The Role of Estrone in Feminizing Hormone Treatment.

CONTEXT: In trans women, hormone treatment induces feminization; however, the degree of feminization varies from person to person. A possible contributing factor could be estrone, a weak estrogen that interferes with the estrogen receptor.
OBJECTIVE: We assessed whether estrone is involved in feminization induced by hormone treatment.
METHODS: This prospective cohort study, with follow-up of 1 year, included 212 adult trans women at a gender identity clinic, who were starting gender-affirming hormone treatment between July 2017 and December 2019, median age 25 years. Change in fat percentage and breast development were assessed.
RESULTS: After 12 months of hormone treatment, estrone concentration was 187 pmol/L (95% CI, 153-220) in transdermal and 1516 pmol/L (95% CI, 1284-1748) in oral estradiol users. Fat percentage increased by 1.2% (interquartile range [IQR], 0.3-4.8) in transdermal and 4.6% (IQR, 2.5-5.9) in oral estradiol users. This was not associated with estrone concentrations in transdermal (+4.4% (95% CI, -4.0 to 13) per 100 pmol/L increase in estrone concentration) nor in oral estradiol users (-0.7% [95% CI, -1.7 to 0.3]). Breast volume increased by 69 mL (IQR, 58-134) in transdermal and 62 mL (IQR, 32-95) in oral estradiol users. This was not associated with estrone concentrations in transdermal (+14% [95% CI, -49 to 156] per 100 pmol/L increase in estrone concentration) nor oral estradiol users (+11% [95% CI -14 to 43]).
CONCLUSIONS: Change in fat percentage and breast development in trans women were not associated with estrone concentrations nor with administration route. Therefore, measurement of estrone concentrations does not have a place in the monitoring of feminization in trans women.

Feminizing hormone treatment is an important part of the transition process of trans women. Combined treatment with estradiol and an antiandrogen induces female secondary sex characteristics. Breast development and gynoid fat deposition are among the most anticipated changes (1). However, the magnitude of these changes varies among trans persons and in some cases the changes are limited, which might cause severe disappointment (2). Previous studies have focused on differences in administration route of estradiol, estradiol concentrations, types of antiandrogens, and lifestyle. However, none of these factors seems to explain the difference in amount of feminization (3, 4).

Another possible factor in the amount of feminization is estrone, an estradiol metabolite and estrogen receptor agonist. Both oral and transdermal administered estradiol are partly converted to estrone (5). However, due to the first-pass effect, oral estradiol causes much higher estrone concentrations than transdermal estradiol, both in the peripheral bloodstream and the portal vein (6-9). Estrone is a weak estrogen agonist, with a relative binding affinity of 4% to 10% for estrogen receptor (ER) α and 2% to 3.5% for ERβ, compared with estradiol (10, 11). Therefore, estrone is mainly considered to be a precursor pool for the more potent estradiol (5, 12, 13). On the other hand, a previous in vitro study suggested that a high estrone/estradiol ratio can interfere with the cooperative binding capacity of estradiol and thereby act as an estrogen antagonist (14). This has led to speculations about a possible antagonistic effect of estrone, discussed on internet fora and in the transgender community.

Clinical data about the role of estrone in feminizing hormone treatment is lacking. Therefore, in this study, the main aim is to assess the association between estrone concentrations and feminization in trans women. We will assess feminization by change in body fat and breast development, since these clinical features are known to change under the influence of estradiol as part of feminizing hormone treatment and are among the most anticipated changes (1, 4, 15).

Methods

Subjects and Study Design

We performed a prospective cohort study, which was part of the European Network for the Investigation of Gender Incongruence (ENIGI) (16). All subjects were 18 years or older, diagnosed with gender dysphoria according to the revised fourth or fifth edition of the Diagnostic and Statistical Manual of Mental Disorders, and planned to start with hormone treatment. Exclusion criteria were previous or current use of hormone treatment, diversion from the local treatment protocol, and inability to understand the patient information and informed consent. A more detailed description of the study protocol is described elsewhere (16). The ENIGI study was approved by the local ethics committee of Ghent University Hospital, Belgium, and the participation of Amsterdam UMC in this study was approved by the local ethics committee of the Amsterdam UMC, location VUmc. The ENIGI partners from the other participating centers have given their approval for publishing this manuscript.

Contrary to the measurement of serum estradiol concentrations, measurement of serum estrone concentrations is not part of the routine follow-up of hormone treatment in transwomen (17). With the switch to tandem-MS based measurements of estradiol in the Amsterdam UMC, estrone concentrations could be easily measured in the same run together with estradiol (18). Therefore, in this study, we analyzed all trans women included in the ENIGI cohort in Amsterdam from the introduction of tandem mass spectrometry–based measurements of estradiol and estrone in our laboratory, in July 2017 until the end of our study in December 2019. Since estrone concentrations were only measured in Amsterdam, in this study, only trans women included in the Amsterdam cohort were analyzed.

