Anna Wiik, Tommy R Lundberg, Eric Rullman, Daniel P Andersson, Mats Holmberg, Mirko Mandić, Torkel B Brismar, Olof Dahlqvist Leinhard, Setareh Chanpen, John N Flanagan, Stefan Arver, Thomas Gustafsson, Muscle Strength, Size, and Composition Following 12 Months of Gender-affirming Treatment in Transgender Individuals, The Journal of Clinical Endocrinology & Metabolism, Volume 105, Issue 3, March 2020, Pages e805–e813, https://doi.org/10.1210/clinem/dgz247
As many sports are divided in male/female categories, governing bodies have formed regulations on the eligibility for transgender individuals to compete in these categories. Yet, the magnitude of change in muscle mass and strength with gender-affirming treatment remains insufficiently explored.
This study explored the effects of gender-affirming treatment on muscle function, size, and composition during 12 months of therapy.
In this single-center observational cohort study, untrained transgender women (TW, n = 11) and transgender men (TM, n = 12), approved to start gender-affirming medical interventions, underwent assessments at baseline, 4 weeks after gonadal suppression of endogenous hormones but before hormone replacement, and 4 and 12 months after treatment initiation.
Knee extensor and flexor strength were assessed at all examination time points, and muscle size and radiological density (using magnetic resonance imaging and computed tomography) at baseline and 12 months after treatment initiation.
Thigh muscle volume increased (15%) in TM, which was paralleled by increased quadriceps cross-sectional area (CSA) (15%) and radiological density (6%). In TW, the corresponding parameters decreased by –5% (muscle volume) and –4% (CSA), while density remained unaltered. The TM increased strength over the assessment period, while the TW generally maintained their strength levels.
One year of gender-affirming treatment resulted in robust increases in muscle mass and strength in TM, but modest changes in TW. These findings add new knowledge on the magnitude of changes in muscle function, size, and composition with cross-hormone therapy, which could be relevant when evaluating the transgender eligibility rules for athletic competitions.
Competitive sports have been divided primarily by the concepts of male/female identity. However, these traditional athletic categories do not account for transgender persons who experience incongruence between the gender assigned at birth and their experienced gender identity. The question of when it is fair to permit transgender persons to compete in sport in line with their experienced gender identity is a delicate issue given the desire to ensure fair, safe, and meaningful competition while at the same time protecting transgender individuals’ rights and autonomy (1–4). This has been a highly controversial topic, not least after the recent International Association of Athletics Federations regulations on testosterone limits for the female category (5,6). The International Olympic Committee stated in 2015 that transgender men (TM, previously termed female-to-male) are allowed to compete in the male category without restrictions, while transgender women (TW, previously termed male-to-female) must have testosterone levels below 10 nmol/L for at least 12 months prior to competition (7). The International Association of Athletics Federations in turn recently stated that the testosterone levels should be below 5 nmol/L for 12 months, and that their expert medical panel will make a comprehensive review of the athlete’s case and medical treatment following sex reassignment (8).
One obvious problem is that there is essentially no data on the magnitude of change in performance indicators with gender-affirming medical treatment. A meta-analysis reported that TM on average gain 3.9 kg of lean body mass whereas TW lose 2.4 kg during the course of 12 months of cross-sex hormone therapy (9). Evidence of lower-limb muscle size changes has been provided by a few studies of the first 12 months of treatment in transgender individuals undergoing hormone therapy. A substantial increase in muscle mass (10–19%) with testosterone administration was reported in TM (10–12), while TW undergoing testosterone suppression and estrogen treatment experienced a 9% reduction in total thigh muscle cross-sectional area (CSA) (10,12). To date, however, there is a complete lack of data on changes in lower-limb muscle strength during gender-affirming medical interventions. The only strength-related parameter previously explored was hand-grip strength in a cohort of TW undergoing 2 years of cross-hormone treatment (13). After the 2-year treatment, hand-grip strength had decreased by 9.5% (from 42 to 38 kg), but there was strikingly little change in lean body mass. Given the established importance of muscle mass and strength, in the lower-limbs in particular, in numerous sports (14,15) further assessments of performance indicators, as well as comprehensive muscle size and quality measures, are highly warranted.
