Do women gain as much from interval training as men?
As far as these trials show, yes: given the same HIIT, VO2max rose alike (g 0.57 each, 19 trials), threshold power too (5). Mostly untrained adults; non-randomised sex comparison.
Lock and colleagues asked whether biological sex changes how much VO2max and performance improve after high-intensity interval training (HIIT) or sprint interval training (SIT).
They included only trials that applied the same protocol to men and women and reported the sexes separately (or a sex × training analysis), so that the training dose was matched within each trial: 33 references from 28 trials, 965 participants (462 women, 503 men), searched to 8 September 2022.
Pre-post changes were pooled separately for women and men (Hedges' g, random effects) and the two pooled estimates compared. VO2max rose alike: g = 0.57 (95% CI 0.44-0.69) in women and 0.57 (0.42-0.72) in men across 19 trials (between-sex p = 0.965); in raw units +0.20 L/min and +3.34 mL/kg/min in women versus +0.32 L/min and +3.50 mL/kg/min in men, also not different.
Threshold power (lactate or ventilatory threshold, 5 trials) rose by g = 0.38 in both; peak power from the incremental test (8 trials) by 0.56 in women and 0.41 in men (p = 0.304). Interventions of 4 weeks or less produced smaller VO2max gains than longer ones in both sexes (p < 0.001), and training status explained much of the heterogeneity.
The one sex-specific signal, a larger peak-power gain in well-trained women (g 0.37 vs 0.17, 2 trials, 19 women and 18 men), rests on a between-subgroup test with p = 0.051 across six subgroups; the abstract calls it significant, the conclusion "borderline".
In the primary studies, every significant sex × training interaction for central cardiorespiratory measures (cardiac output, stroke volume, oxygen pulse, ventilation) favoured men, which the authors read as different routes to the same end.
Our reading: good evidence that a coach need not change an interval prescription because the athlete is a woman, measured mostly on untrained adults; it says nothing about cycle phase and little about trained athletes.
Protocol
Design: systematic review and meta-analyses; PRISMA; PROSPERO CRD42021272615. MEDLINE, SPORTDiscus and the ProQuest Sports Medicine & Education Index searched to 8 September 2022, plus reference lists of earlier HIIT reviews; screening by two independent reviewers in Covidence.
Inclusion: a HIIT or SIT intervention in healthy adults including both sexes; cardiorespiratory or performance outcomes; results presented separately for men and women, individual data with sex, or a sex × HIIT analysis. English-language peer-reviewed papers or theses. Excluded: diagnosed disease as the focus, dietary supplementation, drugs, bed rest, or HIIT not the main intervention. Measured or estimated VO2max, laboratory or field tests all accepted.
Included trials (Table 1): 4 randomised controlled trials, 3 non-randomised controlled, 2 quasi-randomised and 19 uncontrolled trials. Populations: sedentary, overweight or obese, older (three groups with mean ages of about 61-65), recreationally active and physical-education students, plus three trained or well-trained groups. One study with only two women was excluded from the analyses.
Protocols (Table 2): from 4-8 × 30-s all-out sprints and reduced-exertion HIIT (1-2 × 10-20 s) to 4 × 4 min at 85-95% HRmax and 10 × 90 s at 100% of peak power; mostly cycle ergometer, some running; 1 to 4 or more sessions a week (mostly 2-3); 2-3 weeks to 9 months.
Quality: Newcastle-Ottawa Scale (built for case-control studies), with men versus women in place of cases versus controls; 17 trials good, 8 fair, 3 poor; mean inter-rater agreement 82.54%.
Analysis: dependent-groups pre-post meta-analysis (Meta-Essentials), random effects, Hedges' g, with pre-post correlations imputed from the seven studies that allowed their calculation (about 0.8 for VO2max); thresholds 0.20 small, 0.60 moderate, 1.20 large.Subgroups: sex (primary), baseline training status (untrained, moderately trained, well trained), group mean age (<30, 30-45, >45), HIIT vs SIT, length (≤4, 5-9, ≥10 weeks). Sensitivity: poor-quality trials removed; raw mean differences in RevMan. One trial (battling-rope training with an arm-ergometer VO2max test) was removed from all VO2max analyses as an outlier.Publication bias by funnel plots with adjusted estimates.
Funding and interests: open-access publication funded by the Qatar National Library; Hamad Bin Khalifa University funded the Covidence licence and a postgraduate scholarship for the second author, whose master's thesis drew on initial results; the authors declare no conflicts of interest.
