Testosterone use and sexual function among transgender men and gender diverse people assigned female at birth.

In our study, testosterone use among transgender men and gender diverse people AFAB was associated with some domains of positive sexual function (such as a higher interest in sexual activity and ability to orgasm) as well as pain/discomfort during sexual activity. Specifically, we observed a strong, consistent association between current testosterone use and higher interest in sex as well as vaginal/FGO pain during sexual activity. Our findings are consistent with prior studies which found that testosterone use is associated with increased desire and interest in sex. 6 GAHT is associated with significant improvements in overall mental health, quality of life, and body image, 31 – 33 which in turn likely have positive impacts on other areas of wellbeing, including sexual function. For example, other studies have found that access to gender-affirming chest reconstruction surgery is associated with higher sexual function scores, while experiencing barriers to accessing gender-affirming care is associated with lower sexual function scores. 25 Although prior studies have not observed an association between testosterone use and orgasm, 6 testosterone use was associated with a higher ability to orgasm in our study. Notably, we observed that a majority of transgender and gender diverse people AFAB using testosterone (67%) experienced vulvovaginal pain during sexual activity. This prevalence is consistent with prior studies, which have reported that 60-62% of transgender men experience dyspareunia. 7 , 8 Although testosterone was associated with vaginal/FGO pain in all regression analyses, the prevalence of any genital pain among individuals who were testosterone naïve was also quite high (63%). The causes of genital pain during sex (including dyspareunia, vulvodynia, and vaginismus) are multifactorial. 24 Although physiological factors (such as vaginal atrophy, endometriosis, and pelvic floor injury) are associated with genital pain during sex, 5 there are complex associations with psychological and social factors, including co-occurrence with other pain disorders, mental health, substance use, and sexual trauma. 22 , 23 Our sample reported an alarmingly high level of sexual abuse (78%), sexual assault (50%), depression (81%) and PTSD diagnoses (40%)—factors that are correlated with chronic pelvic pain among presumably cisgender women. 17 – 21 This may partially account for the high prevalence of pain in our study. Genital pain, including dyspareunia and chronic vulvodynia, can have a significant impact on people’s well-being and quality of life. 34 , 35 Although one prior study reports that 1 in 5 transgender men reported that pain during sexual activity was causing significant problems in their life or relationship, 7 there has been limited investigation into how this may impact the well-being and quality of life of transgender people. Intravaginal estrogen (delivered via cream, tablets, or a ring) is recommended for transgender men experiencing testosterone-associated dyspareunia, vaginitis, and cervicitis. 36 Locally administered estrogen therapy has been demonstrated to be a safe and effective therapy in post-menopausal cisgender women, who also experience vaginal atrophy associated with estrogen deficiency. 37 – 39 However, few existing reports likewise suggest that a minority of transgender men (<5%) have ever used intravaginal estrogen. 8 Therefore, there may be barriers to the uptake of this intervention among transgender populations, including acceptability, provider awareness, participant/patient knowledge, insurance coverage, and other structural barriers, in addition to the limited evidence base to inform clinical use. To our knowledge, this is the first study to report PROMIS SexFS outcomes for transgender men, and one of few studies to examine the association between testosterone use and vulvovaginal pain. This study has several strengths, including a large, national sample of transgender participants who were diverse with respect to their age, gender identity, and sexual orientations. Our results should be interpreted considering several limitations. The PRIDE Study is a convenience sample of majority White participants that relies on self-reported health outcomes and diagnoses and therefore may be subject to sampling, recall, and social desirability biases. Although our study demonstrates the feasibility of using the PROMIS SexFS with transgender men, the PROMIS SexFS has not been validated in transgender populations. This is notable, given that the item calibration and scoring is stratified by sex assigned at birth. Although scoring for interest in sexual activity, orgasm ability, and orgasm pleasure are identical for male and female populations, the items specific to vaginal/FGO anatomy were calibrated to a sample primarily composed of (presumably) cisgender women. 11 Therefore, there remain opportunities to develop and validate sexual function measures for transgender and gender diverse people. In addition, our survey did not collect data on several important variables. Although most participants reported any receptive oral or receptive vaginal/FGO sex, we did not assess other types of vaginal/FGO sexual activity (e.g., penetration with fingers or sex toys) and acknowledge that some transmasculine people do not use their vagina/FGO at all during sex. We also lacked detailed information of hormone doses and were unable to distinguish between low- and high-dose testosterone. We did not collect data on several factors that also influence genital pain, including intravaginal estrogen use or the use of other medications used to manage symptoms of vaginal atrophy and painful sex ( e.g., topical lidocaine). Lastly, although duration of testosterone use was associated with few outcomes in our study, this may in part be a limitation of our cross-sectional design. Future prospective longitudinal research is important for identifying changes to sexual function over time, accounting for cyclical variability in vaginal symptoms, variation in testosterone access and dosing, as well as for understanding how long-term testosterone use may impact sexual function and vulvovaginal pain.

