BACKGROUND AND OBJECTIVES: The Endocrine Society states that adolescents with gender dysphoria may start cross-sex hormones. The goal of this study was to identify patterns in metabolic parameters in transgender adolescents receiving cross-sex hormones.
METHODS: Data from adolescents aged 14 to 25 years seen in 1 of 4 clinical sites between 2008 and 2014 were retrospectively analyzed. Subjects were divided into affirmed male (female-to-male) patients taking testosterone and affirmed female (male-to-female) patients taking estrogen. Previously recorded measurements of blood pressure, BMI, testosterone, estradiol, prolactin, lipids, electrolytes, liver function tests, hemoglobin/hematocrit, and hemoglobin A1c were reviewed. These values were obtained from before the start of therapy, at 1 to 3 months after initiation, at 4 to 6 months, and at 6 months and beyond. Repeated measures analysis of variance models were used to evaluate changes over time.
RESULTS: One hunderd and sixteen adolescents were included (72 female-to-male subjects and 44 male-to-female subjects). Of the 72 subjects taking testosterone, a significant increase in hemoglobin/hematocrit levels and BMI, as well as a decrease in high-density lipoprotein level, was recorded at each visit. No significant changes in any other parameter tested were found. Of the 44 subjects taking estrogen, no statistically significant changes were noted in the measured metabolic parameters.
CONCLUSIONS: Testosterone use was associated with increased hemoglobin and hematocrit, increased BMI, and lowered high-density lipoprotein levels; estrogen was associated with lower testosterone and alanine aminotransferase levels. Otherwise, cross-sex hormone administration in adolescents was not associated with significant differences in the selected metabolic parameters over time.
- GD —
- gender dysphoria
- HbA1c —
- hemoglobin A1c
- HDL —
- high-density lipoprotein
- LDL —
- low-density lipoprotein
- TG —
What’s Known on This Subject:
Existing data in transgender adults have shown associations between lower blood pressure levels and estrogen use as well as increasing BMI with testosterone; however, the effects of sex hormones in transgender adolescents have not been adequately studied.
What This Study Adds:
Testosterone use was associated with increased hemoglobin and hematocrit, increased BMI, and lowered high-density lipoprotein levels; estrogen use was associated with lowered testosterone and alanine aminotransferase levels. Otherwise, the findings support the short-term safety of cross-sex hormones in transgender adolescents.
Gender dysphoria (GD) is the condition encountered in individuals experiencing significant distress due to persistent cross-gender identification and discomfort with one’s biologic sex or gender role.1 Although the incidence and prevalence of GD in adolescents have been poorly studied, it has been shown that adolescents with GD are now seeking the care of medical and mental health providers at younger ages.2–4 The Endocrine Society guidelines state that adolescents who fulfill eligibility for gender reassignment may undergo treatment to suppress pubertal development upon first exhibiting the physical changes of puberty and may start cross-sex hormone therapy at 16 years of age.1 Affirmed female subjects assigned male at birth (male-to-female) may undergo feminizing therapy with estrogen to develop female secondary sex characteristics. Affirmed male subjects assigned female at birth (female-to-male) have the option of masculinizing therapy with testosterone to develop male secondary sex characteristics. After initiation of puberty induction, the Endocrine Society recommends anthropometric and laboratory evaluations at least every 3 months to monitor pubertal development.
The effects of estrogen and testosterone on lipoprotein metabolism are well known. Estrogen enhances cholesterol transport by raising triglyceride (TG) and high-density lipoprotein (HDL) levels while lowering low-density lipoprotein (LDL) levels. Meanwhile, testosterone has been shown to decrease prolactin, sex hormone–binding globulin, and HDL levels while increasing blood pressure, hemoglobin, and hematocrit levels.5 Existing data in transgender adults receiving long-term cross-sex hormones have associated estrogen with lower blood pressure and testosterone levels and increased BMI6; however, the effects of sex hormones on these parameters have not been adequately studied in adolescents. Risk factors for adult metabolic disease commonly emerge in adolescence, highlighting the importance of determining the effects of sex steroid administration in this population.7
The primary aim of the present study was to identify patterns in metabolic and cardiovascular parameters in transgender adolescents receiving cross-sex hormone therapy in centers located in Washington, DC, Baltimore, MD, and Cincinnati, OH.
