REF-1241 · Jaén-Carrillo et al., 2026 · Scandinavian Journal of Medicine & Science in Sports · Paper · PubMed · Free full text
Do women runners hold up better than men over 3 hours?
In 11 women and 11 men matched for level, uphill TT speed fell 1% vs 10% after 3 h, but the men ran farther. One lab, fed 90 g/h; field data in pro cyclists point the other way.
Jaén-Carrillo et al. asked whether durability, and the physiological, mechanical and neuromuscular changes beneath it, differ between highly trained female and male trail runners.
Eleven women and eleven men, pair-matched by sex-specific ITRA performance level and age, did a graded test, a fresh 12-min uphill time trial at +12%, and on a third visit a 180-min treadmill run at 85% of the speed at lactate threshold + 0.5 mmol/L with the same time trial after every 60 min, taking 90 g/h of carbohydrate gel and water as desired.
Time-trial speed fell as the run went on in both sexes, but after 3 h by 1.1% in women and 9.9% in men (Sex × Time P = 0.003).
The men's carbohydrate oxidation, respiratory exchange ratio and peak lactate in the time trials fell more, their heart rate and perceived effort drifted more in the steady running, and their knee-extensor force was 18% down at 2 h while the women's was unchanged; running economy worsened alike (no interaction), and gait changes were broadly similar.
The men covered more distance (42.2 vs 35.5 km) and spent more energy per kilogram (41.4 vs 34.6 kcal/kg); adding either as a covariate did not change the interactions. The authors conclude that highly trained women show greater durability and physiological resilience than performance-matched men, and that it should be checked at matched distance.
Our reading: the first time-matched laboratory study in the base in which women's performance faded less, and the first in runners and with fuel (REF-1132 found women's thresholds sliding less, with peak power falling alike); the opposite of the work-matched field study in professional cyclists (REF-1240).
Part of the gap may be dose (more distance, faster running) and fuel per kilogram rather than sex.
Protocol
Design: three laboratory visits within 14 days; ethics approval University of Innsbruck (no. 109/2024); no registration is reported. Sample size from earlier sex-difference work: at least 7 per group for 80% power; the authors state ~95% power with 11.
Who (Table 1): n = 22 trail runners, 11 men and 11 women; age 29 ± 6 vs 29 ± 4 y; body mass 72.9 ± 8.7 vs 60.9 ± 6.2 kg; training 88 ± 24 vs 66 ± 23 km and 3309 ± 919 vs 2436 ± 997 m a week; strength training 1.5 vs 1.4 sessions a week; ITRA 703 ± 88 vs 579 ± 64 points ("Advanced 1": 700-725 for men, 575-599 for women).
Visit 1: graded treadmill test (3-min stages, +1 km/h) with earlobe lactate; lactate threshold = highest speed before a systematic rise of about 0.5 mmol/L above rest; speed at LT 13.7 vs 11.5 km/h.
Visit 2 (72 h later): fresh 12-min uphill time trial at +12%, speed self-selected by hand signal to an investigator; gas exchange, heart rate, lactate, RPE and knee-extensor isometric force (iMVC) before and after.
Visit 3: 180 min at 85% of the speed at LT + 0.5 mmol/L (12.1 ± 1.2 km/h men, 10.2 ± 0.8 km/h women), stopping after 60, 120 and 180 min for the same uphill time trial. Gases in 5-min bins at 5, 60, 120 and 180 min and throughout each trial; gait from a foot-mounted Stryd sensor; iMVC and lactate before and after each trial. Carbohydrate gel at 90 g/h, water as desired. 21-24 °C, 45-55% humidity, fan at 2.55 m/s.
Analysis: linear mixed models with Sex, Time and Sex × Time and a random intercept per runner; distance or energy expenditure (kcal/kg) added as a covariate in separate models (supplementary Tables S1-S2, not in the open-access XML); Bonferroni-adjusted contrasts.
Funding and interests: "The authors have nothing to report"; no conflicts declared.