Accordingly, in the current study, we comprehensively examined the effects of gender-affirming treatment, which included inhibition of endogenous sex hormones and subsequent replacement with cross-sex hormones, on muscle function, size, and composition in both TW and TM.
Materials and Methods
This study was part of a single-center observational cohort study (16). Examinations were conducted at 4 time points (Fig. 1): baseline (T00), 4 weeks after initiated gonadal suppression of endogenous hormones but before hormone replacement (T1), 3 months after hormone replacement therapy (T4), and 11 months after the start of hormone replacement therapy (T12). Each time point was divided into 2 separate examination days. On the ﬁrst day, the participant came to the laboratory in the morning after an overnight fast. After at least 5 minutes of rest, blood samples were collected for standard blood biochemistry including sex hormones. On the second day, muscle strength was evaluated using isokinetic dynamometry. During the T00 and T12 visit, the participants also underwent a computed tomography (CT) scan of the lower limbs followed by a whole-body magnetic resonance imaging scan. Levels of physical activity (min/week and type of activity) were assessed through questionnaires.
The study population consisted of 12 TM (age 25 ± 5 years, height 168 ± 5 cm, body mass 66 ± 19 kg) and 11 TW (age 27 ± 4 years, height 180 ± 5 cm, body mass 70 ± 10 kg) that were referred to Andrology Sexual Medicine and Transgender Medicine (ANOVA), at the Karolinska University Hospital (Stockholm, Sweden), for evaluation of gender dysphoria. At the time of inclusion, all participants had been accepted to start gender-afﬁrming medical interventions. Further eligibility criteria can be found in our previous publication (16) and at clinicaltrials.gov (identifier NCT02518009). The participants were informed that their decision to participate or not, or the withdrawal of consent to participate, would not in any way change their treatment. Oral and written informed consent was obtained from all participants. The regional ethics review board in Stockholm (now called the Swedish Ethical Review Authority), approved the study (2014/409–31/4). In one of the data graphs, data from a cohort of recreationally active cisgender men (CM, n = 17, age 27 ± 5 years, height 181 ± 9 cm, body mass 87 ± 19 kg) and cisgender women (CW, n = 14, age 26 ± 4 years, height 163 ± 4 cm, body mass 67 ± 10 kg) are shown for reference comparison. These cisgender individuals were previously assessed in our laboratory using the same measurements techniques (17).
The reversal of the endocrine environment was initiated with injection of a gonadotropin releasing hormone (GnRH) antagonist (degarelix 240 mg subcutaneously). This strategy is known to result in an immediate reduction in gonadotropin secretion and to bring sex hormone levels (estradiol and testosterone) to castrate levels within 24 hours (18). The subsequent cross-hormone treatment was started after postcastration assessments were made. TM were treated with testosterone injections (testosterone undecanoate 1000 mg intramuscularly (IM)) with the first 2 injections given with a 6-week interval and thereafter 1 injection every 10th week, with dose adjustments in order to maintain androgen levels within the normal adult male reference range. Further gonadotropin suppression was maintained with a GnRH analogue administered IM every third month. In TW, estradiol was administered transdermally (gel or patches), p.o., or in a few cases intramuscularly (estrogenpolyphosphate). The estradiol doses given were either 1 to 2 mg as gel applied daily, 100 to 200 μg delivered every 24 hours with a patch, 4 to 8 mg orally, or 80 mg IM every 2 to 4 weeks. The different routes of delivery used during the study reflect continuous changes in clinical practice.
Whole-body magnetic resonance imaging
Body composition was determined by magnetic resonance imaging of the whole body at baseline and after 11 months of cross-sex hormone treatment (T12). Each subject underwent a whole-body magnetic resonance scan in the supine position (arms by their sides) using 2-point Dixon fat/water separated sequences on a 1.5-Tesla magnetic resonance platform (Siemens Aera, Siemens Healthcare, Erlangen, Germany) with the following settings: slice thickness between 3.5 and 4.5 mm, repetition time 6.69, echo times 2.23 and 4.77 ms, number of averages 1, resolution 0.448 pixels per millimeter. The total acquisition time was <10 minutes. After image processing, automated image analysis was performed using segmentation software provided by AMRA™ (AMRA Medical AB, Linköping, Sweden). This analysis allowed for automated segmentation and quantiﬁcation of thigh skeletal muscle volumes (19,20), as well as total adipose tissue (TAT) and total lean tissue (TLT) volumes measured from the most distal point of the thigh muscle up to vertebrae T9, excluding arms (21). The quantification was followed by visual quality review by a trained operator (22).
