The present systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Cumpston et al., 2022). The review protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) under the registration number CRD420251272861. A comprehensive literature search was performed in the following electronic databases: PubMed, Web of Science, Scopus. The search covered the period from the inception of each database to [2025.11.30], with no restriction on publication year.

The search strategy was designed to identify randomized controlled trials (RCTs) examining the effects of high-intensity interval training (HIIT) on cardiorespiratory fitness and physical or functional performance in middle-aged and older women. Search terms were developed using combinations of controlled vocabulary (e.g., MeSH terms) and free-text keywords related to the following domains:

Population (“middle-aged,” “middle age,” “older adult*,” “elderly,” “aged,” “postmenopausal”);

The search strategies were adapted for each database using appropriate syntax and field tags (e.g., title, abstract, and keywords). The detailed search strategies for all databases are provided in Supplementary Material.

The retrieved literature was imported into Zotero software for management. Two researchers independently screened the literature by reviewing titles, abstracts, and full texts, removing duplicate and irrelevant studies, and extracting eligible data. Any disagreements during the screening process were resolved through discussion, and when necessary, consultation with a third reviewer. The literature screening process is illustrated in Figure 1.

The target population of this review consisted of middle-aged and older women, including peri- and postmenopausal women. Middle-aged was operationally defined as women aged 40 years and older, consistent with exercise physiology research definitions. Although participant ages ranged from 44 to 81 years across included studies, this broad inclusion was physiologically justified. From age 40 onward, women experience progressive hormonal changes, declining VO_2max (approximately 10% per decade), muscle mass loss, and increasing cardiovascular risk. These age-related physiological changes, rather than chronological age alone, represent the therapeutic target for HIIT interventions. Subgroup analyses by age (<65 vs. ≥ 65 years) were conducted where feasible to explore potential heterogeneity. There were no restrictions on country, ethnicity, or baseline health status, except for professional or competitive athletes. Only studies involving human participants were included.

The intervention was high-intensity interval training (HIIT), defined as repeated bouts of high-intensity exercise interspersed with recovery periods. The control group included non-exercise control, usual care, moderate-intensity continuous training, or other low-to moderate-intensity exercise interventions.

Only randomized controlled trials (RCTs) published in English or Chinese were eligible for inclusion. Studies were required to report at least one outcome related to cardiorespiratory fitness or physical/functional performance.

The outcome measures included indicators of cardiorespiratory fitness, such as maximal or peak oxygen uptake (VO_2max/VO_2peak), estimated VO_2max; and indicators of physical or functional performance, such as 6-min walk test (6MWT), Timed Up and Go (TUG), sit-to-stand or chair-stand tests, and time to exhaustion, gait speed, handgrip strength, and one-repetition maximum.

The data extraction included the following information: basic characteristics of the included studies (first author’s name, publication year, and country), participant characteristics (mean age, sex, sample size, and health status), intervention characteristics (exercise modality, intensity prescription, interval structure, frequency, and intervention duration), comparator details, outcome measures related to cardiorespiratory fitness and physical or functional performance, and key items for risk of bias assessment.

For the quantitative synthesis, means and standard deviations (SDs) of outcome measures at baseline and post-intervention were extracted for both intervention and control groups. When change-from-baseline values were directly reported, these data were preferentially extracted.

If outcome data were reported in alternative formats (e.g., standard errors, confidence intervals, or medians and interquartile ranges), they were converted to means and SDs using methods described in the Cochrane Handbook (Higgins et al., 2011) and by Wan et al. (Wan et al., 2014) when possible.

In cases of missing or unclear data, attempts were made to contact the corresponding authors to obtain the required information. Studies were excluded from the quantitative analysis if the necessary data could not be retrieved or reliably estimated. The detailed characteristics of the included studies are summarized in Supplementary Table S1.

RCTs were analyzed using the Cochrane Risk of Bias Tool 2.0 (Purgato et al., 2010). There are three levels: low risk, high risk, and uncertain. Two researchers used ReviewManager 5.4.1 software to rigorously evaluate five aspects of randomized allocation methods, allocation concealment of randomized methods, blinding of research subjects and interveners, blinding of outcome evaluators, integrity of outcome data, possibility of selective reporting, and other sources of bias. The risk of biased judgment in each domain was interpreted as low risk, moderate risk, severe risk, borderline risk, or no information. Two reviewers independently assessed the risk of bias, and any disagreements were resolved through a third party. In addition, the risk of publication bias was assessed using funnel plots when the meta-analysis included ≥5 studies.

