Effects of antiandrogens on prolactin levels among transgender women on estrogen therapy: A systematic review

Abstract

Background

Gender-affirming hormone therapy for transgender women includes estrogen and antiandrogens (cyproterone acetate, spironolactone, or gonadotropin-releasing hormone agonists). Both estrogen and antiandrogens are reported to increase prolactin levels. The objective is to systematically review the evidence of the effects of antiandrogens on prolactin levels, hyperprolactinemia, and prolactinomas among transgender women on estrogen therapy.

Methods

We searched PubMed, Embase, and PsycInfo up to May 2020. We included studies with at least 3 months follow-up that evaluated the effects of antiandrogens among transgender women and reported on prolactin levels, hyperprolactinemia, or image-confirmed prolactinomas. Two reviewers independently screened studies for eligibility, serially abstracted data, and independently assessed risk of bias and graded strength of evidence.

Findings

We included 17 studies (16 publications): 8 prospective cohorts, 8 retrospective cohorts, and 1 cross-sectional study, each with a moderate to serious risk of bias. Among transgender women on estrogen, prolactin levels increased by over 100% with cyproterone acetate and by up to 45% with spironolactone. However, we were unable to isolate the effects of antiandrogens from estrogen therapy. We were unable to draw conclusions about effects of antiandrogens on hyperprolactinemia and prolactinomas.

Interpretation

Prolactin levels may be increased in transgender women who are taking both estrogens and an antiandrogen. Future research is needed to determine the effects of different antiandrogens on prolactin levels separately from estrogen therapy. Ideally, future studies would be prospective, provide either a comparison of two different antiandrogens or compare combination of estrogen and antiandrogen therapy to estrogen alone, and control for possible confounders.

Introduction

Transgender individuals can decide to take gender-affirming hormonal therapy with the goal to develop secondary sex characteristics that align with their gender identity while suppressing those associated with the sex they were assigned at birth. For transgender women, gender-affirming hormonal therapy could include an estrogen and an antiandrogen (Deutsch, 2016).

Transgender women are reported to have a possible increased risk for hyperprolactinemia (Coleman et al., 2012a) which, in the general population, can lead to reproductive disorders and bone loss (Samperi et al., 2019). High doses of estrogen can stimulate prolactin production (Ignacak et al., 2012; Samperi et al., 2019). Cyproterone acetate, an antiandrogen commonly prescribed to transgender women in Europe and South America, is reported to cause elevated prolactin levels as well (Defreyne et al., 2017; Nota et al., 2017; Wierckx et al., 2014). Additionally, there have been several case reports of prolactinomas in transgender women; a prolactinoma is a benign tumor of the pituitary gland that increases prolactin levels (Bunck et al., 2009; Cunha et al., 2015; Garcia-Malpartida et al., 2010; Gooren et al., 1988; Serri et al., 1996). Macroprolactinomas can compress the pituitary stalk and the surrounding structures, which can lead to vision problems, decreased production of other hormones produced by the pituitary gland, and headaches.

The objective of this systematic review is to assess the evidence of the effects of antiandrogens, including gonadotropin-releasing hormone agonists (GnRHa), on prolactin levels, hyperprolactinemia, and prolactinomas among transgender women taking estrogen therapy.

Methods

This systematic review is derived from a series of reviews on gender-affirming care conducted for the World Professional Association for Transgender Health (WPATH). We developed and registered a protocol (PROSPERO (CRD42018115379)). The specific questions addressed in the series of reviews can be found in the protocol.

Search strategy and selection criteria

We searched PubMed, Embase, and PyscInfo up to July 2018. We updated the search in May 2020 using a targeted search strategy for the three databases. The original search strategy is available in the Appendix Table 1 (Supplementary material) and the targeted search strategy is in Appendix Table 2 (Supplementary material). We also screened the reference lists of related systematic reviews, reference lists from the WPATH SOC8 Hormone Chapter members, and hand-searched the International Journal of Transgender Health (formerly International Journal of Transgenderism (IJT)).

Table 1.

Studies of the effects of antiandrogens on prolactin levels, hyperprolactinemia, and prolactinomas in transgender women on estrogen therapy.

