal informed consent from the subjects.

Results

Assessment, allocation, and matching

A total of 94,022 TGD individuals from 87 healthcare organizations met the inclusion criteria (ICD-10-CM: F64: gender incongruence) and were included in the study. The GAHT group included trans individuals who had received estradiol in combination with either cyproterone or spironolactone. This cohort consisted of 18,405 individuals (19.6% of the total), with a mean age of 30.0 ± 12.5 years. Based on electronic health record (EHR) data, gender classification comprised 12,578 (68.6%) individuals recorded as male, 5,402 (29.5%) recorded as female, and 352 (1.9%) with unknown gender. The non-GAHT group included individuals with no documented use of estradiol, cyproterone, or spironolactone, totaling 75,617 (80.4%) individuals, with a mean age of 23.9 ± 12.4 years. This group included 24,923 (33.3%) individuals recorded as male, 48,058 (64.3%) as female, and 1,807 (2.4%) with unknown gender.

Following propensity score matching based on age and EHR-recorded gender, 18,332 individuals were retained in each cohort. The matched groups had equivalent age distributions (mean: 30.0 ± 12.5 years) and similar gender proportions: 68.6% recorded as male, 29.5% as female, and 1.9% recorded as unknown in both cohorts (Table 2).

Table 2.

Summarizes the characteristics of GAHT group and non-GAHT group before and after matching.

Before matching					After matching
Individuals (n)	GAHT group	non-GAHT group	p-value	Std diff.	GAHT group	non-GAHT group	p-value	Std diff.
Total	18,405	75,617			18,332	18,332
AI (years)	30.0	23.9	<0.001	0.185	30.0	30.0	1	<0.001
SD	±12.5	±12.4			±12.5	±12.5
EHR-Male	12,578 (68.6%)	24,923 (33.3%)	<0.001	0.754	12,578 (68.6%)	12,583 (68.6%)	0.955	0.001
EHR-Female	5,402 (29.5%)	48,058 (64.3%)	<0.001	0.744	5,402 (29.5%)	5,402 (29.5%)	1	<0.001
EHR-Unknown	352 (1.9%)	1,807 (2.4%)	<0.001	0.034	347 (1.9%)	352 (1.9%)	0.849	0.002

*Abbreviations: AI – Age at Index; SD – Standard Deviation; Std diff. – Standardized Difference

Cancer incidence outcomes

Cancer incidence was evaluated in both cohorts following the exclusion of individuals with documented cancer diagnoses prior to cohort entry. For the breast cancer analysis, 111 individuals in the GAHT group and 70 in the non-GAHT group were excluded due to preexisting breast cancer diagnoses. As a result, 18,221 individuals in the GAHT group and 18,262 in the non-GAHT group were included in the risk analysis. During the follow-up period, 189 cases in the GAHT cohort and 53 in the control cohort were newly diagnosed with breast cancer, yielding incidence rates of 1.0 and 0.3%, respectively (Figure 2A).

Figure 2.

Incidence of breast and prostate cancer among trans individuals receiving GAHT (the GAHT group) and those not receiving GAHT (the non-GAHT group), following exclusion of trans people with preexisting cancer diagnoses.

For the prostate cancer analysis, 38 individuals in the GAHT cohort and 35 in the control cohort were excluded due to preexisting prostate cancer. The remaining analytical cohorts included 22,300 individuals in the GAHT group and 22,303 in the control group. During follow-up, 28 new cases were identified in the GAHT cohort and 14 in the control cohort, corresponding to incidence rates of 0.13 and 0.06%, respectively (Figure 2B).

GAHT exposure was associated with an increased risk of breast cancer, with a risk ratio (RR) of 3.57 (95% CI: 2.64–4.84, p < 0.01) and an absolute risk difference (RD) of 0.70% (95% CI: 0.60–0.90%, p < 0.01). For prostate cancer, the RR was 2.00 (95% CI: 1.05–3.80, p < 0.01) while the RD was 0.06% (95% CI: 0.01–0.12%, p = 0.031). All estimates were adjusted for baseline cancer history, operationalized as the exclusion of individuals with documented preexisting cancer diagnosis. Detailed estimates of OR, RR and RD are presented in Figure 3.

Figure 3.

Risk ratios (RR), odds ratios (or), and absolute risk differences (RD) for breast and prostate cancer between cohorts. Cancer cases diagnosed prior to GAHT exposure or index date were excluded: 111 (GAHT group) and 70 (non-GAHT group) for breast cancer; 38 (GAHT group) and 35 (non-GAHT group) for prostate cancer.

*All percentages and statistical estimates (RR, OR, CI) are rounded to two or three decimal places.

Discussion

This large-scale retrospective cohort study, using data from the TriNetX federated health research network, assessed the association between GAHT and the risk of breast and prostate cancer among transgender individuals recorded with gender incongruence (ICD-10-CM: F64). We observed a 3.6-fold increase in the odds of breast cancer and a 2-fold increase in the odds of prostate cancer among individuals receiving GAHT. Although both associations reached statistical significance, the prostate cancer finding should be interpreted cautiously due to its low absolute incidence and a confidence interval close to 1.0.

