1 INTRODUCTION

Transgender/gender-diverse youth are increasingly seeking care at paediatric clinics specializing in gender dysphoria.^{1, 2} Current guidelines for the diagnosis and management of children and youth with gender dysphoria emphasize the need for mental health and medical interventions but provide limited information on body composition, cardiometabolic parameters and existing lifestyle habits of this unique population.³ We hypothesized that treatment-naïve transgender youth exhibit unfavourable weight status and body composition prior to hormonal transition.

Anthropometric and metabolic characteristics at presentation in treatment-naïve individuals are still unclear, with contradictory findings in the literature. While few studies suggested higher rates of overweight and obesity in transgender youth compared with national values,⁴ others have found no such differences.⁵ Furthermore, previous studies have not differentiated between transgender females and transgender males,⁴ nor have they investigated the association between body composition at presentation and cardiometabolic risk factors.⁵ These deficiencies call for more research to better understand the health needs of transgender youth and thereby provide appropriate care.

Adolescence is a period of physiological and psychological changes that can intensify gender incongruence, particularly as secondary sex characteristics emerge during sexual development.^{3, 6} These changes, coupled with concern regarding body image, low self-esteem, inadequate nutrients intake and sedentary lifestyle can exacerbate the risk of malnutrition, obesity and eating disorders.⁷ Research findings have consistently shown that gender minorities often exhibit unfavourable lifestyle habits, including an unhealthy diet, decreased physical activity,⁸ shortened night sleep⁹ and use of medications for associated conditions,⁴ all of which may contribute to an unhealthy metabolic profile.

Since 2018, one of the goals of our institute has been to screen for modifiable cardiovascular disease (CVD) risk factors, starting at the initial consultation and continuing throughout follow-up as clinically appropriate. Persons are referred for endocrine consultation and undergo a comprehensive metabolic risk assessment, including body composition measurement by bioelectrical impedance analysis (BIA) during clinic visits as part of the routine standard of care.¹⁰ We have previously demonstrated the predictive value of body composition in risk assessment of CVD in children and adolescents living with various endocrine conditions.^11-17

Research is currently focused on the impact of gender-affirming hormone treatment on the health outcomes experienced by transgender individuals.¹⁸ There is a scarcity of published studies examining transgender youth before they undergo hormonal transition. In light of this research gap, our study aims to assess potential associations between body composition parameters, components of MetS and lifestyle habits in transgender youth before they initiate treatment.

2 METHODS

This real-life observational study of youth referred to The Israeli Children and Adolescents Gender Clinic extended from June 2021 to December 2022. Our clinic currently serves as a national referral centre for transgender/gender-diverse youth. As part of the standard of care in the clinic, transgender youth are expected to meet with a multidisciplinary team consisting of paediatric endocrinologists, nurses, dietitians and a psychosocial team.² During routine clinic visits, the staff performs anthropometric measurements and medical interviews to assess the patient's family and medical history, psychiatric morbidities and lifestyle habits (including dietary patterns, physical activity and sleep patterns).

The study population was generated by searching the institutional BIA database for subjects with the terms “gender dysphoria” and “gender diverse” as the reason for referral. The BIA data were linked to the individuals' electronic medical records. Included in the study were individuals aged 8 to 18 years who had not yet initiated hormonal treatment. Excluded were eight individuals, of whom four had BIA measurements by gender identity rather than the designated sex at birth and four with an underlying medical disorder affecting body composition (specifically, type 1 diabetes, inflammatory bowel disease and active Graves' disease). In total, 153 transgender treatment-naïve youth (59 transgender females, 94 transgender males) fulfilled the inclusion/exclusion criteria and were included in the data analysis.

