Shaping the Skeleton: Impact of GnRH Analogue and Sex Hormone Therapy on Skeletal Dimensions in Transgender Individuals

Abstract

Context

Skeletal dimensions vary between sexes. Men typically have broader shoulders and women a wider pelvis. Whether gender-affirming hormone therapy (GAHT) with or without prior puberty suppression (PS) alters these dimensions in transgender individuals remains unclear.

Objective

To investigate impact of PS and GAHT on skeletal dimensions.

Methods

This retrospective cross-sectional study, conducted at a gender identity clinic, included transgender individuals assigned male at birth (AMAB) and assigned female at birth (AFAB) who underwent dual-energy x-ray absorptiometry (DXA) scanning between ages 18 and 28 years and who were divided into 4 groups: Early PS (Tanner G/B2-3) + GAHT, Late PS (Tanner G/B4-5) + GAHT, GAHT-only, and untreated. Shoulder and pelvis dimensions measured by DXA were compared between groups, with adjustment for height.

Results

A total of 121 individuals AMAB and 122 AFAB were included. Only individuals AMAB who underwent early PS had smaller shoulders compared to untreated individuals AMAB (−1.3 cm; 95% CI −2.1; −0.5). In individuals AMAB from both the Early and Late PS groups, pelvic inlet, pubic symphysis width, and interischial distance were greater compared to untreated individuals AMAB, resulting in dimensions comparable to untreated individuals AFAB. Only in Early PS AFAB was pelvic inlet width smaller compared to untreated individuals AFAB (−1.0 cm; 95% CI −1.5; −0.6), and comparable to untreated individuals AMAB.

Conclusion

The study results suggest that skeletal dimensions are only altered by GAHT if endogenous puberty has not yet been completed at start of PS. These findings enhance our understanding of hormonal effects on the skeleton and may hold clinical relevance for body image as well as for forensic anthropology. Future research should evaluate clinical implications for surgical or obstetrical outcomes in transgender individuals.

Transgender and gender diverse individuals experience a mismatch between their sex assigned at birth and their gender identity, known as gender incongruence. The distress that may arise from this incongruence is referred to as gender dysphoria (1). Transgender individuals can be offered gender-affirming hormone treatment (GAHT) to align their body with their experienced gender. Individuals assigned male at birth (AMAB) can receive estrogen and anti-androgens to stimulate feminization, whereas those assigned female at birth (AFAB) can be treated with testosterone to induce masculinization. In adolescents with gender dysphoria, initial treatment usually consists of puberty suppression (PS) with gonadotropin-releasing hormone analogues (GnRHa) to suppress endogenous hormone production. If gender dysphoria persists, GAHT can be introduced from the age of approximately 16 years (2).

Several studies have shown that sex hormones, regardless of sex assigned at birth, can lead to the development of secondary sex characteristics, such as breast development and hair growth, in addition to changes in body composition (3-5). However, there are also studies that demonstrate that certain sexual dimorphisms cannot be altered by sex hormones, such as adult height in transgender adolescents (6, 7). The effects of sex hormones in transgender people on other sex-specific skeletal dimensions are still largely unknown.

Sexual dimorphisms in skeletal dimensions have been extensively documented in literature (8-10). The most substantial variations between male and female individuals are observed in the pelvis (8, 9, 11). Female individuals generally have a larger pelvic inlet width, a broader pubic symphysis width, a greater suprapubic angle, and a greater maximum pelvic width compared to male individuals (12). In addition, males tend to have a broader shoulder width compared to females (9). Furthermore, some of the skeletal differences between sexes appear to be due to a taller height and thus an overall larger skeleton in male individuals (13, 14).

