Do Trans Women Retain an Athletic Advantage After Transition?

The claim in dispute is not whether male puberty confers athletic advantage. It does, and nobody serious disputes the size of it. Handelsman and colleagues put the overall figure at "at least" 8 to 12 percent, and the sport-by-sport picture from Hilton and Lundberg runs from about 11 percent in swimming and running to 31 to 37 percent in Olympic weightlifting and over 50 percent in a field hockey drag flick.

The dispute is what happens to that advantage when an adult suppresses testosterone. One side treats advantage as a function of circulating hormones, so lowering testosterone should remove it. The other treats it as a function of development, so what puberty built stays built. Both are partly right, and the reason the argument never resolves is that each side has picked a different set of measures.

Sorting every physical trait into one bucket is the central mistake, and both sides make it. Some traits track current hormone levels closely. Some track them slowly. Some do not track them at all.

Blood is the clearest case. Testosterone drives red blood cell production, and when it is suppressed, haemoglobin and haematocrit fall to the range typical of cisgender women. Harper and colleagues, reviewing twenty-four studies, put the timeline at about four months.

This is the strongest evidence that hormone-level reasoning works, and it works for endurance events specifically.

Muscle is the contested middle. Across roughly twelve longitudinal studies, twelve months of suppression produces about a 5 percent loss of lean mass. Wiik and colleagues, following eleven trans women, measured thigh muscle volume down 5 percent and quadriceps cross-sectional area down 4 percent at one year, with knee extensor strength holding steady. Over three years Gooren and Bunck recorded a 12 percent fall in thigh muscle area.

Hilton and Lundberg put strength differences between the sexes at 30 to 100 percent depending on the cohort and the task, and the performance gap at 10 to 50 percent depending on the sport. A 5 percent loss against a baseline of that size leaves most of it in place on paper. That subtraction is the heart of the retained-advantage argument, and in absolute terms it holds: both major 2021 reviews concluded that lean mass, muscle area and strength remain above cisgender female means at three years.

Anyone citing a single strength or fitness number in this debate is choosing a denominator, usually without saying so. The absolute figures favour the retained-advantage case. The body-mass-relative figures favour the inclusion case. Both are in the same papers.

Height, limb length, bone geometry, hand and foot size and overall frame are set by puberty and are not altered by adult hormone therapy. Hilton and Lundberg reach this conclusion directly: there is no plausible mechanism by which testosterone suppression would change bone length or hip width, and trans women generally maintain bone mass across at least 24 months of suppression rather than losing it. They name the sports where it matters most, those where height, limb length and handspan are decisive, giving basketball, volleyball and handball as examples.

These matter differently by sport. Reach and leverage are decisive in throwing and striking events. Height matters in basketball, volleyball and handball. In events where performance is measured per kilogram of body mass, they matter much less. This is the least contested part of the retained-advantage case and the part that no amount of hormone therapy addresses.

Put at its strongest: male puberty builds a skeleton that hormone therapy cannot remodel, and it builds muscle that hormone therapy reduces by only about a twentieth in the first year. The decline slows sharply after year one and appears to settle above the cisgender female mean rather than reaching it. Three years in, measured against cisgender women, trans women still carry more lean mass and greater absolute strength.

The field data support this in part. In the US Air Force cohort, the largest with actual performance testing, trans women before hormone therapy ran 1.5 miles 21 percent faster than cisgender women, did 31 percent more push-ups and 15 percent more sit-ups. After one year of testosterone suppression, the period World Athletics then required, they were still 9 percent faster. Two years in, the push-up and sit-up gaps had closed, but the run time advantage was still 12 percent.

The precautionary version of the argument is that where the residual cannot be ruled out, the burden should fall on inclusion rather than on the women already in the category. That is a judgement about which error to risk, not a claim about the data, and it should be argued on those terms.

Put at its strongest: the subtraction that produces the residual is not arithmetic anyone is entitled to do. The baseline gap comes from comparing elite cisgender men with elite cisgender women. The 5 percent decline comes from small clinical studies of mostly sedentary trans women. The two numbers describe different populations measured different ways, and subtracting one from the other assumes trans women start at the cisgender male mean. Wiik's data indicate they do not: muscle volumes were below cisgender male means before treatment began.

