Data

Transgender women on hormones: what changes in a year, and Air Force fitness tests over 2.5 years

Before-and-after measurements of testosterone, hemoglobin, muscle size, lean mass and strength in adult transgender women after about 12 months of hormone treatment, from the study groups in a 2021 systematic review, plus U.S. Air Force fitness-test results of transgender women before and up to 2.5 years into treatment.

Coverage: Studies published 1999-2020; Air Force fitness records 2004-2018

What this dataset measures

These are the two most-cited bodies of evidence on whether testosterone suppression removes the male athletic advantage: before-and-after clinical studies, gathered in the Harper et al. (2021) systematic review, and the only study of real fitness-test performance over time (Roberts et al. 2021). Using every adult 12-month study group in the review, rather than one study, avoids picking a favorable result.

Curated data tables

These related tables form this versioned dataset release. Each table has its own downloadable file and field definitions.

Before-and-after measurements in transgender women, study groups in Harper et al. (2021)

One row per study group and measure, as listed in Tables 2-5 of the Harper et al. (2021) systematic review: testosterone, hemoglobin or hematocrit, muscle cross-sectional area, total lean body mass and strength. Testosterone values come from each study's before-and-after testosterone column. Rows marked included=no are adolescent or young-adult cohorts, or follow-ups that were not about 12 months; they are kept so readers can see what was left out.

Unit of observation: One study group and one measure

View this table

harper-12-month-study-groups.csv text/csv · 4.5 KB (4,574 bytes)
15 field definitions
FieldTypeMeaningUnitsValue type
measure string testosterone, hemoglobin (hemoglobin or hematocrit), muscleArea, leanMass or strength. — observed
study string First author and year, as cited by Harper et al. — observed
harperTable integer Table number in Harper et al. (2021) where the row appears. — observed
group string Sub-group or body region, where the study reported more than one. — observed
n integer Number of transgender women in the group. people observed
detail string What was measured. — observed
baseline number Mean (or median) before hormone treatment. see units observed
followUp number Mean (or median) at follow-up. see units observed
units string Units of baseline and followUp. — observed
months string Months of treatment at follow-up; a range where the study pooled follow-ups. months observed
pctChangeReported number Percentage change as printed by Harper et al.; blank for testosterone (not printed) and one lean-mass row whose value was not printed. % observed
testosteroneBefore number Group testosterone before treatment, from Harper et al.'s testosterone column. nmol/L observed
testosteroneAfter number Group testosterone at follow-up. nmol/L observed
included string yes if the row is an adult study group measured at about 12 months; no otherwise. — derived
exclusionReason string Why a row was left out of the medians. — derived

U.S. Air Force fitness tests of transgender women, by time on hormones (Roberts et al. 2021, Table 4)

Mean push-ups and sit-ups in one minute and 1.5-mile run time of 46 transgender women in the U.S. Air Force before hormones and at 0-1, 1-2 and 2-2.5 years of treatment, with the difference from the average Air Force woman under 30 (2004-2014) and 95% confidence intervals, exactly as printed.

Unit of observation: One fitness measure in one time band

View this table

air-force-fitness-trans-women.csv text/csv · 1,007 bytes
10 field definitions
FieldTypeMeaningUnitsValue type
measure string pushups, situps or run. — observed
timeOnHormones string Time band as printed: Before hormones, Under 1 year, 1 to 2 years, 2 to 2.5 years. — observed
binOrder integer Order of the time band, 0 = before hormones. — derived
transWomenMean number Mean score of transgender women in the band. see unit observed
transWomenCiLow number Lower end of the 95% confidence interval for the mean. see unit observed
transWomenCiHigh number Upper end of the 95% confidence interval for the mean. see unit observed
diffVsAverageWoman number Mean difference from the average Air Force woman under 30; for the run, negative means faster. see unit observed
diffCiLow number Lower end of the 95% confidence interval for the difference. see unit observed
diffCiHigh number Upper end of the 95% confidence interval for the difference. see unit observed
unit string repetitions in 1 minute, or seconds for 1.5 miles. — observed

Supporting files

Source snapshots, manifests, quality reports, documentation, and other materials used to produce or review this release.

  • analyses.json Derived table · application/json · 16.7 KB (17,090 bytes)

    Every saved-analysis result in this dataset, as computed by build.js.

  • build.js Transformation manifest · text/javascript · 4.2 KB (4,317 bytes)

    Node.js script that computes every saved analysis from the two curated tables.

Saved analyses

Median change after about 12 months of hormone treatment, adult transgender women

Question: After about a year of testosterone suppression and estrogen, how much do testosterone, hemoglobin, muscle size, lean mass and strength change in transgender women?

Result: Median change after about 12 months: testosterone -95.6% (12 groups), hemoglobin -7%, muscle size -6.4%, lean body mass -4%, strength +0.5% (range -7.1% to +2%).