Limited research has been performed on the effect of sex hormones on skeletal dimensions in transgender individuals. Sitek et al conducted a study that explored biometric differences in pelvic structures between individuals AFAB with masculinizing treatment and cisgender individuals aged 25 to 39 years (12). They found that certain pelvic dimensions, such as the height of the pelvis and the ilium, remained more consistent with female dimensions. Other dimensions, like the greatest width of the pelvis, pelvic inlet, and interacetabular distance started to resemble male controls. However, the study did not investigate the effects of estrogen on pelvic dimensions in transgender women, and it did not include participants who had undergone PS.

It is conceivable that, besides GAHT, (early) intervention with PS might also have impact on skeletal dimensions in transgender individuals, as puberty is a critical phase of bone remodeling (8, 15). Van der Loos et al reported that changes in hip bone geometry in transgender adolescents who started PS in early puberty were aligning with the experienced gender (ie, a greater subperiosteal width and endocortical diameter in men compared to women) (16). However, if PS was initiated later in puberty, hip bone dimensions continued to correspond to the sex assigned at birth.

Apart from its impact on external physical features, the influence of hormone therapy on skeletal dimensions may also be pertinent to obstetrical outcomes in trans masculine individuals (17). Furthermore, in trans feminine persons, alterations in pelvic dimensions may be significant for surgical outcomes of vaginoplasty procedures (18).

This study aims to evaluate the influence of GAHT, with or without prior treatment with GnRHa, on pelvic dimensions and shoulder width in transgender individuals. Furthermore, it seeks to determine whether the effects are related to the timing of PS, comparing initiation in early vs late puberty. We hypothesized that individuals who underwent early puberty suppression with GnRHa followed by GAHT would exhibit skeletal dimensions more aligned with their affirmed gender than those who only received GAHT after completing their endogenous puberty. From a wider perspective, the study will provide insights into the impact of sex hormones on the skeleton.

Methods

Participants

For this study, data from the Amsterdam Cohort of Gender Dysphoria (ACOG) database were used (19). This dataset contains data from individuals who were seen at the Center of Expertise on Gender Dysphoria in Amsterdam from 1972 to December 2018.

Participants were included in the current study if a dual-energy x-ray absorptiometry (DXA) scan was performed between the age of 18 to 28 years. Subsequently, they were divided into 4 groups based on treatment history: Untreated, no PS or GAHT (yet) at the time of DXA scan; Early PS, PS initiated at Tanner breast (B) or genital (G) stages 2 or 3 followed by ≥ 2 years of GAHT; Late PS, PS initiated at Tanner B or G stage 4 or 5 followed by ≥ 2 years of GAHT; and GAHT-only, ≥ 2 years of GAHT without prior treatment with PS. The terms AFAB and AMAB are used throughout this manuscript to simplify group comparisons.

Medical Treatment Protocol

All individuals were assessed by a mental health professional to confirm gender dysphoria according to the Diagnostic and Statistical Manual of Mental Disorders, fourth and later fifth edition (20, 21). PS could be initiated if individuals had reached at least Tanner stage G/B2 and consisted of subcutaneous or intramuscular triptorelin (Decapeptyl-CR 3.75 mg every 4 weeks or Pamorelin 11.25 mg every 10-12 weeks). GAHT was initiated from the age of 15 to 16 years. Feminizing treatment consisted of oral 17β-estradiol with a starting dosage of 5 microgram/kilogram/day (μg/kg/d) which was increased until an adult dose of 2 to 4 mg/d was reached. Some individuals AMAB were temporarily treated with a higher dosage of 6 mg estradiol or with ethinyl estradiol 100 to 200 mcg to reduce adult height (6). Masculinizing treatment consisted of intramuscular testosterone esters with a starting dosage of 25 mg/m² every 2 weeks and a final dose of 250 mg every 3 to 4 weeks. Individuals in the GAHT-only group, who started treatment as adults, had a different treatment schedule. Individuals AMAB started with estrogens and anti-androgens simultaneously. Estrogens most often consisted of oral 17β-estradiol at a dosage of 2 to 4 mg/day, but in some individuals, estradiol patches (50-150 μg/day twice a week) or estradiol gel (0.75-1.5 mg per day) were used. The anti-androgens consisted mainly of cyproterone acetate (25-100 mg/ day), but in some, spironolactone or a GnRHa was used. Individuals AFAB were treated with intramuscular testosterone esters (125-250 mg every 2-3 weeks), intramuscular testosterone undecanoate (1000 mg every 12 weeks), or transdermal testosterone gel (20-60 mg/day). In individuals AMAB, GnRHa or other anti-androgens were discontinued if a gonadectomy was performed. GnRHa was discontinued in individuals AFAB when the final dosage of testosterone was reached.

