Cardiometabolic Benefits and Risks for Transgender Youth with the Initiation of Gender-Affirming Hormone Therapy: a case series

Abstract

Background.

Interventions for transgender youth include puberty suppression and gender affirming hormone therapy (GAHT). Studies have begun to show changes in cardiometabolic outcomes in adolescent transgender males (TGM) and females (TGF). We hypothesized that in TGM and TGF youth, GAHT would increase total cholesterol (TC), LDL cholesterol (LDL-C), insulin resistance (IR) and adiposity, while decreasing HDL cholesterol (HDL-C) and endothelium-dependent vasodilation.

Methods.

We measured IR (oral glucose tolerance test), lipid profile and endothelial function (flow mediated vasodilation) prior to and following six months of GAHT in TGM (n=3) and TGF (n=2).

Results.

This case series shows attenuated endothelium-mediated vasodilation in TGM, but improved endothelium-mediated vasodilation in TGF. Our data also support earlier findings showing reduced HDL-C with GAHT in TGM. Two of three TGM showed increased while one TGF showed decreased IR following GAHT.

Conclusion.

These data indicate rapid, but divergent, changes in lipids and endothelial function in TG youth with GAHT.

NEW AND NOTEWORTHY

While these data are preliminary, they show changes in FMD in transgender youth and suggest divergent responses, indicating attenuated FMD in and lower HDL-C in males, while some improvement in endothelial function and little impact on lipids in females.

INTRODUCTION

The practice of gender affirming hormone therapy (GAHT) is increasing in pediatric patients (1). In the United States approximately 0.7–1.8% of youth identify as transgender (2). Gender affirming hormone therapy reduces natal sex characteristics and induces those of the desired sex. GAHT is the primary medical intervention sought by transgender people and is recognized by multiple academic medical organizations, including the Endocrine Society, Pediatric Endocrine Society (3), and the World Professional Association for Transgender Health as medically necessary. The Endocrine Society (4) and World Professional Association for Transgender Health (5) guidelines recommend an approach of the timing of initiation of GAHT based on the individual’s emotional and intellectual maturity, capacity to understand risks and benefits, and balancing potential benefits of early medical intervention on physical and mental health outcomes. In transgender male (TGM) and female (TGF) youth, the goals of therapy are to achieve estrogen and testosterone levels comparable to their cis gender (CG) counterparts, achieve the physical and physiological changes that are consistent with their affirmed gender, and optimize safety.

In adults, some studies have found that GAHT in both transgender men and women can increase cardiometabolic risk (6–8), including dyslipidemia, increased blood pressure (BP) (9–11) and compromised endothelial function (8). In transgender women, GAHT can increase thrombosis risk (7). However, in both sexes these lipid and vascular outcomes are variable across studies (8, 12, 13), estrogen may improve vascular function in males (12) and show little change in insulin resistance (IR) associated with GAHT in adults (14). Several studies in transgender youth have shown increased low density lipid cholesterol (LDL-C) (6, 9), and IR (7) in adolescent TGM and TGF. However, recent investigations demonstrated little impact of testosterone therapy on IR or endothelial function in TGM youth, and that these effects may be associated with puberty blocking (15, 16).

Our goal was to assess the cardiometabolic effects of GAHT on TGM and TGF over the early period of GAHT. We tested the hypothesis that GAHT would increase serum total cholesterol (TC), LDL-C, IR and adiposity, while decreasing HDL cholesterol (HDL-C) and endothelium-dependent vasodilation in TGM and TGF youth at six months of GAHT. In this case series, we present data on three TGM and two TGF individuals and three cisgender (CG) adolescents for comparison. While these conclusions are preliminary, we consider it important to share our findings.

MATERIALS AND METHODS

Ethical Approval.

All participants (and in the case of minors, a parent/guardian) gave written informed consent to participate, which conformed to the standards set by the Declaration of Helsinki. The study was approved by the Human Investigation Committee of Yale School of Medicine (HIC Protocol no. 2000030999).

Participants

We recruited three TGM, two TGF and three CG individuals (2 females, 1 male), all in Tanner Stage 5, age 16–19 yrs. All our participants identified as non-Hispanic White, indicated no health conditions when interviewed about their medical history, and females reported regular menstrual cycles (26–32 days) with no gynecological abnormalities. We excluded participants who smoked or vaped, had diabetes or hypertension, or were taking insulin sensitizing, cardiovascular medications or hormonal contraception for at least four weeks prior to participating in these protocols.