Operation: For each measure, the median of the percentage changes across included study groups; range = lowest and highest group. Testosterone change computed from each group's before and after values; other changes as printed by Harper et al., or computed from before and after means where not printed.

Scope: Study groups of adult, mostly untrained transgender women in Europe and the U.S., 1999-2020.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
Testosterone -95.6 % change after about 12 months (range -97.8 to -48.8) range Median of the percentage changes in the included study groups; the range is the lowest and highest study group. 12
Hemoglobin (red blood cells) -7 % change after about 12 months (range -14 to -4.6) range Median of the percentage changes in the included study groups; the range is the lowest and highest study group. 5
Muscle size -6.4 % change after about 12 months (range -9.5 to -1.5) range Median of the percentage changes in the included study groups; the range is the lowest and highest study group. 4
Lean body mass -4 % change after about 12 months (range -5.4 to -0.8) range Median of the percentage changes in the included study groups; the range is the lowest and highest study group. 8
Strength 0.5 % change after about 12 months (range -7.1 to 2) range Median of the percentage changes in the included study groups; the range is the lowest and highest study group. 5

View this analysis

  • Participants were adults and mostly untrained; no study measured athletes or adolescents through this design.
  • Small groups (11 to 249 women) and few groups per measure (4 to 12).
  • About 12 months only; Harper et al. found strength, lean mass and muscle area still above women's after 36 months, but with too few studies to chart.

How much faster transgender women in the Air Force ran 1.5 miles than the average woman, by time on hormones

Question: How much faster than the average Air Force woman under 30 did transgender women run 1.5 miles before and during hormone treatment?

Result: Transgender women ran 1.5 miles 147 seconds faster than the average Air Force woman before hormones (20.8% faster), 97 s after under a year, 64 s at 1-2 years and 90 s at 2-2.5 years (95% CI 10 to 169; 11.8% faster).

Operation: Table 4 differences from the average woman, sign reversed so faster is positive; speed advantage = average woman's time / transgender women's time - 1.

Scope: 46 transgender women in the U.S. Air Force, 2016-2018 records.

Estimates with uncertainty
EstimateValueUncertaintyMethodn
averageWomanSeconds: 855, speedAdvantagePct: 20.8, timeOnHormones: Year 0 (before), transWomenSeconds: 708 147 seconds faster than the average Air Force woman over 1.5 miles (95% CI 121 to 173) 95% confidence interval Roberts et al. (2021) Table 4 mean difference from the average Air Force woman under 30, sign reversed so faster is positive; 95% CI from the paper. —
averageWomanSeconds: 855, speedAdvantagePct: 12.8, timeOnHormones: Year 0-1, transWomenSeconds: 758 97 seconds faster than the average Air Force woman over 1.5 miles (95% CI 70 to 124) 95% confidence interval Roberts et al. (2021) Table 4 mean difference from the average Air Force woman under 30, sign reversed so faster is positive; 95% CI from the paper. —
averageWomanSeconds: 855, speedAdvantagePct: 8.1, timeOnHormones: Year 1-2, transWomenSeconds: 791 64 seconds faster than the average Air Force woman over 1.5 miles (95% CI 26 to 101) 95% confidence interval Roberts et al. (2021) Table 4 mean difference from the average Air Force woman under 30, sign reversed so faster is positive; 95% CI from the paper. —
averageWomanSeconds: 855, speedAdvantagePct: 11.8, timeOnHormones: Year 2-2.5, transWomenSeconds: 765 90 seconds faster than the average Air Force woman over 1.5 miles (95% CI 10 to 169) 95% confidence interval Roberts et al. (2021) Table 4 mean difference from the average Air Force woman under 30, sign reversed so faster is positive; 95% CI from the paper. —

View this analysis

  • Air Force members, not competitive athletes; fitness tests measure endurance, not maximal speed.
  • The comparison is an Air Force-wide average for women under 30, not matched controls; training was not recorded.
  • Few women reached the 2-2.5 year band, so its confidence interval is wide (10 to 169 seconds).
  • Results may differ for people who began treatment early in puberty, as the authors note.

Push-ups, sit-ups and run: transgender women vs the average Air Force woman, by time on hormones

Question: How did transgender women's push-up, sit-up and run advantages over the average Air Force woman change with time on hormones?

Result: Push-ups: 14.8 more than the average woman before hormones, 2.1 more (not significant) at 2-2.5 years. Sit-ups: 7.9 more before, 0.9 fewer (not significant) at 2-2.5 years. Run: 147 seconds faster before, 90 seconds faster (significant) at 2-2.5 years.

Operation: For each measure and time band: transgender women's mean, the implied average woman's value, the printed difference and its 95% CI, the advantage in percent (run: % higher speed; push-ups and sit-ups: % more repetitions than the average woman), the paper's own percentage convention (difference / transgender women's mean), and whether the 95% CI excludes zero.

