A unique pseudo-eligibility analysis of longitudinal laboratory performance data from a transgender female competitive cyclist

Introduction

Today’s transgender population is ∼3 % of the world’s population [n=∼241,359,343 [1]] and the prevalence of elite athletes in the general population is estimated to be ∼0.0003 % [2]], therefore, there is potential for ∼72,408 transgender athletes globally. The question of how and where to integrate transgender athletes into every level of competitive sport is being debated rigorously [3], [4], [5], [6], [7], [8], [9], [10], [11], and the International Olympic Committee’s Framework, although criticised [4, 12], states that any eligibility decision should be “largely based on data collected from a demographic group that is consistent in gender and athletic engagement with the group that the eligibility criteria aim to regulate [13]”. The regulations set by the Union Cycliste Internationale (UCI) [14], which is the world governing body for cycling sports, stipulates a specific ineligibility for transgender women athletes in the female category. The UCI cites concerns about potential unfair performance advantages despite the lack of supporting athlete data despite using cisgender performance differences as a proxy for transgender performance [6, 7]. We investigated the likely effects on laboratory-based performance indicators in response to these policy changes cross-sectionally [15], but the design of this study does not show the effect of gender-affirming hormone treatment (GAHT) over time. Our research group also pioneered a decision-making framework for transgender athletes’ eligibility in 2021 [5]; however, the framework was demonstrated by employing cisgender archery and shooting performance as a proxy for transgender performance in those sports [6, 7]. To address the shortcomings of the cross-sectional analysis and the decision-making framework, and that it is vital to understand the level of performance transgender athletes possess relative to their cisgender counterparts [7]; the primary aim of this manuscript is to perform a pseudo-eligibility analysis of a longitudinally assessed transgender athlete using laboratory measures of athletic performance under the same conditions, compared with cross-sectionally assessed cisgender and transgender athletes [15]. The primary objective is to present a case study of data from a transgender female athlete over one year of GAHT; the secondary objective is to analyse where this athlete sits within a group of cross-sectionally assessed transgender and cisgender women athletes. The final objective is to use our laboratory’s decision-making framework [5] to perform a pseudo-eligibility analysis for the sport of cycling, demonstrating the efficacy of the process. Any recommendations should be treated cautiously as sport’s governing bodies are best placed to determine their priorities and eligibility policies, and the data fuelling it is limited [15]. The summary of this article is presented in Figure 1.

Figure 1:

Graphical representation of this article. Figure created with BioRender.

Methods

Results

Handgrip

Cross-sectional Transgender women also had greater absolute right handgrip strength than cisgender women (t ₍₆₆₎=3.2, p<0.05, Figure 2a) The transgender athlete exhibited a reduction in absolute (7 %, Figure 2a) over one year to exhibit a similar handgrip profile to cross-sectional transgender women. No differences were found cross-sectionally in the relative handgrip to fat-free mass transgender and cisgender women and the longitudinal athlete showed a reduction of 13 % after one year (Figure 2b). After 3 months, the absolute handgrip remained 24 % above the mean of cisgender women (Figure 2a) and relative handgrip was 2 % below cisgender women (Figure 2b). After 6 months, the absolute handgrip remained 21 % above the mean of cisgender women (Figure 2a) and relative handgrip was 6 % below cisgender women (Figure 2b). After 12 months, the absolute handgrip remained 12 % above the mean of cisgender women (Figure 2a) and relative handgrip was 17 % below cisgender women (Figure 2b).

Figure 2:

Comparisons of absolute (a) and relative (b) handgrip performance measures in cross-sectional athletes and during one year of GAHT in a transgender woman athlete. Notes: kg, kilogram; FFM, fat-free mass; CM, cisgender men; TM, transgender men; CW, cisgender women; TW, transgender women; LTWB, longitudinal transgender women baseline; LTW3, longitudinal transgender women 3 months; LTW6, longitudinal transgender women 6 months; LTW12, longitudinal transgender women 12 months. The grey dots represent individual data points. The grey lines indicate the range. The top and bottom of the blue box indicate the upper and lower quartile values and the bold black dash indicates the median value. The black box indicates the mean value.