Author, year Study name Location	Study design	Start year	Length of followup	Transfeminine population	Type of antiandrogen	Route of administration for estrogen	Mean age (age range)	Mean BMI (BMI range)	Risk of bias
Defreyne et al., 2017 ENIGI Europe	Prospective cohort	2010	18 months	Hormone therapy naïve	Cyproterone acetate	Oral, transdermal patch, or transdermal gel	30.75	23.12	Serious
					Cyproterone acetate	Oral, transdermal patch, or transdermal gel +  orchiectomy	32.35	23.26
Nota et al., 2017* ENIGI Europe	Prospective cohort	2012	12 months	Hormone therapy naive	Cyproterone acetate	Oral or transdermal patch	34 (19 to 69)	NR	Serious
Wierckx et al., 2014 ENIGI Europe	Prospective cohort	2010	12 months	Hormone therapy naïve	Cyproterone acetate	Oral	19.3 to 31.7†	22.9 to 23.9†	Moderate
					Cyproterone acetate	Transdermal patch or transdermal gel	NR	NR
Giltay et al., 2005 Europe	Prospective cohort	1996	4 months	NR	Cyproterone acetate	Oral	32.4 (20 to 43)	22.8	Moderate
Manieri et al., 2014 Europe	Prospective cohort	NR	12 months	NR	Cyproterone acetate	Oral	32.7	NR	Serious
Asscheman et al., 1989 Europe	Retrospective cohort	1972	3.6 to 4.4 years	NR	Cyproterone acetate	Oral or IM	Median, 32 (16 to 67)	NR	Moderate
Cunha et al., 2018 South America	Retrospective cohort	NR	6 months	Hormone therapy naïve	Cyproterone acetate	Oral	38.3 (23 to 58)‡	NR	Serious
					None	Oral	NR	NR
Fung et al., 2016 North America	Retrospective cohort	2009	12 months	Hormone therapy naïve or newly starting hormone therapy	Cyproterone acetate	Oral or transdermal	38.2	25.5	Moderate
					Spironolactone	Oral or transdermal	32.9	25.0
Nota et al., 2017* Europe	Retrospective cohort	2009	2.2 years	Aged > 18 years, hormone therapy naïve	Cyproterone acetate	Oral or transdermal	37 (19 to 63)	NR	Serious
Tack et al., 2017 Europe	Retrospective cohort	2008	12 to 16 months	Tanner stage G4 or higher	Cyproterone acetate	NR	16.5	NR	Serious
van Kesteren et al., 1997 Europe	Retrospective cohort	1975	NR	Hormone therapy > 2 months	Cyproterone acetate	Oral or transdermal patch	41 (18 to 86)	NR	Moderate
Becerra Fernandez et al., 1999 Europe	Cross-sectional	1993	NA	Hormone therapy > 6 months	Cyproterone acetate	IM	29.9 (18 to 42)	NR	Serious
Olson-Kennedy et al., 2018 North America	Prospective cohort	2011	21 to 31 months	Aged 12 to 24 years, hormone therapy naïve or < 3 months of previous use	Spironolactone or GnRH analogue	NR	(12 to 23)	24.60	Moderate
Bisson et al., 2018 North America	Retrospective cohort	2015	6 years	NR	Spironolactone	Oral or transdermal	41 (20 to 68)	28 (19 to 44)	Serious
Jain et al., 2019 North America	Retrospective cohort	2011	3.4 years	Eugonadal	Spironolactone +  medroxyprogesterone acetate	Sublingual or transdermal	31 (14 to 69)‡	NR	Moderate
					Spironolactone	Sublingual or transdermal	NR	NR
Dittrich et al., 2005 Europe	Prospective cohort	NR	24 months	Hormone therapy naïve, eugonadal	Goserelin acetate (SC)	Oral	38.37	24.19	Serious
Hannema et al., 2017 Europe	Prospective cohort	1998	3 years	Treated with estrogen for > = 12 months	Triptorelin (IM)	Oral	Median, 16 (13.9 to 18.9)	20.8	Serious

BMI = body mass index; ENIGI = European Network for the Investigation of Gender Incongruence; GnRH = gonadotropin-releasing hormone; IM = intramuscular; NA = not applicable; NR = not reported; SC = subcutaneous.

*This publication included both a prospective study and a retrospective study.

†Range in mean values stratified by study center.

‡Mean age and age range for the total sample.