This study analyzed preexisting databases maintained for the care of patients with GD at MedStar Washington Hospital Center and Children’s National Medical Center (both, Washington, DC). Outpatient records at the University of Maryland Medical Center (Baltimore, MD) and Cincinnati Children’s Hospital Medical Center (Cincinnati, OH) were also retrospectively reviewed. Each study site serves a substantial transgender adolescent population. All research methods were reviewed and approved by the respective institutional review board at each institution.
Outpatient data from adolescents aged 14 to 25 years diagnosed with GD (International Classification of Diseases, Ninth Revision codes 302.85 and 302.50) and receiving cross-sex hormone therapy from 2008 to 2014 were included in this study. Subjects were divided into 2 treatment groups: affirmed male subjects taking testosterone and affirmed female subjects taking estrogen with or without testosterone blockers (ie, spironolactone). Testosterone was most commonly initiated at a weekly dose of 25 mg, with effective weekly subcutaneous doses of 25, 50, or 100 mg noted at subsequent visits. Estrogen was administered more variably, given either orally at 1, 2, 3, 4, 6, and 8 mg daily; intramuscularly at 20, 40, or 80 mg monthly; or transdermally at 0.025, 0.05, 0.100, or 0.200 mg weekly. It was common practice to incrementally increase or adjust testosterone and estradiol doses at each follow-up visit, which accounts for the wide range of doses and administration methods. Previously recorded measurements of testosterone, estradiol, prolactin, and lipids (total cholesterol, LDL, HDL, TG, and TG/HDL ratio) were reviewed, as were the levels of electrolytes, liver enzymes, hemoglobin/hematocrit, and hemoglobin A1c (HbA1c). Anthropometric measurements (height/weight, BMI, and blood pressure) were also compared. Values reviewed were obtained at or immediately before initiation of therapy (baseline), at 1 to 3 months after initiation, at 4 to 6 months after initiation, and at 6 months and beyond.
Continuous variables are described by using means and SDs; categorical variables are described by using frequencies and percentages. Repeated measures analysis of variance models were used to evaluate changes in laboratory values and metabolic markers over time for each individual subject. SAS version 9.3 (SAS Institute, Inc, Cary, NC) was used to perform the analysis.
From 2008 to 2014, a total of 116 transgender adolescents were recorded as having received cross-sex hormone therapy; this total included 72 affirmed male subjects and 44 affirmed female subjects (Table 1). The mean age of the affirmed male and female subjects was 16 and 18 years, respectively. Depression was the most common medical comorbidity, with 35 subjects (30%) reportedly undergoing treatment for depression during hormone use. Ten (23%) affirmed female subjects were undergoing medical therapy for HIV; no concurrent HIV was reported among the affirmed male subjects.
The availability and completeness of follow-up data varied for each subject largely because of gaps in patient compliance and transfer out of practices, as well as provider differences regarding which measurements were recorded at each visit. Subjects with missing data points were excluded from the analysis for each time period in which that particular value of interest was unavailable. Baseline values served as reference points for statistical analysis where available. The longest follow-up time noted was 35 months.
Affirmed Male Subjects
For the 72 subjects taking testosterone, the mean baseline testosterone level was well within the adolescent female range (29.4 ng/dL); however, a number of subjects were noted to have levels above what is expected for adolescent female subjects, with a maximum baseline testosterone level of 75.0 ng/dL. Seven affirmed male subjects reported undergoing puberty suppression with gonadotropin-releasing hormone agonists before treatment, and 2 reported hormone use outside the practice of their medical providers (ie, street hormones). Mean testosterone levels significantly increased at each visit, falling within adolescent male levels by 6 months (462 ng/dL). Testosterone therapy was associated with increasing BMI, changing from a mean baseline BMI of 26.0 to 27.3 after 6 months (P < .0001). Mean baseline systolic and diastolic blood pressures were 118 and 71 mm Hg, respectively. A reduction in mean diastolic blood pressure was noted at 6 months (67 mm Hg; P = .02); however, values returned to baseline levels at subsequent visits. Otherwise, both systolic and diastolic blood pressures remained within normal limits.
