Time Course of Body Composition Changes in Transgender Adolescents During Puberty Suppression and Sex Hormone Treatment

Abstract

Context

Transgender adolescents can undergo puberty suppression (PS) and subsequent gender-affirming hormone therapy (GAHT) but little information is available on the expected rate of physical changes.

Objective

To investigate the time course of body composition changes during PS and GAHT.

Methods

In this study, retrospective data of 380 trans boys and 168 trans girls treated with PS prior to GAHT from a gender identity clinic were included. Total lean and fat mass Z-scores using birth-assigned sex as reference were determined using dual-energy X-ray absorptiometry.

Results

In trans boys, lean mass Z-scores decreased (−0.32, 95% CI −0.41; −0.23) and fat mass Z-scores increased (0.31, 95% CI 0.21; 0.41) in the first year of PS and remained stable thereafter. Lean mass Z-scores increased (0.92, 95% CI 0.81; 1.04) and fat mass Z-scores decreased (−0.43, 95% CI −0.57; −0.29) only during the first year of testosterone,. In trans girls, both lean and fat mass Z-scores gradually changed over 3 years of PS (respectively −1.13, 95% CI −1.29; −0.98 and 1.06, 95% CI 0.90; 1.23). In the first year of GAHT, lean mass Z-scores decreased (−0.19, 95% CI −0.36; −0.03) while fat mass Z-scores remained unchanged after 3 years (−0.02, 95% CI −0.20; 0.16).

Conclusion

Compared with peers, trans girls experienced ongoing lean mass decrease and fat mass increase during 3 years of PS while in trans boys smaller changes were observed that stabilized after 1 year. A large increase in lean mass Z-scores occurred only during the first year of testosterone treatment. In trans girls, body composition changed only slightly during GAHT. This information can improve counseling about treatment effects.

Adolescents who experience gender dysphoria can start with medical treatment when eligibility criteria are met as outlined in the Endocrine Society guideline and Standards of Care 8 (1, 2). This treatment is typically initiated with puberty suppression (PS), consisting of treatment with gonadotropin-releasing hormone agonists (GnRHas) to suppress the production of endogenous sex hormones (3, 4). When gender dysphoria persists, gender-affirming hormone treatment (GAHT) with testosterone or estradiol can be offered from the age of approximately 16 years to induce secondary sex characteristics consistent with the gender identity (1).

One of these sex characteristics is body composition, which is well known to notably differ between men and women in the general population. Men tend to have a greater proportion of lean mass, while women have more fat mass. These differences are largely attributed to sex hormones as they become apparent during puberty (5, 6). In addition, various studies have suggested that changes in sex hormones can lead to alterations in body composition. For instance, a study conducted on adult men with acquired hypogonadism revealed a higher total body fat percentage than eugonadal men. Testosterone administration resulted in an increase in lean mass and a decrease in body fat percentage (7).

Studies of hormone treatment in the transgender population showed similar results (8, 9). In trans men, testosterone resulted in an increase in lean body mass and a decrease in fat mass (9). Trans women experienced an increase in total body fat and fat accumulation in the hip region during estradiol treatment (8).

In transgender adolescents undergoing medical treatment, body composition is not only affected by GAHT but also by PS. Several studies have been carried out on body composition in transgender youth, revealing an increase in fat percentage among both transgender boys and girls during PS (10-12). Klaver et al also reported changes in body composition between initiation of GAHT and age 22 and found results similar to what has been observed in the adult population (12). One might hypothesize that initiation of PS in early puberty prior to GAHT will result in a body composition that is more consistent with the experienced gender compared with those who started PS later in puberty and have had more exposure to their endogenous sex hormones. However, the study by Klaver et al did not confirm this hypothesis. Furthermore, it could be suggested that alterations in body composition might also be influenced by the concentrations of serum estradiol and/or testosterone given the recently reported dose-dependent effect of estrogen on bone mineral density (13).

Most of the studies on body composition in transgender youth report absolute changes rather than comparing changes with those in cisgender peers, which makes it difficult to put the changes into perspective as growth, regardless of treatment, also causes alterations in body composition. Furthermore, most studies describe the total change in body composition after several years of treatment, but up to this point there is a lack of data regarding the pace at which these changes occur and the duration until these changes stabilize in transgender youth. There are some studies that describe annual changes in body composition in transgender adults (8, 14), but the differences in treatment such as the use of PS prior to GAHT and the incremental dosage of GAHT in adolescents might lead to different findings. A study examining the time course of changes during PS and GAHT can optimize counseling and management of expectations regarding body composition in transgender youth.