All statistical analyses were conducted using Review Manager (RevMan) version 5.4.1. Meta-analyses were performed when at least two studies reported comparable outcomes. Outcomes related to cardiorespiratory fitness (as VO_2max and VO_2peak) and physical or functional performance (as 6-min walk test, Timed Up and Go, sit-to-stand performance, gait speed, muscle strength, and flexibility) were analyzed separately.

For continuous outcomes measured using the same unit, mean difference (MD) with 95% confidence intervals (CIs) was calculated. When outcomes were assessed using different scales or measurement methods, standardized mean difference (SMD) with 95% CIs was used. Given the expected clinical and methodological heterogeneity across studies (as differences in training protocols, intervention duration, participant characteristics, and outcome assessment methods), a random-effects model was applied for all meta-analyses. Statistical heterogeneity was assessed using the I² statistic and the Chi-square test, with I² values of approximately 25%, 50%, and 75% indicating low, moderate, and high heterogeneity, respectively.

Prespecified subgroup analyses were conducted to explore potential sources of heterogeneity according to age (<65 vs. ≥ 65 years), population characteristics (obese vs. non-obese), intervention duration, and training frequency. Sensitivity analyses were performed by sequentially excluding individual studies to assess the robustness of the pooled estimates. When at least five studies were included in a meta-analysis, publication bias was assessed visually using funnel plots.

The overall certainty of evidence for primary outcomes was evaluated using the GRADE approach (Prasad, 2024), considering risk of bias, inconsistency, indirectness, imprecision, and publication bias. A two-sided P value <0.05 was considered statistically significant for all analyses. Meta-regression was not performed due to insufficient studies (<10) for each outcome (Thompson and Higgins, 2002).

Figure 1 presents the flowchart of the literature screening process. A total of 288 relevant records were identified through searches of three databases (PubMed:47 articles, Web of Science: 175 articles, Scopus: 72 articles). After removing 225 duplicate publications, 63 articles proceeded to the screening process. During the title and abstract screening phase, 16 articles were excluded. During the full-text screening phase, 28 articles were excluded. Finally, 19 studies were included in the meta-analysis as Table 1 (Gliemann et al., 2020; Teixeira Do Amaral et al., 2024; Ballesta-García et al., 2019; Chen et al., 2024; Coswig et al., 2020; Dupuit et al., 2020a; Dupuit et al., 2022; Dupuit et al., 2020b; Henke et al., 2018; Kazemi et al., 2023; Klonizakis et al., 2014; Marcotte-Chénard et al., 2021; Noorbakhsh and Dabidi Roshan, 2023; Norling et al., 2024; Pan et al., 2025; Steckling et al., 2019; Twerenbold et al., 2023; Yu et al., 2025; Zanini et al., 2025).

The included studies involved a total of 646 middle-aged and older women, with mean ages ranging from approximately 44–81 years. Across the studies, participants included healthy women as well as those who were overweight or obese, postmenopausal, or with cardiometabolic risk factors. The intervention group consisted of participants receiving high-intensity interval training (HIIT), while control groups included non-exercise controls, moderate-intensity continuous training, resistance training, or combined exercise interventions.

The training frequency ranged from 2 to 4 sessions per week, and the intervention duration varied from 2 weeks to 9 months. HIIT protocols differed in exercise modality, including cycling, treadmill walking or running, Nordic walking, and sport-based interval training, with exercise intensity generally prescribed using heart rate, oxygen uptake, power output, or perceived exertion. Outcome measures primarily assessed cardiorespiratory fitness, such as VO_2max or VO_2peak, and physical or functional performance, including the 6-min walk test, Timed Up and Go, sit-to-stand performance, gait speed, muscle strength, and flexibility.

Most studies reported supervised exercise interventions; however, the level of detail regarding supervision varied across trials. Participant withdrawals were reported in several studies, commonly due to personal reasons or non-exercise-related factors.

Risk of bias assessment using Cochrane RoB 2 showed variable quality across 19 RCTs (Figures 2, 3). Most studies had low risk for randomization (63%), missing data (74%), and selective reporting (84%). Overall, 42% had low risk of bias, 53% raised some concerns, and 5% had high risk.