Hematocrit levels increased significantly after initiation of testosterone from a mean baseline hematocrit of 39.4% to 44.5% after 6 months of therapy (P ≤ .0001). A simultaneous increase in hemoglobin levels was noted. Of note, 2 subjects had supraphysiologic hematocrit levels (>50%) after 3 months of treatment, whereas another subject maintained elevated hematocrit levels after 6 and 9 months (51.0% and 52.7%, respectively).
Mean baseline total cholesterol levels were within normal limits at 151 mg/dL, with a corresponding baseline LDL level of 84 mg/dL. There was an overall increase in mean total cholesterol and LDL levels noted after initiation of testosterone; however, these changes were not statistically significant and plateaued after 3 months. Six subjects had cholesterol levels >200 mg/dL over the course of study treatment. Subsequent measurements showed lower levels despite continued testosterone use. A significant decrease in mean HDL level over time was observed, from a mean baseline of 50.2 to 45.0 mg/dL after 6 months (Table 2). No significant changes in levels of aspartate aminotransferase/alanine aminotransferase, serum urea nitrogen/creatinine, estradiol, prolactin, TG, HbA1c, and TG/HDL ratio were observed.
Weekly testosterone doses ≥50 mg were associated with higher hematocrit levels and systolic blood pressure compared with 25-mg doses of testosterone. Similarly, HDL levels were found to be lower at testosterone doses ≥50 mg weekly.
Affirmed Female Subjects
Baseline estradiol and testosterone levels among the 44 subjects taking estrogen were within the expected range for pubertal male subjects, with a mean baseline estradiol level of 21.6 pg/dL and a mean baseline testosterone level of 391.7 ng/dL. Previous puberty suppression with gonadotropin-releasing hormone agonists was noted in 2 affirmed female subjects, whereas 5 subjects reported exogenous street hormone use. Physiologic levels of estradiol were achieved after 6 months, with mean estradiol levels of 96.4 ng/dL (P < .0001). Although mean testosterone levels significantly decreased after 3 months (256.3 ng/dL; P = .01), nearly all of the affirmed female subjects maintained levels above the pubertal female range within the time period included in the study. Mean baseline BMI was 23.7 and remained stable during treatment. No statistically significant changes in systolic and diastolic blood pressures were noted.
Baseline hemoglobin and hematocrit were within the male range (14.5 g/dL and 43.8%, respectively), with no significant changes noted during treatment. Estrogen was not associated with any changes in aspartate aminotransferase, total cholesterol, LDL, HDL, TG, or the TG/HDL ratio (Table 3). Alanine aminotransferase levels were found to have significantly decreased during therapy; however, the change was not clinically significant because the levels remained within normal range. Furthermore, there were no significant differences noted in these lipid parameters between subjects with testosterone levels adequately suppressed to pubertal female levels versus those whose levels remained above the female range. Prolactin levels increased from a mean baseline level of 12.0 to 20.7 ng/mL after 6 months. This change was not statistically significant (P = .18). Prolactin levels remained within the female range throughout the course of the study.
There was no statistically significant difference in measured metabolic parameters among the various methods of estrogen administration (patch, oral, or intramuscular). There was also no significant change in potassium levels among the 38 subjects taking spironolactone. However, 1 affirmed female subject had a sodium level of 132 mEq/L and a potassium level of 5.1 mEq/L while taking spironolactone 100 mg daily. Her care was complicated by multiple medical comorbidities, including highly active antiretroviral therapy for HIV and antibiotics for secondary syphilis. Electrolyte levels normalized after cessation of spironolactone and remained within the normal range after resumption at 50 mg daily. Spironolactone doses ranged from 50 to 200 mg daily.
Current data support the safety of cross-sex hormone therapy in the treatment of GD; however, the full extent of the effects of cross-sex hormones is not well known.8 Although studies have supported the appropriateness and efficacy of the medical treatment of GD in adolescents, less is known regarding the effects of cross-sex hormones in this population.9,10 This study found patterns in metabolic and cardiovascular parameters that support the short-term safety of cross-sex hormone therapy in adolescents.