Key Results
| Outcome (change from fresh) | Women | Men | P (Sex × Time, or the sex contrast at that time point) |
|---|---|---|---|
| Uphill time-trial speed, after 3 h | −1.1% | −9.9% | P = 0.003 |
| Steady-state carbohydrate oxidation at 3 h | −9% | −29% | P < 0.05 |
| Steady-state RER at 3 h | −2.1% | −5.6% | contrast P < 0.05 |
| Steady-state RPE at 3 h | +30% | +47% | P = 0.03 |
| Knee-extensor iMVC at 2 h | 0% | −18% | P = 0.02 |
| Running economy at 3 h (mL/kg/km) | +2.8 ± 4.2% | +1.4 ± 5.2% | P = 0.24 (time effect P = 0.03) |
| Steady-state blood lactate at 3 h | +34.3 ± 26.7% | +20.9 ± 56.4% | P = 0.14 |
| Steady-state contact time / leg stiffness (changes as reported; time point shown in a figure) | +2.8% / −5.8% | +0.7% / −1.5% | P = 0.03 / 0.04 |
| Time-trial peak lactate, after 3 h | −27% (5.54 to 3.82 mmol/L) | −53% (7.44 to 3.38 mmol/L) | P = 0.006 |
| Time-trial RER, after 3 h | −4.3% (0.91 to 0.87) | −9.0% (0.96 to 0.87) | P = 0.02 |
| Time-trial carbohydrate oxidation, after 3 h | −16% | −36% | P < 0.001 |
| Time-trial stride length, after 2 h | −0.8% | −8.7% | P = 0.04 |
| Distance covered in the session | 35.5 ± 2.6 km | 42.2 ± 3.4 km | P < 0.001 (between sexes) |
| Total energy expenditure | 34.6 ± 3.8 kcal/kg | 41.4 ± 4.2 kcal/kg | P < 0.001 (between sexes) |
| Body mass change | −1.7 ± 0.8% | −2.7 ± 1.3% | P < 0.001 (between sexes) |
Heart rate: steady-state heart rate drifted upward with a Sex × Time interaction (smaller drift in women); the values are shown only in a figure. In the time trials peak heart rate and peak VO2 did not change in either sex.
Covariates: adding total distance or energy expenditure per kilogram left the Time and Sex × Time effects unchanged (authors' statement; supplement not read, and the captions of Tables S1-S2 name only total distance as the covariate).
Numbers that do not add up:
The text says the groups did not differ in training experience or strength-training frequency "(p ≥ 0.66)"; Table 1 gives p = 0.47 and 0.58. The text gives p = 0.03 for weekly volume and elevation; Table 1 gives 0.04 and 0.05.
The text says sex-specific performance status "did not differ" on ITRA points and level, while Table 1 gives p = 0.001 for the points; what was matched is the level band ("Advanced 1", defined separately for each sex). The standard deviations (±88 and ±64 points) are much wider than those bands (700-725 and 575-599), so not every runner can have been inside them (our reading).
Total energy for the men is printed as 3073 ± 38 kcal, but its two parts carry standard deviations of 303 and 92 kcal, so ±38 cannot be right (our arithmetic; probably a typo).
The Results place the sex difference in stride length in the post-2 h trial; the Discussion speaks of the post-3 h trial.
Against REF-548 (same first author, two shared co-authors and the same protocol; 23 runners including 11 women): its abstract reports energy cost and lactate stable in the steady running and oxygen cost and leg stiffness preserved across the trials, whereas this paper reports running economy and lactate rising and stiffness falling with time; and REF-548's 6.6% loss of trial distance compares with an equal-weight mean of about 5.5% here, or about 6% if the sexes are weighted by fresh time-trial speed (our arithmetic), so this gap alone is not decisive.Neither paper states the overlap.
Mechanisms
Measured: the women's smaller loss of uphill speed came with smaller falls in carbohydrate oxidation, RER and peak lactate in the time trials, smaller heart-rate and RPE drift in the steady running and preserved knee-extensor force; running economy, most gait measures, peak heart rate and peak VO2 changed alike or not at all.
Authors' interpretation: a more fatigue-resistant female profile — a larger share of type I fibres, more capillaries and more fat oxidation sparing glycogen, less muscle deoxygenation — gives metabolic and neuromuscular resilience; the mechanical changes are not sex-specific. They concede the extra distance may have partly emphasised the gap and call for distance-matched work.
Inference (dose): the men ran 19% farther than the women (42.2 vs 35.5 km, our arithmetic) at a higher absolute speed, with more load per stride; a single covariate for total distance cannot remove a dose that grows through the session, as the authors note.
Inference (fuel): 90 g/h for everyone was about 1.48 g/kg/h for the women and 1.23 g/kg/h for the men (mean body mass 60.9 vs 72.9 kg), and the 270 g over 3 h (about 1,100 kcal at the paper's 4.07 kcal/g) equalled about half the women's total expenditure (2147 kcal) but about a third of the men's (3073 kcal; our arithmetic). Relatively more exogenous carbohydrate may explain part of the women's steadier carbohydrate oxidation; the paper does not discuss it.
Inference: the men trained more (88 vs 66 km a week), so the gap is not explained by the women being better trained.