Key Results
Pooled pre-post effects (Table 3; % change is the unweighted mean across trials):
| Outcome | Women: trials, n, % change, g (95% CI) | Men: trials, n, % change, g (95% CI) | Between sexes |
|---|---|---|---|
| VO2max | 19, 265, 11.16%, 0.57 (0.44-0.69) | 19, 273, 10.90%, 0.57 (0.42-0.72) | p = 0.965 |
| VO2max, relative (mL/kg/min) | 16, 205, 0.57 (0.43-0.71) | 16, 249, 0.55 (0.38-0.71) | p = 0.845 |
| VO2max, absolute (L/min) | 7, 87, 0.56 (0.28-0.83) | 7, 91, 0.51 (0.32-0.70) | p = 0.805 |
| Peak power, incremental test | 8, 99, 11.16%, 0.56 (0.32-0.80) | 8, 143, 8.22%, 0.41 (0.22-0.59) | p = 0.304 |
| Threshold power | 5, 79, 8.07%, 0.38 (0.13-0.64) | 5, 123, 7.09%, 0.38 (0.11-0.64) | p = 0.964 |
Raw units (sensitivity analysis): baseline VO2max higher in men by 1.06 L/min and 5.88 mL/kg/min; change +0.32 vs +0.20 L/min (p = 0.38) and +3.50 vs +3.34 mL/kg/min (p = 0.88), men vs women. The authors warn that this analysis assumes independent data and probably underestimates heterogeneity (I² 0%).
Block length (Table 4, VO2max): ≤4 weeks g = 0.25 women and 0.16 men; 5-9 weeks 0.55 and 0.67; ≥10 weeks 0.79 and 0.66; between subgroups p < 0.001. The authors note that the moderately trained groups were almost all in 2-4-week trials, so length and training status are confounded.
Training status (VO2max): untrained g 0.63 women / 0.66 men; moderately trained 0.33 / 0.18 (8.40% vs 4.37%); well trained 0.72 / 0.86 (3 trials, 29 and 29). Peak power: untrained 0.77 / 0.60; moderately trained 0.35 / 0.27; well trained 0.37 / 0.17 (2 trials, 19 women, 18 men); the between-subgroup test across all six cells was p = 0.051.
HIIT vs SIT: no difference (VO2max p = 0.716; women 0.52 HIIT, 0.59 SIT; men 0.60, 0.46).
Age: no difference by group mean age; the authors say the ">45" groups were studies with mean ages of 59 or more, but the subgroup's 91 women and 44 men add up only with Dalzill 2014 (55 people, mean age 51.3 y) as the fourth trial (our arithmetic from Tables 1 and 3), and the 30-45 men's group hinged on one study.
In the primary studies: of 32 sex × training interaction tests for VO2max, 26 were not significant; 2 favoured women (one study) and 4 favoured men (four studies, all absolute VO2max).
Trained athletes, one trial (cyclists and triathletes, 8 women, 9 men, 10 sessions of 10 × 90 s at peak power over 6 weeks, Tables 6-7): 40-km time trial 2.69% faster in women and 1.70% in men (interaction p = 0.10); power at the second lactate threshold not significantly changed in either sex.
Publication bias: funnel plots for VO2max and peak power were asymmetric; adjusted effects 0.32 (vs observed 0.49) for VO2max and 0.25 (vs 0.41) for peak power; threshold power unchanged at 0.37. The abstract still says publication bias is unlikely to have significantly influenced the VO2max results, although the adjustment cut that effect by about a third (our arithmetic).
Numbers that do not add up:
The abstract says "significant differences were present for PPO" in well-trained women vs men; Table 3 gives p = 0.051 for the between-subgroup test, and that test spans six subgroups (three training levels × two sexes), so it is not a direct test of well-trained women against well-trained men; the Results say "p = 0.05" and the Conclusions "borderline significant".
Well-trained men's VO2max effect is g = 0.86 with a 95% CI of 0.86-0.87 (3 trials, 29 men), and their peak-power effect g = 0.17 with a 95% CI of 0.15-0.19 (2 trials, 18 men); intervals that narrow are not plausible for so few people, and the second is what makes the well-trained peak-power gap look significant (Table 3).
The funnel-plot section gives a combined observed VO2max effect of 0.49 (0.43-0.55), though women and men are each 0.57; the 0.49 and its narrow interval fit a fixed-effect or differently weighted pool rather than the random-effects model used elsewhere (our reading); the paper does not say which.
The Results say no relationships with time-trial change were present for men "(r2 = 0.01-0.21; p all < 0.05)"; a "no relationship" result should read p > 0.05, most plausibly a typo.
Mechanisms
Measured (in the primary studies, qualitative synthesis): every significant sex × training interaction for central cardiorespiratory measures favoured men: peak cardiac output, peak stroke volume and cardiac index after SIT (one study), oxygen pulse, maximal ventilation and carbon dioxide output, accumulated oxygen uptake.Men also showed larger gains in mitochondrial respiratory capacity, beta-HAD activity and GLUT4 in single studies, and in mitochondrial biogenesis and muscle protein synthesis in one more (called significant in the text, p = 0.056 in Table 8); women showed larger gains in type II fibre area and muscle glycogen in one study.
Authors' interpretation: men may rely more on improved oxygen delivery and uptake, women on other (possibly peripheral) routes, reaching similar fitness gains; in one trial of trained cyclists (measured, Table 7) the change in threshold power explained 77% of the variance in time-trial improvement in women (r2 = 0.77) and showed no significant relationship in men (r2 = 0.07, p = 0.49). They call these "different means to the same end".