The Endocrine Society guidelines recommend that adolescents with GD can start cross-sex hormone therapy at 16 years of age.1 The mean age of affirmed male and female subjects included in this study was 16 and 18 years, respectively, reflecting the increasingly more common practice of starting cross-sex hormones before 16 years of age. The prevalence of depression and anxiety among the adolescents in this study is consistent with previous data, highlighting the increased mental health burden in transgender adolescents compared with their peers.3 Although there are few data regarding the risk of HIV in transgender male subjects, transgender female subjects have been shown to have a greater risk of acquiring HIV than their peers,11 consistent with the findings in our cohort.
Affirmed Male Subjects
Testosterone in our adolescent cohort was associated with a gradual increase in BMI, consistent with previous data associating testosterone with increasing BMI in transgender adults.12 Although the role of testosterone in fat redistribution and muscle mass expansion has been reported, the exact role of testosterone in increasing BMI in transgender patients is unclear.8 Similar to previous data showing no significant change in blood pressure after 6 months of testosterone in adults, this study found no associations between testosterone and short-term changes in blood pressure.6
The mean baseline testosterone level was within the expected adolescent female range; however, levels were higher than expected for a number of the affirmed male subjects. Hyperandrogenic states frequently encountered in adolescent female subjects, including polycystic ovarian syndrome, were not taken into account in the present study. Furthermore, surreptitious testosterone use and self-prescription of masculinizing hormones have been shown to be more common in affirmed male subjects.13 These factors may have contributed to the higher testosterone levels in a number of participants before therapy.
Testosterone has been associated with increased hemoglobin and hematocrit levels in transgender adults.12 This study found a similar increase in hemoglobin and hematocrit in adolescents. In Jacobeit et al,12 testosterone was associated with decreased plasma cholesterol and LDL levels, with no changes in HDL levels in affirmed male adults. Conversely, Meyer et al reported a small but significant increase in total cholesterol levels in affirmed male adults followed up at 3- to 6-month intervals.14 This study noted only a statistically significant decrease in HDL level among affirmed male subjects taking testosterone. Longer term studies and larger cohorts are needed to determine the full effect on lipids in transgender adolescents receiving testosterone, especially given the inconclusive changes noted in adults.
These findings support the monitoring of hematologic and metabolic parameters in affirmed male adolescents as early as 3 months after initiation of therapy, as significant changes may be evident at that time. This study also suggests that frequent laboratory surveillance beyond blood counts and lipids may be of little benefit because no significant changes were found beyond these parameters to support this practice. Because some evidence points to an increased prevalence of polycystic ovarian syndrome and insulin resistance in transgender adults taking testosterone, providers should keep in mind the future metabolic implications of testosterone in adolescents.15 This study also suggests that closer surveillance may be warranted at doses ≥50 mg weekly because effects may be more apparent than at lower doses.
Affirmed Female Subjects
Estrogen therapy in adolescents was not associated with significant changes in BMI and blood pressure, consistent with adult data noting minimal change in weight, BMI, and blood pressure after 6 months of estrogen administration.6 Similar to testosterone, the role of estrogen in the redistribution of fat from the viscera to the subcutaneous tissue has been reported; however, its exact role in weight and BMI has not been established.16
Baseline estradiol and testosterone levels were within the range for pubertal male subjects, albeit on the lower end of expected levels for testosterone. Testosterone levels decreased in all affirmed female subjects after initiation of estrogen, likely secondary to suppression of gonadotropin release leading to decreased endogenous testosterone production. No significant change in hemoglobin/hematocrit, liver function, or serum urea nitrogen/creatinine was noted. These findings are similar to previous literature describing transgender adults taking estrogen.17 Some retrospective data suggest that estrogen therapy in transgender adults is associated with TG levels higher than in both male and female adults. Our study found no significant changes in TG, total cholesterol, LDL, or HDL levels in affirmed female adolescents regardless of whether testosterone was suppressed to pubertal female levels. These findings are comparable to longitudinal data in adults showing no changes in serum lipid levels after 6 months of estrogen administration.15
Spironolactone was not shown to be associated with changes in potassium levels in the doses used by the patients in this study. An isolated but clinically important event involving hyponatremia and mild hyperkalemia was observed in 1 affirmed female subject whose care was complicated by multiple medical comorbidities, which may have contributed to these metabolic derangements. It is within reason to conclude that, in the absence of preexisting medical conditions and medications, hyperkalemia and hyponatremia are unlikely with feminizing therapy with spironolactone.