Computed tomography scan
The CSA and radiological density of the thigh muscles were assessed by bilateral CT scans covering the lower limbs at baseline and T12. To minimize the eﬀect of ﬂuid shifts, the participants rested in the supine position for 30 minutes before the scan (23). Preliminary scout images were obtained to ensure accurate positioning. All scans were obtained using a second-generation 64-slice dual source CT system (SOMATOM Deﬁnition Flash, Siemens Healthcare, Forchheim, Germany) operating at 120 kV and a ﬁxed ﬂux of 100 mA. The area of interest was manually deﬁned on 5-mm-thick axial slices and measured using manual planimetry with associated imaging software (Image J, National Institutes of Health, Bethesda, MD). The image selected for analysis was from the right leg, 50 mm proximal to the image where the muscle belly of rectus femoris disappeared.
Muscle strength assessment
Isokinetic and isometric peak torque were determined for the knee extensors and ﬂexors using isokinetic dynamometry (Biodex System 4 Pro, Medical Systems, Shirley, NY) at all 4 time points. The chest, hip, and thigh were stabilized using straps, and the ankle was strapped to the lever arm, which was aligned with the axis of rotation of the knee joint. Peak isometric (0°/s) and isokinetic torque (60°/s and 90°/s) were measured for both the right and left leg. The participant performed 3 maximal repetitions for each leg, alternating between knee extension and flexion, at each angular velocity. A 30-second rest period was employed between the trials. For the isometric test, performed at the ﬁxed knee angle of 120° (ie, 60° from full extension), the participants were instructed to apply as much force as possible for 5 seconds. Three trials were administered for each leg interspersed with 30 seconds of recovery.
The participants rested for at least 5 minutes prior to the blood collection. All blood samples were taken by conventional clinical procedures (sitting position, no tourniquet around the arm) from the antecubital vein. All samples were analyzed at the Karolinska University Laboratory using established procedures. The hemoglobin measurement was based on photometry measurement of oxyhemoglobin at 2 wavelengths. Levels of testosterone and estradiol were measured through liquid chromatography coupled with mass spectrometry.
Data were analyzed using a 2-way repeated measures analysis of variance with 1 group factor (TM vs TW). The changes over time were then compared in graphs displaying the mean change with 95% confidence intervals (CI). Selected absolute and height-adjusted mean values with 95% CI were also compared with the cisgender controls. Here, we conducted independent t-tests to compare the TW12 timepoint with the TM12, CM, and CW examinations. Lastly, Pearson’s R value was computed in a linear regression analysis displaying the relationship between muscle volume and strength. The significant alpha level was set at 5%. All statistical analyses were performed using Prism 7 for Mac OSX (GraphPad Software Inc, San Diego, CA). Data are presented as means ± standard deviation (SD) unless otherwise stated. Two participants in the group of TM missed the T4 assessment for muscle strength. Data points for these 2 participants were imputed in the main statistical analysis by taking the average value of the T1 and T12 timepoint. While we performed all lower-limb muscle volume and strength assessments for both the right and the left leg, the results confirmed that the 2 limbs responded very similarly and could be seen as replicates of each other. Thus, to reduce the overall volume of data presented, we report the average right/left data in the results and graphs.
Participant characteristics and blood biochemistry
In the TM, body mass was 66 ± 19 kg at T00 and 66 ± 10 kg at T12 (P > .05). In the TW, body mass was 70 ± 10 at T00 and 73 ± 10 at T12 (P = .10). TAT volume increased from 12.7 ± 6.8 to 16.8 ± 6.0 L in TW (P = .014). In contrast, TAT decreased in TM from 20.1 ± 11.2 to 16.9 ± 9.4 L (P = .043). TLT volume increased in TM from 19.1 ± 3.1 to 21.6 ± 3.0 L (P < .0001). In TW, TLT did not change significantly (P = .15). Values were 25.0 ± 2.4 L at T00 and 24.5 ± 2.3 at T12. The reported levels of physical activity did not change during the intervention (P > .05 in both groups). The reported activity consisted mainly of low to moderate intensity aerobic activities such as brisk walking. The TM reported 216 ± 67 min/week of physical activity at T00 and 246 ± 128 min/week at T12. The corresponding values for the TW were 236 ± 206 and 230 ± 163 min/week. The blood concentrations of hemoglobin, testosterone, and estradiol during the course of the study are represented in Fig. 2.