Our analysis included 1,219 participants ages 18-72 years old (median age 27.1; Table 1 ). Most participants (61.1%) endorsed more than one gender identity: most commonly non-binary (54.2%), transgender man (46.0%), genderqueer (33.6%), and man (21.5%). Participants were diverse in sexual orientation, although most identified as queer (65.1%). The majority (80.6%) of participants only reported White race and/or ethnicity and 12.0% of participants selected more than one racial and/or ethnicity. There were 516 (42.3%) participants who had never used testosterone, 602 (49.4%) who currently used testosterone, and 76 (6.2%) former testosterone users. Twenty-five participants were missing information about GAHT use. The median duration of GAHT with testosterone use was 37.7 months (range 7 days to >27 years). Current testosterone users were significantly more likely to identify as a man or transgender man, and less likely to identify as agender, genderqueer, nonbinary, or questioning ( Supplemental Table 1 ). There were no differences in testosterone use by age, race, or ethnicity. Most participants reported having sex with another person in the past year (68.9%), were in a relationship (59.4%), and reported any sexual activity (including masturbation) in the past 30 days (89.5%; Table 2 ). Among participants who reported having sex with another person in the past year, 88.5% reported any receptive oral sex or receptive vaginal/FGO sex. Participants currently using testosterone were more likely to be sexually active in the past year (73.3% v. 63.4%, p=0.001) and in the past 30 days (93.9% v. 84.7%, p<0.001). Table 2 reports participants’ pelvic health histories. Current testosterone users were more likely to have a hysterectomy (17.9% v. 5.4%, p<0.001) and oophorectomy (14.5% v. 3.3%, p<0.001), and less likely to currently use hormonal contraceptives (5.8% v. 17.8%, p<0.001) compared to participants who never used testosterone. Current testosterone users were also more likely to have a past year bacterial STI diagnosis (3.0% v. 1.6%, p=0.001). Testosterone use was not associated with pelvic inflammatory disease, polycystic ovary syndrome, uterine fibroids, inflammatory bowel disease, irritable bowel syndrome, pregnancy history, or intrauterine device use. Participants reported very high levels of lifetime experiences of sexual abuse, sexual assault, and ever having received a depression and PTSD diagnoses ( Table 2 ). Testosterone use was associated with lower scores ( i.e., better mental health) for current depressive and PTSD symptoms. Testosterone use was also associated was substance use, including current smoking (7.5% v. 4.3%, p=0.033) and higher AUDIT scores. Among the 1,091 sexually active participants, most (n=693, 64.6%) reported genital pain/discomfort during sexual activity in the past 30 days, most commonly in the vagina/FGO (52.2%), followed by the clitoris (29.1%) and labia (24.5%; Table 3 ). There were 103 (9.6%) participants who reported pain at all three genital sites. T-scores for pain/discomfort were higher than the population mean (p-values<0.001) while T-scores for orgasm pleasure (mean T-score 45.0, p-value<0.001) and satisfaction with sex life (mean T-score 45.6, p-value<0.001) were lower than the population mean. Figure 1 shows the correlation between each sexual function T-score among participants. Interest in sexual activity, satisfaction, lubrication, orgasm ability and pleasure were positively correlated. Measures of pain/discomfort were negatively correlated with all other domains. Compared to participants who never used testosterone, current testosterone users were less likely to report difficulty with lubrication (58.5% v. 66.7%, p=0.01), more like to report any vaginal pain/discomfort (56.0% v. 49.2%, p=0.04), and more likely to achieve orgasm (p=0.003; Table 3 ). In the minimally adjusted regression models ( Table 4 ), current testosterone use was associated with higher interest in sexual activity (β=6.32, 95% CI: 4.91-7.74), higher ability to orgasm (β=1.50, 95%CI: 0.19-2.81), pain/discomfort during sexual activity in the vaginal/FGO (β=1.80, 95% CI: 0.61-3.00) and in the clitoris (β=1.20, 95%CI: 0.10-2.30). The association between testosterone use, ability to orgasm, and clitoral pain/discomfort were not statistically significant in the robustly adjusted model. No associations were observed between current testosterone use and satisfaction with sex life, lubrication, labial pain/discomfort, or orgasm pleasure. Duration of testosterone use was associated with increased interested in sexual activity, but not with any other outcomes ( Supplemental Table 4 ). In sensitivity analyses, we did not observe any evidence of effect modification by asexual sexual orientation ( Supplemental Figure 2 ), although asexual identity was independently associated with most sexual function domains.