In this study, we aim to describe the time course of changes in body composition during the first 3 years of PS and subsequently GAHT in a large cohort of transgender adolescents. In addition, differences in these changes between those who started PS in early and late puberty are described.

Materials and Methods

Study Design and Population

The study is part of the Amsterdam Cohort of Gender dysphoria (ACOG) study (15). All participants were seen at the Center of Expertise on Gender Dysphoria at the Amsterdam UMC in the Netherlands between 1972 and 2018. For the current analyses, people were included when they were diagnosed with gender dysphoria according to DSM IV or 5 criteria, aged <18 years old when PS was initiated, and at least 1 whole body dual-energy X-ray absorptiometry (DXA) scan was performed during the first 3 years of treatment with either PS or GAHT. Participants were excluded if they received treatment with a progestin or cyproterone acetate.

Treatment

Participants were treated according to the clinical practice guideline from the Endocrine Society (1). Puberty suppression was achieved through suppression of gonadotropins with intramuscular or subcutaneous injections of GnRHa, triptorelin 3.75 mg 1 × per 4 weeks, or 11.25 mg 1 × per 10 to 12 weeks. Treatment with GnRHa in both transgender boys and girls was initiated after puberty had started (eg, establishment of at least Tanner breast or genital stage 2) (3, 4).

When gender dysphoria persisted, puberty was induced from the age of 15 to 16 years. In trans boys, puberty induction was obtained through 2-weekly intramuscular injections of testosterone esters. In trans girls, puberty induction was obtained through daily oral supplementation of estradiol or, in a few very tall individuals who wished growth reductive therapy, with ethinylestradiol (16). Dosage was increased in both trans boys and trans girls every 6 months until adult dosage was attained. Treatment with GnRHa was continued throughout treatment with GAHT until a gonadectomy was performed. Visits were scheduled on a regular base, every 6 to 12 months.

Anthropometric Measurements and Body Composition

At every visit, standing height was measured using a wall-mounted stadiometer. Body weight was measured in indoor clothing without shoes using a digital body weight scale. Height SD score (SDS) was calculated for sex assigned at birth according to Dutch reference data (17) and body mass index (BMI) SDS was calculated according to reference data from Cole et al (18).

DXA scans were routinely performed before start of PS, before start of GAHT, and at 1- to 2-year intervals during treatment. These DXA scans were used to measure whole-body fat mass, lean mass, and total mass. The amount of android and gynoid fat was calculated determined based on predefined regions using the software provided by Hologic. Hologic Delphi was used from December 1998 up to January 2011, and Hologic Discovery A from February 2011 up to December 2018. Z-scores of the sex assigned at birth were calculated using reference data from National Health and Nutrition Examination Surveys (NHANES) (19).

Biochemical Determinations

To determine serum testosterone concentrations, a competitive immunoassay (Architect, Abbott, Abbott Park, IL, USA) was used (interassay coefficient of variation [CV] 6-10%, lower limit of quantitation [LOQ] 0.1 nmol/L) from January 2013. Serum testosterone concentrations from before this date were converted to Architect as described by Wiepjes et al (20). From October 2018 liquid chromatography–tandem mass spectrometry (LC-MS/MS) was used to determine serum testosterone concentrations with an interassay CV of 4% to 9% and a LOQ of 0.1 pmol/L. The use of a conversion formula was not necessary according to the local laboratory.

Serum estradiol concentrations were measured using LC-MS/MS with an interassay CV of 7% and a LOQ of 20 pmol/L from July 2014. Serum estradiol concentrations from before this date were converted to LC-MS/MS values as described by Wiepjes et al (20).

Statistical Analysis

All data were analyzed using Stata software (StataCorp. 2015. Stata Statistical Software: Release 14; StataCorp LP, College Station, TX). Continuous data were expressed as mean ± SD when normally distributed, or as median (interquartile range) when not normally distributed.

