Impact of Exogenous Testosterone on Reproduction in Transgender Men

Abstract

Studies show that a subset of transgender men desire children; however, there is a paucity of literature on the effect of gender-affirming testosterone therapy on reproductive function. In this manuscript, we will review the process of gender-affirming hormone therapy for transgender men and what is known about ovarian and uterine consequences of testosterone exposure in transgender men; draw parallels with existing animal models of androgen exposure; summarize the existing literature on parenting experiences and desires in transgender people; discuss considerations for assisted reproductive technologies and fertility preservation; and identify gaps in the literature and opportunities for further research.

Recent data from population-based surveys estimate there are approximately 1.4 million self-identified transgender adults living in the United States (1, 2), many of whom experience gender dysphoria, or distress stemming from a discrepancy between their gender identity and their birth-assigned sex (3). To bring their secondary sex characteristics more in line with their gender identity, transgender people will often undergo gender-affirming treatment with hormones and/or surgery. The World Professional Association for Transgender Health (4), American Society for Reproductive Medicine (ASRM) (5), and Endocrine Society (6) have all put forth practice guidelines that recommend fertility preservation counseling before the initiation of gender-affirming therapy (medical or surgical); however, ASRM also notes that there are relatively few data on the impact of gender-affirming hormone therapy on future fertility (5). As such, it is unclear whether fertility preservation counseling is warranted in this setting.

Despite these practice guidelines, fertility preservation often poses challenges in this population, particularly transgender men, because the process of ovarian stimulation is often costly, time-consuming, and physically invasive. Given these challenges, transgender men often do not elect fertility preservation, but may present subsequent to initiating testosterone (T) therapy for management of pregnancy, to pursue assisted reproductive technologies (ART) for current or future fertility, or for ovarian tissue cryopreservation (OTC) (7). Unfortunately, the effects of long-term T therapy on reproductive function in transgender men, as well as the reversibility of any T-induced changes, are largely unknown. The current knowledge gaps preclude evidence-based counseling for transgender men about fertility treatment options. The objectives of this scoping review are to describe the process of gender-affirming T therapy and what is known about reproductive consequences of T exposure in transgender men; draw parallels with existing animal models of androgen exposure; summarize the existing literature on parenting experiences and desires in transgender people; discuss considerations for ART and fertility preservation in transgender men; and identify gaps in the literature and opportunities for further research.

Methods

A literature search was conducted in PubMed, Google Scholar, Medline, Web of Science, and PsycInfo, without restrictions on date of publication, using the following search terms: transgender, gender identity disorder, gender dysphoria, trans men, transsexual, testosterone, reproductive health, parenting desires, fertility, fertility preservation, oocyte cryopreservation, ovary, uterus, pregnancy, ovarian stimulation, gender-affirming hormones, cross-sex hormones, animal models, and assisted reproductive technologies. The references of each selected article were also manually searched. All authors agreed on which articles to include. Given the relative paucity of literature in this subject area, all relevant studies in English were included.

Gender-Affirming Hormone Therapy in Transgender Men

Hormone therapy for transgender men includes exogenous T administration, with the goal of inducing more masculine secondary sex characteristics and the regression of female secondary sex characteristics (4, 6). The Endocrine Society recommends titrating T doses to serum T levels within the typical range of cisgender men, generally 320 to 1000 ng/dL (6), although multiple protocols for T management exist, some with less reliance on serum levels. There are few data to guide clinicians providing T therapy specifically to transgender men, with the majority of protocols being derived from experience with androgen replacement in hypogonadal cisgender men (8). Studies have suggested similar hormone profiles between transgender men and hypogonadal cisgender men taking exogenous T (9); however, the goals of treatment differ between the 2 populations necessitating further research on the most appropriate protocols for the transgender population. The most common route of administration is intramuscular or subcutaneous injections with T enanthate or cypionate, but other options include intramuscular injections of mixed T esters or longer-acting T undecanoate, T implants, or transdermal administration with patches or gel (4, 6, 8). Studies comparing subcutaneous to intramuscular T injections suggest equivalent effectiveness and safety, with better patient acceptance of subcutaneous injections (10). Oral preparations of T are rarely used because of required frequent dosing and unpredictability of serum levels (8). Other hormonal agents, such as gonadotropin-releasing hormone agonists (GnRHa) or progestins, are rarely required, and are generally used for menstrual suppression in transgender men who do not achieve amenorrhea with T therapy alone (4, 6). Peripubertal transmasculine adolescents may also be treated with GnRHa to prevent further pubertal progression until the time they are ready for masculinizing treatment (4, 6).