Experimental Design

We tested TGM and TGF at baseline (BSL), prior to starting and following six months of GAHT; CG came back to the laboratory six months after their first test. BSL TGM and CG were tested in females in the early follicular phase of their menstrual cycle. The GAHT medication regimens varied among the participants depending on their treatment goals (4). All TGM subjects were treated with a GnRH agonist for menstrual suppression (beginning with hormone treatment), and one TGM also received norethindrone. Testosterone doses were 38 mg weekly in one TGM and 44 mg weekly in the other two TGM. Both TGF received GnRH agonist (beginning with hormone treatment); estradiol doses were 1 mg daily in one participant and 2 mg in the other.

Oral glucose tolerance test

Participants reported to the laboratory in the morning after an overnight fast. We placed an intravenous catheter in an antecubital vein with the participant semi-recumbent. Following baseline (−15, 0 min) samples for glucose and insulin, the participant ingested 75 g of glucose; glucose and insulin samples were then collected at 30, 45, 60, 90, 120- and 180-min.

Physiology Studies

On a separate day from the OGTT, participants reported to the laboratory in the morning after an overnight fast. The participants avoided alcohol and strenuous activity/exercise during the 24 hours prior to the studies. We confirmed hydration status with a urine sample (urine specific gravity, USG ≤ 1.020). Height was recorded, and body mass measured on a beam balance.

Body Composition

We used the BodPod^® (Cosmed, Rome, Italy) to determine body fat and lean mass. The BodPod^® is an air displacement plethysmograph that uses whole body densitometry to determine body composition (fat vs. lean). The BodPod^® is based on the underwater weighing technique (17) but uses air instead of water to measure body volume and is comparable to Dual x-ray absorptiometry and reliable across multiple days under the same conditions (18).

Endothelial Function

We used flow mediated vasodilation (FMD) to assess brachial artery endothelial function. Prior to baseline measures, participants performed three isometric maximal voluntary contractions using a handgrip dynamometer to determine maximum voluntary contraction (MVC). Continuous measures of brachial artery diameter and blood flow velocity were simultaneously measured using the Sonoscope S2 ultrasound imaging system. After 15 minutes of supine rest, a two-minute longitudinal image of the brachial artery was recorded. Forearm ischemia was then induced by rapidly inflating a blood pressure cuff to supra-systolic pressure (i.e. ≥160–180 mmHg). Upon cuff release, blood flow increases to the ischemic tissue (reactive hyperemia, RH-FMD). The increase in shear stress associated with this rapid blood flow increase stimulates nitric oxide release from endothelial cells, causing vasodilation. We recorded the change in brachial artery diameter 15 seconds prior to cuff deflation and continuously thereafter for 2 minutes. Brachial artery measures were then obtained continuously for 3 minutes following cuff deflation. Participants completed a standard 5-minute occlusion followed by reactive hyperemia, and an enhanced stimulus of a 3-minute occlusion concomitant with ischemic handgrip exercise at 20% MVC (IE-FMD) (19). The order of the RH-FMD and IE-FMD protocols were randomized across participants, and participants rested supine for 15 minutes between trials.

Data Analysis for FMD

Machine-learning software (Brachial Analyzer for Research, Medical Imaging Applications, USA) was used to analyze ultrasound video files and %FMD was calculated using standardized protocols and equations as described Adler et al (19).

Microvascular endothelial function

Following FMD measurements, we used local forearm skin heating combined with Laser Doppler flowmetry (LDF) to measure skin blood flow (SkBF). Participants rested semirecumbent for at least 30 min during instrumentation for beat-to-beat BP measures with a Finometer Pro (Finapres, Amsterdam, The Netherlands). Blood pressure and SkBF were recorded with a 16-channel computerized data-acquisition system at a sampling Speed of 400 Hz (ADI instruments PowerLab 9, Castle Hill, Australia). We measured resting SkBF using two LDF probes equipped with local heaters, which both control and measure local skin temperature. After 10 min of rest, the temperature of the local heaters was increased to 42°C to induce local vasodilation for 30 min. After this heating period, the temperature of the local heaters was increased to a local skin temperature of 43°C, where it was clamped ~10 min combined with transdermal-nitroglycerin patch (0.6 mg/hr) to induce maximal vasodilation. Cutaneous vascular conductance (CVC=LDF/MAP) was calculated and normalized to %CVC_max (20).