All analyses were performed separately for trans boys and trans girls. Body composition throughout treatment was examined using linear mixed model regression, with measurements clustered within participants. DXA scans were regarded as baseline if within 100 days from start of PS or GAHT. Scans acquired 101 to 180 days after start of treatment were not used for analysis (this amounted to scans of 101 trans boys and 33 trans girls). Dates of follow-up scans were rounded to the nearest whole year treatment duration. All scans acquired after >3 years of PS or GAHT were not used for analysis. Longitudinal analyses were performed and repeated for subgroups based on Tanner stage at start of PS with Tanner stage B or G 2-3 considered “early pubertal” and Tanner stage B or G 4-5 considered “late pubertal.” For subanalyses based on testosterone concentrations in transgender boys, average serum testosterone concentrations during GAHT were calculated for each individual and 3 groups were used: group 1, <10 nmol/L (n = 45); group 2, ≥10 nmol/L and <30 nmol/L (n = 113); and group 3, >30 nmol/L (n = 54). In trans girls the following groups were used for subanalyses with average serum estradiol concentrations during GAHT: group 1, <200 pmol/L (n = 61); group 2, ≥200 pmol/L and <400 pmol/L (n = 30); and group 3, >400 pmol/L (n = 19).

Ethics

The data collection protocol for the ACOG data set underwent evaluation by the local medical ethics committee, which concluded that the Medical Research Involving Human Subjects Act (WMO) did not apply to this particular data collection and the need for informed consent was waived.

Results

The ACOG data set contained 8831 individuals, of whom 674 were ≤18 years at start of PS and had at least 1 DXA scan. A total of 126 individuals were excluded, 37 because of the use of progestins or cyproterone acetate, 79 because of no available DXA during the period of interest (in 73 cases scans had been made 101-180 days after the start of treatment), and 10 because of missing Tanner stage at start of PS.

Baseline Characteristics

A total of 548 individuals (99% white) were included, of whom 380 trans boys and 168 trans girls. Baseline characteristics are shown in Table 1. For 463 individuals, DXA scans were available during PS and 346 individuals had available DXA scans during GAHT.

Table 1.

Characteristics of the study cohort

	Transgender boys n = 380	Transgender boys Early start n = 71	Transgender boys Late start n = 309	Transgender girls n = 168	Transgender girls Early start n = 92	Transgender girls Late start n = 76
Start GnRHa, n	309	69	240	154	88	66
Age, years	14.4 ± 2.1	12.1 ± 0.8	15.1 ± 1.9	14.1 ± 1.8	13.1 ± 0.9	15.4 ± 1.7
Weight, kg	53 (48 to 63)	44 (40 to 48)	56 (50 to 66)	54 (43 to 66)	44 (42 to 56)	62 (54 to 73)
Height, cm	162.5 ± 8.6	155.8 ± 6.5	164.4 ± 8.1	167.1 ± 9.9	161.1 ± 6.7	174.5 ± 7.9
Height SDS	−0.14 ± 1.15	−0.09 ± 1.06	−0.16 ± 1.18	−0.11 ± 0.92	−0.20 ± 0.69	−0.00 ± 1.15
BMI, kg/m2	20.3 (18.4 to 22.6)	17.6 (16.5 to 19.6)	20.9 (19.1 to 23.1)	19.1 (16.9 to 21.8)	17.8 (16.0 to 20.6)	20.4 (18.4 to 22.6)
BMI SDS	0.57 ± 1.08	0.16 ± 1.03	0.69 ± 1.07	0.41 ± 1.33	0.25 ± 1.41	0.61 ± 1.20
Menarche, n (%)a	170 (65)	1 (2)	169 (84)
Testis volume, mL				12 (8 to 20)	8 (6 to 10)	20 (20 to 25)
Start GAHT, n	235	33	202	111	55	56
Age, years	16.7 ± 0.9	15.6 ± 0.8	16.8 ± 0.9	16.2 ± 1.0	15.7 ± 0.6	16.8 ± 1.1
Duration PS monotherapy, years	0.6 (0.5 to 2.5)	3.2 (2.9 to 3.7)	0.6 (0.5 to 1.3)	2.3 (1.0 to 2.8)	2.8 (2.1 to 3.1)	1.0 (0.6 to 2.3)
Weight, kg	60 (54 to 68)	55 (51 to 63)	61 (54 to 68)	60 (53 to 70)	55 (50 to 65)	64 (57 to 75)
Height, cm	166.8 ± 6.6	166.7 ± 7.5	166.9 ± 6.4	174.6 ± 7.0	172.4 ± 6.2	176.9 ± 7.2
Height SDS	−0.30 ± 0.99	−0.17 ± 1.08	−0.32 ± 0.98	−0.58 ± 0.93	−0.72 ± 0.86	−0.44 ± 1.00
BMI, kg/m2	21.7 (19.6 to 24.2)	20.2 (18.5 to 22.3)	21.9 (19.6 to 24.5)	20.0 (18.1 to 22.6)	18.5 (17.2 to 21.5)	20.8 (19.4 to 23.3)
BMI SDS	0.58 ± 1.22	0.19 ± 1.14	0.65 ± 1.23	0.14 ± 1.32	−0.10 ± 1.37	0.39 ± 1.24