Authors' interpretation: an earlier meta-analysis found larger VO2max gains in men after moderate-intensity endurance training; the authors suggest that the sex difference may depend on the protocol or on analysing absolute rather than relative change.
Inference: the mechanism findings come from single small studies selected because they reported a significant interaction; they are hypotheses, not a basis for different prescriptions.
Inference: a relative effect size that is equal in both sexes means a smaller absolute gain in L/min for most women because their baseline is lower; the raw-unit difference (0.20 vs 0.32 L/min) was not significant, and per kilogram the gains were almost identical (3.34 vs 3.50 mL/kg/min).
Practical Application
Do women gain as much from interval training as men? As far as these trials show, yes: given the same HIIT, VO2max rose alike (g 0.57 each, 19 trials), threshold power too (5). Mostly untrained adults; non-randomised sex comparison.
Prescribe intervals by the athlete, not by sex: with dose matched inside each trial, relative gains in VO2max, peak power and threshold power showed no evidence of a difference between women and men (A-, mostly Tier 0-1). The pooled per-kilogram VO2max gain was 3.34 mL/kg/min in women and 3.50 in men.
Give a VO2max block more than 4 weeks before judging it: blocks of 4 weeks or less gave small gains in both sexes (g 0.16-0.25) against 0.55-0.79 for longer ones; the base's practice of a retest after 3-6 weeks (SYNTH-33) should expect the smaller number at the short end (inference; length was confounded with training status here).
Trained women: in the two trials with well-trained participants, peak power rose more in women than in men (g 0.37 vs 0.17, 19 and 18 people, p = 0.051 for the whole subgroup test) — no basis for a different plan, at most a reason not to expect less.
Absolute numbers will differ: a woman's VO2max in L/min and her watts rise by less in absolute terms because they start lower (baseline gap 1.06 L/min); compare her gain with her own baseline, not with a male training partner's (inference).
Limitations
Observational sex comparison inside mostly uncontrolled trials. Only 4 of 28 trials were randomised, and none could randomise sex. Women and men were pooled as separate subgroups and their pooled estimates compared, although they come from the same trials; a within-trial difference was not analysed (our reading of the methods).
Who. Mostly sedentary to recreationally active adults, many under 30, mostly on cycle ergometers. Trained athletes are 2-3 trials per outcome; the well-trained peak-power signal rests on 19 women and 18 men.
Heterogeneity and small trials. I² 62-79% for VO2max and peak power; most trials small; pre-post correlations imputed for most studies; quality assessed with a tool built for case-control studies.
Size of the gains. Uncontrolled pre-post effects; funnel-plot asymmetry; the bias-adjusted VO2max effect is 0.32 rather than 0.49.
Hormones not assessed. Menstrual phase, contraception and menopause were outside the review's scope and largely ignored by the included trials.
Reporting. The abstract's "significant" peak-power difference is p = 0.051 across six subgroups; the 0.86-0.87 interval for well-trained men and the combined 0.49 effect do not fit the other numbers (see Key Results).
Interests. No conflicts declared; university and national-library funding only.
How this entry was checked and changed
Why this level
systematic review and meta-analysis read from the full text (open access, CC BY 4.0). The rubric's A- line is an SR/MA "with serious heterogeneity or mostly small trials": I² was 62-79% for VO2max and peak power, most trials are small, and only 4 of the 28 were randomised (between training arms or to a control, never by sex). The sex comparison itself is observational, but each trial gave both sexes the same protocol, which is the right design for this question; the absolute size of the gains (uncontrolled pre-post changes, funnel-plot asymmetry) is weaker evidence than the sex comparison
Full text (checked 2026-10-27)
Thirty-three references from 28 trials were included in the review (n = 965; 462 women and 503 men) (full text)
Similar increases in VO2max in women (g = 0.57; 95% CI 0.44-0.69) and men (g = 0.57; 95% CI 0.42-0.72), and powerAT in women (g = 0.38; 95% CI 0.13-0.64) and men (g = 0.38; 95% CI 0.11-0.64) (full text)
Raw mean differences for change in VO2max were 0.32 L/min and 3.50 mL/kg/min in men versus 0.20 L/min and 3.34 mL/kg/min for women (full text)
Peak power by baseline training status: between-group Q 11.04, df 5, p 0.051; well-trained women g 0.37 and well-trained men g 0.17, two studies of 18 men and 19 women (full text)
VO2max by intervention length: 4 weeks or less g 0.25 (women) and 0.16 (men); 10 weeks or more 0.79 and 0.66; between groups p < 0.001 (full text)
Of 32 sex × HIIT/SIT interaction analyses for VO2max, 26 were not significant; of the six that were significant, two favored females and four favored males (full text)
Adjusted effect size for VO2max 0.32 (95% CI 0.27-0.38) compared with the observed effect size of 0.49 (95% CI 0.43-0.55) (full text)
Well-trained men, VO2max: g 0.86, 95% CI 0.86-0.87, 3 trials, 29 participants (full text)