Although physical outcomes vary from patient to patient, some masculinizing effects from exogenous T therapy can start to be seen as early as 1 or 2 months after treatment initiation, including voice deepening, acne, skin oiliness, clitoromegaly, and cessation of menses (4, 6). Other changes, such as facial/body hair growth, changes in fat distribution, and increased muscularity, take longer to fully develop, in some cases taking up to 5 years (6). A Japanese study suggests that higher initial doses of testosterone (T enanthate 250 mg every 2 weeks) may result in a faster onset of physical changes, but that physical results are similar with lower doses (250 mg every 3 weeks and 125 mg every 2 weeks) at 6 months of treatment (11). Potential risks of gender-affirming T therapy are generally extrapolated from studies on T therapy in hypogonadal cisgender men, and include erythrocytosis, liver dysfunction, and cardiovascular risk (6). The studies on long-term risks that have been performed in transgender men have been overall reassuring, suggesting that adverse effects are rare (12-14); however, there are limited to no data on the risks of GnRHa for puberty blockade specifically in transgender adolescents. Notably, there have been no prospective studies to date evaluating the effect of long-term hormone therapy on fertility, although there are a few reports in the literature of transgender men who have carried pregnancies and given birth (15, 16).

Clinical Data on Reproductive Consequences of Gender-Affirming Testosterone Therapy

Ovarian findings

Ovarian studies in transgender men treated with T have yielded conflicting results regarding the association of long-term T therapy with polycystic ovarian morphology on ovarian histopathology (Table 1). Broadly, many studies have found increases in collagenization of the tunica albuginea and stromal hyperplasia (17-23). Increased luteinization of stromal cells was also commonly seen in higher percentages of transgender men as compared to controls (17-21). Some conflict arises, however, in how studies defined follicular changes, leading to differences in whether an individual study characterized ovarian changes as consistent with polycystic ovarian morphology. Several studies reported multiple cystic follicles in transgender men, antral follicle counts (AFCs) of more than 12 follicles per ovary, or multifollicular ovaries (17-19, 21, 23-25). In contrast, other studies reported similar follicular counts between transgender men and controls, although higher percentages of atretic follicles have been noted in transgender men (20, 26, 27). Two studies evaluating serum anti-müllerian hormone (AMH) levels as a serum marker for follicular changes before and during T therapy in transgender men also have conflicting results, with one study showing a decrease in AMH (28), and the other reporting no change from baseline (29). Moreover, participants in both studies were on other hormone-modifying medications (GnRHa, aromatase inhibitor, and/or progestin) in addition to T, making it impossible to isolate the effect of T (28, 29). Studies evaluating oocyte function have shown that oocytes isolated from T-treated transgender men at the time of oophorectomy can be matured in vitro to develop normal metaphase II meiotic spindle structure (30, 31).

Table 1.

Summary of findings from histopathological examination of ovaries from transgender men