Subgroup analyses were performed for walking ability as Table 2, VO_2peak, and VO_2max as these outcomes included sufficient studies to allow meaningful stratified comparisons. For walking ability, participants ≥65 years showed a borderline significant improvement (5 studies, SMD 0.65, 95% CI: 0.00 to 1.30, p = 0.05, I² = 74%). Training frequency analysis revealed a significant effect with 2 sessions per week (4 studies, SMD 0.93, 95% CI: 0.29 to 1.58, p = 0.005, I² = 68%). Other subgroups showed no significant effects. For VO_2peak, only participants <64 years showed significant improvements (4 studies, SMD 0.50, 95% CI: 0.13 to 0.88, p = 0.009, I² = 0%).

For VO_2max, significant improvements were observed across all subgroups: age (<65 years: 5 studies, SMD 1.19, 95% CI: 0.51 to 1.87, p = 0.0006, I² = 60%; ≥65 years: 3 studies, SMD 1.23, 95% CI: 0.67 to 1.80, p = 0.0001, I² = 0%); population (non-obese: 2 studies, SMD 0.74, 95% CI: 0.11 to 1.38, p = 0.02, I² = 32%; obese: 6 studies, SMD 1.38, 95% CI: 0.98 to 1.78, p < 0.0001, I² = 17%); and intervention duration (≤10 weeks: 4 studies, SMD 1.06, 95% CI: 0.60 to 1.53, p < 0.0001, I² = 22%; ≥12 weeks: 4 studies, SMD 1.35, 95% CI: 0.85 to 1.84, p < 0.0001, I² = 47%).

Sensitivity analyses indicated that the pooled effects for most outcomes were robust to the exclusion of individual studies, with the exception of one outcome that demonstrated substantial dependence on a single influential study. Overall, these findings suggest that the main conclusions of the meta-analysis are generally stable, although results for outcomes with high heterogeneity and limited numbers of studies should be interpreted with caution (As Supplementary Figure S1).

Due to the limited number of studies included in each meta-analysis (<10 studies for all outcomes), statistical tests for asymmetry (e.g., Egger’s test) were not performed. Publication bias was assessed using funnel plots for outcomes with ≥5 studies (Sterne et al., 2011). Funnel plots for VO_2max (Figure 11) and VO_2peak (Figure 12) showed relatively symmetrical distribution of studies. The funnel plot for walking ability (Figure 13) showed some asymmetry with fewer studies on the left side, suggesting potential publication bias for this outcome.

The strength of evidence according to GRADE criteria generated by the GRADEpro website (https://www.gradepro.org/) is listed on Table 3.

For cardiorespiratory fitness outcomes, VO_2max was rated as high-quality evidence, showing a significant improvement with HIIT (SMD 1.20, 95% CI: 0.86–1.54) and low heterogeneity (I² = 31%). VO_2peak was rated as low-quality evidence, downgraded by two levels due to moderate inconsistency (I² = 51%) and imprecision (wide confidence interval crossing the line of no effect).

For physical performance outcomes, all indicators demonstrated very low to low-quality evidence. Walking ability was rated as very low-quality evidence, downgraded by three levels for serious inconsistency (I² = 73%), imprecision (confidence interval touching null), and suspected publication bias (asymmetric funnel plot). Timed Up and Go, sit-to-stand performance, and muscle strength were all rated as very low-quality evidence, primarily downgraded for very serious inconsistency (I² = 89%–94%) and imprecision. Flexibility was rated as low-quality evidence, downgraded by two levels for serious inconsistency (I² = 75%) and imprecision.

Sex-specific considerations. Postmenopausal estrogen decline accelerates cardiovascular aging and impairs muscle protein synthesis, potentially modifying HIIT responses compared to men. Previous mixed-sex meta-analyses may mask these differences. Our women-focused approach addresses this gap, though direct sex comparisons await future research.

Meta-analysis showed HIIT significantly improved VO_2max (SMD 1.20, 95% CI: 0.86–1.54, I² = 31%), as detailed in Results. These cardiorespiratory gains are clinically meaningful given the well-established association between fitness level and reduced mortality risk in women (Womack et al., 2003). HIIT produced greater VO_2max improvements than moderate-intensity training across diverse populations (Milanović et al., 2015).