These results indicate that affirmed female adolescents taking estrogen may not require laboratory testing as frequently as their affirmed male counterparts, at least in the short-term. Closer surveillance may be more prudent in transgender adolescents with multiple medical issues complicating hormone therapy. There is also compelling literature to suggest that there is a higher prevalence of venous thrombosis, myocardial infarction, and cardiovascular disease in transgender adults undergoing long-term estrogen therapy, highlighting the importance of clinical follow-up and lifestyle modification over laboratory testing beginning from adolescence.18
To the best of our knowledge, the present study involved the largest cohort of transgender adolescents to assess the short-term effects of cross-sex hormone therapy. No studies have previously examined the metabolic effects of such therapy in this population. As hormone therapy for GD becomes standard of care at an increasingly younger age, data such as those presented in the present study will prove valuable in ensuring the safety of medical intervention in transgender adolescents. Having 4 clinical sites allowed for a greater breadth of data accessible for review; however, this design comes with its own set of limitations. Baseline values were not recorded in some participants, and some laboratory parameters were either not measured or not readily available for some follow-up visits. This scenario was due, in part, to differences in practice among the providers across the sites as to what testing was performed, especially because some of the visits reviewed predated the publication of the Endocrine Society guidelines in 2009. Financial considerations, such as lack of insurance coverage, limited the availability of laboratory testing. Accordingly, the limited availability of data for some parameters at certain time points may have underpowered the study to show statistical significance or otherwise.
Because this study was a retrospective analysis, it was difficult to control for the doses and methods of administration of both testosterone and estrogen, accounting for the wide range of treatment doses. This study allowed only assessment of the effects of exposure to testosterone and estradiol. It was not designed to identify any dose-dependent effects that would intuitively follow treatment dose adjustments. There was also no way to control for any medical confounders affecting laboratory results, such as depression, anxiety, or HIV. Certain medications such as antipsychotic drugs and psychotropic agents have been associated with an increased risk of developing obesity and diabetes in adolescents.19 Although use of these medications was not controlled for in the present study, any potential effects on laboratory testing were mitigated by using each patient as his or her own control and looking at individual data over time. Previous and continuing exposure to masculinizing or feminizing hormones outside the setting of a medical provider was also difficult to ascertain retrospectively and may have caused aberrations in recorded laboratory values. Transgender adolescents receiving cross-sex hormone therapy are repeatedly counseled regarding the importance of diet and exercise. It would be difficult to ascertain to what degree lifestyle modifications affected results in laboratory testing.
Finally, this study used BMI as a clinical indicator of potential metabolic change while receiving cross-sex hormone therapy. Waist circumference, which was not available for analysis, has been shown to be a better predictor of metabolic risk in adolescents.20 Nevertheless, the findings of this study support the short-term safety of cross-sex hormone therapy in transgender adolescents. This patient population will benefit from larger scale prospective studies to determine the long-term risks of cross-sex hormone therapy in adolescents.
Funding for statistical analysis was supported by the Graduate Medical Office of MedStar Washington Hospital Center. We thank our colleagues from MedStar Washington Hospital Center, Children’s National Medical Center, the University of Maryland, and Cincinnati Children’s Hospital Medical Center who provided insight and expertise that greatly assisted the research, although they may not agree with all of the interpretations/conclusions of the paper. We thank Leah Orta-Nieves, MD, for her work during the formative stages of the protocol.
- Accepted February 14, 2017.
- Address correspondence to Jason Jarin, MD, Obstetrics and Gynecology, University of Texas Southwestern, 5323 Harry Hines Blvd, Mail Code 9032, Dallas, TX 75390. E-mail: email@example.com
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: No external funding.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
- Zucker KJ
- Asscheman H,
- Giltay EJ,
- Megens JA,
- de Ronde WP,
- van Trotsenburg MA,
- Gooren LJ
- Bunck MC,
- Toorians AW,
- Lips P,
- Gooren LJ
- Weinand JD,
- Safer J
- Jacobeit J,
- Gooren L,
- Schulte H.
- Baba T,
- Endo T,
- Honnma H, et al
- Elbers M,
- Asscheman H,
- Seidell C,
- Gooren J
- Copyright © 2017 by the American Academy of Pediatrics