Effects of exercise snacks on cardiorespiratory fitness, body composition, and blood lipids in adults across different age groups: a systematic review and meta-analysis

Graphical abstract

Abstract

Objective

To evaluate the effects of exercise snacks on cardiorespiratory fitness, body composition, and blood lipids among adults of different age groups.

Methods

PubMed, Web of Science, CINAHL, Embase, the Cochrane Library, and Scopus were searched from inception to April 15, 2026. Randomized controlled trials evaluating exercise snacks in adults were included. Two reviewers independently performed study selection, data extraction, and risk-of-bias assessment. Subgroup analyses, meta-regression, sensitivity analyses, and publication bias assessments were conducted.

Results

Twenty-one studies involving 921 participants were included. Exercise snacks significantly improved maximal oxygen uptake (VO₂max; g = 1.04, 95% CI = 0.68–1.39; moderate-certainty evidence) and peak power output (PPO; g = 0.68, 95% CI = 0.25–1.10; moderate-certainty evidence), whereas body fat percentage showed an overall increase (g = 0.51, 95% CI = 0.01–1.02; low-certainty evidence). No statistically significant effects were observed for total cholesterol, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, or triglycerides. VO₂max improved in studies with mean participant ages <25 and 25–50 years, but not >50 years, whereas PPO improved across all three age strata. Further analyses suggested that a greater number of exercise bouts per day, an intervention frequency of 4–5 days per week, and an intervention duration of 7–9 weeks may be associated with greater improvements in cardiorespiratory fitness.

Conclusion

Exercise snacks may improve cardiorespiratory fitness in adults, although effects may vary by age and intervention characteristics. Evidence for body composition and blood lipids remains inconsistent, and the age-based findings require confirmation using individual participant data.

1. Introduction

The World Health Organization has identified insufficient physical activity as an important global health risk. In 2022, approximately 31% of adults worldwide, equivalent to about 1.8 billion people, did not meet the recommended levels of physical activity. This proportion increased by approximately 5 percentage points between 2010 and 2022 and may reach 35% by 2030 if current trends continue [1]. Insufficient physical activity is associated with increased risks of cardiovascular disease, type 2 diabetes, certain cancers, and premature mortality [2]. Although regular exercise can improve cardiorespiratory fitness and help prevent chronic disease [3], many adults find it difficult to allocate uninterrupted time for exercise because of work demands, family responsibilities, and other daily commitments [4,5]. Developing flexible and time-efficient strategies to increase physical activity participation has therefore become an important priority [6].

Exercise snacks, defined as multiple brief bouts of physical activity performed throughout the day, have received increasing attention [7]. Common modalities include stair climbing, short sprints, cycling, and resistance exercise, which may be integrated into work, commuting, or home-based routines [8]. Their low time requirement and flexibility may facilitate physical activity participation and help interrupt prolonged sedentary behavior [9]. Repeated brief exercise bouts may improve cardiorespiratory fitness through adaptations in oxygen delivery and utilization and may influence body composition and blood lipid regulation through repeated skeletal muscle activation and increased energy expenditure [10,11]. One meta-analysis, for instance, reported improvements in maximal oxygen uptake and peak power-related outcomes and further examined whether intervention effects varied across age groups [12]. However, the effects on body composition and blood lipid outcomes remain less clear and have not yet been investigated from an age-stratified perspective.

The present review was therefore designed to update and extend the existing evidence in three respects. First, in addition to cardiorespiratory fitness outcomes, we quantitatively synthesized body fat percentage and blood lipid outcomes. Second, age-related variation was examined using both study-level age-stratified analyses and meta-regression. Third, we explored whether bout duration, daily bout frequency, weekly intervention frequency, and intervention duration contributed to differences in effect estimates. Accordingly, this systematic review and meta-analysis aimed to evaluate the effects of exercise snacks on cardiorespiratory fitness, body composition, and blood lipid outcomes across adult age strata.

2. Methods

This study was conducted in strict accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) statement [13], and was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) (registration number: CRD420261371903).