Practical Application
Do women runners hold up better than men over 3 hours? In 11 women and 11 men matched for level, uphill TT speed fell 1% vs 10% after 3 h, but the men ran farther. One lab, fed 90 g/h; field data in pro cyclists point the other way.
Both directions are in the base: two time-matched laboratory studies favoured women (REF-1132 cycling, water only: thresholds slid less, peak power fell alike; this one running, fuelled: uphill speed fell less), one work-matched field study of professionals found women faded more (REF-1240).Matching by time gives the men more work; matching by kJ/kg gives the women more time. No sex rule follows for an individual, which supports SYNTH-32's line not to assume a woman will fade less because of her sex.
Durability tests inside long runs: the sex difference showed more clearly in the repeated uphill trials than in steady-running heart rate or economy; the authors suggest brief maximal efforts inside long training runs as a more sensitive durability check than submaximal monitoring (their suggestion; the same protocol as REF-548).
Fuelling women by the gram per hour: a fixed 90 g/h is a larger dose per kilogram for a lighter athlete; runners should expect to tolerate less than riders (inference in SYNTH-02). The tolerance of 90 g/h here was not reported.
Sex-specific pacing (men more conservative on early climbs, women more even) is the authors' suggestion from these data, not a tested strategy.
Limitations
Small, one laboratory, one protocol: 22 runners on a treadmill with fixed +12% climbs and no downhill; level and downhill running, and outdoor terrain, may behave differently.
Matched by time and race rating, not by work: the women covered 16% less distance, spent 16% less energy per kilogram and ran 1.9 km/h slower; the authors' covariate adjustment cannot fully remove a dose that grows through the run.
Fuel not scaled: 90 g/h for all, a larger relative dose for the lighter women (inference above); gut tolerance and intake compliance not reported.
Female physiology not described: menstrual phase and hormonal contraception are not reported.
Sensor limits (authors): Stryd underestimates ground contact time and its stiffness model is not validated for steep uphill running.
Probable cohort overlap with [REF-548], not disclosed, and with differing conclusions on economy, lactate and stiffness.
Reporting: p-values in the text disagree with Table 1 for training experience, strength frequency, volume and elevation; the matching claim sits uneasily with the ITRA spread; the men's total-energy SD is implausible.
How this entry was checked and changed
Why this level
n = 22 (11 pairs); sex cannot be randomised; one laboratory protocol with a self-paced main outcome; matched by sex-specific race rating, not by work, so the men ran farther and spent more energy per kilogram; very probably the same runners as REF-548 (same first author, two shared co-authors, the same protocol), which this paper does not cite; read from the full text (open access, CC BY 4.0)
Full text (checked 2026-10-26)
Eleven male and eleven female highly-trained trail runners were pair-matched based on the ITRA performance index level and age (full text)
The prolonged run was set at 85% of the speed corresponding to LT + 0.5 mmol/L for 180 min, with a 12-min uphill TT at + 12% gradient every 60 min; a CHO gel solution was provided at a rate of 90 g/h (full text)
TT speed showed a Sex × Time interaction (p = 0.003), with females displaying smaller decreases in the post-3 h TT (−1.1% vs. −9.9%) (full text)
Total distance covered 42.2 ± 3.4 km (males) vs 35.5 ± 2.6 km (females); total EE 41.4 ± 4.2 vs 34.6 ± 3.8 Kcal/kg (full text)
At 3 h CHOox changed −29% vs. −9% (males vs females); RPE +30% vs. +47% (females vs males); relative iMVC at 2 h −18% (males) vs. 0% (females) (full text)
Running economy change at 3 h: 2.8 ± 4.2% (females) and 1.4 ± 5.2% (males); Sex × Time p = 0.24 (full text)
Peak B[La−] fell from unfatigued (7.44 vs. 5.54 mmol/L) to post-3 h TT by −53 vs. −27% (males vs females), reaching 3.38 vs. 3.82 mmol/L (full text)
ITRA points 703 ± 88 (males) vs 579 ± 64 (females), p = 0.001; Advanced 1 level points 700–725 for males and 575–599 for females (full text)
Training volume 88 ± 24 vs 66 ± 23 km/wk (p = 0.04) and training experience p = 0.47, ST frequency p = 0.58 in Table 1, while the text gives p ≥ 0.66 and p = 0.03 (full text)
Total EE (Kcal) 3073 ± 38 (males) vs 2147 ± 187 (females); steady-state EE 2376 ± 303 and TT EE 697 ± 92 in males (full text)
Ethics Committee of the University of Innsbruck (no. 109/2024); funding: the authors have nothing to report; no conflicts of interest declared (full text)