Measurements

All DXA scans were conducted as part of routine follow-up visits in the context of monitoring bone health and body composition. If multiple DXA scans were available, the scan closest to the age of 22 years was used. Before February 2011, DXA scans were performed using a Hologic Delphi system (Hologic Inc., Bedford, MA, USA), which underwent an update in 2004. In February 2011, this system was replaced with a Hologic Discovery system. Subsequently, the software on the Discovery system was updated in 2012 and 2015. In December 2020, a Hologic Horizon device was installed. Based on previous literature on sex dimorphisms of the skeleton (9, 12), 5 skeletal dimensions were selected for analysis. APEX software (version 4.0) was used to conduct the measurements using clear landmark points, see Fig. 1. The researcher who conducted all measurements was unaware of the sex and treatment group of the individuals. Measurements on the first 20 scans were conducted by 2 researchers independently, to assess the interobserver variability.

The following measurements were conducted (Fig. 1):

• A. Shoulder width: The distance between the most lateral points of the shoulders. The tuberculum majus was used as reference point.

• B. Greatest pelvic width: The distance between the 2 most lateral points of the iliac crest.

• C. Width of pelvic inlet: The distance between the most lateral points of the pelvic inlet.

• D. Width of pubic symphysis: The distance between the most medial points of the obturator foramen

• E. Interischial distance: The distance between the ischial spines. A straight horizontal line was drawn and the lowest point on each ischial spine that intersected the line was used as a landmark. The distance between the landmarks was the interischial distance.

Statistics

The statistical analyses were performed using STATA 15.1 (StataCorp). Both absolute and relative differences between the 20 measurements performed by 2 researchers were calculated (22). Furthermore, the intraclass correlation coefficient (ICC) was used to assess the interobserver variability. Data are presented as mean ± SD or as median (interquartile range [IQR]) if the data are not normally distributed. Linear regression was used to compare the continuous baseline characteristics and the skeletal dimensions between the 2 untreated groups. For the comparison between different treatment groups, linear regression was used with dimensions as outcome variable and the groups as categorical determinant. Additional correction for height was performed.

Ethics

The Medical Ethics Review Committee of VU University Medical Center in Amsterdam, the Netherlands, assessed the ACOG protocol and determined that the Medical Research Involving Human Subjects Act (WMO) did not apply to this data collection. Informed consent was waived due to the retrospective design and the size of the cohort. In order to expand the group with prior use of GnRHa, 62 eligible individuals with DXA scans performed after 2018 were included. These individuals had participated in previous studies and had given informed consent for the use of their clinical data, including their DXA scans (6, 23).

Results

The ACOG data set contained 8831 individuals, of whom 629 had at least one DXA scan between the age of 18 to 28 years; among these, 182 adults had undergone a scan before the start of GAHT. As the untreated group was relatively large, only 100 individuals from this group were randomly selected, of whom 49 were individuals AMAB and 51 were individuals AFAB. Of the remaining treated individuals, 81 had an available DXA scan ≥ 2 years after the start of GAHT. This group was combined with 62 individuals from previous study cohorts. In total, 46 individuals were included in the Early PS group, 57 in the Late PS group, and 40 in the GAHT-only group. Due to poor scan quality, the interischial distance and the pubic symphysis width could not be measured 6 and 5 times, respectively.