2.1. Literature search strategy

A systematic search was conducted in PubMed, Web of Science, CINAHL, Embase, the Cochrane Library, and Scopus from database inception to April 15, 2026. The search strategy combined Medical Subject Headings (MeSH) and free-text terms to identify randomized controlled trials (RCTs) examining the effects of exercise snacks on cardiorespiratory fitness, body composition, and blood lipid outcomes in adults as comprehensively as possible. To minimize the risk of missing eligible studies, the reference lists of the included studies and relevant systematic reviews were also manually screened. The detailed search strategy is provided in Supplementary Table S1.

2.2. Eligibility criteria

The eligibility criteria were predefined according to the PICOS framework [14]. Studies were included if they met the following criteria: (1) participants were adults, regardless of sex; (2) the intervention involved exercise snacks or other fragmented exercise patterns with similar characteristics, defined as physical activity accumulated through multiple brief, low-volume exercise bouts, with no restriction on exercise modality; (3) the control condition involved maintaining the original lifestyle, receiving no structured exercise intervention, receiving usual health advice, or another comparable control condition; (4) at least one prespecified outcome related to cardiorespiratory fitness, body composition, or blood lipids was reported. Cardiorespiratory fitness outcomes included maximal oxygen uptake (VO₂max) and peak power output (PPO), body composition was represented by body fat percentage (BF%), and blood lipid outcomes included total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG); and (5) the study design was an RCT, with no language restriction. Studies were excluded if they met any of the following criteria: (1) animal experiments, in vitro studies, reviews, conference abstracts, dissertations, commentaries, or methodological papers; (2) interventions combined exercise snacks with diet, medication, or conventional exercise training, such that the independent effect of exercise snacks could not be distinguished; (3) participants were children or professional athletes; or (4) the prespecified outcomes were not reported, or the data were incomplete and effect sizes could not be extracted or calculated.

2.3. Study selection and data extraction

All retrieved records were first imported into EndNote X9 for duplicate removal [15]. Subsequently, two reviewers independently screened the studies according to the predefined eligibility criteria. The screening process consisted of two stages. First, titles and abstracts were screened to exclude clearly irrelevant studies. Second, the full texts of potentially eligible studies were obtained and assessed in detail to determine whether their participants, interventions, control conditions, outcomes, and study designs met the inclusion criteria. Any disagreement between the two reviewers was resolved through discussion, and when necessary, a third reviewer was consulted for adjudication.

Data extraction was also performed independently by two reviewers and cross-checked afterward to ensure accuracy and completeness. Extracted information included basic study characteristics, participant characteristics, intervention protocols, control conditions, and outcome measures. Specifically, the following data were collected: first author, year of publication, country, sample size, sex distribution, age, body mass index, exercise modality, duration of each exercise session, number of bouts per day, weekly intervention frequency, and intervention duration. When multiple follow-up time points were reported in the same study, data at the end of the intervention were preferentially extracted. If original data were incompletely reported, the necessary information was extracted or calculated from the text, tables, or figures whenever possible.

2.4. Risk of bias and certainty of evidence assessment

The risk of bias and certainty of evidence for the included studies were independently assessed by two reviewers, with disagreements resolved by discussion or, when necessary, by consultation with a third reviewer.

Risk of bias was assessed using the Risk of Bias 2 tool (RoB 2) [16], which evaluates the following domains: the randomization process, deviations from intended interventions, missing outcome data, measurement of the outcome, and selection of the reported result. Each domain was judged as low risk of bias, some concerns, or high risk of bias.

The certainty of evidence for each outcome was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach [17]. The certainty of evidence was rated as high, moderate, low, or very low, taking into account study limitations, inconsistency, indirectness, imprecision, and publication bias.