Data are presented as mean ± SD or median (IQR).

Abbreviations: GnRHa, gonadotropin-releasing hormone agonists; GAHT, gender affirming hormone treatment; Tanner B, Tanner breast stage, Tanner G, Tanner genital stage.

^a Percentage shown is based on the total number of individuals with available data on menarche.

Puberty Suppression

Transgender boys

Data from 309 transgender boys were used for longitudinal analysis of body composition during PS. Over the course of 3 years, BMI increased by 2.0 kg/m² (95% CI 1.6 to 2.5) to 22.8 kg/m² (SE 0.3), while BMI SDS using female references remained stable during 3 years of PS (0.05, 95% CI −0.15 to 0.20). Mean lean mass Z-score at baseline was 0.60 (SE 0.05) which declined by 0.32 (95% CI −0.41 to −0.23) in the first year and remained stable in the following 2 years (3 years PS vs baseline −0.40, 95% CI −0.52 to −0.28). Z-scores of total fat mass increased during PS, with the greatest increase during the first year of PS by 0.31 (95% CI 0.21 to 0.41) and a total increase of 0.47 (95% CI 0.34 to 0.60) in the first 3 years of PS. Android/gynoid fat ratio increased during the first 3 years of PS by 0.07 (95% CI 0.04 to 0.09) to 0.79 (SE 0.01) with the greatest increase in the first year of 0.04 (95% CI 0.02 to 0.06).

As shown in Fig. 1A and 1B, trends differed between the early and late pubertal starters. Z-scores of lean and fat mass were significantly higher at baseline in the late pubertal than in early pubertal trans boys. However, after 3 years of PS lean mass Z-scores were comparable due to a larger decrease in the late pubertal group than in the early pubertal group (difference –0.49, 95% CI −0.75 to −0.22). The increase in total fat mass Z-score and android/gynoid fat ratio (Fig. 2A) was comparable between the early and late pubertal trans boys.

Figure 1.

Body composition during puberty suppression. Absolute values and Z-scores of total lean mass (A and C) and total fat mass (B and D) during PS by transition type and subdivided by early pubertal and late pubertal individuals. Abbreviations: early pubertal, Tanner genital/breast stage 2 or 3 at start of PS; late pubertal, Tanner genital/breast stage 4 or 5 at start of PS; PS, puberty suppression.

Figure 2.

Android/gynoid fat ratio during puberty suppression and gender-affirming hormone treatment. Android/gynoid fat ratio during PS (A and C) and GAHT (B and D) by transition type and subdivided by early pubertal and late pubertal individuals. Abbreviations: early pubertal, Tanner genital/breast stage 2 or 3 at start PS, GAHT, gender-affirming hormone therapy; late pubertal, Tanner genital/breast stage 4 or 5 at start of PS; PS, puberty suppression.

Trans girls

Data from 154 trans girls contributed to longitudinal analysis of body composition during PS. During 3 years of PS, BMI increased by 1.8 kg/m² (95% CI 1.3 to 2.4) to 21.1 kg/m² (SE 0.4), while BMI SDS using male references did not change (−0.08, 95% CI −0.29 to 0.14). As shown in Fig. 1C, the decrease in lean mass Z-scores persisted during the first 3 years of PS resulting in a total decrease of 1.13 (95% CI −1.29 to −0.98). The increase in fat mass Z-scores also continued throughout the first 3 years of PS with a total increase of 1.06 (95% CI 0.90 to 1.23). Android/gynoid fat ratio remained stable during the first 3 years of PS when those who started in early and late puberty were analyzed as 1 group.