These findings partially align with established biological understanding, particularly regarding the influence of estrogen on breast tissue (Berliere et al., 2022; Nikolić et al., 2018; Tangpricha & den Heijer, 2017). Interestingly, the increased incidence of prostate cancer among individuals receiving GAHT contrasts with expectations based on androgen deprivation (Chen et al., 2017), which is generally thought to reduce prostate cancer risk. This discrepancy may stem from ascertainment bias, incomplete data on sex assigned at birth, or unmeasured confounding factors. Moreover, it highlights the need for more mechanistic studies that can disentangle the effects of GAHT components on prostate physiology in transgender populations. Estradiol, as an estrogen, increases the risk of ER-positive breast cancer by binding to estrogen receptors in breast cells and activating signaling pathways that affect cell growth and differentiation (Chen et al., 2017; Deng et al., 2014; Leśniewska et al., 2009; Yum et al., 2009). Cyproterone, a steroidal antiandrogen, primarily inhibits androgen receptors and has been shown to suppress the proliferation of androgen-dependent prostate cancer cells (Chen et al., 2017). Its influence on breast cancer risk remains less well understood and may depend on factors such as estrogen receptor (ER) status and hormonal balance (Yuk, 2024). Spironolactone, another anti-androgenic agent commonly used in feminizing regimens, may also affect hormone-sensitive tissues (Burinkul et al., 2021; Rice et al., 2019), although its effects on tumor development are likely mediated by hormonal rather than metabolic pathways. The combined use of these three agents may contribute to hormonal imbalances, alterations in receptor sensitivity, or other complex physiological changes that could potentially influence the growth and differentiation of breast and prostate tissue. While the exact mechanisms remain unclear, such effects may be associated with an increased risk of cancer and warrant further investigation.

Our study’s primary strength is its large sample size (>90,000 participants) drawn from a diverse network of 87 healthcare organizations. This provides substantial statistical power to detect even modest increases in cancer risk. Furthermore, the use of propensity score matching balanced the two cohorts across key demographic variables (age and sex as recorded), minimizing potential confounding effects. This is particularly relevant given the increasing legal recognition of gender identity since 2000 (2003), which has likely influenced the documentation of gender incongruence (ICD-10-CM: F64) in electronic health records. Trans people included in the study were identified from these records, matched by unique identifier, and had received hormone therapy as listed.

We acknowledge several important limitations. This study is retrospective and relies on electronic health records, which inherently limit data completeness and granularity. Although there is evidence that certain lifestyle factors—such as alcohol consumption, smoking, and obesity—increase the risk of both breast and prostate cancer, key covariates including smoking status, alcohol intake, physical activity, family history, and detailed hormone therapy parameters (e.g. dosage, formulation, treatment duration, and adherence) were not consistently available across participating healthcare organizations. Attempting to incorporate these variables would have substantially reduced the sample size due to extensive missing data, thereby limiting statistical power and representativeness. The lack of adjustment for these covariates may introduce residual confounding, potentially affecting both the magnitude and direction of the observed associations. For example, individuals who opt for GAHT may differ systematically from those who do not in ways that influence cancer risk independent of hormone treatment itself. Although propensity score matching balanced key demographic factors such as age and recorded sex, these unmeasured confounders warrant caution in causal interpretation.

Although this study included a relatively large cohort, the absolute number of breast and prostate cancer cases remains limited due to the rarity of these outcomes, which may affect the precision of risk estimates. Nevertheless, these preliminary findings contribute valuable early evidence regarding the potential association between GAHT and cancer risk in transgender individuals. While exploratory in nature, the results highlight the importance of nuanced clinical awareness and should not be interpreted as grounds to restrict access to GAHT, which remains critical and often life-saving for transgender populations. To strengthen future research, comprehensive clinical data including sex assigned at birth, detailed information on GAHT exposure, and key confounding factors such as smoking and alcohol use will be essential to confirm diagnoses, refine exposure classification, and reduce residual confounding. In the interim, clinicians should integrate these insights into shared decision-making processes, individualized care planning, and the development of inclusive cancer screening protocols, while ensuring ongoing surveillance and evidence-based practice that respects the dignity and healthcare needs of transgender individuals.

Conclusions

These findings suggest a potential association between GAHT and elevated cancer risk in this population. While the absolute incidence remains low, the results should be interpreted with caution given the retrospective design and the lack of adjustment for key potential confounders such as smoking status, alcohol consumption, physical activity, family history, and detailed GAHT parameters (e.g. dosage, formulation, treatment duration, and adherence). Individualized risk–benefit assessments should also consider the presence or absence of relevant organs, alongside proactive screening strategies and long-term monitoring for transgender and gender diverse individuals undergoing hormone therapy. Further prospective research with comprehensive clinical data—including detailed GAHT exposure information, key lifestyle and medical covariates—will be essential to confirm diagnoses, refine exposure classification, reduce residual confounding, and improve the precision of risk estimates. Such efforts will support the development of evidence-based, inclusive clinical guidelines that promote both safety and equitable access to essential gender-affirming care.

Funding Statement

The author(s) reported there is no funding associated with the work featured in this article.

Disclosure statement

No potential conflict of interest was reported by the author(s).