Study	Population	T Exposure, mo	Ovarian Findings With T Exposurein Trans Men
Amirikia et al, 1986	10 trans men (mean age, 29 y); compared to 3 PCO and 3 controls. Trans men prior cycles 8/10 regular, 2/10 irregular	14 to 84 (mean, 35)	Thicker tunica albuginea, thickened basal membrane of atretic follicles, no noticeable increase in stroma or theca cell hyperplasia, no recent corpora lutea noted
Futterweit et al, 1986	19 trans men (mean age, 27 y); 12 age-matched controls. In trans men, pre-T hirsutism (8/19) and oligomenorrhea (5/19)	12 to 120 (mean, 37)	PCOM in 13/19 trans men (3 of 4: multiple cystic follicles [17/19], collagenization outer cortex [13/19], stromal hyperplasia [16/19], luteinized stromal cells [5/19]), 0/12 controls (cystic follicles [4/12], stromal hyperplasia [2/12], luteinized stromal cells [0/12]). Post-T corpora lutea (3/19) and corpora albicantia (5/19) in trans men, in controls corpus lutea (7/12), corpus albicantia (10/12)
Miller et al, 1986	32 trans men (mean age, 29 y); 36 controls	12 to 96	All trans men had primordial follicles, follicular cysts, corpora albicantia, 4/32 current or recent corpora lutea
Spinder et al, 1989	26 trans men (mean age, 26 y); 9 age-matched controls; 9 trans men before and after 6-mo T for hormonal studies	9 to 36 (mean, 18)	PCOM in 18/26 trans men (3 of 4: multiple cystic follicles [18/26], collagenization tunica albuginea [25/26], stromal hyperplasia [21/26], luteinization of stromal cells [7/26]), 0/9 controls (multiple cystic follicles [1/9], collagenization tunica albuginea [3/9], stromal hyperplasia [0/9], luteinized stromal cells [2/9]). In trans men corpora lutea (4/26), corpora albicantia (26/26); for controls corpora lutea (3/9) and corpora albicantia (9/9).
Pache et al, 1991	17 trans men (mean age, 25 y); 13 controls (mean age, 29 y). Prior regular menstrual cycles in 13/17 trans men	11 to 72 (mean, 21)	2× cystic follicles, 3.5× atretic follicles, 3× thicker ovarian cortex (93%), stromal hyperplasia (100%), luteinization stromal cells (clusters: 35%), theca interna hyperplasia (100%) and luteinization
Chadha et al, 1994	Ovaries: 11 trans men (mean age, 25 y); uteri: 6 trans men (mean age, 27 y); 5 premenopausal (age, 32-51 y) and 7 postmenopausal (age, 56-78 y) controls, 10 ovaries PCOS patients. Prior regular menstrual cycles in 8/11 trans men used for ovaries. [overlap Pache?]	Ovaries 11 to 72 (mean, 21); uteri: 24-96	Increased cystic/atretic follicles, collagenized thicker cortex, theca interna hyperplasia/luteinization, stromal hyperplasia. More intense AR staining ovarian stroma
Van den Broecke et al, 2001	1 trans man (age, 21 y)	12	A total of 98.6% primordial follicles and 1.4% primary in cortex. Follicles developed with FSH stimulation (grafted to mice)
Mueller et al, 2008	35 trans men (mean age, 31 y)	12	No ovarian pathology noted on vaginal ultrasound
Grynberg et al, 2010	112 trans men (mean age, 29 y). 98/112 prior regular cycles	24 to 108 (mean, 44)	More than 12 antral follicles/ovary (89/112), stromal hyperplasia (112/112)
Ikeda et al, 2013	11 trans men (mean age, 33 y), 10 age- and BMI-matched controls (mean age, 34 y) with gynecologic malignancies. No prior PCOS in trans men or controls (Rotterdam criteria)	17-164 (mean, 70)	Thicker tunica albuginea collagenization (10/11), more stromal hyperplasia (8/11) and luteinization (10/11), similar preantral/antral follicles, greater atretic follicles in trans men, no corpus lutea in trans men (vs 7/10 controls), 3/11 trans men had corpora albicantia (vs 3/10 controls).
Loverro et al, 2016	12 trans men (20-32 y)	Mean, 32	Multifollicular (10/12: active endometrium), corpus lutea (2/10: secretory endometrium)
Caanen et al, 2017	56 trans men (mean, 23 y), 80 controls (mean, 34 y)	≥ 12 (median, 29.5). GnRHa use (current 28.6% [16/56], past 48.2% [27/56])	PCOM defined as AFCs ≥ 12 follicles (2-10 mm) in at least 1 ovary by TVU: trans men 17/53, controls 23/75
De Roo et al, 2016	40 trans men (mean age, 24 y)	Mean, 14	69% primordial follicles, 31% primary; 87% normal spindle structure after IVM. AMH predicted oocytes.
Lierman et al, 2017	16 trans men (mean age, 24 y) [overlap with De Roo?]	Mean, 13	COCs in vitro matured (rate 38.1% MII), 85.7% normal spindle after IVM
Khalifa et al, 2018	27 trans men (mean age, 29 y), all G0P0, 1 with congenital adrenal cortical hyperplasia, 2 with prior oophorectomy, 5 history of dysmenorrhea, 1 AUB	24 received androgen 19 to 288	Bilateral cystic follicles in 23/23 with prior androgen exposure, follicular density mean 10.7 primordial and primary follicles/mm2 (range, 1.5-32.5 follicles/mm2). Two incidental mature cystic teratomas, 2 hemorrhagic corpus lutea (1 with no prior androgen exposure)
De Roo et al, 2017	3 trans men and 3 age-matched oncology patients (mean age, 23 y)	13, 16, 27	Superficial part of cortical fragments (< 1.4 mm) stiffer than oncology patients’ in texture profile analysis, although thickness, hardness, cohesiveness, and springiness did not differ.
Adeleye et al, 2019	13 trans men for ovarian stimulation (6 before T, 7 after discontinuing T) (age 14.6-37.1 y, median, 22.4)	Median 46, discontinued time for stimulation median 6 (range, 1-13)	Lower peak estradiol and fewer oocytes retrieved (median, 12 vs 25.5 oocytes) in trans men with vs without prior T therapy (differences not detected when 2 trans men with initial AFCs < 5 were removed from analysis). No detectable differences between trans men with or without prior T therapy in follicles at cycle start, estradiol per oocyte, meiosis II oocyte yield, or maturity rate (MII/oocyte)