In trans girls, a difference in body composition changes between early and late pubertal starters was seen as shown in Fig. 1C and 1D. During 3 years of PS, the decrease in lean mass Z-scores was greater in the late pubertal group than in the early pubertal group by −0.63 (95% CI −0.91 to −0.34). The increase in fat mass Z-scores in the late pubertal group compared with the early pubertal group was significantly larger in the first year of PS (0.55, 95% CI 0.31 to 0.79), after which the increase between both groups was comparable. As shown in Fig. 2C, android/gynoid fat ratio decreased in the first 3 years of PS in late pubertal trans girls by 0.12 (95% CI −0.20 to −0.04) while it remained stable in early pubertal girls (change in 3 years PS of 0.03 95% CI −0.01 to 0.08).

Hormone Therapy

Transgender boys

Data of 235 trans boys were used for longitudinal changes in body composition during treatment with testosterone. BMI increased by 1.0 kg/m² (95% CI 0.5 to 1.5) during the first year, and remained stable afterwards. BMI SDS using female references increased after 1 year of GAHT by 0.27 (95% CI 0.10 to 0.45) but was comparable with values at the start of GAHT and again after 3 years of testosterone. Lean body mass Z-scores increased by 0.93 (95% CI 0.81 to 1.04) during the first year of treatment and remained stable thereafter (3 years GAHT vs baseline 0.79, 95% CI 0.66 to 0.93). Fat mass Z-scores mainly decreased in the first year of testosterone therapy by −0.43 (95% CI −0.57 to −0.29) with a total decrease of 0.58 (95% CI −0.74 to −0.42) after 3 years of testosterone. Android/gynoid fat ratio slightly increased in the 2 years of GAHT by 0.08 (95% CI 0.04 to 0.12) and remained stable thereafter. As shown in Fig. 3B no differences between early and late pubertal trans boys were observed in the change of android/gynoid fat ratio. No differences in course over time of either lean mass or fat mass Z-scores were observed when comparing groups by pubertal stage at start of PS (Fig. 3A and 3B) and by serum testosterone concentrations. After 3 years of treatment the fat percentage was comparable with cisgender men (24.3% ± 5.5 vs 23.4% ± 8.4) using reference data from NHANES at age 20 years (19).

Figure 3.

Body composition during gender affirming hormone treatment. Absolute values and Z-scores of total lean mass (A and C) and total fat mass (B and D) during GAHT by transition type and subdivided by early pubertal and late pubertal individuals. Abbreviations: early pubertal, Tanner genital/breast stage 2 or 3 at start of PS; GAHT, gender-affirming hormone therapy; late pubertal, Tanner genital/breast stage 4 or 5 at start of PS; PS, puberty suppression.

Trans girls

Analyses of 111 trans girls showed an increase in BMI in the second and third year after the start GAHT of 1.7 kg/m² (95% CI 1.0 to 2.3) to a BMI of 22.3 (SE 0.4). However, BMI SDS using male references slightly decreased after 1 year of GAHT by 0.18 (95% CI −0.38 to 0.03) and remained stable in the following years. Lean mass Z-scores only decreased during the first year of estrogen therapy by 0.17 (95% CI −0.31 to −0.03) and remained stable thereafter (3 years of GAHT vs baseline −0.30, 95% CI −0.45 to −0.15). After 3 years of estradiol, fat mass Z-scores were comparable with the start of GAHT (−0.02, 95% CI −0.20 to 0.16). As shown in Fig. 2D, and Fig. 3C and 3D, no differences in course over time of either lean mass and fat mass Z-scores and android/gynoid fat ratio were observed between early and late pubertal starters. When comparing groups by serum estradiol concentrations, no differences in change over time of lean mass and fat mass Z-scores were found either. The fat percentage after 3 years of treatment was still lower than in cisgender women (29.4% ± 8.4 vs 35.1% ± 7.2) using reference data from NHANES at age 20 years (19).

Discussion

This study was the first to analyze the annual effects of PS and subsequent GAHT on body composition Z-scores in a large cohort of transgender adolescents. In transgender boys, changes during PS stabilized after 1 year, and the most substantial effects of testosterone were also seen within the first year. In transgender girls, the transition to a female pattern of body composition largely occurred during PS.

Puberty Suppression

During PS, a decrease in lean mass and an increase in fat mass Z-scores was observed in both transgender boys and girls, which has also been observed in cisgender adolescents with central precocious puberty during treatment with GnRHa (21).