Table 2.

Results of studies that evaluate the effects of antiandrogens on prolactin levels in transgender women on estrogen therapy.

Author, year	Study design	Length of followup	Arm Description	N	Baseline prolactin level, µg/L*	Followup prolactin level, µg/L*	Within-arm comparison	Between-arm comparison
Defreyne et al., 2017 ENIGI	Prospective cohort	18 months	Cyproterone acetate + oral or transdermal estrogen	49	Median, 7.84 (IQR, 6.25 to 9.80)	12 months followup Median, 23.05 (IQR, 13.38 to 29.85) 18 months followup Median, 14.10 (IQR, 11.78 to 18.65)	% change in median, 194%; P < 0.05 12 months vs. baseline % change in median, −39%; P < 0.001 18 months vs. 12 months	P = 0.008 18 months vs. postorchiectomy
Defreyne et al., 2017 ENIGI	Prospective cohort	18 months	Cyproterone acetate† + oral or transdermal estrogen + orchiectomy	58	9.42 (8.64)	Preorchiectomy 23.72 (13.52) Postorchiectomy 10.17 (5.64)	% change, 152%; P < 0.05 preorchiectomy vs. baseline
Nota et al., 2017 ENIGI	Prospective cohort	1 years	Cyproterone acetate + oral or transdermal estrogen	61	7.05	20.68	% change, 193% (95% CI, 156% to 219%)	NA
Wierckx et al., 2014 ENIGI	Prospective cohort	12 months	Cyproterone acetate + oral estrogen	40	Median, 8.1 (IQR, 6.3 to 11.3)	Median, 20.6 (IQR, 16.7 to 28.9)	% change in median, 154%; P < 0.001	NR
Wierckx et al., 2014 ENIGI	Prospective cohort	12 months	Cyproterone acetate + transdermal estrogen	13	Median, 6.6 (IQR, 4.1 to 9.3)	Median, 26.2 (IQR, 18.6 to 36.2)	% change in median, 297%; P = 0.001	NR
Giltay et al., 2005	Prospective cohort	4 months	Cyproterone acetate + oral estrogen	15	6.11 (1.4)	20.68 (7.99)	% change, 238%; P < 0.0005	NA
Manieri et al., 2014	Prospective cohort	12 months	Cyproterone acetate + oral estrogen	56	Median, 10.5	Median, 29.5	% change in median, 181%; P is NR	NA
Cunha et al., 2018	Retrospective cohort	6 months	Oral estrogen alone	8	Median, 7.6 (IQR, 3.9 to 17.9)	Median, 16.3 (IQR, 3.4 to 48.9)	% change in median, 114%; NS
Cunha et al., 2018	Retrospective cohort	6 months	Cyproterone acetate + oral estrogen	43	Median, 7.6 (IQR, 3.9 to 17.9)	Median, 16.3 (IQR, 8.0 to 48.9)	% change in median, 114%; P = 0.009	P = 0.126 vs. oral estrogen alone
Fung et al., 2016	Retrospective cohort	12 months	Cyproterone acetate + oral or transdermal estrogen	31	7.92 (4.18)	19.7 (10.29)	% change, 149%; Mean change, 11.8 (SD, 8.63), P < 0.001 vs. baseline	P < 0.001 vs. the change in prolactin among those on spironolactone
Fung et al., 2016	Retrospective cohort	12 months	Spironolactone + oral or transdermal estrogen	82	8.73 (4.13)	11.8 (6.96)	% change, 35%; Mean change, 3.10 (SD, 5.70)
Nota et al., 2017	Retrospective cohort	38 months	Cyproterone acetate + oral or transdermal estrogen	38	7.5	20 months followup 16.9 38 months followup§ 7.1	% change from baseline to 20-month followup, 125%; % change from baseline to 38-month followup, −6% (95% CI, −24% to 15%; P > 0.05)	NA
Tack et al., 2017	Retrospective cohort	12 months	Cyproterone acetate	27	7.3 (3.2)	17.5 (5.2)	% change, 140%; P = 0.003	NA
Tack et al., 2017	Retrospective cohort	12 months	Cyproterone acetate + estrogen	21	16.7 (7.4)	10.6 (5.3)	% change, −37%; P = 0.011	NA
Olson-Kennedy et al., 2018	Prospective cohort	24 months	Spironolactone or GnRH analogue + estrogen	13	8.27 (5.98; Range, 1 to 22)	11.99 (5.44; Range, 3.2 to 19.7)	% change, 45%; P = 0.047	NA
Bisson et al., 2018	Retrospective cohort	6 years	Spironolactone + oral or transdermal estrogen	26	Median, 8.5 (IQR, 6.5 to 12.1)	Median, 9.7 (IQR, 8.5 to 14.5)	% change in median, 14%; NS	NA
Jain et al., 2019	Retrospective cohort	3.4 years	Medroxyprogesterone acetate + spironolactone + sublingual or transdermal estrogen (8 mg)‡	40	8.7 (4.2)	10.7 (5.5)	% change, 23%; NS	NS vs. spironolactone +  estrogen
			Spironolactone + sublingual or transdermal estrogen (8 mg)‡	49	8.7 (4.2)	11.8 (4.7)	% change, 36%; NS
Dittrich et al., 2005	Prospective cohort	24 months	Goserelin acetate + oral estrogen	60	11.5 (9.7)	13.5 (4.6)	% change, 17%; P = 0.41	NA