Abbreviations: AFC, antral follicle counts; AMH, anti-müllerian hormone; AR, androgen receptor; AUB, abnormal uterine bleeding; COC, cumulus-oocyte complex; FSH, follicle-stimulating hormone; GnRHa, gonadotrophin-releasing hormone agonists; IVM, in vitro maturation; MII, meiosis II; PCOM, polycystic ovarian morphology; PCOS, polycystic ovary syndrome; T, testosterone; TVU, transvaginal ultrasound.

Although the aforementioned studies provide important preliminary human data, there are also several limitations that make it unclear whether they can be used to counsel transgender men about the long-term effects of T on ovarian function. All the studies were observational in nature, and the T regimens used and serum levels achieved were not consistent. Additionally, the average duration of T therapy was less than 4 years in the majority of the studies, with wide ranges of duration reported. Changes to the ovaries with T treatment are also complicated by the reportedly high rates of PCOS (Rotterdam 2003 criteria) in transgender men even before hormone therapy (15%-58%) (32-34). Finally, few studies have examined the ovaries of transgender men after discontinuation of T to investigate whether any T-related changes are reversible, and there are only limited data on fertilization or embryogenesis from oocytes previously exposed to T.

Uterine findings

Studies that include an examination of uterine histopathology in transgender men treated with T (Table 2) also have conflicting results. A study of uteri removed from 12 transgender men found endometria to be either active or secretory (25), whereas another study of 112 transgender men yielded endometria that were split nearly evenly between active and atrophic (23). In further contrast, all endometria were deemed histologically inactive in a third study of 27 transgender men in comparison to 12 premenopausal and 30 postmenopausal women (35). Other uterine changes reported in transgender men on T therapy include atrophy of cervical mucosa, eosinophilic infiltration of myometrium, increased androgen receptor expression, and decreased expression of proliferation markers (21, 24, 25, 35). The limitations of these studies are identical to those discussed with the ovarian studies discussed previously. Reassuringly, most transgender men who had regular menses before starting T therapy are reported to resume menses if T is discontinued (6, 15, 36, 37).

Table 2.

Summary of findings from histopathological examination of uteri from transgender men