Demographics

A total of 243 individuals (97% White) were included in the study, of whom 121 (50%) were AMAB and 122 (50%) were AFAB. Demographics of the study population are presented in Table 1. As expected, the Early PS group initiated PS and GAHT at a younger age than the Late PS group in both transition types. In individuals AMAB, the height of the Early PS group was greater than the untreated group. In individuals AFAB, the height from both the early and the Late PS group was greater than the untreated group.

Interobserver Variability

The mean differences of the measurements between the 2 observers for the 5 dimensions are shown in Table 2. All intraclass correlation coefficients exceeded 0.96, indicating a strong positive correlation. Additionally, Supplementary Fig. 1 (24) shows the level of agreement between the 2 observers using Bland-Altman plots.

Untreated Individuals (no PS or GAHT)

Measurements from the untreated individuals were compared to assess differences in the 5 skeletal dimensions between AMAB and AFAB without exposure to PS or GAHT.

The findings revealed that individuals AMAB had a larger shoulder width (A) compared to AFAB (difference 2.8 cm [95% CI, 2.0 to 3.6]; after correction for height this difference was 1.2 cm [95% CI, 0.4 to 2.1]). Additionally, untreated individuals AFAB had a similar greatest pelvic width (B) (0.4 cm [95% CI, −0.2 to 1.1]), a greater width of the pelvic inlet (C) by 0.4 cm (95% CI, 0.1 to 0.7), a greater width of the pubic symphysis (D) by 0.7 cm (95% CI, 0.4 to 1.0), and a greater interischial distance (E) by 2.1 cm (95% CI, 1.4 to 2.8) compared to untreated individuals AMAB. After correction for height, differences compared to AMAB of all pelvic dimensions increased and were significantly greater in AFAB (see Supplementary Table 1) (25). Because correction for height substantially influenced the differences of the various dimensions between the sexes, this correction was performed in all further analyses of the treated groups.

Treated Individuals

Feminizing treatment

Figure 2 and Table 3 show mean skeletal dimensions per treatment group. In individuals AMAB, shoulder width (A) was comparable between the untreated, GAHT-only, and Late PS groups. Shoulder width from individuals in the Early PS group was smaller compared to untreated individuals AMAB (−1.3 cm [95% CI, −2.1 to −0.5]) and comparable to untreated individuals AFAB (0.1 cm [95% CI, −0.9 to 1.1]). Greatest pelvic width (B) in all 3 treated AMAB groups was comparable to both untreated individuals AMAB and to untreated individuals AFAB. Pelvic inlet (C) of both the early and late PS group was greater compared to untreated individuals AMAB by 1.0 cm (95% CI, 0.7 to 1.4) and 0.6 cm (95% CI, 0.2 to 0.9), respectively. They were both comparable to untreated individuals AFAB (Early PS, 0.1 cm [95% CI, −0.3 to 0.5]; Late PS, −0.3 cm [95% CI, −0.8 to 0.1]). Pelvic inlet of the GAHT-only group was comparable to untreated individuals AMAB and smaller by 0.9 cm (95% CI, −1.4 to −0.4) compared to untreated individuals AFAB. When comparing to untreated individuals AMAB, the width of the pubic symphysis (D) was greater in Early PS by 0.9 cm (95% CI, 0.7 to 1.2) and in Late PS by 0.7 cm (95% CI, 0.4 to 1.0). These 2 groups had a similar width of the pubic symphysis as the untreated AFAB group (Early PS, −0.1 cm [95% CI −0.4 to 0.3]; Late PS, −0.3 cm [95% CI, −0.6 to 0.03]). No significant difference was observed between the untreated individuals AMAB and GAHT-only group. Interischial distance (E) was greater in all treated AMAB groups compared to untreated individuals AMAB (Early PS, 2.2 cm [95% CI, 1.5 to 3.0]; Late PS, 2.6 cm [95% CI, 1.8 to 3.4]; GAHT-only, 1.7 cm [95% CI, 0.7 to 2.7]). There were no differences in interischial distance between the treated individuals AMAB and the untreated individuals AFAB.