Blood parameters

We collected blood samples for 17 β-estradiol (E₂), testosterone, glucose, insulin, hemoglobin A1c (HbA_1C), LDL-C, HDL-C, and TC at the time of the baseline sample for the OGTT. An aliquot was transferred into a tube without anticoagulant for the determination of lipids and T (S_[T]) and E₂ (S_E2). S_[T], S_E2, using LC/MS; TC, HDL-C were analyzed with enzymatic colorimetric methods (Department of Laboratory Medicine, Yale School of Medicine). LDL-C was calculated using the Martin-Hopkins calculation (21). Lipid profiles were not collected for CG. Blood for HbA_1C was collected in EDTA and assayed using immunoturbidimetry.

Data analysis

Due to the low sample size, we show individual data points in graphs and effect sizes using Cohen’s d_z for paired measures (the mean of the difference scores/standard deviation the difference scores, G-Power 3.1, Heinnrich Heine Universtat, Dusseldolf). We did not calculate d_z in TGF because our sample of n=2 did not allow standard deviation calculation. Total and incremental areas under the glucose and insulin response curves over 180 min were calculated by the trapezoidal method.

RESULTS

We collected complete data on two CG, three TGM and two TGF (Table 1). The changes in FMD were variable in CG participants on follow-up (d_z=0.74 ± 0.98, Fig 1) with two of three TGM participants showing declines in FMD (d_z=1.79 ± 1.22) during testosterone treatment, while one TGM showed no change. In contrast, both TGF showed increased FMD (Δ 7.1, 6.4). Skin blood flow showed similar trends in both the TGM and TGF, although one TGM (the same TGM who showed no change in FMD) showed increased peripheral endothelial mediated vasodilation (Fig 1). Area under the curve for IR appeared increased in two of three of our TGM (from 6699 to 15702, and 24835 to 38089 IU) and decreased in our TGF (from 13907 to 6926 IU) (Fig 2). Finally, there was little consistent change in lipid profile across participants except HDL-C decreased in TGM (Table 1).

Table 1. Subject Characteristics.

Mean (± SD) data are from the oral glucose tolerance test (OGTT) experiment in Cisgender Controls, Transgender male and female participants. Body mass index (BMI), glucose and insulin are area under the curve (AUC) are from the OGTT testing day. Systolic (SBP), Diastolic (DBP), and Mean (MAP) arterial pressures as measured by Finipres during skin blood flow measurements. Hemoglobin A1c (HbA1c), total cholesterol (TC), low density lipoproteins (LDL-C), high density lipoproteins (HDL-C) are from the OGTT testing day. Fat and lean weight are based on measurements from the BodPod^® measurement.

	Cisgender n=3		Transgender male n=3		Transgender female n=2
	Pre	Post	Pre	Post	Pre	Post
Age, y	18 ± 1		17 ± 2		16, 17
Height, cm	171.9 ± 16.9		164.3 ± 6.4		175.3, 180.3
Weight, kg	67.3 ± 9.3	64.8 ± 7.6	65.6 ± 22.0	70.0 ± 17.4	63.5, 68.8	61.2, 62.7
BMI kg/m2	23.0 ± 3.1	22.2 ± 2.7	24.2 ± 7.4	25.6 ± 7.0	20.7, 21.1	19.9, 19.3
Fat weight, kg	18.5 ± 13.3	16.0 ± 7.2	24.2 ± 13.8	22.9 ± 11.2	7.7, 14.1	10.4, 9.6
Lean weight, kg	48.8 ± 13.5	48.9 ± 8.4	41.2 ± 8.2	46.2 ± 6.9	55.9, 54.4	50.8, 53.2
% Body fat	27.3 ± 17.5	24.6 ± 9.8	34.7 ± 10.8	31.5 ± 11.2	12.2, 20.6	17.1, 15.2
HbA1c, mmol/mol			4.9 ± 0.9	5.2 ± 0.6	5.4, 5.2	5.6,5.1
SBP, mm HG	114 ± 6	113 ± 10		117 ± 6	127 ,120	138,117
DBP, mm HG	61 ± 5	60 ± 10	67 ± 3	60 ± 3	72,64	70,60
MAP mm HG	79 ± 5	78± 10	82 ± 4	79 ± 3	90,80	93,79
AUC glucose, g/dl	18030, 24413	21405,26318	23177 ± 2534	26601 ± 3662	20407,18615	21413, 18420
AUC insulin, UI	7153, 29429	7706,24944	22458 ± 7781	28749 ± 3983	Missing,13907	5718, 6926
TC, mg/dl			138, 173	170,123	171,165	111, 162
LDL-C, mg/dl			52, 108	115,70	99, 104	63, 95
HDL-C mg/dl			72, 82	41, 35	59, 51	37, 58

Figure 1: Flow mediated vasodilation (FMD, % change) and changes in skin blood flow (SkBF) expressed as relative cutaneous vascular conductance (%CVC_max) in Cisgender (CG) Controls, Transgender male (TGM) and female (TGF).