In transgender boys, the decrease in lean mass Z-scores was mild and stabilized after the first year of PS. A study by Nokoff et al in transgender individuals who received GnRHa described no significant changes in lean mass compared with age- and BMI-matched cisgender girls (11). Another study by Navabi et al also reported no significant changes in lean mass Z-score compared with baseline in transgender boys on GnRHa (22). As estradiol, though to a lesser extent than testosterone, is also known to positively influence muscle mass (23), a decrease in lean mass Z-scores after GnRHa initiation would not be unexpected. The fact that this decrease was observed in our cohort but not in the previously mentioned studies might be explained in particular by the late pubertal group. In this group, lean mass was above average at baseline whereas lean mass Z-score was −1 in the study by Navabi et al (22). Therefore, the effects of a drop in estradiol concentrations by GnRHa will be more pronounced in this late pubertal group, as was also seen when comparing this group with the early starters in this study. The higher lean mass Z-scores may be an effect of life style. The late pubertal starters from this cohort might have been more engaged in physical activity than cisgender girls, either because of an increased interest in sports and outdoor activities, or in a conscious attempt to build muscle to have a more masculine appearance. However, earlier studies described less physical activity in American transgender youth than cisgender youth, although this was most evident in transgender girls (24). Unfortunately, physical activity was not reported in this cohort. The increased lean mass Z-score could not be explained by taller stature, as height SDS in the late pubertal group was just below average (−0.16).

The increase in fat mass Z-score by 0.5 in transgender boys, with the greatest increase in the first year of PS, was similar to findings by Boot et al in 23 girls and 2 boys with central precocious puberty (21). They described an increase in fat mass Z-score in the first year of treatment with GnRHa from 0.38 to 0.98 but no further increase in the second year of treatment.

In contrast to transgender boys, lean mass Z-scores significantly decreased throughout 3 years of PS in transgender girls. A significant decrease in lean mass Z-scores in transgender girls was also reported by Navabi et al and the study by Nokoff et al also reported significant differences in lean mass percentage in transgender girls during GnRHa compared with cisgender boys (11, 22). The greater impact of GnRHa on lean mass in transgender girls may be explained by the influence of testosterone on lipolysis and muscle fiber growth (25). While cisgender boys experience an increase in muscle mass during puberty due to testosterone exposure, these changes are smaller in early pubertal and absent in late pubertal transgender girls receiving GnRHa, resulting in a substantial difference in Z-scores. This also explains the greater decrease in lean mass Z-score in the late pubertal trans boys compared with the early pubertal starters as the former group has already accumulated a greater amount of muscle mass due to a longer period of testosterone exposure.

Fat mass Z-scores increased by 1.06 throughout the first 3 years of PS in transgender girls. The increase was similar to the 1.05 increase reported by Navabi et al in transgender girls (22). The time course of these changes cannot be compared with previous findings in transgender girls, as data on annual changes in body composition during GnRHa are, to date, only available in cisgender children with central precocious puberty, who are mainly girls (21). However, during physiological puberty an increase in total body fat occurs from 8 to 14 years followed by a small decrease in the following 2 years after which a plateau is reached (26). PS prevented this decrease and plateau phase leading to higher fat mass Z-scores in transgender girls.

Hormone Therapy

In trans boys, the greatest changes in body composition were observed in the first year of testosterone treatment. This is somewhat surprising, since current guidelines indicate that maximum effect of testosterone on muscle mass and strength are achieved after 2 to 5 years (1). However, the limited data that are available on longitudinal changes in body composition in transgender boys or men are in line with the current findings. One study in 45 transgender men treated with testosterone undecanoate reported an increase in lean mass of 1.7 kg in the first year of treatment which remained stable in the following year (14). Furthermore, there is 1 study in hypogonadal men undergoing testosterone treatment, indicating a decrease in body fat percentage in the initial 6 months that stabilized in the following 12 months of treatment without significant changes in body weight (7). A similar trend was observed in a randomized controlled trial involving healthy cisgender men aged over 60 years receiving testosterone treatment (27). These findings confirm that the impact of testosterone on lean mass is achieved in the short-term. However, muscle strength might still increase in the following years as several studies showed that muscle strength can increase independent of muscle mass up until the age of 30 years in cismen (28, 29). No associations between changes in body composition and serum testosterone levels were observed. However a study by van Velzen et al reported that a lack of change in body composition in transgender men was associated with serum testosterone levels (30). This disparity might be explained by use of intramuscular testosterone esters in our population and the untimed blood sampling resulting in highly variable serum testosterone levels. In contrast, in the study by van Velzen a significant proportion used testosterone gel (46%), resulting in more stable serum testosterone levels.