Lower-limb muscle volume, CSA, and radiological attenuation
There were group × time interactions (all P < .0001) for all muscle volume measurements (Fig. 3). The volume of the anterior thigh (15%), posterior thigh (15%), and total thigh (15%) increased (P < .0001) in TM. In contrast, these muscle volumes decreased (P < .01) during the intervention in TW: anterior thigh (–5%), posterior thigh (–4%), and total thigh (–5%). There were group × time interactions (all P < .0001) for both quadriceps CSA and radiological density (Fig. 3). In TM, CSA increased 15%, while the radiological density increased 6%. In TW, the corresponding parameters decreased by –4% for CSA, and –2% (P > 0.05) for radiological density.
Lower-limb muscle strength
Strength data and key statistics are displayed in Fig. 4. Isometric torque increased over the intervention for both knee extension (12%) and flexion (26%) in TM. In TW, isometric strength levels were maintained over the intervention for both extension and flexion. Isokinetic strength at 60°/s showed a main effect of time (no interactions), for both knee extension and flexion. Strength at 90°/s knee extension showed an interaction effect (P = .006) since strength increased in TM but was maintained over the 12 months in TW. For 90°/s knee flexion, there was a main effect of time (P < .0001) since both groups increased peak torque. Since strength and muscle mass increased in parallel in the TM, the specific strength (torque/quadriceps area) remained unchanged. In contrast, the specific strength increased slightly in TW (Fig. 4).
Fig. 5 summarizes the absolute and height-adjusted levels of isometric muscle strength and anterior thigh volume in relation to the intervention and in comparison with the cisgender control groups. At T12, the absolute levels of strength and muscle volume were greater in TW than in TM and CW. The height-adjusted value for muscle strength was also greater in TW than CW and TM at T12 (both P < .05). However, for muscle volume, this difference remained (P = .014) only in the TW versus CW comparison (Fig. 5). While there were strong correlations (R>0.77, P < .006) between muscle size and strength at baseline (Fig. 6), the correlation between the change in size and change in strength was more moderate in TW (R = 0.62, P = .04), and absent in TM (R = 0.31, P = .33).
The current study was conducted to comprehensively assess muscle function, size, and composition in both TW and TM undergoing 1 year of gender-affirming medical treatment. Our results show that thigh muscle volume increased in the TM (15%), which was paralleled by increased quadriceps CSA and radiological density. In TW, the change in muscle volume was considerably smaller, where the total loss of thigh muscle volume was 5%, which was paralleled by decreased quadriceps CSA but not radiological density. The TM increased strength over the assessment period, while the TW generally maintained their strength levels. Despite the robust changes in lower-limb muscle mass and strength in TM, the TW were still stronger following 12 months of gender-affirming hormone treatment, both in absolute and height-adjusted values. There have been very few investigations on whether any physical advantages of many years of male testosterone levels, as untreated adult TW have experienced, remain after gender-affirming interventions. To add further complexity to the issue, the magnitude of physical advantage of both endogenous and exogenous testosterone levels is highly debated (24–29). Therefore, the findings reported herein add important new information that increase our understanding of the effects of cross-sex hormone treatment on muscle strength, mass, and composition in transgender individuals.
The increase in muscle volume and quadriceps CSA in the TM was expected given the well-known effect of testosterone administration on muscle mass (27,30). Moreover, the results and magnitudes of the gains in muscle mass were in line with previous data on TM undergoing cross-sex hormone treatment (10,11). A more novel finding was that the contractile density of the quadriceps also increased, as evident by the measure of radiological attenuation, a valid proxy of fat infiltration in skeletal muscle (31,32). This suggests that the skeletal muscles of the TM not only increase in size but also in contractile density. While this theoretically could amplify the specific tension (force produced per area) of the muscle (15), our data indicated that this did not occur in the current cohort of TM.