CI = confidence interval; ENIGI = European Network for the Investigation of Gender Incongruence; IQR = interquartile range; NA = not applicable; NR = not reported; NS = not statistically significant; SD = standard deviation.

* Results are presented as mean (standard deviation), unless stated otherwise. Units are µg/L.

† Cyproterone acetate was stopped when the orchiectomy was performed.

‡ Results were stratified by dose of estrogen. Prolactin levels were significantly greater than baseline when transgender women were taking 6 mg of estrogen and spironolactone with or without medroxyprogesterone acetate. At all other doses of estrogen (2 mg, 4 mg, 8 mg, 10 mg, and 12 mg), there were no significant differences in baseline and follow-up prolactin levels. For all doses of estrogen, there were no significant differences in prolactin levels comparing those taking medroxyprogesterone acetate and those not taking medroxyprogesterone acetate.

§ Patients had underwent gender affirmation surgery between 20 months and 38 months of followup.

We included studies that evaluated the effects of antiandrogens among transgender women on prolactin levels, hyperprolactinemia, or prolactinoma confirmed by imaging. Participants on gender-affirming hormone therapy must have been followed for at least 3 months. We included studies regardless of language. We included any study design, except for case reports.

Two reviewers independently screened titles, abstracts, and full text articles. Differences between reviewers were resolved through consensus. We used DistillerSR (Evidence Partners, 2010) to manage the screening process.

Data analysis

We created and pilot tested standardized forms for data extraction. One reviewer completed data abstraction, and the second reviewer confirmed the data abstraction for completeness and accuracy. We completed the data abstraction process using the Systematic Review Data Repository^TM (SRDR).

Two reviewers independently assessed the risk of bias in included studies. We used the Cochrane tool for assessing the risk of bias of randomized controlled trials (RCTs) (Higgins et al., 2011). For non-randomized studies, we used the Cochrane Risk of Bias Assessment Tool for Non-Randomized Studies of Interventions (ROBINS-I tool) (Sterne et al., 2016).

We graded strength of evidence for outcomes using the grading scheme recommended by the U.S. Agency for Healthcare Research and Quality Methods Guide for Conducting Comparative Effectiveness Reviews (Berkman et al., 2015). For each outcome, we graded the strength of the evidence using the best available evidence. We graded the strength of the evidence considering the domains of study limitations, directness, consistency, precision, and reporting bias. The overall strength of evidence grades was then determined to be “high,” “moderate,” “low,” or “insufficient.”

Role of the funding source

WPATH provided the initial review questions and reviewed the protocol. WPATH had no role in study design, data collection, data analysis, data interpretation, or writing of this manuscript. The corresponding author had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Results

We retrieved 2609 records (Figure 1) and included 17 studies (published in 16 papers) (Asscheman et al., 1989; Becerra Fernandez et al., 1999; Bisson et al., 2018; Cunha et al., 2018; Defreyne et al., 2017; Dittrich et al., 2005; Fung et al., 2016; Giltay et al., 2005; Hannema et al., 2017; Jain et al., 2019; Manieri et al., 2014; Nota et al., 2017; Olson-Kennedy et al., 2018; Tack et al., 2017; van Kesteren et al., 1997; Wierckx et al., 2014). Table 1 details the study design and participant characteristics of the included studies. Eight studies were prospective cohorts, eight were retrospective cohorts, and one was a cross-sectional study. One publication reported both a prospective cohort and a retrospective cohort (Nota et al., 2017).