Study	Population	T Exposure	Uterine Findings With T Exposurein Trans Men
Futterweit et al, 1986	19 trans men (mean age, 27 y); 12 age-matched controls. In trans men, pre-T hirsutism (8/19) and oligomenorrhea (5/19)	12 to 120 mo (mean, 37 mo)	Endometrium in trans men: 12/19 proliferative; 7/19 inactive, 4/19 leiomyomata. In controls: 6/12 proliferative, 4/12 secretory, 2/12 inactive
Miller et al, 1986	32 trans men (mean age, 29 y); 36 controls	12 to 96 mo	Endometrium: 26/32 inactive, 6/32 atrophic, 5/32 leiomyomata. Mucosal atrophy of cervix (24/32), myometrium with focal interstitial infiltration by eosinophils (9/32)
Roy et al, 1989	10 trans men (age, 20-28 y), 10 pregnant women (age, 25-40 y) at elective cesarean section, 10 nonpregnant women (age, 22-35 y) at hysterectomy for fibromata or other nonmalignant condition	At least 1 y	Approximately 10-fold increase in uterine NAD+-dependent prostaglandin 15-hydroxydehydrogenase (prostaglandin metabolizing enzyme) as compared to pregnant and nonpregnant women
Chadha et al, 1994	Ovaries: 11 trans men (mean age, 25 y); uteri: 6 trans men (mean age, 27 y); 5 PM (age, 32-51 y) and 7 M (age, 56-78 y) controls, 10 ovaries PCOD patients. Prior regular menstrual cycles in 8/11 trans men used for ovaries. [overlap Pache?]	Ovaries 11 to 72 mo (mean, 21 mo); uteri: 24 to 96 mo	Endometrium: 67% inactive, 33% atrophic. More intense AR staining myometrial/endometrial stroma in trans men
O’Hanlan et al, 2007	41 trans men (mean age, 32 y) 68% on T, 552 controls (mean age, 51 y)	6 to 168 mo (mean, 50 mo)	Lower uterine weight (on laparoscopic hysterectomy)
Mueller et al, 2008	35 trans men (mean age, 31 y)	12 mo	Decreased endometrial thickness
Perrone et al, 2009	27 trans men (mean age, 31 y), 13 PM (mean, 34 y) and 30 M (mean age, 65 y) controls	12 to 72 mo, (mean, 34 mo)	Endometrium: trans men 27/27 inactive; Nonfunctional polyps 5/27, PM proliferative endometrium 13/13, M atrophic endometrium 30/30. Lower Ki-67 (proliferation marker) in uterine glands/stroma of trans men vs PM and similar to M, trans men with polyps higher BMI than those without
Grynberg et al, 2010	112 trans men (mean age, 29 y); 98/112 prior regular cycles	24 to 108 mo (mean, 44 mo)	Endometrium: 54/112 proliferative, 50/112 atrophic; 19/112 leiomyomata
Loverro et al, 2016	12 trans men (age, 20-32 y)	Mean, 32 mo	Endometrium: 10/12 active; 2/12 secretory; Myometrium: 5 fibrosis, 5 normal, 2 hypertrophic. Endometrial epithelial cells ER (54%), PR (59%), AR (24%), Ki67 (8%). Endometrial stroma ER (40%), PR (40%), AR (39%), Ki67 (5%). Myometrium ER (17%), PR (68%), AR (69%), Ki67 (2%). Amenorrhea 8 to 12 mo after starting therapy
Khalifa et al, 2018	27 trans men (mean, 29 y), all G0P0, 1 with congenital adrenal cortical hyperplasia, 2 with prior oophorectomy, 5 history of dysmenorrhea, 1 AUB	24 received androgen 19 to 288 mo	Endometrial glands inactive 20/27, proliferative 5/27, and secretory 2/27. Focal decidua-like endometrial stromal change with glandular paucity 16/27 associated with inactive endometrial glands. Ectocervical or transformation zone transitional cell metaplasia in 17/27; 3 benign endometrial polyps, 3 leiomyomas, 1 adenomyosis
Grimstad et al, 2018	94 trans men (mean age, 30 y); 13 had documented estrogen and/or progesterone therapy while on T (7 COCs, 4 Depo-medroxyprogsterone, 3 levonorgestrel 52 ng IUD), majority nulliparous (76/94), 24/94 documented PCOS or AUB before initiation of T, 1 pubertal suppression before T.	Mean, 36.7 ± 36.6 mo 0 to 1 y 23.4% (22/94), > 1 to 2 y 31.9% (30/94), > 2 to 4 y 20.2% (19/94), > 4 y 24.5% (23/94)	Atrophic 23/94, secretory 4/94, proliferative 61/94, endometrial polyps or fibroids 9/94, adenomyosis 7/94, complex hyperplasia without atypia 1/94, other benign pathology 4/94. Mean endometrial thickness 2.0 mm ± 1.3 (paper includes further breakdown). Uterine weight mean, 77.2 g.

Abbreviations: AMH, anti-müllerian hormone; AR, androgen receptor; AUB, abnormal uterine bleeding; COC, cumulus-oocyte complex; ER, estrogen receptor; FSH, follicle-stimulating hormone; GnRHa, gonadotrophin-releasing hormone agonists; IUD, intrauterine device; IVM, in vitro maturation; M, menopausal; NAD, nicotinamide adenine dinucleotide; PCOM, polycystic ovarian morphology; PCOS, polycystic ovary syndrome; PM, premenopausal; PR, progesterone receptor; T, testosterone; TVU, transvaginal ultrasound.