Data collected from the participants' medical files at the time of BIA assessment included designated sex at birth, age, anthropometric measurements, pubertal stage according to Marshall and Tanner^{19, 20} and laboratory evaluations. The clinical evaluation of subjects and their parents included measurement of height while standing barefoot on a commercial Harpenden stadiometer (Holtain Ltd., Crosswell, UK). The subjects' height and body mass index (BMI) values were converted to sex designated at birth- and age-specific standard deviation scores (z-scores) according to the CDC 2000 growth charts.²¹ Blood pressure (BP) measurements were conducted with an automated device (Welch Allyn—52000, Tycos Instruments, Inc., NC, USA) using the appropriate cuff for age according to AHA guidelines.²² Systolic BP (SBP) and diastolic BP (DBP) percentiles were calculated by means of an online age-based paediatric BP calculator.²³

Body composition was indirectly measured by BIA (Tanita Body-Composition Analyser, Tanita MC-780 MA and GMON Professional Software) adjusted for sex designated at birth, age, height and race (white), according to reference ranges.^{24, 25} The measurement required the subject to stand barefoot on the analyser and grip the handles, and it took approximately 1 min to perform. The BIA measures whole body and segmental fat and muscle. The calculated variables include appendicular skeletal muscle mass (the sum of muscle mass of four limbs), and muscle-to-fat ratio (appendicular skeletal muscle mass [kg]/fat mass [kg]). The z-scores for muscle-to-fat ratio (sex designated at birth) were calculated according to BIA paediatric reference curves.²⁶

The socioeconomic position (SEP) as determined by home address was analysed based upon the Israel Central Bureau of Statistics' Classification.²⁷ The neighbourhoods or localities are divided into SEP clusters, with 1 representing the lowest rating and 10 representing the highest rating. The SEP index is an adjusted calculation of 14 variables that measure social and economic levels in the domains of demographics, education, standard of living and employment (−2.797 to 2.590).

3 RESULTS

The cohort was comprised of 153 transgender treatment-naïve youth (59 transgender females, 94 transgender males) with a higher proportion of transgender males than transgender females (61.4% vs. 38.6%, respectively, p = 0.006). The median age of the transgender youth was 15.7 years [range 8–18]. Over one-half of the individuals (58.8%) resided in affluent neighbourhoods, yielding a high median socioeconomic position cluster by locality of residence of 8 (range 2–10).

Most of the individuals in our transgender cohort were born after spontaneous conceptions (88.9%) and singleton pregnancies (96.7%), with exposure to gestational diabetes in 3.9%. The median gestational age for births in this group was 39 weeks [IQR 38, 40] with 7.8% of the cohort having been born preterm.

The anthropometrics and body composition of the transgender cohort are presented in Table 1. The height of the transgender cohort according to designated sex at birth was similar to their genetic height potential (the difference between the height z-score of the individual and their mid-parental height z-score: 0.07 ± 0.74 [range: −1.90 to 2.20]).

The overall weight status (mean BMI z-score and SD) of the cohort was 0.29 ± 1.38 (range: −4.44 to 2.49). Weight status differed in a sex-specific manner: it was average for transgender females (H₀: BMI z-score = 0, p = 0.192) and above average for transgender males (H₀: BMI z-score = 0, p < 0.001). The mean BMI z-scores of transgender females were lower than those of the transgender males (p < 0.001). The distribution of weight categories differed between genders, with a greater proportion of underweight subjects among the transgender females and a greater proportion of overweight/obese subjects among the transgender males (p = 0.035).

The median muscle-to-fat ratio z-score of the transgender cohort was −0.66 ([IQR −1.38, −0.10], range: −2.36 to 3.33). The muscle-to-fat ratio z-score differed in a sex-specific manner: it was average for transgender females (H₀: muscle-to-fat ratio z-score = 0, p = 0.125) and below average for transgender males (H₀: muscle-to-fat ratio z-score = 0, p < 0.001). Transgender females had significantly higher muscle-to-fat ratio z-scores than the transgender males (p = 0.002). The latter finding was more pronounced when calculating gender identity muscle-to-fat ratio z-scores (p < 0.001).

There were strong negative correlations between the muscle-to-fat ratio z-score and the BMI z-score in both transgender females and transgender males (r = −0.737, p < 0.001 and r = −0.863 and p < 0.001, respectively). No significant correlations were found between birthweight z-score and either BMI z-scores or muscle-to-fat ratio z-scores in both of those groups.