Figure 1.

Summary of the literature search.

*Total exceeds the number of citations in the exclusion box, because citations could be excluded for more than one reason

Three of the eight prospective cohort studies were conducted among the European Network for the Investigation of Gender Incongruence (ENIGI) cohort (Defreyne et al., 2017; Nota et al., 2017; Wierckx et al., 2014). Two of these studies recruited participants from the same study center, so some patients may be represented in both papers (Defreyne et al., 2017; Wierckx et al., 2014). Participants received cyproterone acetate in 5 studies (Defreyne et al., 2017; Giltay et al., 2005; Manieri et al., 2014; Nota et al., 2017; Wierckx et al., 2014), goserelin acetate in one study (Dittrich et al., 2005), triptorelin in one study (Hannema et al., 2017), and either spironolactone or a GnRHa in one study (Olson-Kennedy et al., 2018). Two studies recruited participants who were adolescents or young adults (Hannema et al., 2017; Olson-Kennedy et al., 2018). The remaining studies recruited adults (mean age ranged from 30.8 to 38.4 years old). We assessed each of the eight prospective cohort studies as having moderate (Giltay et al., 2005; Olson-Kennedy et al., 2018; Wierckx et al., 2014) to serious (Defreyne et al., 2017; Dittrich et al., 2005; Hannema et al., 2017; Manieri et al., 2014; Nota et al., 2017) risk of bias because of concerns about the selection of participants into the study and the selection of reported results (Appendix Figure 1, Supplementary material).

One of the eight retrospective studies included adolescents (Tack et al., 2017), with the other seven recruiting adults. Participants received cyproterone acetate in five retrospective studies (Asscheman et al., 1989; Cunha et al., 2018; Nota et al., 2017; Tack et al., 2017; van Kesteren et al., 1997), spironolactone in one study (Bisson et al., 2018), and spironolactone with or without medroxyprogesterone acetate in one study (Jain et al., 2019). One study compared transgender women taking spironolactone with those taking cyproterone acetate (Fung et al., 2016). In one study, participants received estrogen therapy after a minimum of 6 months of therapy with cyproterone acetate (Tack et al., 2017). We considered these retrospective studies to have a moderate (Asscheman et al., 1989; Fung et al., 2016; Jain et al., 2019; van Kesteren et al., 1997) to serious (Bisson et al., 2018; Cunha et al., 2018; Nota et al., 2017; Tack et al., 2017) risk of bias because of concerns regarding confounding, selection of participants into the study, missing data, and selection of reported results.

The cross-sectional study included adults who had self-administered hormonal therapy for at least 6 months (Becerra Fernandez et al., 1999). Therapy included cyproterone acetate, in addition to estradiol and algestone. This study was considered to have a serious risk of bias because of concerns with confounding, selection of participants into the study, and classification of the interventions.

Prolactin levels

Seven prospective (N = 303–356) and six retrospective (N = 347) cohort studies reported prolactin levels among transgender women taking both an estrogen and an antiandrogen (Table 2) (Bisson et al., 2018; Cunha et al., 2018; Defreyne et al., 2017; Dittrich et al., 2005; Fung et al., 2016; Giltay et al., 2005; Jain et al., 2019; Manieri et al., 2014; Nota et al., 2017; Olson-Kennedy et al., 2018; Tack et al., 2017; Wierckx et al., 2014). The follow-up ranged from 4 to 24 months in the prospective studies and from 6 months to 6 years in the retrospective studies. Most of the prospective studies compared prolactin levels in transgender women before and after starting a combination of an antiandrogen and estrogens (Bisson et al., 2018; Cunha et al., 2018; Dittrich et al., 2005; Fung et al., 2016; Giltay et al., 2005; Manieri et al., 2014; Nota et al., 2017; Olson-Kennedy et al., 2018; Tack et al., 2017). With the exception of one study in which the participants received goserelin acetate and oral estrogen (Dittrich et al., 2005), all of the prospective cohort studies reported statistically significant increases in prolactin levels among transgender women. The percentage change from baseline in prolactin levels ranged from 152% to 297% in the studies where participants took cyproterone acetate and was 45% in the study where participants took spironolactone; this difference is not statistically significant (P = 0.10). During the follow-up in one study, some of the patients had an orchiectomy and ceased taking cyproterone acetate (Defreyne et al., 2017). Prolactin levels were significantly higher in transgender women who had not had an orchiectomy and remained on cyproterone acetate compared to those who had an orchiectomy and ceased cyproterone acetate (P < 0.008).