The association between estrone concentrations and change in fat percentage or breast development was assessed at 12 months. The association between estrone concentrations and administration route or sex hormone binding globulin (SHBG) concentrations was assessed at 3 months, since estrone, estradiol, and SHBG concentrations were stable between 3 and 12 months and no changes in administration form or dosage occurred before the 3 months measurement.

Treatment Protocol

Trans women were treated with estradiol, either oral (estradiol valerate 2 to 6 mg daily) or transdermal (estradiol patches 50 to 100 mcg/24 hours twice weekly), mostly in combination with an antiandrogen, either cyproterone acetate (10 to 50 mg daily), triptorelin (3.75 mg once every 4 weeks) or spironolactone (100 to 150 mg daily). The dosage of either estradiol or antiandrogens could be adjusted to achieve adequate serum estradiol (200-600 pmol/L) and testosterone concentrations (<2.5 nmol/L) during follow-up.

Data Collection

Bioelectrical impedance analysis measurements were performed using a Tanita MC-780 MA (Tanita Europe B.V., Amsterdam, The Netherlands) at baseline, after 3 months, and after 12 months of hormone treatment. This device measures whole body and segmental bioelectrical impedance using 8 electrodes. Participants were measured in underwear and barefoot. The machine output of fat percentages was used. A more detailed description of the bioelectrical impedance analysis protocol is described elsewhere (3).

Breast volume was measured with 3-dimensional imaging, using the VECTRA XT scanner (Canfield Inc.). For this study, measurements at baseline and 12 months were included for analyses. Breast volume was measured by manual placement of landmarks and the assessment software (version 5.7.1.) of the VECTRA XT. A more detailed description of this protocol can be found elsewhere (19).

Hormone Analyses

Blood samples were collected at baseline, after 3 months, and after 12 months of hormone treatment. Estradiol and estrone concentrations were measured in serum using an in-house developed liquid chromatography–tandem mass spectrometry (LC-MS/MS) method in our endocrine laboratory (18). In short, stable, isotopically labeled internal standards (13C3-labeled estradiol; obtained from IsoSciences, King of Prussia, PA, USA, and 13C3-labeled estrone; obtained from Sigma Aldrich, Saint Louis, MO, USA) were added to every specimen (sample, control, calibrator). Estrogens were extracted from 150 μL of sample using a 4:1 (volume/volume) mixture of hexane and ether. The supernatant was subsequently dried under a stream of nitrogen and reconstituted in 75 μL of a 1:1 (volume/volume) mixture of methanol and water. Analysis of the samples was performed using tandem mass spectrometry with a 2-dimensional Xevo TQ-S mass spectrometer (Waters Corp., Milford, MA) with electrospray in negative mode using an injection volume of 50 μL. Separation was achieved on 2 analytical columns, a C4 Acquity UPLC column (Waters Corp), 2.1 × 50 mm, 1.7 μm particle size, and a Phenomenex Kinetex Biphenyl, 2.1 × 50 mm, 1.7 µm column (Torrance, CA, USA), with a methanol:water gradient elution at a flow rate of 0.6 and 0.4 mL/minute (water containing 100 µM NH4F). Total analysis time was 8.2 minutes. The monitored transitions were m/z 271 to 145 and 183 and m/z 274 to 148 and 186 for estradiol and its internal standard and m/z 269 to 143 and 145 and m/z 272 to 146 and 148 for estrone and its internal standard, respectively.

The LC-MS/MS method has a lower limit of quantification of 10 pmol/L for estradiol and 10 pmol/L for estrone. For estradiol, the intra-assay variation was < 5% between 20 and 1700 pmol/L and the interassay variation was 9% at 21 pmol/L and < 7% at 179 and 760 pmol/L. For estrone, the intra-assay variation was < 10% over the whole concentration range and the interassay variation was 10% at 30 pmol/L and < 6% at 220 and 640 pmol/L. All samples were run in duplicate. SHBG was measured using an automated immunoassay (Architect, Abbott Diagnostics, Chicago) with an interassay coefficient of variation of < 8% over the whole concentration range.

Statistical Analysis

Baseline characteristics are presented as mean (± SD), or as median (interquartile range [IQR]) when not normally distributed. The use of tobacco and type of hormone treatment at 3 or 12 months are presented as number and percentage of users. All outcome variables were log-transformed before analyses, back-transformed to ratios, and converted to percentage change. The independent variables estrone and estradiol concentrations are presented as percentage change per 100 pmol/L increase in estrone or estradiol. This was calculated by multiplying the regression coefficients and 95% CIs by 100 before back-transformation.