This analysis used data from The Population Research in Identity and Disparities for Equality (PRIDE) Study, an online, prospective, longitudinal cohort study of sexual and/or gender minority people in the US. We conducted a cross-sectional analysis using 2019-2021 data from a questionnaire administered annually to study participants among adult transgender men and gender diverse participants AFAB. We included participants who self-reported currently having a vagina or frontal genital opening (FGO) and who completed the Patient-Reported Outcomes Measurement Information System (PROMIS) Sexual Function and Satisfaction (SexFS) items. 11 We excluded participants who did not self-report having a vagina/FGO or who reported having a phalloplasty or vaginectomy. Participants self-reported data on race and ethnicity, current gender identity, current sexual orientation, and sex assigned at birth. Participants could select multiple response options for race, ethnicity, gender, and sexual orientation. Participants self-reported the gender(s) of people they had any sexual activity within the past year. We categorized participants as never, current, and former testosterone users. Current testosterone use was assessed on the 2019-2021 annual questionnaires and included participants who, at the time, indicated they were currently taking testosterone (of any type in any formulation such as gel, injection or patch), testosterone cypionate, testosterone enanthate, or testosterone undecanoate for gender affirmation. To differentiate participants who had never used testosterone from participants who formerly used testosterone, we also incorporated participants’ responses to a baseline survey that assessed participants’ lifetime testosterone use. Duration of testosterone use was calculated based on participant’s self-reported month and year of initiating testosterone and the date of survey completion. We did not collect information regarding the dose of testosterone. The PROMIS SexFS assesses self-reported sexual function and satisfaction over the past 30 days. This instrument was originally developed for cancer populations and was subsequently validated among a broad group of sexually active adults in the US who were not cancer survivors. 11 Although sexual minority ( e.g., bisexual, gay, and lesbian) individuals were involved in the development of the PROMIS SexFS, transgender people were not. The PRIDE Study implemented a modified version of the PROMIS SexFS that allowed participants to select their preferred anatomical language ( i.e. , “vagina” or “frontal genital opening”). 12 , 13 In completing the survey, each participant’s selection for their preferred terminology was propagated throughout subsequent survey items. Our analysis included eight PROMIS SexFS domains: interest in sexual activity, satisfaction with sex life, vaginal/FGO lubrication, ability to orgasm, orgasm pleasure, as well as vaginal/FGO, labial, or clitoral pain/discomfort during sexual activity. Interest in sexual activity was assessed for all participants while all other domains were only assessed for participants who reported any sexual activity in the past 30 days (which was broadly defined and included masturbation). For each domain, we calculated each participant’s raw score and T-scores. A T-score of 50 represents the mean for the US population and has a standard deviation of 10. Higher T-scores indicate more of the construct measured by the domain; for example, more interest in sexual activity, increased vaginal/FGO lubrication, and more pain/discomfort relative to lower scores. We considered covariates that are associated with sexual function and pelvic pain based on prior literature. These included standardized clinical assessment tools and self-reported lifetime diagnoses. We broadly considered five categories of covariates that have been previously associated with sexual function among cisgender women and transgender men: structural changes to the pelvis, inflammatory conditions in the pelvis, hormonal influences, mental health, and substance use. 17 – 30 Structural changes to the pelvis included prior pregnancies, hysterectomy, and uterine fibroids. Inflammatory conditions in the pelvis included pelvic inflammatory disease, inflammatory bowel disease (including Crohn’s disease and ulcerative colitis), irritable bowel syndrome, and bacterial sexually transmitted infections (STI; defined as chlamydia, gonorrhea, or syphilis diagnosis in the past year). Hormonal influences include oophorectomy, polycystic ovary syndrome, current hormonal contraceptive use (including oral contraceptives, transdermal patch, vaginal rings, medroxyprogesterone acetate injections, and etonogestrel implants), and intrauterine device use. Mental health measures included lifetime diagnoses of depression and post-traumatic stress disorder (PTSD) as well as lifetime experiences of sexual abuse, rape, and sexual assault. Depressive symptoms in the past two weeks were assessed using the Patient Health Questionnaire-9 (PHQ-9; score range 0-27 with higher scores indicating more depressive symptoms) 14 and PTSD symptoms in the last month were assessed with the brief 6-item version of the PTSD Check List (PCL-6; score range 6-30). 15 Substance use variables included current smoking and current alcohol consumption behaviors; the latter was assessed using the Alcohol Use Disorders Identification Test (AUDIT; score range 0-40 with higher scores indicating more disordered alcohol use). 