Countermovement jump

No differences were found between cross-sectionally assessed transgender and cisgender women in absolute countermovement jump performance (Figure 3a) The longitudinal transgender athlete exhibited a reduction in absolute (23 %, Figure 3a) after one year. Cross-sectional transgender women were found to have lower countermovement jump height relative to fat-free mass than cisgender women (t ₍₆₆₎=−5.3, p<0.001) and the longitudinal athlete exhibited a 29 % reduction after one year (Figure 3b). After 3 months, the absolute countermovement jump was 5 % below the mean of cisgender women (Figure 3a) and the relative countermovement jump was 32 % below cisgender women (Figure 3b). After 6 months, the absolute countermovement jump was 8 % below the mean of cisgender women (Figure 3a) and the relative countermovement jump was 36 % below cisgender women (Figure 3b). After 12 months, the absolute countermovement jump was 19 % below the mean of cisgender women (Figure 3a) and the relative countermovement jump was 48 % below cisgender women (Figure 3b).

Figure 3:

Comparisons of absolute (a) and relative (b) counter movement jump performance measures in cross-sectional athletes and during one year of gaht in a transgender woman athlete. Notes: cm, centimetre; FFM, fat-free mass; CM, cisgender men; TM, transgender men; CW, cisgender women; TW, transgender women; LTWB, longitudinal transgender women baseline; LTW3, longitudinal transgender women 3 months; LTW6, longitudinal transgender women 6 months; LTW12, longitudinal transgender women 12 months. The grey dots represent individual data points. The grey lines indicate the range. The top and bottom of the blue box indicate the upper and lower quartile values and the bold black dash indicates the median value. The black box indicates the mean value.

Power

Cross-sectional cisgender women were found to have reduced peak power compared with transgender women (t ₍₆₆₎=−3.6, p<0.01, Figure 4a) with the longitudinal transgender athlete exhibiting a reduction of 3 % peak power over one year (Figure 4a). Cross-sectional transgender women were found to have no difference in peak power relative to fat-free mass than cisgender women with the longitudinal transgender athlete exhibiting a reduction of 8 % peak power relative to fat-free mass over one year (Figure 4b). Cross-sectional cisgender women were found to have reduced average power compared with transgender women (t ₍₆₆₎=−3.0, p<0.01, Figure 4c) with the longitudinal transgender athlete exhibiting a reduction of 3 % average power over one year (Figure 4c). Cross-sectional transgender women were found to have no difference in average relative to fat-free mass than cisgender women with the longitudinal transgender athlete exhibiting a reduction of 9 % peak power relative to fat-free mass over one year (Figure 4d). After 3 months, peak power was 26 % above the mean of cisgender women (Figure 4a) and relative peak power 1 % above (Figure 4b). The average power was 26 % above cisgender women (Figure 4c) and relative average power 3 % above (Figure 4d). After 6 months, peak power was 29 % above the mean of cisgender women (Figure 4a) and relative peak power was matched (Figure 4b). The average power was 30 % above cisgender women (Figure 4c) and relative average power was matched (Figure 4d). After 12 months, peak power was 31 % above the mean of cisgender women (Figure 4a) and relative peak power 1 % above (Figure 4b). The average power was 32 % above cisgender women (Figure 4c) and relative average power 3 % above (Figure 4d).

Figure 4:

Comparisons of peak power (a) peak power relative to fat-free mass (b) average power (c) and average power relative to fat-free mass (d) in cross-sectional athletes and during one year of GAHT in a transgender woman athlete. Notes: W, watts; FFM, fat-free mass; CM, cisgender men; TM, transgender men; CW, cisgender women; TW, transgender women; LTWB, longitudinal transgender women baseline; LTW3, longitudinal transgender women 3 months; LTW6, longitudinal transgender women 6 months; LTW12, longitudinal transgender women 12 months. The grey dots represent individual data points. The grey lines indicate the range. The top and bottom of the blue box indicate the upper and lower quartile values and the bold black dash indicates the median value. The black box indicates the mean value.