Scope: 46 transgender women in the U.S. Air Force, 2016-2018 records.

View this analysis

  • Push-up and sit-up advantages were no longer statistically significant at 2-2.5 years; the run advantage was.
  • Weight gain on hormones can lower push-up scores, which the authors say may understate the remaining advantage.

Every study group behind the 12-month medians

Question: Which study groups and values make up each 12-month median, and which were left out and why?

Result: 34 included rows and 8 excluded rows across five measures.

Operation: Lists every row of the Harper et al. table with its percentage change and inclusion flag.

Scope: All rows of harper-groups.

View this analysis

  • Values are as printed in the review; we did not re-extract every one from the original studies. Scharff et al. (2019) and Wiik et al. (2020) were cross-checked against the original papers.

Sources

  1. Harper J, O'Donnell E, Sorouri Khorashad B, McDermott H, Witcomb GL (British Journal of Sports Medicine) Systematic review, Tables 2-5: per-study changes in lean body mass, muscle cross-sectional area, strength and hemoglobin or hematocrit, with each study's testosterone before and after treatment, in transgender women on hormones. Free copy

    Retrieved November 5, 2026. CC BY 4.0. Open access under CC BY 4.0; FactHub republishes the extracted numbers with attribution.

    Reliability: Peer-reviewed study. Peer-reviewed systematic review of 24 studies of non-athletic transgender women, from Loughborough University and partners; no specific funding and no competing interests declared.

  2. Roberts TA, Smalley J, Ahrendt D (British Journal of Sports Medicine) Table 4: U.S. Air Force fitness-test push-ups, sit-ups and 1.5-mile run times of 46 transgender women before and during hormone treatment, compared with the average Air Force woman and man under 30.

    Retrieved November 5, 2026. Copyright BMJ; numbers reported as facts. FactHub republishes the table's means, differences and confidence intervals as facts, with attribution; the article text is not redistributed.

    Reliability: Peer-reviewed study. Peer-reviewed study of U.S. Air Force medical and fitness records, 2004-2018, by military and academic pediatricians; no specific grant. Fitness tests measure endurance, not maximal strength, and the comparison is to an Air Force average, not matched controls.

  3. Scharff M, Wiepjes CM, Klaver M, Schreiner T, T'Sjoen G, den Heijer M (Endocrine Connections) Cross-check of the largest 12-month study group: 249 transgender women in the European ENIGI cohort, grip strength 41.8 kg before and 40.0 kg after 1 year; median testosterone 17.9-18.5 nmol/L before and 0.4-0.8 nmol/L during treatment by clinic (Table 1). Free copy

    Retrieved November 5, 2026. CC BY 4.0. Open access; numbers republished with attribution.

    Reliability: Peer-reviewed study. Multicenter prospective clinical cohort in Amsterdam, Ghent and Oslo; authors declare no conflict of interest.

  4. Wiik A, Lundberg TR, et al. (Journal of Clinical Endocrinology & Metabolism) Cross-check of the Wiik study group: thigh muscle volume fell 5% and quadriceps cross-sectional area 4% in 11 transgender women after 12 months, while strength was maintained.

    Retrieved November 5, 2026. Copyright Endocrine Society; numbers reported as facts. Numbers republished as facts with attribution.

    Reliability: Peer-reviewed study. Prospective study of 11 untrained transgender women at Karolinska Institutet, Stockholm; funded by Stockholm County Council and two foundations; no conflicts declared.

License: FactHub's selection, cleaning and combined tables are licensed under CC BY 4.0. Source data keeps the terms listed above (terms).

Limitations and reproducibility

Reproducible from public sources: Both curated tables are transcribed from published tables (Harper et al. 2021, Tables 2-5; Roberts et al. 2021, Table 4). build.js computes every saved analysis from them. The largest 12-month group (Scharff 2019) and the Wiik 2020 group were cross-checked against the original papers.

  • Nearly all participants were adults, and none were competitive athletes. No before-and-after study has measured athletic performance in transgender girls.
  • The 12-month medians count each study group once, regardless of size, and pool different muscles and strength tests.
  • The Air Force comparison is to an average woman, not to matched women with the same training.
  • Changes after 12 months are not charted; the review found too few studies to say whether losses continue.

Publications using this dataset

Version details

The exact citing versions, and the version of this dataset each cites.

  • Article trans-girls-athletic-advantage — version 3 (published) cites version 1
  • Article trans-girls-athletic-advantage — version 2 (published) cites version 1
  • Chart hormones-year-testosterone-vs-strength — version 1 (published) cites version 1
  • Chart air-force-run-gap-on-hormones — version 1 (published) cites version 1

Node: dataset:testosterone-suppression-athletic-measures:1