There were a variety of comparisons in the retrospective cohort studies. One of the retrospective cohorts compared transgender women on estrogens plus cyproterone acetate with transgender women on estrogens alone (Cunha et al., 2018). There was a statistically significant increase in prolactin levels among transgender women taking estrogens plus cyproterone acetate (P = 0.009), but a nonsignificant increase among transgender women taking estrogens alone. The prolactin levels at follow-up did not significantly differ between those taking cyproterone acetate and estrogens and those taking estrogens alone (P = 0.126). Another retrospective cohort reported a 125% increase in prolactin levels while transgender women were taking cyproterone acetate plus oral or transdermal estrogens prior to gender affirmation surgery, but the study did not provide enough information to determine if this was statistically significant (Nota et al., 2017). After gender affirmation surgery, which included a gonadectomy, prolactin levels decreased to baseline levels (Nota et al., 2017). In another study, prolactin levels increased significantly while transgender women were on cyproterone acetate (P = 0.003), but they significantly decreased when estrogen was added to their regimen (P = 0.011) (Tack et al., 2017).

One retrospective cohort in which participants took spironolactone with either oral or transdermal estrogens reported a change in prolactin levels that was not statistically significant (Bisson et al., 2018). Another retrospective cohort reported a significant increase in prolactin levels among transgender women taking cyproterone acetate plus estrogen compared to those taking spironolactone plus estrogen (P < 0.001) (Fung et al., 2016). The mean change from baseline in prolactin levels was 11.8 µg/L (standard deviation [SD], 8.63) in the cyproterone acetate group and 3.10 µg/L (SD, 5.70) in the spironolactone group. One retrospective cohort included transgender women who were taking sublingual or transdermal estradiol and spironolactone, comparing the effects on prolactin levels with and without medroxyprogesterone acetate stratified by the dose of estradiol (Jain et al., 2019). There was no statistically significant difference between baseline and follow-up levels of prolactin with or without medroxyprogesterone acetate for most doses of estradiol. There were no statistically significant differences in prolactin levels comparing regimens with and without medroxyprogesterone acetate.

Prolactin levels may be increased in transgender women who are taking both estrogens and an antiandrogen (low strength of evidence). Among transgender women on estrogens, prolactin levels increased by over 100% with cyproterone acetate and by up to 45% with spironolactone. However, we are unable to isolate the effects of antiandrogens from estrogen therapy as most patients were on combination therapy.

Hyperprolactinemia

Four prospective cohort (N = 186), four retrospective cohort (N = 1119), and one cross-sectional studies (N = 31) with moderate to serious risk of bias reported on hyperprolactinemia among transgender women taking both estrogens and an antiandrogen (Table 3) (Asscheman et al., 1989; Becerra Fernandez et al., 1999; Cunha et al., 2018; Hannema et al., 2017; Manieri et al., 2014; Nota et al., 2017; van Kesteren et al., 1997; Wierckx et al., 2014). The follow-up ranged from 12 months to 3 years in the prospective studies and from 6 months to 4.4 years in the retrospective studies. The definition for hyperprolactinemia varied across the studies; the cutoff values for prolactin levels ranged from 0.94 μg/L (Becerra Fernandez et al., 1999) to 47 μg/L (Asscheman et al., 1989; van Kesteren et al., 1997). One study reported the incidence of hyperprolactinemia using twice the upper limit of normal for the male and female reference, but they did not specify the values for the upper limits (Wierckx et al., 2014). Another study did not specify criteria for diagnosing hyperprolactinemia (Hannema et al., 2017).

Table 3.

Results of studies that evaluate the effects of antiandrogens on hyperprolactinemia in transgender women on estrogen therapy.