Pregnancy after testosterone exposure

Pregnancies and live births have been reported in transgender men (15, 16, 38), but the total number of transgender men who have attempted pregnancy is unknown, and therefore a fecundity rate cannot be calculated. In a cross-sectional survey of 41 transgender men who had been pregnant and delivered, 25 respondents had used T before pregnancy (76% intended pregnancies) (15). Of these 25 individuals, 20% conceived while amenorrheic and 80% had resumption of menses within 6 months of stopping T (15). The majority of individuals used their own oocytes (84%), or those of a significant other (16%) (15). In this small, self-reported sample, there were no statistically significant differences in perinatal complications between transgender men with and without prior T use (15). These survey data cannot be generalized because participants were specifically screened for successful births, and those with difficulty conceiving or carrying to term were therefore excluded. Additionally, 64% of the T-treated individuals had been on T for less than 2 years, so results may not accurately reflect the effects of more long-term treatment on future fertility. There are currently no prospective studies on pregnancy in transgender men previously on T therapy.

Reports of fertility in women with endogenous hyperandrogenism, such as with congenital adrenal hyperplasia (CAH) or androgen-producing tumors, are limited. Rates of infertility and miscarriage risk increase substantially with the degree of enzyme dysfunction in CAH, with reported live birth rates of 0% to 10% in classical salt-wasting CAH patients to 63% to 90% in nonclassical CAH (39-43). Interestingly, studies have also shown that ovulation and menses do not always resume in women with CAH, even after androgens have been adequately suppressed—often hypothesized to be due to persistently elevated progesterone levels (43-46). In addition to progesterone, there are likely other confounding factors related to fertility in women with CAH, including endometrial defects, genital surgery, neuroendocrine factors, and reportedly low compliance with treatment (39, 40, 43, 45, 47), making it difficult to isolate the effect of T on fertility from these data. Finally, in a case report of a woman with an androgen-producing ovarian tumor removed by unilateral oophorectomy, menses returned in 4 months and pregnancy occurred 9 months from surgery (48). However, she likely only had male-range T levels for a short period of time, so these findings cannot be directly translated to longstanding gender-affirming T therapy.

Testosterone is contraindicated in pregnancy because of potential virilization of a female fetus to varying degrees, most profound with first trimester exposure (49). However, the reproductive health of offspring resulting from T-exposed oocytes, without direct intrauterine exposure, is unknown. There is evidence that offspring from women with PCOS may have impaired fertility and PCOS-like phenotypes (48, 50), although these outcomes have not been studied in the offspring of T-treated transgender men. Given the multifactorial nature of PCOS, however, the findings may not be applicable to gender-affirming T therapy. Additionally, transgender men are advised to stop T before attempting conception, or have a partner or gestational surrogate carry the pregnancy, therefore their offspring do not have the same in utero exposure to elevated androgens as do PCOS offspring. Additionally, there are case reports of healthy offspring being born to women with a history of CAH or androgen-producing ovarian tumors, suggesting that there may not be effects on offspring from T-exposed oocytes (48, 51); however, these studies do not specifically evaluate fertility in the offspring.

Relevant Animal Models

Animal models can provide an avenue to better understanding the reproductive impact of androgen treatment in transgender men. The historical paradigm for sex-steroid–induced changes held that permanent or organizational changes were possible during development (prepubertal, prenatal), and transient or activational changes took place during adulthood (52). In their 1985 analysis, Arnold and Breedlove challenged the strict organizational/activational dichotomy and noted that this framework has notably limited research on changes induced by long-term sex-steroid administration in adult vertebrates because of the assumption that persistent changes would not occur (53).

To date, 2 groups have published animal data intentionally mimicking gender-affirming hormone therapy, although 1 was investigating the effect of T on bone health and atherosclerosis in adult female mice that were ovariectomized, preventing analysis of reproductive changes (54, 55). The second proposes a mouse model in which the reproductive effects of exogenous testosterone can be studied, by administering twice weekly testosterone injections to adult female mice for 6 weeks (56). Compared to controls, these mice exhibit cessation of estrous cycles, T elevation, suppression of luteinizing hormone, clitoromegaly, increased atretic late antral follicles, and lack of corpora lutea. Primordial, primary, secondary, and total AFCs were similar between T-treated mice and controls, suggesting that T therapy did not affect overall ovarian reserve in this model (56).