The median SBP and DBP percentiles in the transgender females and the transgender males were above average, albeit within the normal range. Categorical stratification of the BP levels revealed that 25.5% were elevated, while the distributions of BP categories were similar between genders.

Exploration of the lifestyle habits of the transgender youth revealed that most of the cohort consumed an unhealthy diet (64.1%) and were not engaged in sufficient physical activity (77.8%) and that the number of hours of sleep was inadequate for almost one-half (49.0%) of them. Overall, 21 (35.6%) of the transgender females and 29 (30.9%) of the transgender males did not meet any of the three lifestyle recommendations, and only 16 (10.5%) of the transgender youth engaged in all three recommended practices.

The medical characteristics of the transgender cohort are presented in Table 2. The overall rate of individuals affected by psychiatric morbidities was 59.5%, with psychiatric medication usage in 39.2%. Atypical antipsychotic drug usage was more common among transgender males than transgender females (p = 0.008).

The laboratory evaluations of the transgender cohort are presented in Table 3. The transgender males had higher TG levels (p = 0.016), higher non-HDL-c levels (p = 0.032), higher TG:HDL-c ratios (p = 0.006), higher total cholesterol:HDL-c ratios (p = 0.014) and higher LDL-c:HDL-c ratios (p = 0.020) than the transgender females.

The distribution of the type and number of MetS components in the transgender cohort is presented in Figure 1A,B. The transgender males had higher frequencies of MetS components than the transgender females (n = 66/94, 70.2% vs. n = 22/59, 37.3%, respectively, p < 0.001). Among the individuals with healthy weight status (n = 91/153), the transgender males (n = 25/53, 47.2%) were more likely to have at least one MetS component than the transgender females (n = 8/38, 21.1%, p = 0.011), while no differences were found in transgender individuals with overweight/obese weight status (p = 0.763). The median muscle-to-fat ratio z-score for the cohort as a whole was lower in those individuals with MetS components than those without (−1.19 vs −0.29, respectively, p < 0.001). The median muscle-to-fat ratio z-score of the transgender males with at least one MetS component was significantly lower than those without (−1.22 vs −0.44, p < 0.001), while the median muscle-to-fat ratio z-score of transgender females with at least one MetS component tended to be lower than those without (−0.87 vs −0.12, p = 0.061).

Multivariable linear regression models for muscle-to-fat ratio z-scores were performed with the following variables: transgender males, age, SEP index, single caregiver household, birthweight z-score, psychiatric conditions, dietary pattern physical activity and sleep. The muscle-to-fat ratio z-score was negatively associated with being designated female at birth (Β = -0.55, standard error [SE] = 0.17, p < 0.001) and psychiatric morbidity (B = -0.42, SE = 0.16, p = 0.012) and positively associated with meeting physical activity recommendations (Β = 0.43, SE = 0.19, p = 0.028). The BMI z-score was positively associated with designated female at birth (B = 0.87, SE = 0.22, p < 0.001) and single caregiver household (B = 0.45, SE = 0.23, p = 0.048).

The final logistic regression model revealed that designation as female at birth increased the odds for MetS components by 3.49 (odds ratio [OR] = 3.49 (95%CI:1.63–7.44), p = 0.001]. Additionally, the odds for evidence of a MetS component(s) was increased by 2.2 (1/OR) for every 1SD decrease in the muscle-to-fat ratio z-score [OR = 0.46 (95%CI:0.31–0.69), p < 0.001].

4 DISCUSSION

The findings of this observational study of a cohort of treatment-naïve transgender youth from a national referral centre provide a picture of the health parameters and lifestyle of this unique population. Our results revealed gender dimorphism in weight status, body composition and cardiometabolic derangements in this population. Transgender adolescents exhibited body compositions that differed from their gender identity, with transgender males showing greater disparities. Taken together, our observations suggest that treatment-naïve transgender male adolescents have an imbalance between muscle and adipose tissue, which places them at increased risk for MetS components even prior to hormonal therapy. Importantly, these findings held true even after adjustment for social circumstances, family history of cardiovascular disease, medical and psychiatric morbidities and lifestyle-related factors.