Author, year	Study design	Length of follow-up	Arm description	Outcome	n / N (%)
Becerra Fernandez et al., 1999	Cross-sectional	NA	Cyproterone acetate + intramuscular estrogen	Hyperprolactinemia (prolactin level > 0.94 μg/L)	24 / 31 (77)
Nota et al., 2017	Retrospective cohort	2.2 years	Cyproterone acetate + oral or transdermal estrogen	Hyperprolactinemia (prolactin level > 9.4 μg/L)	0 / 38 (0)
Cunha et al., 2018	Retrospective cohort	6 months	Oral estrogen +/- cyproterone acetate	Hyperprolactinemia (prolactin level > normal range [2 to 15 μg/L])	4 / 51 (8)
Manieri et al., 2014	Prospective cohort	12 months	Cyproterone acetate + oral estrogen	Hyperprolactinemia (prolactin level > 20 μg/L)	24 / 56 (44)
Nota et al., 2017 ENIGI	Prospective cohort	12 months	Cyproterone acetate + oral or transdermal estrogen	Hyperprolactinemia (prolactin level ≥ 28.2 μg/L)	18 / 61 (29)
Asscheman et al., 1989	Retrospective cohort	3.6 to 4.4 years	Cyproterone acetate + oral or intramuscular estrogen	Hyperprolactinemia (prolactin level > 47 μg/L)	46 / 214 (21)
van Kesteren et al., 1997	Retrospective cohort	NR	Cyproterone acetate + oral or transdermal estrogen	Hyperprolactinemia (prolactin level > 47 μg/L)	115 / 816 (14)
Wierckx et al., 2014 ENIGI	Prospective cohort	12 months	Cyproterone acetate + oral or transdermal estrogen	Hyperprolactinemia (prolactin level > 2x ULN according to male reference range)	8 / 53 (16)
Wierckx et al., 2014 ENIGI	Prospective cohort	12 months	Cyproterone acetate + oral or transdermal estrogen	Hyperprolactinemia (prolactin level > 2x ULN according to female reference range)	2 / 53 (4)
Hannema et al., 2017	Prospective cohort	3 years	Triptorelin + oral estrogen	Hyperprolactinemia (NR)	1 / 16 (6)

ENIGI = European Network for the Investigation of Gender Incongruence; NR = not reported; ULN = upper limit of normal.

The incidence of hyperprolactinemia ranged from 0% to 77%. One study had a very low cutoff value for defining hyperprolactinemia (Becerra Fernandez et al.,1999). The range in the incidence of hyperprolactinemia is 0–44% with this study removed. Among the five studies that used a cutoff of at least 20 µg/L of prolactin, the range in the incidence of hyperprolactinemia was 4% to 44% (Asscheman et al., 1989; Manieri et al., 2014; Nota et al., 2017; van Kesteren et al., 1997; Wierckx et al., 2014).

Because of the concern for study limitations and the imprecise results, we were unable to draw a conclusion of the effects of a combination of estrogens and antiandrogen therapy on hyperprolactinemia (insufficient strength of evidence).

Prolactinoma

One retrospective cohort study (N = 98) with a serious risk of bias reported on prolactinoma among transgender women taking both estrogens and spironolactone (Appendix Table 3, Supplementary material) (Bisson et al., 2018). After 6 years of follow-up, none of the 98 participants was diagnosed with prolactinoma. This was confirmed by radiologic evaluation of the pituitary and laboratory testing whenever prolactin levels were elevated.

Because of study limitations and the imprecise results, we were unable to draw a conclusion of the effects of a combination of estrogens and antiandrogen therapy on prolactinomas (insufficient strength of evidence).

Discussion

Based on the evidence we identified in this systematic review, there may be an increase in prolactin levels in transgender women taking an antiandrogen and estrogens (low strength of evidence). Among transgender women on estrogens, prolactin levels increased by over 100% with cyproterone acetate; prolactin levels increased to a lesser extent with spironolactone. However, it is unclear if the increase in prolactin levels is due to estrogen therapy or to antiandrogen therapy. Most of the studies compared prolactin levels before and after starting estrogens and antiandrogen therapy. Only one study provided a head-to-head comparison of two different antiandrogens (Fung et al., 2016), and three studies compared estrogen therapy alone to estrogen therapy plus cyproterone acetate (Cunha et al., 2018; Defreyne et al., 2017; Nota et al., 2017).