Masculinizing treatment

In individuals AFAB, shoulder width (A) was not significantly different from untreated individuals AFAB in all 3 treated groups (Early PS, 0.6 cm, [95% CI, −0.5 to 1.7]; Late PS, 0.1 cm [95% CI, −0.7 to 0.9]; GAHT-only, 0.4 cm [95% CI, −0.5 to 1.2]). However, only in the Early PS group, shoulder width was comparable to untreated individuals AMAB (−0.6 cm [95% CI, −1.7 to 0.5]). No significant differences were observed in the greatest pelvic width (B). The pelvic inlet (C) in the Early PS group was smaller by 1.0 cm (95% CI, −1.5 to −0.6) compared to the untreated individuals AFAB and comparable to untreated individuals AMAB (0.1 cm [95% CI, −0.3 to 0.6]). Compared to untreated individuals AFAB, the width of the pubic symphysis (D) was smaller in both the early and the late PS group by −1.1 cm (95% CI, −1.5 to −0.8) and −0.4 cm (95% CI, −0.7 to −0.1), respectively and similar in the GAHT-only group (0.0 cm [95% CI, −0.3 to 0.3]). Only the Early PS group had a width of the pubic symphysis comparable to untreated individuals AMAB (−0.1 cm [95% CI, −0.5 to 0.3]). Interischial distance (E) was not significantly different to untreated AFAB in the early and late PS group and even slightly greater in the GAHT-only group by 0.9 cm (95% CI, 0.1 to 1.7). Interischial distance was significantly greater in all treated groups compared to untreated individuals AMAB (Early PS, 2.0 cm [95% CI, 0.9 to 3.0]; Late PS, 2.9 cm [95% CI, 2.1 to 3.7]; GAHT-only, 3.2 cm [95% CI, 2.3 to 4.0]).

Early vs late puberty suppression

In individuals AMAB, only shoulder width was different between these 2 groups, which was 1.2 cm smaller (95% CI, −2.3 to −0.1) in the Early PS group. Fourteen individuals were treated with ethinyl estradiol, of whom 12 in the Early PS group and 2 in the Late PS group. No differences in the skeletal dimension were found between individuals treated with ethinyl estradiol compared to those who were treated following the regular treatment protocol.

For individuals AFAB, only pelvic inlet and width of the pubic symphysis were smaller in the Early PS group compared to the Late PS group by −0.8 cm (95% CI, −1.2 to −0.3) and −0.8 cm (95% CI, −1.1 to −0.5), respectively.

Discussion

This study investigated the effects of both the inhibition of endogenous sex hormones with GnRHa and the supplementation of gender-affirming hormones on skeletal dimensions in transgender individuals. By dividing individuals into 4 treatment groups, the effects of the duration of endogenous hormone exposure, that is, to what extent they had undergone their endogenous puberty, on the skeleton was also examined. As hypothesized, skeletal dimensions from individuals who started PS in early puberty were most similar to those of the affirmed gender.

To assess the skeletal dimensions, DXA scans were used. Our first aim was to determine the accuracy of these measurements. Independent assessment of the first 20 scans by 2 observers showed excellent interobserver agreement for all 5 skeletal dimensions. The gold standard to assess skeletal dimensions is cadaver analysis (26). As no gold standard is available for measuring skeletal dimensions in living individuals, previous studies used different imaging techniques, such as roentgen pelvimetry, magnetic resonance imaging (MRI) or DXA scans. A study conducted by Decrausaz et al compared DXA and MRI measurements of 3 pelvic dimensions and described correlations ranging from R² = 0.77 to R² = 0.96. The study concluded that DXA stands as a feasible method for obtaining in vivo measurements of the pelvis (27).