Socioeconomic circumstances are widely recognized as a key determinant of health conditions in children, with a strong inverse relationship between socioeconomic position and obesity.^{34, 35} Notably, this relationship appears to be stronger in females.³⁵ In our study cohort, although the majority of individuals came from high socioeconomic backgrounds, nearly one-half of them were classified as having unhealthy weight categories. These findings are consistent with previous research on weight status in transgender youth and adults,^{36, 37} which has shown a higher prevalence of overweight/obesity compared with the cisgender populations, particularly among transmales with higher BMI z-scores.^{4, 8} Our study further reveals a higher proportion of individuals falling within the underweight classification among transgender females compared with that in a national representative sample of adolescent males.³² Interestingly, our findings suggest that socioeconomic circumstances during childhood and adolescence may not be the cause of the gender-based disparities in weight status that we observed, since the sociodemographic background was comparable between genders.

The intrauterine environment has a critical influence on an individual's metabolic health, especially regarding their susceptibility to later-life obesity and dyslipidaemia.^{38, 39} The majority of pregnancies were uncomplicated and resulted from spontaneous/natural conception in healthy mothers of our study participants. The median birth weight z-scores were within the average range and normally distributed, suggesting that adverse conditions during foetal development are unlikely to have played a significant role in the participant's current weight status. Furthermore, no significant associations were found between birth weight z-scores and either BMI z-scores or muscle-to-fat ratio z-scores in both transgender females and males. One could therefore speculate that postnatal medical conditions and environmental factors contributed to the weight status during childhood and adolescence.

Several factors may explain the gender differences in weight distribution among transgender individuals, including distorted body image, pathological eating behaviours and intentional weight manipulation for gender-affirming purposes.⁴⁰ For some transgender youth, these behaviours may serve as a coping mechanism for gender-related distress. Studies have consistently shown that transgender individuals engage in food restriction and/or compensatory eating behaviours to prevent or halt puberty onset.^40-42 Additionally, individuals with gender dysphoria have been reported to have higher rates of weight manipulation to accentuate desired gender characteristics.^{40, 43-47} For example, transgender males may use weight manipulation to improve gender congruence by increasing weight to conceal feminine features.⁴¹ Psychiatric conditions and medications prescribed for them may impact appetite and result in changes in body weight.^48-50 In the current study, over one-half of the participants had a diagnosis of psychiatric morbidities, and almost one-half of them reported having used psychiatric medications prior to treatment. Although the prevalence of psychiatric conditions was similar across genders, transgender males were found to have greater use of atypical antipsychotic drugs, which are known to have appetite-inducing effects.⁵¹ It is possible that these medications contributed to the higher rates of obesity observed among our transgender male participants. Interestingly, the presence of psychiatric morbidities was linked to an unfavourable body composition, as demonstrated by lower muscle-to-fat ratio z-scores.

BMI is widely used as a predictor of metabolic complications, but it has limitations in distinguishing between muscle and adipose tissue and in reflecting adipose distribution. To address these limitations, body composition assessment, particularly through the use of muscle-to-fat ratio z-scores, can identify those at risk for sarcopenic obesity.^{52, 53} Sarcopenic obesity, which is characterized by excessive fat mass and reduced muscle mass, is associated with chronic low-grade inflammation and may have more severe health consequences than obesity or sarcopenia alone.⁵⁴ Although the co-occurrence of sarcopenia and obesity in adults has been shown to amplify the risk of adverse health outcomes, its consequences in the paediatric population have yet to be fully elucidated.⁵⁵ Our findings indicate that treatment-naive transgender adolescent males are more likely to experience sarcopenic obesity. We have previously demonstrated the predictive value of muscle-to-fat ratio z-scores as a surrogate marker of sarcopenic obesity in assessing CVD risk factors in youth.^{11, 16, 17} Our previous and current studies collectively show that muscle-to-fat ratio z-scores can predict CVD risk factors in transgender male adolescents.