GAHT did not result in large alterations in body composition in transgender girls. Fat mass Z-scores remained stable and lean mass Z-scores only decreased slightly in the first year of treatment. This is not in line with studies in transgender women that describe significant increase in fat mass and a decrease in lean mass (8, 31). We hypothesize that this difference is due to the fact that the transgender girls from this cohort all received GnRHa prior to GAHT, which has similar effects on body composition as estradiol, so that most of the expected effects of estradiol had already occurred during PS. The lack of a relation between serum estradiol levels and body composition changes could be explained by these similar effects of PS and estrogen on body composition. In contrast, the recently described dose-dependent effect of estrogen on bone mineral density may be due to the opposing impacts of PS and estrogen on bone mineral accrual (13). Absolute lean mass increased slightly with estradiol, which may be due to remaining growth in these adolescents, but Z-scores still declined as this increase was smaller compared with cisgender boys in whom this increase is continuous until the age of 18 (26). The fat mass distribution in trans girls also remained stable although in trans women, an increase in fat mass in the leg and gynoid region has been reported (8). This might be explained by the fact that PS already promotes fat mass distribution to these specific regions.

In contrast to transgender boys that had a fat percentage similar to age matched cisgender men after 3 years of testosterone treatment, the fat percentage in transgender girls of 29% after 3 years of estradiol treatment was still low compared with cisgender women. This is in line with earlier studies in trans women that describe a lean body mass that is higher than in cis women, both at short-term and long-term follow-up (32). An explanation for this might be that prepubertal and/or early pubertal testosterone exposure causes an irreversible change in body composition, as a study in XY women with complete androgen insensitivity syndrome and a study in androgen receptor–deficient XY mice report a body composition similar to females, suggesting that in the absence of androgen action estradiol can completely feminize body composition (33, 34).

Android/gynoid fat ratio decreased in trans girls and increased in trans boys, aligning with expectations, since men typically have a higher android/gynoid fat ratio than women. The higher prevalence of cardiovascular events in cisgender men might be partly attributed to a higher android/gynoid fat ratio as various studies have linked this ratio to increased cardiovascular risk by establishing its close association with insulin resistance and dyslipidemia (35, 36). Therefore, the increase in the android/gynoid fat ratio in trans boys, resulting in a ratio similar to those in cis boys, may contribute to an elevated cardiovascular risk. However, previous research has been inconclusive about an elevated risk of cardiovascular events in transgender men (37, 38). Strengths of this study are the large cohort of transgender adolescents and the number of DXA scans, which made it possible to describe the time course of the changes in body composition. By using Z-scores, changes in body composition could be compared with age- and sex-matched peers. However, one might argue that the use of sex assigned at birth as reference value is arbitrary as lifestyle factors such as physical exercise and eating behavior might differ. Unfortunately, due to the retrospective character of this study, data on these lifestyle factors were missing. The cohort was relatively homogenous as the majority consisted of white adolescents. However, the findings may not apply to populations of other ethnicity.

To conclude, in both transgender boys and transgender girls treatment with GnRHa resulted in a decrease in lean mass Z-scores and an increase in fat mass Z-scores. In transgender boys, these changes were small and inverse effects were seen after initiation of testosterone, which occurred primarily in the first year of treatment. In transgender girls the changes occurred throughout all 3 years of PS and stabilized during the first year of estradiol. These findings are useful when counseling transgender adolescents on expected changes in body composition during their transition.

Abbreviations

ACOG

Amsterdam Cohort of Gender dysphoria

BMI

body mass index

CV

coefficient of variation

DXA

dual-energy X-ray absorptiometry

GAHT

gender-affirming hormone treatment

GnRHa

gonadotropin-releasing hormone agonist

LC-MS/MS

liquid chromatography–tandem mass spectrometry

LOQ

limit of quantitation

PS

puberty suppression

Funding

None.

Disclosures

No disclosures.

Data Availability

Restrictions apply to the availability of some or all data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.