2.5. Statistical analysis

Because all included outcomes were continuous variables and differences existed across studies in measurement methods, units, and participant characteristics, pooled effect sizes were calculated using a random-effects model and expressed as Hedges’ g (g) with 95% confidence intervals (95% CI) [18]. Between-study heterogeneity was assessed using Cochran’s Q test and the I² statistic. The between-study variance (τ²) was estimated using the restricted maximum likelihood (REML) method. According to conventional criteria, g < 0.20 was considered a trivial effect, 0.20 ≤ g < 0.50 a small effect, 0.50 ≤ g < 0.80 a moderate effect, and g ≥ 0.80 a large effect.

Pooled standard deviation:

$${SD}_p=\sqrt{\frac{(n_e-1)S{D}_e^2+(n_c-1)S{D}_c^2}{n_e+n_c-2}}$$ (1)

where n_e and n_c denote the sample sizes of the experimental and control groups, and SD_e and SD_c represent their respective standard deviations.

Cohen’s d:

$$d=\frac{Mea{n}_e-Mea{n}_c}{S{D}_p}$$ (2)

where Mean_e and Mean_c denote the means of the experimental and control groups.

Hedges’ g correction:

$$g=d\times \left(1-\frac{3}{4(n_e+n_c)-9}\right)$$ (3)

where d is Cohen’s d, and n_e and n_c are the sample sizes of the experimental and control groups.

According to the mean age reported in each included study, studies were categorized into a lower-age stratum (<25 years), an intermediate-age stratum (25–50 years), and a higher-age stratum (>50 years). To examine the robustness of the findings using a more conventional threshold for older populations, an additional sensitivity analysis was conducted by restricting the analysis to studies with a reported mean participant age of ≥60 years, consistent with the threshold adopted in the World Health Organization’s healthy ageing framework [19]. Meta-regression was also performed to explore the association between study-level mean age and intervention effects [20]. Subgroup analyses were conducted according to session duration, number of exercise bouts per day, weekly intervention frequency, and intervention duration.

To examine the robustness of the main findings, leave-one-out sensitivity analyses were performed [21]. When a sufficient number of studies were available for a given outcome, funnel plots, Egger’s regression test, and Begg’s rank correlation test were used to assess publication bias [22]. In interpreting the results, positive effect sizes for VO₂max, PPO, and HDL-C indicated improvement, whereas negative effect sizes for BF%, TC, LDL-C, and TG indicated improvement. The minimum important difference (MID; g = 0.50), as recommended in the Cochrane Handbook for Systematic Reviews of Interventions (version 6.5), was also used as an auxiliary threshold for interpreting the clinical relevance of effect sizes [23]. All statistical analyses were conducted using R software (version 4.4.2; R Foundation for Statistical Computing) [24]. All tests were two-sided, and P < 0.05 was considered statistically significant.

3. Results

3.1. Search results

A total of 6,934 records were identified, including 6,872 from database searches and 62 from other sources. After duplicates were removed, 2,115 records remained for title and abstract screening. Following the initial screening, 72 full-text articles were assessed for eligibility, including 68 identified through database searches and 4 from other sources. After full-text review, 51 studies were excluded for the following reasons: study design not meeting the inclusion criteria (n = 19), control condition not meeting the inclusion criteria (n = 12), intervention not meeting the inclusion criteria (n = 15), participants not meeting the inclusion criteria (n = 2), and outcomes not meeting the inclusion criteria (n = 3). Ultimately, 21 studies were included in the systematic review and meta-analysis (Fig. 1).

Fig. 1.

Flow diagram of the literature search and study selection process.

3.2. Characteristics of the included studies

A total of 21 studies were included in the systematic review and meta-analysis. These studies were published between 1994 and 2025 and were conducted in the United States, the United Kingdom, Denmark, Singapore, Canada, Serbia, the Netherlands, Germany, China, and Korea. The mean age of participants ranged from 18.9 ± 0.6 to 75.7 ± 8.9 years, and the reported body mass index ranged from 20.9 ± 1.5 to 39.6 ± 7.8 kg/m². The exercise-snack interventions mainly included stair climbing, resistance exercise, cycling, and sprint interval exercise.