The majority of studies using animal models to investigate the consequences of androgen therapy on the reproductive function of female animals have been carried out in the context of PCOS, and have been extensively reviewed elsewhere (57-61). There are also multiple comprehensive reviews delineating what is known regarding the influence of androgens on ovarian function, with particular attention to androgen receptor knockout models (62-64). In brief, PCOS models have been developed in monkeys, sheep, rats, and mice, and include androgen treatment (T, 5-dihydrotestosterone [DHT], dehydroepiandrosterone), estrogen treatment, aromatase inhibitors, antiprogesterones, exposure to constant light, and multiple transgenic models (57-59). Although these studies have demonstrated changes both in ovarian histopathology and function, androgens were generally administered in the prenatal and prepubertal developmental stages (57, 61) and are therefore not translatable to gender-affirming T therapy initiated postpubertally. Of the studies that did examine postpubertal administration, one study noted that adult female mice treated with DHT stopped cycling and had reduced fertility compared to controls; however, comparisons after cessation of DHT were not performed (65). Another study investigated mouse ovaries after removal of DHT that was initiated in midpubertal mice, noting reversibility of certain DHT-induced differences, including a DHT-induced lack of corpora lutea (66). However, unlike T, DHT is not aromatizable, and it is not used clinically. Another study demonstrated reduced numbers of mature oocytes on superovulation after adult female mice were treated for a week with T undecanoate, but the persistence of these changes with cessation was not examined and the short treatment duration limits generalizability (67).

Unrelated to reproduction, several studies of differential Plasmodium chabaudi malarial susceptibility have investigated adult female mice treated for 3 weeks with high levels of T (aiming for male levels) (68-73). Although these studies did not investigate reproductive changes, they reported persistent alterations in hepatic gene expression and promoter methylation when measured 12 weeks after T cessation (72, 73). These findings raise the question as to whether T can also induce persistent alterations in gene expression in reproductive organs.

Parenting Desires and Experiences

In a single-center Dutch study from 2012 analyzing the parenting desires of 50 transgender men via questionnaire, Wierckx et al found that 54% of transgender men had a current desire to have children, and an additional 8% had experienced this desire in the past (74). Of note, the study did not specifically ask about desire for genetically related children. Interestingly, the wish to parent was not significantly associated with the current presence of children (74). Further highlighting the importance of discussing fertility desires before initiating treatment, 38% of participants mentioned they would have considered oocyte cryopreservation if this technique had been available at the time of gender-affirming hormone therapy initiation; this desire was significantly more present in transgender men without children compared to participants with children (74). A mixed-methods study from Fenway Health in 2019 explored the reproductive planning of 150 transmasculine adults (of whom 121 had used gender-affirming hormone therapy). They found that 9.3% were planning to become pregnant in the future (19.3% did not know) and 12.0% were planning to use their oocytes with a surrogate (32.0% did not know) (75). Although 2 survey studies of transgender adults reveal that many believe medical professionals should offer fertility preservation before initiating gender-affirming hormone therapy (76, 77), other studies suggest extremely low use of fertility-preservation technologies in transgender men (78-80).

There is no evidence to suggest that having a transgender parent affects a child’s gender identity, sexual orientation, or developmental milestones (81, 82). On the contrary, citing a transgender parent as a reason to separate parent and child has been documented to be psychologically traumatizing for involved children (83). Moreover, Wierckx and colleagues found that transgender men with children scored significantly higher on self-perceived mental health status and vitality compared to those without children, even when correcting for relationship status and age (84).

Assisted Reproductive Technologies and Fertility Preservation in Transgender Men

As with all patients pursuing ART for current or future fertility, the transgender patient’s partnership status, the reproductive organs of their partner, and the willingness/ability of the patient or their partner to carry a pregnancy will ultimately determine options available to them for family building (85). Particularly for fertility preservation patients, it is important to clarify plans for future pregnancy so the appropriate US Food and Drug Administration–required screening can be performed for those planning to use a gestational carrier. Additionally, consideration should be given to the fact that the required medications and procedures for harvesting oocytes may create discomfort or distress for transgender men. In a prospective study among 15 adult transgender men referred for fertility preservation, Armuand et al found that many transgender men were averse to pelvic examinations required for ultrasound monitoring and oocyte retrievals and were particularly distressed by the cessation of T therapy and/or hormonal stimulation and the subsequent physical effects, including the resumption of menses, perceived voice heightening, and swelling of the hips and chest (86).