The same analysis found the gaps against cisgender men large and statistically significant, which places trans women between the two groups rather than with either. Hamilton's study of trained athletes points the same way on the measures that scale with body mass: lower relative jump height, lower relative oxygen uptake, lower lung function ratio than cisgender women athletes.

Download this conclusion: the meta-analysis

This section as one image: what the largest synthesis found when it compared trans women on hormone therapy with cisgender women, and the two cautions its authors attach.

Download the card (PNG, 1,200 × 630)

Policy has moved further and faster than the evidence. In 2021 the International Olympic Committee dropped its single testosterone threshold, established that there should be no presumption of advantage, and pushed eligibility down to individual federations. Most federations then went the other way.

World Aquatics in June 2022 restricted the women's category to those who had not experienced male puberty beyond Tanner stage 2 or age 12, whichever came later, and created an open category. When that category was piloted at a World Cup event in 2023, nobody entered. Cycling's UCI followed in July 2023, having concluded that its previous testosterone threshold was not sufficient to eliminate the benefits of puberty and that lasting differences in bone and limb length could not be ruled out. World Athletics announced on 23 March 2023 that it would exclude trans women who had been through male puberty from female world ranking competition, effective 31 March. Sebastian Coe said that where the science was insufficient to justify keeping a testosterone rule, the council had to be guided by its overarching principle, which was to protect the female category, and World Athletics said that with no transgender athletes then competing internationally there was no athletics-specific evidence either way, so it chose fairness over inclusion. The NCAA changed its policy in February 2025, one day after a presidential executive order on the subject.

This literature is thin enough that anyone claiming it settles the question is overstating it. Sample sizes run from single digits to about fifty. Wiik studied 11 trans women, Alvares 15, Hamilton 23, Roberts 46. Most of the longitudinal physiology comes from clinical populations who were not athletes, which is the wrong population for a question about competitive sport. Hamilton's study is the main one using trained athletes, and it is cross-sectional, so it cannot establish a time course.

The reviews rest heavily on the same small studies, so agreement between them reflects shared source data rather than independent confirmation. And the designs disagree systematically: small longitudinal studies emphasise retention, matched cross-sectional comparisons emphasise convergence.

Several things are simply unstudied. Muscle fibre type change over time. Tendon and connective tissue. Anything beyond four years. Cardiac and respiratory remodelling. Competition outcomes at elite level.

1. The answer is different for different measures, and no single verdict covers them all. Blood oxygen capacity normalises to cisgender female levels within about four months. Muscle mass and strength fall roughly 5 percent in the first year and remain above cisgender female means in absolute terms at three years. The skeleton does not change at all.

2. Whether the retained muscle translates into competitive advantage is not settled. The largest synthesis, 50 studies and 6,099 people in its corrected form, found no statistically significant difference in strength or aerobic fitness between trans and cisgender women, at very low to low certainty, with a confidence interval on upper-body strength wide enough to include a large advantage and a deficit. Studies of trained athletes find absolute measures higher and body-mass-relative measures equal or lower.

3. The common subtraction, baseline male advantage minus the hormone therapy decline, is sound for skeletal traits and directionally sound for absolute muscle. It is not reliable as arithmetic, because the two figures come from different populations measured different ways.

4. Nobody has the data that would decide this. There is no large longitudinal study of trained trans women with matched cisgender female athletes and actual competition outcomes. Every position currently held is held on evidence graded low certainty.

5. Most major federations now exclude post-pubertal trans women from elite women's categories. Those that gave reasons cited precaution and residual uncertainty rather than evidence of disproportionate winning, which nobody has collected.

Download the conclusion

The verdict above as one image, with the Air Force cohort’s before-and-after figures as its chart. It is 1,200 by 630 pixels, the size social networks use for link previews, and anyone may download it and post it.

Download the card (PNG, 1,200 × 630)