Maximal oxygen uptake (V̇O₂max)

No differences were found cross-sectionally in absolute V̇O₂max between cisgender and transgender women (Figure 5a) and the longitudinal transgender athlete exhibited a reduction of 15 % in V̇O₂max over 1 year (Figure 5a). Cross-sectionally, transgender women had lower relative V̇O₂max than cisgender women (t ₍₆₆₎=−3.3, p=0.01, Figure 5b) and after one year of GAHT, the longitudinal athlete presented with a 30 % loss in relative V̇O₂max to body mass (Figure 5b). After 3 months, V̇O₂max was 30 % above the mean of cisgender women (Figure 5a) and relative V̇O₂max was 2 % above cisgender women (Figure 5b). After 6 months, V̇O₂max was 13 % above the mean of cisgender women (Figure 5a) and relative V̇O₂max was 19 % below cisgender women (Figure 5b). After 12 months, V̇O₂max was 15 % above the mean of cisgender women (Figure 5a) and relative V̇O₂max was 25 % below cisgender women (Figure 5b).

Figure 5:

Comparisons of absolute (a) and relative (b) V̇O₂max in cross-sectional athletes and during one year of GAHT in a transgender woman athlete. Notes: ml·min⁻¹, millilitres per min; ml·kg⁻¹·min⁻¹, millilitres per kilogram of body mass per minute; CM, cisgender men; TM, transgender men; CW, cisgender women; TW, transgender women; LTWB, longitudinal transgender women baseline; LTW3, longitudinal transgender women 3 months; LTW6, longitudinal transgender women 6 months; LTW12, longitudinal transgender women 12 months. The grey dots represent individual data points. The grey lines indicate the range. The top and bottom of the blue box indicate the upper and lower quartile values and the bold black dash indicates the median value. The black box indicates the mean value.

Discussion

The findings of this pilot longitudinal study provide insights into changes in laboratory-based performance measurements of a transgender athlete undergoing 1-year GAHT. Given the purpose of GAHT [23] and that this study addressed the cross-sectional limitations of Hamilton [15]; this pilot investigation corroborated that transgender women’s oestradiol was higher than that of cisgender women after one year of GAHT (Table 1). Demonstrating longitudinally that this transgender women athlete shares a similar endocrine profile with cisgender women and that these high oestradiol concentrations could be consistent with GAHT in transgender women athletes. Notably, the [Hb] profiles of the transgender women athlete were lower than those of the cisgender female athlete references (Table 1); However, a 7 % decline in transgender women athletes’ [Hb] after one year of GAHT is consistent with the work of Wiik [24] and Olson-Kennedy [25] in sedentary transgender women.

It was hypothesised in the cross-sectional analysis that the differences observed in body composition in the cross-sectional sample indirectly showed the potential role of androgens in body composition [15]. Specifically, oestradiol’s role in fat accumulation [26] as the transgender women’s oestradiol concentrations and fat mass were greater than those of cisgender women. Although this pilot longitudinal study has no statistical power, a strong correlation (r=0.91) between oestradiol and fat mass accumulation was found over one year of GAHT. With testosterone reductions not resulting in declines in fat-free mass after one year in this transgender woman athlete (Table 1) and the athlete training intensity was consistent (Table 1), the hypothesis that oestradiol is the primary androgen in altering body composition in transgender women is strengthened, though not confirmed, as research with greater statistical power is required to confirm this hypothesis.

Investigations using more precise strength assessments are necessary for transgender athletes [22] as increased hand size predicts handgrip strength [15]. However, this pilot’s longitudinal nature allows the assessment of GAHT’s effects on handgrip strength over one year. The loss of absolute (Figure 2) handgrip strength without negative changes in exercise intensity and exercise time (Table 1), the 4.5 to 7.1 % reductions in handgrip strength seen in sedentary transgender women [8, 27, 28] and relative handgrip loss exceeds 7.1 % (Figure 2). These results conflict with the maintenance of strength reported by Wiik [24], although Wiik [24] measured strength with the more accurate isokinetic knee dynamometer.