Linear regressions were performed to analyze the associations between estrone concentrations and change in fat percentage or breast development at 12 months. Mean estrone and estradiol concentrations were calculated by taking the mean of the serum concentrations at 3 and 12 months. If only 1 measurement was performed, the only available concentration was used in the analyses (13 for estrone and 15 for estradiol). Other linear regressions were performed to analyze the associations between estrone concentrations and SHBG concentrations at 3 months. Finally, linear regressions were performed to analyze the effect of either transdermal or oral estradiol on the estrone/estradiol ratio, SHBG concentrations, change in fat percentage, and breast development.

The regression models for change in fat percentage, breast development, and SHBG were stratified by route of estradiol administration at the time of the measurement. Possible confounders were added to the model and considered a confounder if the regression coefficient increased or decreased by 10% or more.

STATA Statistical Software, version 15.1 (Statacorp, College Station, Texas, USA) was used to perform analyses and create graphs.

Results

Between July 2017 and December 2019, 258 trans women were included in the ENIGI study. After applying our inclusion criteria, 212 trans women were eligible for analyses (Fig. 1). The baseline characteristics are presented separately for trans women who were prescribed transdermal estradiol and oral estradiol (Table 1). The median age at baseline was higher for transdermal estradiol users than for oral estradiol users. The mean estradiol and estrone concentrations were similar at 3 and 12 months of hormone treatment. In transdermal estradiol users, the estradiol concentration was 298 pmol/L (95% CI, 240-355) at 3 months and 360 pmol/L (95% CI, 276-443) after 12 months. The estrone concentration was 200 pmol/L (95% CI, 159-240) after 3 months and 187 pmol/L (95% CI, 153-220) after 12 months. In oral estradiol users, the estradiol concentration was 239 pmol/L (95% CI, 215-263) at 3 months and 241 pmol/L (95% CI, 210-271) after 12 months. The estrone concentration was 1443 pmol/L (95% CI, 1274-1612) after 3 months and 1516 pmol/L (95% CI, 1284-1748) after 12 months. This led to a mean estrone/estradiol ratio of 0.96 ± 1.06 in transdermal estradiol users and of 6.04 ± 1.84 in oral estradiol users (see Fig. 2).

Figure 1.

Study flowchart.

Table 1.

Baseline characteristics

	Transdermal (n = 88)	Oral (n = 124)
Age, years	43 [27 to 54]	22 [20 to 26]
BMI, kg/m2	24.3 ± 3.6	22.9 ± 4.7
Current smoker, n (%) yes	19 (22)	25 (21)
Estrone, pmol/L	119 ± 52	106 ± 32
Estradiol, pmol/L	82 ± 28	77 ± 22
Estrone/estradiol ratio	1.5 ± 0.4	1.4 ± 0.3
Testosterone, nmol/L	19 ± 7	18 ± 7
Antiandrogens
CPA, n (%)	72 (83)	113 (92)
Triptorelin, n (%)	-	1 (1)
Spironolactone, n (%)	-	1 (1)
None, n (%)	15 (17)	8 (6)

Baseline characteristics, displayed per prescribed estradiol administration route at baseline. Data are presented as mean ± SD, median [interquartile range], or number (percentage).

Abbreviations: BMI, body mass index, CPA, cyproterone acetate.

Figure 2.

The serum (A) estrone and (B) estradiol concentrations, and (C) estrone/estradiol ratio in trans women after 3 months of hormone treatment. Panel A and B are in logarithmic scale.

The mean SHBG concentrations after 3 months were 37 nmol/L (95% CI, 33-40) in transdermal estradiol users and 44.67 nmol/L (95% CI, 40-49) in oral estradiol users. After adjustment for age, body mass index (BMI) at baseline, and estradiol concentrations, SHBG concentrations at 3 months were 45% (95% CI, 23%-71%) higher in trans women using oral estradiol compared with transdermal estradiol. In oral estradiol users, SHBG concentrations increased by 2.7% (95% CI, 1.9%-3.6%) for every 100 pmol/L increase in estrone concentration (Table 2). After adjustment for estradiol concentrations and BMI at baseline this increase was 1.9% (95% CI, 0.5%-3.2%) for every 100 pmol/L increase in estrone concentration. In transdermal estradiol users, SHBG concentrations at 3 months were not associated with estrone concentrations.

Table 2.