We observed a higher frequency of MetS components among transgender males compared with transgender females in our cohort. This observation remained consistent among adolescents of healthy weight in our cohort and was particularly pronounced among those who were living with overweight or obesity. One possible explanation for this observation is the lower muscle-to-fat ratio z-scores seen in transgender males. This aligns with previous studies that have shown a positive association between low muscle-to-fat ratio z-scores and MetS components.^{11, 56}

Dyslipidaemia, particularly atherogenic dyslipidaemia, as indicated by the TG:HDL-c ratio, emerged as the predominant MetS component in our cohort, with elevated rates detected among transgender males. It is important to acknowledge that future testosterone treatment for transgender males could exacerbate the atherogenic lipid profile, leading to increased LDL-c and TG levels and decreased HDL-c levels. This potential scenario might place transgender males at a heightened risk for cardiovascular disease. Given these findings, addressing modifiable risk factors, such as dyslipidaemia, through lifestyle and medical interventions becomes crucial, particularly for transgender individuals who may commence testosterone treatment.

Various environmental factors, such as poor nutrition, lack of physical activity, and inadequate sleep, are well-known to contribute to weight status in adolescents. Previous studies have demonstrated that transgender youth engage in unhealthy dietary choices, sedentary recreational activities, and reduced physical activity, and that they have insufficient sleep and poor sleep hygiene compared with their cisgender peers.^{4, 57, 58} In particular, transgender youth have been found to consume more processed food and sugary beverages and less fruit, vegetables, and water than their cisgender peers.⁸ Unhealthy lifestyle habits in our transgender cohort before they embark upon treatment may contribute to the higher rates of overweight/obesity observed in this population compared with the general Israeli population. Adequate sleep was reported in only 51% of our transgender cohort in comparison with 84.2% of youth in Israel.⁵⁹ Israeli youth also reported better compliance with physical activity recommendations compared with transgender youth (37.7% vs 22.2%, respectively).⁵⁹ However, our study revealed that both transgender males and females exhibit unfavourable lifestyle habits at similar rates so that lifestyle factors alone cannot fully account for gender differences in weight status.

It is noteworthy to consider that seemingly modifiable behaviours, traditionally perceived as under individual control, may be significantly influenced by broader structural factors, disproportionately affecting vulnerable segments of the population, including transgender individuals. Hence, rather than solely attributing unfavourable lifestyle behaviours to individual choices, it is crucial to acknowledge the role of systemic determinants. Given this perspective, it is pertinent to initiate a discussion on systems-level interventions aimed at addressing these health behaviours within the target population. Such interventions could encompass policies, community-based initiatives and healthcare system reforms designed to mitigate the impact of structural inequalities and promote healthier lifestyles more equitably.

5 CONCLUSION

Treatment-naïve transgender-male adolescents have an imbalance between muscle and adipose tissue, which may increase their susceptibility for MetS components even prior to hormonal treatment. In addition, our findings suggest that treatment-naïve transgender youth exhibit unfavourable lifestyle behaviours. The effects of such unhealthy climates on weight status and body composition are apparently gender-specific. Personalized medical and nutritional interventions are warranted in this unique group of individuals in the attempt to promote good cardiometabolic health outcomes.

AUTHOR CONTRIBUTIONS

Ms. Ophir Borger conceptualized and designed the study, collected data, carried out the initial analyses, drafted the initial manuscript and critically reviewed and revised the manuscript for important intellectual content. Dr Anat Segev-Becker collected data, drafted the initial manuscript and critically reviewed and revised the manuscript for important intellectual content. Dr Liat Perl conceptualized the study, collected data, carried out the initial analyses and critically reviewed and revised the manuscript for important intellectual content. Dr Asaf Ben Simon carried out the initial statistical analyses and critically reviewed and revised the manuscript. Dr Michal Yackobovitch-Gavan designed the data collection instruments, carried out the statistical analyses and critically reviewed and revised the manuscript for important intellectual content. Prof Avivit Brener and Dr Asaf Oren collected data, and critically reviewed and revised the manuscript for important intellectual content. Ms. Tamar Sheppes collected data and critically reviewed and revised the manuscript. Prof Yael Lebenthal conceptualized and designed the study, coordinated and supervised data collection and critically reviewed and revised the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to declare.