Shoulder Width

As expected, shoulder width was nearly 3 cm greater in untreated individuals AMAB compared to untreated individuals AFAB, a finding consistent with prior studies in cisgender individuals (9, 28). However, after correction for height, the difference decreased to only 1 cm. This indicates that shoulder width is not solely linked to sex but also associated with height, which was previously reported by Völgi et al (13).

In individuals AMAB, only those who started PS in early puberty had smaller shoulders than the untreated AMAB group. This finding underscores the impact of testosterone exposure on shoulder width during puberty. Notably, when adjusting for height, this specific group showed shoulder width similar to individuals AFAB, revealing that the remaining difference of approximately 2 cm can be attributed to the taller height in the treated individuals AMAB. The feminine proportions of shoulder width to height in the Early PS individuals AMAB may not only result in a more feminine figure but may also affect the position and appearance of the breasts. Transgender women who started GAHT in adulthood have a more lateral position of the breasts on the chest than women from the general population, due to anatomical differences such as a wider sternum in individuals AMAB (4). Further studies will have to investigate if this is different in those who start PS in early or late puberty prior to GAHT.

Testosterone treatment in individuals AFAB did not result in a greater shoulders width compared to the untreated AFAB group when adjusted for height. However, when compared to untreated individuals AMAB, only the Early PS group had a similar shoulder width.

Overall, hormonal treatment can alter shoulder width in transgender individuals toward their experienced gender, but this was only observed in those who started treatment in early puberty. This suggests that testosterone can only alter shoulder width if significant growth potential is left, and that once shoulder width has increased during puberty, it is an irreversible process. In addition, overall skeletal growth (as indicated by height) also plays a significant role in determining shoulder width. As described in previous research, PS and GAHT have minimal effect in this regard (6, 7).

Pelvic Dimensions

In contrast to shoulder width, all pelvic dimensions except the greatest pelvic width were greater in untreated individuals AFAB compared to untreated individuals AMAB. This finding aligns with prior research and underscores the importance of a wider pelvis for obstetric outcomes (8, 11, 12). Interestingly, several studies describe changes in the female pelvis across different life stages, such as puberty and menopause, possibly mediated by sex hormones (8). This suggests estradiol treatment may induce obstetrically favorable pelvic changes, even after finishing puberty.

Indeed, remarkable differences away from the untreated individuals AMAB toward the untreated individuals AFAB are seen in 3 out of 4 pelvic dimensions in treated individuals AMAB, with the greatest differences in the Early PS and Late PS groups. These findings confirm the obstetrically beneficial effects of estradiol on the pelvis previously described in humans and primates (8, 29), but the effects are more limited after endogenous puberty has taken place and linear growth has finished.

In individuals AFAB, 2 out of 4 pelvic dimensions changed toward the untreated individuals AMAB. However, only in the Early PS group, width of the pelvic inlet and of the pubic symphysis were comparable to the untreated individuals AMAB. This implies an irreversible impact of endogenous estradiol exposure on pelvic dimensions. However, a previous study by Sitek et al using x-rays of the pelvis described that greatest pelvic width and width of the pelvic inlet in individuals AFAB who started testosterone treatment in adulthood were smaller compared to cis females and comparable to cis males (12). These results are in line with the hypothesis that pelvic dimensions can decrease after menopause due to a drop in serum estradiol levels (8), resulting in a restricted birth canal. The fact that the changes observed in the study by Sitek et al were absent or only visible in individuals that received PS in the current study, might be explained by the different imaging methods resulting in different positioning and thereby noncomparable outcomes. In contrast to greatest pelvic width and width of the pelvic inlet, no difference in the width of the pubic symphysis was observed by Sitek et al, while it was smaller in both the early and late PS groups compared to untreated individuals AFAB in the current study. This might indicate that this dimension is irreversibly influenced by estradiol exposure.