Even ultra-short interventions demonstrate rapid adaptations. A 2-week low-volume HIIT protocol (six sessions, 10 × 1-min intervals at 100% peak power) increased VO_2peak by 2.2 mL·kg⁻¹·min⁻¹ (11%, P = 0.01) in postmenopausal women with half the training volume (558 vs. 1,237 kJ) and time (2.5 vs. 5 h) versus moderate-intensity training (Klonizakis et al., 2014). An 8-week Tabata-style HIIT protocol (twice weekly at 80%–90% HRmax) significantly improved VO_2max and oxygen pulse in overweight elderly women (Noorbakhsh and Dabidi Roshan, 2023). In sedentary older women (60–75 years), 8-week HIIT increased VO_2max by 20% (19.36–23.25 mL·kg⁻¹·min⁻¹, p < 0.001) with 97% adherence, alongside significant cognitive improvements (Norling et al., 2024).

In older women with type 2 diabetes, 12-week low-volume walking HIIT (75 min/week: 6 × 1-min at 90% HRR) produced similar VO_2peak improvements (+7.6%) as moderate-intensity training (150 min/week, +7.0%) despite 50% less training time, with superior improvements in 6-min walk distance (HIIT: +98 ± 56 m vs. MICT: +28 ± 70 m, P = 0.01) and grip strength (HIIT: +3.8 ± 5.5 kg vs. MICT: +0.1 ± 1.9 kg, P = 0.02) (Marcotte-Chénard et al., 2021). Subgroup analyses showed consistent VO_2max benefits across ages when sessions met vigorous-intensity guidelines (≥75 min·week⁻¹) (Dupuit et al., 2020a; Dupuit et al., 2022).

High compliance rates (>85%) suggest HIIT feasibility in this population, though some women report higher perceived exertion during intervals versus continuous exercise (Jung et al., 2014).

HIIT did not significantly improve muscle strength, with only borderline effects on walking ability. Low muscle strength is sarcopenia’s main diagnostic criterion and predicts disability, falls, and mortality (Cruz-Jentoft et al., 2019; Newman et al., 2006). Age-related muscle loss involves neuromuscular, hormonal, and inflammatory changes not fully addressed by HIIT alone.

A 12-week trial comparing HIIT-based Nordic walking with strength training in postmenopausal women (60–79 years) showed both interventions prevented sarcopenia: HIIT Nordic walking improved lower limb strength (chair stand, knee flexor strength, timed up-and-go) and skeletal muscle index, while strength training enhanced upper limb strength (hand grip, arm curl) and reduced body fat (Pan et al., 2025). Notably, HIIT interventions appear to predominantly benefit lower-body strength, with less consistent effects on upper-body strength measures such as handgrip strength (Pan et al., 2025), suggesting potential site-specific adaptations to this training modality.

High-intensity and moderate-intensity circuit training produced similar lower-limb strength and balance improvements, with additional upper-limb gains in the high-intensity group (Ballesta-García et al., 2019).

In elderly nursing home residents, 8-week HIIT (4 × 4-min at 85%–95% HRmax) improved chair stand and 6-min walk tests more than moderate-intensity training, with sustained benefits after 2–4 weeks detraining while other groups declined below baseline (Coswig et al., 2020). Walking performance improvements are clinically relevant for independence and survival (Studenski et al., 2011), but larger trials are needed to determine if HIIT alone reliably improves gait speed without targeted strength and balance training (Garber et al., 2011).

Beyond cardiorespiratory benefits, HIIT favorably modulates metabolic and inflammatory profiles. In postmenopausal women with metabolic syndrome, 12-week HIIT (4 × 4-min at 90% HRmax, 3×/week) improved VO_2max (+27.7%), body composition, glucose control, and blood pressure while reducing inflammatory cytokines (IL-6, IL-18, TNF-α, IFN-γ) and increasing anti-inflammatory IL-10, alongside beneficial adipokine shifts (increased adiponectin; decreased resistin, leptin, ghrelin) (Steckling et al., 2019). Even 8-week HIIT demonstrated microvascular endothelial improvements in hypertensive patients, with significant increases in retinal arteriolar flicker-induced dilation independent of blood pressure changes (Twerenbold et al., 2023).

Combined interventions may amplify benefits. Tabata-HIIT plus nanocurcumin supplementation produced superior improvements in body composition, VO_2max, and inflammasome suppression versus exercise alone (Noorbakhsh and Dabidi Roshan, 2023). In obese middle-aged women with prehypertension, 6-week concurrent training combining resistance exercise with HIIT showed superior blood lipid improvements (LDL-C, HDL-C, LDL-C/HDL-C ratio) compared to resistance plus moderate-intensity training (Yu et al., 2025).