16 We assessed cross-sectional associations between current testosterone use and sexual function. For participants who completed more than one annual questionnaire, we restricted our analysis to only include their first year of responses. We calculated descriptive statistics stratified by testosterone use and conducted Chi-squared tests. We conducted one sample t-tests to test if the SexFS domain T-scores differed from the population mean of 50 and calculated the Pearson correlation coefficient to estimate the strength and direction of the relationship between PROMIS SexFS domains. We then used multivariable linear regression to estimate the association between current testosterone use (relative to never testosterone use) and the T-scores for each sexual function domain. We used causal diagrams to select covariates to include in our model ( Supplemental Figure 1 ). We estimated three models: an unadjusted model, a minimally adjusted model, and a robustly adjusted model. We chose our primary analysis to be a minimally adjusted model that included confounding variables ( i.e., covariates that were associated with both the exposure [testosterone use] and outcomes [sexual function] in our sample). The minimally adjusted model included age, current depression symptoms (PHQ-9 scores), current PTSD symptoms (PCL-6 scores), alcohol use (AUDIT score), current smoker, hysterectomy, oophorectomy, and hormonal contraception use. We conducted secondary analyses with a robustly adjusted model that included all covariates that are associated with sexual function and dyspareunia in prior literature, but were not associated with testosterone use in our sample, including history of sexual assault, prior pregnancy, inflammatory bowel disease, irritable bowel syndrome, uterine fibroids, pelvic inflammatory disease, polycystic ovary syndrome, and intrauterine devices. 17 – 30 Our primary analysis considers current versus never testosterone users. Results comparing former to current users are provided in the Supplementary Materials . We hypothesized that asexual identity may be a potential effect modifier, as asexual participants may differ from non-asexual participants with respect to interest in sexual activity. Therefore, we conducted sensitivity analyses stratified by self-reported asexual sexual orientation. All analyses were conducted in R version 4.2.1. This study received ethical approval from University of California, San Francisco, Stanford University School of Medicine, and WCG Institutional Review Boards.

Using data from a large national sample of transgender men and gender diverse people AFAB, we observed that testosterone use was both associated with positive sexual function and dyspareunia. The relationship between receipt of testosterone GAHT and sexual function is complex, and likely includes both physiological ( e.g., vaginal atrophy) and psychological factors ( e.g., affirmation). However, given the high burden (>60%) of dyspareunia observed among transgender people AFAB, there is a need to assess its impact on the overall quality of life, identify effective and acceptable interventions, and reduce barriers to accessing treatment for transgender people experiencing dyspareunia.

At least 1.6 million transgender adults and adolescents live in the United States, 1 among whom an estimated 70% of transgender men have ever used testosterone as gender-affirming hormone therapy (GAHT). 2 Vaginectomy is rare (<3%) in this population, and the majority of transgender men and gender diverse people retain their vagina. 2 Testosterone GAHT likely impacts sexual function via several complex pathways: Testosterone GAHT is associated with vaginal atrophy, which may be associated with decreased lubrication and/or discomfort during sexual activity. 3 – 5 At the same time, increased gender affirmation through testosterone use may be associated with improved sexual function. There is limited research on the sexual function of transgender men and gender diverse people assigned female at birth (AFAB). What does exist suggests that, although testosterone GAHT is associated with increased desire and arousal, 6 a high proportion of transgender men also reported dyspareunia (painful sex), a common symptom of vaginal atrophy. The prevalence of dyspareunia may be as high as 60-62% among transgender men, 7 , 8 markedly higher than the prevalence reported among cisgender women (3-48%). 9 Several studies also suggest that transgender men using testosterone may experience chronic genital pain and discomfort, with one study reporting that 10-16% of transgender men had been diagnosed with vulvodynia (defined as chronic burning, stinging, or irritating vulvovaginal pain for three consecutive months or longer). 7 , 8 Only one prior small study directly assessed the impact of testosterone GAHT on genital pain during sexual activity. Although 30% of transgender men reported that testosterone had caused genital dryness and 14% experienced genital tearing since initiating testosterone, they did not observe an association between testosterone use and vulvodynia or dyspareunia symptoms. 10 The present study aimed to improve our understanding of the impact of testosterone on the sexual health of transgender people, with a focus on symptoms that are commonly associated with vaginal atrophy, including decreased lubrication, and pain during sexual activity. Using data from a large national online sample of sexual and/or gender minority adults in the US, we examined the association between current testosterone use and self-reported measures of sexual function and satisfaction experienced by transgender men and gender diverse people.