Besides a valuable insight into hormonal effects on the skeleton, this study also has several clinical implications. Firstly, the results demonstrate that initiation of PS in early puberty, followed by GAHT, has the most impact on skeletal dimensions when aiming for skeletal outcomes consistent with the affirmed gender. Although the absolute differences are relatively small, the altered proportions may result in a different body shape and a more positive body image in transgender individuals, preventing the need for surgery to feminize body figure such as surgical shoulder width reduction (30). These findings are not only valuable for counseling transgender individuals about expected body changes with hormone treatment but also hold potential implications for surgical procedures. For instance, in individuals AMAB, the induced changes in the pelvic bone could be advantageous during vaginoplasty, as they might provide increased space for constructing the vaginal canal. A wider pelvis would certainly be important if uterine transplantation is considered in transgender women, a potential treatment option that has been suggested (31). In contrast, in transmasculine individuals desiring pregnancy, these alterations in the pelvic structure may decrease the size of the birth canal, potentially resulting in adverse obstetric outcomes. However, earlier studies were unable to identify pelvic dimensions that could predict labor outcomes (17). In addition, the hormonal milieu of pregnancy may induce further changes in the pelvis. Future studies should focus on obstetrical outcomes in transmasculine individuals who received treatment with GnRHa in adolescence and subsequent testosterone therapy to determine the clinical significance of these skeletal changes.

The findings may also be relevant to forensic anthropology. The pelvis and the skull are the part of the skeleton used most frequently to determine sex from skeletal remains, with the pelvis being the most reliable single area (32). The described changes due to hormonal therapy may complicate sex determination (33).

A strength of this study is the use of a method with a confirmed good interobserver variability, ensuring accuracy of the collected data. A limitation is the fact that we could not study the anteroposterior dimension of the pelvis. However, previous research found strong correlation between iliac crest and pelvic inlet area (34). Dividing the individuals into 4 treatment groups and also comparing to treatment-naïve individuals from both sexes provided detailed insight into the effects of hormone exposure on skeletal dimensions as well as the importance of the timing of hormone exposure. However, this did result in relatively small subgroups while the groups in total were of reasonable size. Despite the retrospective nature of this study, there were minimal missing data except for a few skeletal dimension measurements that could not be performed due to poor scan quality.

In summary, hormonal treatment can affect certain skeletal dimensions in transgender people. Shoulder width is only affected when PS is initiated in early puberty, while pelvic dimensions may be sensitive to hormonal changes even after puberty has ended. The effects were most prominent in those who started PS in early puberty, suggesting irreversible skeletal changes during endogenous puberty. This conclusion aligns with the findings of van der Loos et al, who observed that hip bone geometry in those who start PS early in puberty resembles that of their experienced gender (16). We hypothesize that the observed effects are reversible up until initiation of GAHT. The different effects on different skeletal dimensions suggest that the development of some skeletal structures is less sensitive to sex hormone influences and may be regulated by genetic factors, or alternatively, by sex hormone exposure during prenatal or early postnatal development. These findings enhance our understanding of hormonal effects on the skeleton and may have clinical significance for body image and surgical and obstetric outcomes in transgender individuals, as well as for forensic anthropology. Follow-up studies could investigate hormone-induced changes in other skeletal regions, such as facial skeletal dimorphisms, including the jawline, bridge of the nose, and cheekbones.

Funding

No grants.

Disclosures

No disclosures.

Data Availability

The data set generated during and analyzed during the present study is not publicly available because of privacy regulations.

References

Abbreviations

 
• ACOG

  Amsterdam Cohort of Gender Dysphoria

 
• AFAB

  assigned female at birth

 
• AMAB

  assigned male at birth

 
• DXA

  dual-energy x-ray absorptiometry

 
• GAHT

  gender-affirming hormone therapy

 
• GnRHa

  gonadotropin-releasing hormone agonist

 
• PS

  puberty suppression