In socioeconomically vulnerable older women, 6-month community-based concurrent training (HIIT + RT, MICT + RT, or RT alone, twice weekly) yielded comparable improvements in lower limb strength, mobility, aerobic performance, and mood profile regardless of exercise intensity, demonstrating accessible community programs can improve physical function and mental health (Zanini et al., 2025).

Implementation strategies should consider individual preferences, time constraints, facility access, and supervision requirements to maximize adherence. Optimal prescription parameters—session frequency, interval duration, intensity targets, progression schemes—remain to be fully elucidated for diverse subpopulations.

Substantial heterogeneity was observed for muscle strength (I² = 89%), sit-to-stand performance (I² = 91%), Timed Up and Go (I² = 94%), and flexibility (I² = 75%). High I² values (>75%) indicate considerable inconsistency that may reflect true differences rather than sampling variation alone (Higgins et al., 2003).

Methodological sources. In our meta-analysis of muscle strength outcomes, handgrip dynamometry was the primary assessment method due to its standardization and prevalence across included studies (Dabidi Roshan et al., 2024). While other strength measures (e.g., one-repetition maximum testing, isokinetic dynamometry) capture distinct neuromuscular aspects (Roberts et al., 2011), insufficient data precluded their inclusion in the pooled analysis.

Sit-to-stand tests employed different protocols (30-s chair stand versus five-repetition timed tests), assessing different capacities (muscular endurance versus power).

HIIT protocols differed substantially in modality. Cycle ergometry-based HIIT (Dupuit et al., 2020a; Dupuit et al., 2022; Klonizakis et al., 2014) may elicit different neuromuscular adaptations than weight-bearing modalities such as treadmill walking (Marcotte-Chénard et al., 2021) or Nordic walking (Pan et al., 2025), as non-weight-bearing exercise reduces mechanical loading. Additionally, intensity prescription methods, interval structure (work-to-rest ratios: 1:1 to 1:4), and session duration varied, influencing adaptation magnitude and specificity (Buchheit and Laursen, 2013).

Beyond heterogeneity issues, several methodological considerations warrant attention. The limited number of studies per outcome (fewer than 10) precluded meta-regression analyses. Qualitative examination suggests the high heterogeneity in physical function outcomes (I² = 75–89%) may reflect protocol differences, with HIIT combined with resistance training showing greater strength benefits than aerobic HIIT alone. First, most studies had short intervention durations (median 12 weeks), potentially insufficient for detecting changes in vascular function. While short-term HIIT (2 weeks) rapidly improves cardiopulmonary function, macrovascular and microvascular function remained unchanged, suggesting vascular adaptations may require longer periods (>12–24 weeks) (Klonizakis et al., 2014). Second, most studies used cycle ergometry, with limited data on other modalities. Third, we did not assess quality of life, psychological wellbeing, or long-term adherence patterns. Fourth, potential publication bias cannot be ruled out. The limited number of studies precluded meta-regression to formally explore sources of heterogeneity.

Future studies should prioritize adequately powered randomized controlled trials comparing different HIIT protocols with moderate-intensity training and combined training. Standardized reporting of exercise dose (intensity, duration, time near VO_2max), adherence, and adverse events would enable precise dose-response modeling (Papadopoulou, 2020; MacInnis and Gibala, 2017). Longer follow-up is needed to determine whether VO_2max improvements translate into reduced cardiovascular events, disability, and mortality. Combining physiological and imaging measures with functional and patient-reported outcomes could clarify mechanisms linking HIIT to muscle, adipose tissue, and cardiometabolic health. Pragmatic studies are warranted to evaluate HIIT integration into community and clinical programs, including strategies to maximize adherence addressing established correlates of physical activity participation (Trost et al., 2002).

This systematic review provides high-quality evidence that HIIT elicits clinically meaningful improvements in cardiorespiratory fitness in middle-aged and older women with a time-efficient format. Current evidence does not support meaningful benefits of HIIT alone for muscle strength or physical function, outcomes central to sarcopenia prevention and disability reduction. HIIT should be incorporated as one element within comprehensive exercise programs including resistance and balance training to address the multidimensional needs of aging women.