In contrast to the TM, the TW experienced reductions in muscle mass over the intervention. However, it is worth noting that the reduction in muscle mass in TW was smaller than the corresponding increase in TM, both in terms of relative and absolute changes. At the 12-month follow-up, the TW still had larger muscle volumes and quadriceps area than the TM. This is somewhat in contrast to two earlier reports where total thigh muscle area was similar between the TM and TW at the 12-month follow-up (10,12). This seems to have been driven mainly by the fact that the relative changes in muscle area in those studies were larger both in TM and TW than in the current study. On the contrary, our findings in the cohort of TW are in line with a previous study reporting very modest changes in lean body mass following 1 and 2 years of cross-hormone treatment (13). Thus, our findings open up for the speculation that the estradiol treatment itself had a protective effect against lean mass loss, which is supported by research conducted on estrogen replacement therapy in other targeted populations (33–35) and in mice after gonadectomy (36).
The TM increased strength over the assessment period, both in knee extension and knee flexion. In contrast, the TW maintained knee extension strength and knee flexor isometric strength, but there were some improvements in the dynamic knee flexor measurements. Given that there was no structured training performed during the intervention, the improvements in some of the knee flexion measurements in the TW most likely arose from learning effects from repeating the test at 4 occasions. This is supported by reliability data showing that the learning effect is of greater concern during knee flexion than during knee extension (37). Nevertheless, since the potential learning effect arguably would be very similar between the TM and TW, the more substantial strength increase in TM is most likely attributed to the testosterone administration (30). At the 12-month follow-up, absolute and height-adjusted strength levels were still greater in the TW than in TM and CW. This was, however, only true for knee extension and not knee flexion. Based on the data, the most obvious explanation for this was the relatively smaller differences between groups already at baseline for knee flexion strength, and the more substantial increase in knee flexion than in knee extension strength in the group of TM. Taken collectively, the most robust and reproducible strength measure in this study, namely the isometric knee extension (37), showed that the TM increased strength by around 12%, whereas the TW maintained their strength levels.
We acknowledge that this study was conducted with untrained individuals and not transgender athletes. Thus, while this gave us the important opportunity to study the effect of the cross-hormone treatment alone, and as such the study adds important data to the field, it is still uncertain how the findings would translate to transgender athletes undergoing advanced training regimens during the gender-affirming intervention. It is also important to recognize that we only assessed proxies for athletic performance, such as muscle mass and strength. Future studies are needed to examine a more comprehensive battery of performance outcomes in transgender athletes. Given the marked changes in hemoglobin concentration in the current study, it is possible that gender-affirming treatment also has effects on endurance performance and aerobic capacity. Furthermore, since the TM and to some extent also the TW demonstrated progressive changes in muscle strength, and strikingly some of the TW individuals did not lose any muscle mass at all, follow-ups longer than 12 months are needed to better characterize the long-term consequences and individual responsiveness to gender affirming interventions. Future studies should also include age- and body size-matched cisgender control groups undergoing the same assessment timepoints without the therapies.
In summary, cross-sex hormone treatment markedly affects muscle function, size, and composition in transgender individuals. While the TM experienced robust changes in lower-limb muscle mass and strength, the corresponding changes in TW were modest. The question of when it is fair to permit a transgender woman to compete in sport in line with her experienced gender identity is challenging and few data have been provided to add clarity on the potential remaining physical advantage for TW after medical interventions. As such, these findings add new knowledge that could be relevant for sport governing bodies when evaluating the eligibility of transgender individuals to compete in athletic events in line with their experienced gender identity.
gonadotropin releasing hormone
total adipose tissue
total lean tissue
Financial support: This work was supported by the Stockholm County Council grant numbers 20160337 and K0138-2015, the Thuring foundation and the 1.6 Million Club.
Clinical Trial Information: clinicaltrials.gov (Identifier NCT02518009, Registered 12 May 2015).
Disclosure Summary: The authors declare no conflicts of interest.
Data Availability: The datasets generated and/or analyzed during the current study are not publicly available but are available from the corresponding author on reasonable request.