We found insufficient evidence to determine if antiandrogens increase the risk of hyperprolactinemia or prolactinomas in transgender women taking estrogens. The definition of hyperprolactinemia differed substantially across the studies (from prolactin levels > 0.94 µg/L to prolactin levels > 47 µg/L). Only five of the included studies defined hyperprolactinemia using a cutoff value for prolactin levels that was higher than the upper limit of normal (Asscheman et al., 1989; Manieri et al., 2014; Nota et al., 2017; van Kesteren et al., 1997; Wierckx et al., 2014). Furthermore, it is impossible to separate the effects of antiandrogens agonists from estrogens. Only one retrospective study reported on prolactinomas confirmed by imaging, and it reported no cases among transgender women on spironolactone and estrogens (Bisson et al., 2018).

To our knowledge, this is the first systematic review to assess the effect of antiandrogens on prolactin levels, hyperprolactinemia, or prolactinoma in transgender women taking estrogen therapy. Another systematic review assessed the prevalence of brain tumors, including prolactinomas, among transgender women, but did not focus on antiandrogens (McFarlane et al., 2018). Like that systematic review, we found too few cases of prolactinoma in transgender women on gender-affirming hormone therapy to draw a conclusion. That review identified some case reports of transgender women developing prolactinoma after starting gender-affirming therapy, but it did not include any large cohort studies that reported on prolactinomas. Further, the McFarlane review could not attribute the role of estrogens or antiandrogens in the development of prolactinomas. Current guidelines from WPATH and the Endocrine Society suggest that the risk of prolactinoma is low and recommend periodically monitoring prolactin levels in transgender women (Coleman et al., 2012c; Hembree et al., 2017). Other guidelines for transgender people recommend screening for prolactin levels only in symptomatic transgender women (i.e., visual disturbances, excessive galactorrhea, or new onset headaches) (Deutsch, 2016).

Our review has some limitations. First, our review may be subject to publication bias. We may have missed some studies that were published as only as conference abstracts or were never published. Second, we excluded studies in which not all participants were taking an antiandrogen. Our search captured two of these studies: a retrospective study reporting a non-statistically significant increase in prolactin levels after 6 months of follow-up among transgender women taking estrogens with or without spironolactone (Jarin et al., 2017), and a retrospective study reporting an increased incidence of total prolactinomas among transgender women on estrogens, and, if desired, an antiandrogen (either cyproterone acetate or spironolactone) compared to the general Dutch cisgender male and female populations (Nota et al., 2018). Third, we did not consider prolactinomas unless confirmed by imaging. We excluded two studies from our analysis that reported no prolactinomas among their participants but did not state if this was confirmed by imaging or not (Defreyne et al., 2017; Jain et al., 2019).

Several limitations with the evidence base hindered our ability to conduct syntheses and draw conclusions. Many of the included studies did not report on whether they controlled for possible confounders, such as use of other pharmacologic therapies that could influence prolactin levels. In particular, antidepressants and antipsychotics are thought to elevate prolactin levels (Samperi et al., 2019), and many transgender persons have some type of psychiatric comorbidity (Gomez-Gil et al., 2009; Hepp et al., 2005; Heylens et al., 2014; Reisner et al., 2016). There was significant heterogeneity in the study design, therapies used, and outcome definitions, limiting our ability to pool results. Lastly, it is impossible to separate the effects of antiandrogen therapy from estrogen therapy in most of the studies.

Although our results suggest that prolactin levels are elevated in transgender women taking both estrogen and antiandrogen therapy, future research is needed to confirm this and to determine the comparative effects of different antiandrogens on prolactin levels. Ideally, studies should be prospective, provide either a comparison of two different antiandrogens or compare combination of estrogen and antiandrogen therapy to estrogen alone, and control for possible confounders. Additionally, future research would benefit from an agreed definition of hyperprolactinemia in transgender women.

Disclosure statement

All authors declare no competing interests.

Disclaimer

This review was partially funded by the World Professional Association for Transgender Health (WPATH). The authors of this manuscript are responsible for its content. Statements in the manuscript do not necessarily reflect the official views of or imply endorsement by WPATH.