According to the cross-sectional analysis, transgender women have less lower-body anaerobic power than cisgender women [15]. After one year of GAHT, the transgender female sub-elite athlete was found to have 17 % less countermovement jump height than cisgender women (Figure 3), a 1.5-fold greater rate of decrease than what was found cross-sectionally [15], and the transgender athlete’s countermovement jump height had decreased by 19 % from baseline (Figure 3). This longitudinal data suggests that the changes in absolute and relative countermovement jump performance are brought on by increased fat mass owing to the increased oestrogen brought on by GAHT (Table 1). This hypothesis is strengthened by the moderate reductions in peak and average power (Figure 4). This leads to the notion that the increasing inertia caused by the increase in non-contractile fat mass is the primary limiting factor in this transgender woman's poor countermovement jump performance. High oestrogen levels have been found to decrease power and performance in cisgender women [29], and more work is needed to confirm or refute this hypothesis in transgender women athletes.

In the cross-sectional analysis [15], no differences were found between transgender women athletes and cisgender women athletes in absolute V̇O₂max. Transgender women exhibited lower relative V̇O₂max compared to cisgender women (Figure 5). This sub-elite transgender women cyclist exhibited a higher absolute V̇O₂max than reference cisgender women after one year of GAHT, with a 14 % reduction from baseline (Figure 5). Relative to body mass, this transgender women athlete showed a 26 % decline in relative V̇O₂max from baseline that was notably lower than the reference cisgender women (Figure 5). O₂ delivery to tissues is a limiting factor in V̇O_2max proficiency, as [Hb] plays a significant role in O₂ transport [30] and endurance sports like cycling [31]. Table 1 shows a decrease in [Hb], which correlates (r=0.74) with a reduction of V̇O_2max after one year of GAHT. This suggests that loss of [Hb] during GAHT is crucial in reducing V̇O_2max performance. The association between these variables, along with lung function measurements [15], should be investigated with greater statistical power to determine which is more damaging to performance in transgender female athletes transitioning with GAHT.

To fulfil the third objective of this study, a pseudo-decision-making process for this athlete’s eligibility for the sport of cycling will be made using the decision-making tools set out in Hamilton [5]. Deciding on this athlete’s eligibility in cycling requires understanding the sport’s indices. Power output [32], [33], [34], [35], [36], V̇O₂max [37], [38], [39], [40], [41], [42], [43] and, to a lesser extent, strength [44], [45], [46] are crucial to the sport of cycling. Isometric strength doesn’t predict cyclist performance, whereas higher maximal V̇O₂max and peak power do [47]. The power test this athlete undertook (Countermovement Jump) is a reliable measure of power [48], however, a measure of direct power via cycle ergometer would have been better in judging this case [49]. The athlete only undertook handgrip strength and, again, a better measurement of strength, such as a leg press [48], would have been preferable. However, grip strength seems to be a temporary factor in cycling performance when riding for extended periods [50]. Nevertheless, GAHT had a significant effect on this athlete’s laboratory performance after one year of GAHT, as their grip strength (Figure 2), countermovement jump (Figure 3), and V̇O₂max (Figure 5) declined.