Data show individual values pre and post six months of gender affirming hormone therapy (GAHT) in the transgender participants and repeat testing in the cisgender controls (Pre, Post).

Figure 2. Area under the curve (AUC) for insulin changes during an oral glucose tolerance test in Cisgender (CG) Controls, Transgender male (TGM) and female (TGF) participants.

Data show individual values pre and post six months of gender affirming hormone therapy (GAHT) in the transgender participants and repeat testing in the cisgender controls (Pre, Post).

DISCUSSION

An earlier study (15) using a larger population over twelve months demonstrated that decreased endothelial function with testosterone administration to TGM youth was dependent on blocking puberty using a GnRH analog. Our case series supports these findings in that all pediatric transgender participants were taking GnRH analogs with testosterone treatment and showed changes in brachial artery and peripheral microcirculatory endothelial function. However, we saw such changes despite our subjects having passed through puberty. Endothelial dysfunction is a predictor of atherosclerosis, hypertension and diabetes, and an important marker for clinical cardiovascular risk; thus, a reduction in endothelial function is striking to be noted in these young TGM after only six months of GAHT. In contrast, we saw improved endothelial function in our TGF participants similar to longer term TGF (12). Endothelial mediated vasodilation in the brachial artery (22, 23) and microcirculation in cisgender females and males occurs primarily through nitric oxide (NO) dependent mechanisms (20). Thus, the providing estrogen to the TGF may have contributed to vasodilation during FMD by increasing availability of NO. Studies that determine the mechanism involved in the impact of estrogen on male vasculature has not been explored in humans, although estrogen appears to induce vasodilation in both males and females in a mouse model through an NO pathway (24). In both TGM and TGF we saw similar changes in the peripheral microvasculature as those noted changes in FMD, indicating that changes in the brachial artery were reflected in the peripheral microvasculature with GAHT.

Our findings in these younger participants are consistent with earlier data showing decreased HDL-C associated with GAHT in TGM. We did not see consistent changes in other cardiometabolic variables, including TC or LDL-C also consistent with a recent report (15). Earlier studies have shown adverse effects of testosterone on lipid levels and blood pressure in adults (9, 25) and youth (26), raising concern for increased cardiometabolic risk with chronic testosterone exposure in TGM. However, while adult TGM lipid levels may increase with testosterone treatments they often stay within normal limits (8) and typically are not associated with IR (14).

Our case series supports cross-sectional studies showing early changes in HDL-C. In these earlier studies, dyslipidemia appears driven by the androgen therapy (27) and is the most consistent finding across studies (9, 28). Testosterone exposure is also a risk factor for dyslipidemia in cisgender females (29). Importantly, our we found similar changes in lipids in this younger population than found in earlier studies in young TGM in which puberty was blocked using a GnRH analog (15). In the present study, although we used a GnRH analog, all the subjects were post-puberty, and the TGM were menstruating regularly prior to beginning the study. It bears noting that other factors, such as BMI, family history of cardiometabolic risk, social determinants of health, and acute and chronic mental health stressors may also contribute significantly to the risk of adverse cardiometabolic outcomes (7, 11, 30).

There is thus a delicate balance to the determination of timing for initiation of GAHT: definitive larger-scale clinical trials in youth have clearly demonstrated that earlier initiation of GAHT can improve mental health outcomes including depression, anxiety and scores of body congruence (31), but increases the duration of exposure to the potentially adverse effects of hormone therapy, compared to starting in adulthood. Our case series suggests that the impact of GAHT on both TGM and TGF appear mild in the early months, but it is crucial to understand the cardiovascular and metabolic changes associated with GAHT in youth and evaluate optimal monitoring and intervention strategies, including exercise and medication as needed, to optimize health and safety in youth undergoing GAHT.