Reducing the duration of T cessation before ovarian stimulation, or not stopping T at all, could also help lessen any emotional or physical discomfort that transgender men might experience when using ART. Currently, there are no data regarding the optimal management of T around ovarian stimulation, except for anecdotal reports from clinics requiring T discontinuation 1 to 6 months before stimulation. Many clinics also use hormonal contraceptives or wait for return of menses to time cycles (7, 86-88). There are anecdotal reports, including the authors’ personal experience, with successful ovarian stimulation and blastocyst development without stopping T therapy during stimulation, but no large studies have been performed. In an attempt to minimize any potential discomfort triggered by ovarian stimulation and its physical and emotional effects, regardless of whether the patient is currently on T, clinicians should consider the use of aromatase inhibitors concurrent with gonadotropin stimulation to limit estradiol elevations, similar to protocols used for fertility preservation in breast cancer patients (86, 89). Additional considerations include using transabdominal monitoring when feasible, and properly educating all clinic staff to avoid misgendering of the patient during clinical encounters.

Although multiple organizations recommend fertility preservation counseling before starting gender-affirming therapy, until recently there were no data comparing ART outcomes between transgender men who had previously been on T to those who had not, or comparing outcomes in transgender men to cisgender women. In 2019, Adeleye and colleagues reported retrospective ovarian stimulation outcomes for 13 transgender men, 7 of whom had previously been on T (90). They found no differences in total oocytes retrieved or number of mature oocytes between transgender men and age- and BMI-matched cisgender women; however, they did report reduced number of oocytes retrieved in transgender men who had previously been on T vs those who had never been exposed to T (25.5 vs 12, P = .038). This difference was no longer significant when 2 outlier transgender men with low initial AFCs less than 5 were removed from the analysis (90). Subsequently, Leung et al published a similar retrospective report comparing age-, BMI-, and AMH-matched cisgender women to 26 transgender men, 16 of whom had previously been on T, but they did not directly compare transgender men with and without prior T exposure (91). Compared to cisgender women, transgender men had significantly more oocytes retrieved (19.9 vs 15.9, P = .04), but there was no statistically significant difference when cisgender women were compared only to transgender men who had previously been on T. Total gonadotropin requirements were higher in transgender men in both comparisons (91). There was a 58.3% live birth rate in couples who proceeded to embryo transfer (into the patient or a female partner), with 85.7% of the transgender men in those couples reporting prior T exposure (91). T was discontinued before stimulation in both studies.

OTC may also be an option for transgender men pursuing fertility preservation that would not require T cessation and could be performed if the ovaries were being removed as part of a gender-affirming surgery (30, 31). Although OTC is currently considered experimental, more than 130 live births have been reported from previously cryopreserved ovarian tissue in cisgender women (92-94). With in vitro maturation technique advancements, OTC may become a viable option for transgender men to have offspring from their gametes without necessitating ovarian stimulation (31, 95).

Conclusions

Unfortunately, in regards to best practices surrounding fertility treatment and/or fertility preservation in transgender men, there seem to be more questions than answers from the available literature. Although some inferences might be considered from existing animal models and clinical studies of hyperandrogenic disorders in cisgender women, there is a paucity of data directly translatable to the high levels of and prolonged postpubertal exposure to T that is characteristic of gender-affirming hormone therapy. As such, the current status quo is to recommend fertility preservation before initiation of T therapy and, for patients presenting subsequent to T therapy, cessation of T before ovarian stimulation. Animal models directly mimicking gender-affirming T therapy can provide preliminary data about the effects, and reversibility, of T on fertility and may be able to fill in gaps in knowledge when a randomized controlled trial would be unethical. More clinical outcome data are also desperately needed, however, to ensure that we are providing appropriate care to this patient population, and will likely require multi-institutional collaboration. Moreover, there is a gap in the literature assessing regret surrounding family-building decisions in transgender individuals, as well as decision tools in reproductive decision making, rendering more qualitative research on this topic also critical. The long-term goal should be to equip medical providers with the information necessary to provide high-quality, data-driven counseling regarding fertility options for transgender men.

Glossary

Abbreviations

AFC

antral follicle count

ART

assisted reproductive technologies

BMI

body mass index

CAH

congenital adrenal hyperplasia

DHT

dihydrotestosterone

GnRHa

gonadotropin-releasing hormone agonists

OTC

ovarian tissue cryopreservation

PCOS

polycystic ovary syndrome

T

testosterone

Additional Information

Disclosure Summary: The authors have nothing to disclose.