Association between serum estrone or estradiol and SHBG concentrations at 3 months of hormone treatment

	Transdermal estradiol users	Oral estradiol users
SHBG concentration
Estrone (per 100 pmol/L)	2.5% (−4.0 to 9.3)	2.7% (1.9 to 3.6)
Estradiol (per 100 pmol/L)	4.5% (−0.01 to 9.3)	18.5% (11.4 to 26.0)
Estrone adjusted for estradiol	−1.7% (−9.0 to 6.3)	2.3% (0.8 to 3.8)
Estrone adjusted for estradiol and BMI	−1.4% (−8.4 to 6.1)	1.9% (0.5 to 3.2)

Regression analyses, crude analyses for estrone and estradiol and adjusted analyses. Data in percentage change with (95% CI). Estrone and estradiol in 100 pmol/L, SHBG in nmol/L and BMI in kg/m2.

Estrone and Change in Fat Percentage

After 12 months, the increase in fat percentage was 1.2% (IQR, 0.3-4.8) in trans women using transdermal estradiol and 4.6% (IQR, 2.5-5.9) in trans women using oral estradiol. This change in fat percentage over 12 months in trans women was not associated with estrone concentrations in transdermal (+4.4% [95% CI, −4.0 to 13]) for every 100 pmol/L increase in average estrone concentration) and oral estradiol users (−0.7% [95% CI, −1.7 to 0.3]). This did not change after adjustment for estradiol concentrations and other possible confounders, (−9.4% [95% CI, −22 to 4.8] and −0.8% [95% CI, −2.2 to 0.7], for transdermal and oral estradiol users, respectively) (Table 3). Change in fat percentage, adjusted for age and BMI at baseline, did not differ between oral and transdermal estradiol users (1.1% [95% CI, −13 to 18]).

Table 3.

Association between serum estrone concentrations and change in fat percentage and breast development after 12 months of hormone treatment

	Transdermal estradiol users	Oral estradiol users
Change in fat percentage
Estrone (100 pmol/L)	4.4% (−4.0 to 13.4)	−0.7% (−1.7 to 0.33)
Estradiol (100 pmol/L)	3.0% (−0.4 to 6.5)	−3.7% (−11.3 to 4.5)
Estrone adjusted for estradiol	−4.6% (−17.2 to 10.1)	−0.7% (−2.2 to 0.8)
Estrone adjusted for confounders	−9.4% (−21.7 to 4.8)	−0.8% (−2.2 to 0.7)
Breast development
Estrone (100 pmol/L)	14.5% (−49.7 to 155.6)	10.9% (−14.0 to 43.0)
Estradiol (100 pmol/L)	14.1% (−20.1 to 62.9)	−16.7% (−85.1 to 366.4)
Estrone adjusted for estradiol	8.5% (−54.8 to 160.4)	62.5% (5.5 to 150.4)
Estrone adjusted confounders	−3.7% (−79.7 to 356.2)	49.3% (−7.3 to 140.4)

Regression analyses, crude analyses for estrone and estradiol and adjusted analyses. Data in percentage change with (95% CI). Estrone and estradiol in 100 pmol/L, age in years, and BMI in kg/m2. Change in fat percentage and breast development were measured over 12 months of hormone treatment, estrone and estradiol concentrations are average concentrations over 3 and 12 months. Confounders change in fat percentage: estradiol concentration, testosterone concentration, age, BMI at baseline. Confounders breast development: estradiol concentration, age, BMI at baseline.

Estrone and Breast Development

After 12 months, the breast development in trans women was 69 mL (IQR, 58-134) in trans women using transdermal estradiol and 62 mL (IQR, 32-95) in trans women using oral estradiol. Breast development in trans women was not associated with estrone concentrations in trans women using transdermal (+14% [95% CI, −49 to 156]) for every 100 pmol/L increase in average estrone concentration) or oral estradiol (+11% [95% CI, −14 to 43]). Adjustment for estradiol, age, and BMI did not change this (Table 3). Breast development, adjusted for age, BMI at baseline, and estradiol and testosterone concentrations, did not differ between oral and transdermal estradiol users (−33% [95% CI, −85 to 209]).

Discussion

The main result of this study is that we found no relationship between estrone concentrations and change in body fat or breast development after 12 months of hormone treatment. Further, we found no association between administration route and these feminization markers, despite the fact that oral estradiol was clearly associated with much higher estrone levels compared with transdermal estradiol.

Effect of Administration Route on Estrone

In line with previous research in trans and postmenopausal women, we found a higher estrone/estradiol ratio in trans women using oral estradiol compared with transdermal estradiol (6-8, 20, 21). Orally administered estradiol is already partly metabolized into estrone in the intestines. The ileum produces more estrone after incubation with estradiol than any other tissue except for the placenta (22). In contrast, transdermal administered estradiol bypasses the gastrointestinal mucosa and the hepatic system and is slowly metabolized into estrone (7).