The question at hand is whether one year of GAHT is enough to negate any advantages of a high testosterone puberty [51] in this athlete to participate in the sport of cycling in the female category. To answer this and fulfil the request of the International Olympic Committee Framework [13], we must compare data generated from the transgender athlete with cisgender women athletes tested under the same conditions and measures of elite cisgender cyclist cohorts in the literature. Compared to the cisgender women of Hamilton [15], this transgender woman athlete maintained an advantage in 54 % (6:11) of laboratory performance measures (measured from hand grip, countermovement jump, and V̇O₂max). First, the transgender athlete has greater absolute and relative grip strength (Figure 2). Therefore, a grip strength advantage is apparent after one year of GAHT. Countermovement jump performance was absolutely and relatively less than cisgender women (Figure 3). Therefore, a jump performance disadvantage exists after one year. Peak power was above that of cisgender women (Figure 4a). Therefore, a peak power advantage exists after one year. The average power was higher than that of cisgender women (Figure 4b). Therefore, an average power advantage exists after one year. V̇O₂max performance was absolutely above cisgender women (Figure 5a), although relatively (Figure 5b) below cisgender women. Therefore, a V̇O₂max performance advantage is likely not to be present. Compared to cisgender women cyclists in the literature, this transgender athlete has a 14 % disadvantage (48.9 mL·kg⁻¹·min⁻¹ [52] vs. 41.9 mL·kg⁻¹·min⁻¹ [LTW12 Figure 5]) in V̇O₂max, likely confirming the lack of V̇O₂max performance advantage in this athlete. Comparing countermovement jump performance with a cohort of eight elite sprint track cyclists (Male=4, Female=4), the transgender women athletes’ average power was similar to elite sprint (2039 [53] vs. 2003 [LTW12 Figure 4c) but above elite endurance (1,668 [53]vs. 2003 [LTW12 Figure 4c] W) performance. Therefore, a disadvantage or advantage in average power for this transgender female athlete in the elite category of women’s cycling is unlikely to exist. However, the transgender women athletes’ peak power did match those of sprint cyclists (4,110 [53] vs. 4,160 [LTW12 Figure 4a] W) and was above endurance cyclists (3,184 [53] vs. 4,160 [LTW12 Figure 4a] W). Therefore, a disadvantage or advantage peak average power for this transgender female athlete in the elite category of women’s cycling is unlikely to exist.

The observed performance metrics of the athlete, particularly the increased peak power compared to cisgender women, may be attributed to their specialisation as a sprint cyclist. However, it is essential to note that this hypothesis remains speculative, as the current data do not allow for definitive conclusions. Notably, the athlete’s overall performance trajectory indicates a downward trend. Further, longitudinal studies are necessary to ascertain the specific factors contributing to these performance changes and to determine the extent to which they may be influenced by the athlete’s training regimen, physiological adaptations, or other external variables. Given the data analyses and using the tools of Hamilton et al. [5] it can clearly be defined as a fairness analysis (Figure 6a) as there is no contact with the opponent in cycling to provide an athlete-to-athlete contact safety concern, and there are some metrics that advantages are still held after one year of GAHT, which highlights fairness as the motivating factor (Figure 6b) behind the eligibility decision. Given that the transgender women cyclist in this analysis holds an advantage in 33 % (5:15) measures analysed, it could be warranted that this transgender cyclist be allowed to compete in the female category after one year of GAHT. A follow-up analysis should be performed in 12 months to see if 2 years of GAHT has a more significant effect, and the athlete’s eligibility should be re-evaluated. As this transgender cyclist matches sprint cyclists from the literature [53] appropriate future testing measures could be the Wingate anaerobic test or a six-second ‘all-out’ test as a more valid measures of peak power in cycling [54]. Nevertheless, the current data suggest that, even after one year of GAHT, some transgender female athletes may be fairly integrated into the female category on a provisional basis. Should subsequent competition data indicate unfair performance advantages, this decision could be reversed, or the GAHT requirement could be extended to two years or more. Conversely, an outright ineligibility for the female category on transgender women athletes in a specific sport precludes the opportunity to ascertain the true impact of GAHT on performance fairness, leaving the issue subject to speculation.

Figure 6:

The sliding scale (a) and declared weightings model of fairness (b) adapted from Hamilton [5].

Summary

In summary, this study aimed to assess the impact of one year of GAHT on a longitudinally examined transgender athlete’s laboratory performance. The findings revealed that GAHT declined grip strength, countermovement jump, and V̇O₂max performance. However, compared to cross-sectionally assessed cisgender women athletes, the transgender athlete still maintained an advantage in 54 % (6:11) of laboratory performance measures (measured from hand grip, countermovement jump, power, and V̇O₂max). For cycling events, a pseudo-eligibility analysis suggests that this transgender female athlete could be licensed to participate in cycling events in the female category after one year of GAHT. Moving forward, more accurate measurements of power and strength are essential for authentic transgender eligibility assessments in cycling sports. A follow-up analysis of this athlete is recommended in 12 months.