Sex Hormone Binding Globulin

In addition to the clinical outcomes of breast development and body fat, we assessed the relation between estrone and SHBG, a binding globulin that is produced in the liver and found to be the most sensitive estrogenic marker in postmenopausal women (23, 24). We found a positive association between estrone and SHBG concentrations in oral estradiol users, in line with a previous in vivo study in postmenopausal women (25). Furthermore, the positive association between estradiol and SHBG concentrations is also in line with previous in vivo studies in postmenopausal women (20, 25-28). This finding is further supported by previous in vitro studies confirming that estrogens increase the production of SHBG in hepatocytes (29, 30). However, no association was found between estrone and SHBG concentrations in transdermal estradiol users. This could suggest that both the increase in estrone and SHBG concentrations are a result of the oral administration route. Orally administered estradiol leads to high estradiol concentrations in the portal vein and consequently the hepatic microvasculature which stimulates both SHBG production and conversion of estradiol to estrone in the liver (26, 31, 32). In addition, around 90% of estrone excreted in the bile is reabsorbed in the intestines, stimulating protein synthesis in the liver even more (33). Finally, cyproterone acetate is known to decrease the secretion of SHBG in vitro (30); however, in our study, adjustment for cyproterone acetate did not change the association between estrone and SHBG concentrations and our SHBG concentrations match those of postmenopausal women using only estradiol and not cyproterone acetate in other studies (25, 26).

Change in Body Fat and Breast Development

We found no association between estrone concentrations and change in fat percentage or breast development after 12 months of hormone treatment. Estradiol concentrations were also not associated with these outcomes. This is in line with previous studies in trans women, which found no association between estradiol concentrations or administration route and change in fat percentage or breast development (3, 19, 34). Our results are in contrast with other studies suggesting that high estrone levels can act as an estrogen antagonist (14, 35-37). An in vitro study found that a high excess of estrone interferes with the positive cooperative binding of estradiol and its receptor, by inducing a change in the estradiol-receptor complex that favors the nonactive confirmation of this receptor (14). This nonactive conformation results in lower affinity for estradiol and DNA (38). This in vitro study is reinforced by 2 clinical studies, showing that lowering the upper estrone limit strengthened the association between estradiol and clinical outcome, increase in mammographic density (36), and collagen turnover in urogenital tissue (37). In our study, lowering the upper limit of estrone did not influence the association between estradiol concentrations and our outcomes (data not shown). This study demonstrates that higher estrone concentrations or higher estrone/estradiol ratios are not associated with antagonistic effect on feminization in trans women. Therefore, estrone has no role in monitoring feminization in trans women using hormone treatment.

Estrone and Long-Term Effects

In this study, we found no association between estrone concentrations and feminization markers in trans women. However, in previous studies in postmenopausal women, estrone concentrations have been associated with an increased risk of diseases such as thrombosis and breast cancer (9, 39, 40). Bagot et al not only found significantly increased thrombin generation in women using oral estradiol, but furthermore, that estrone concentrations in these women were correlated with peak thrombin generation (9). Other previous studies in postmenopausal women found that high estrone concentrations were associated with a higher risk of ER-positive breast cancer (39, 40). Therefore, estrone concentrations might be associated with increased risks in trans women and future research is needed to assess this.

Strengths and Limitations

As far as we know, this is the first study to assess the role of estrone concentrations in feminization of the body in trans women. Due to collaborations within the ENIGI cohort, we were able to assess 2 markers of feminization, making our conclusions stronger. This study has some limitations as well. Due to our local protocol, almost all women were treated with cyproterone acetate, which makes it impossible to exclude effects of this antiandrogen on our primary outcome parameters. In addition, in accordance with our local protocol, we advised trans women above the age of 40 years to choose transdermal estradiol, while younger transwomen were able to choose either transdermal or oral estradiol. This has led to a significant difference in age between users of both administration forms. Although we have adjusted our models for age, this might still have an effect on the outcomes.

Conclusion

Estrone concentrations are not associated with breast development or change in body fat in trans women receiving either oral or transdermal estradiol. Moreover, no difference in clinical outcomes was observed between oral and transdermal administration of estradiol, despite the fact that oral estradiol was clearly associated with much higher estrone levels compared with transdermal estradiol. This suggests that estrone has neither an agonistic nor an antagonistic effect on feminization in trans women. Therefore, this study shows that estrone has no role in the monitoring of feminization in trans women.