3.3. Meta-analysis

Compared with the control group (Fig. 2), exercise snacks significantly improved VO₂max (g = 1.04, 95% CI = 0.68–1.39, P < 0.01, I² = 74; moderate-certainty evidence) and PPO (g = 0.68, 95% CI = 0.25–1.10, P < 0.01, I² = 63; moderate-certainty evidence), whereas BF% showed an overall increase (g = 0.51, 95% CI = 0.01–1.02, P = 0.04, I² = 84; low-certainty evidence). By contrast, no statistically significant overall effects were observed for TC (g = −0.10, 95% CI = −0.48 to 0.28, P = 0.60, I² = 72; low-certainty evidence), HDL-C (g = −0.26, 95% CI = −0.75 to 0.22, P = 0.28, I² = 82; low-certainty evidence), LDL-C (g = 0.05, 95% CI = −0.37 to 0.46, P = 0.82, I² = 75; low-certainty evidence), or TG (g = 0.18, 95% CI = −0.26 to 0.62, P = 0.42, I² = 78; low-certainty evidence).

Fig. 2.

Meta-analysis of cardiorespiratory fitness, body composition, and blood lipid outcomes across age strata.

After stratification by age, improvements in VO₂max were primarily observed in the lower-age stratum (g = 1.20, 95% CI = 0.57–1.83, P < 0.01, I² = 72; moderate-certainty evidence) and the intermediate-age stratum (g = 1.14, 95% CI = 0.58–1.71, P < 0.01, I² = 76; moderate-certainty evidence), whereas no statistically significant effect was found in the higher-age stratum (g = 0.56, 95% CI = −0.17 to 1.30, P = 0.13, I² = 78; low-certainty evidence). PPO increased in the lower-age stratum (g = 0.83, 95% CI = 0.16–1.51, P = 0.01, I² = 69; moderate-certainty evidence), the intermediate-age stratum (g = 0.65, 95% CI = 0.26–1.05, P < 0.01, I² = 0; moderate-certainty evidence), and the higher-age stratum (g = 0.52, 95% CI = 0.07 to 0.97, P = 0.02, I² = 41; moderate-certainty evidence). BF% increased only in the intermediate-age stratum (g = 1.06, 95% CI = 0.08–2.05, P = 0.03, I² = 92; low-certainty evidence). TC decreased only in the lower-age stratum (g = −0.92, 95% CI = −1.48 to −0.35, P < 0.01, I² = 0; moderate-certainty evidence) and the higher-age stratum (g = −0.31, 95% CI = −0.48 to −0.14, P = 0.04, I² = 0; low-certainty evidence). No statistically significant differences were observed for the remaining outcomes across age groups.

3.4. Meta-regression

The meta-regression results (Fig. 3) showed that the effect size for VO₂max was associated with age and generally declined with increasing age. Both the linear model (R² = 0.53, P < 0.01) and the quadratic model (R² = 0.60, P < 0.01) were statistically significant, suggesting a possible nonlinear relationship between age and the intervention effect on VO₂max. The effect size for BF% also showed an overall downward trend, with the linear model reaching statistical significance (R² = 0.31, P = 0.05), whereas the quadratic model was not statistically significant (R² = 0.34, P = 0.13). No clear significant associations were observed for the remaining outcomes. Overall, age may be an important moderator of the intervention effect on VO₂max, whereas no definite significant associations were identified between age and the effect sizes of the other outcomes.

Table 1.

Basic characteristics of the included studies.

Study	Country	Age (years)	BMI (kg/m2)	Participants, n (M/F)	Intervention context	Reported exercise intensity	Exercise protocol	Reported outcomes
Allemeier et al., 1994 [25]	United States	22.7 ± 5.0	–	17 (17/0)	Laboratory setting	Supramaximal intensity (Wingate; resistance equivalent to 7.5% of body mass)	Sprint interval, 0.5 min/session, 3 bouts/d, 3 d/wk, 6 wk	VO2max, PPO, BF%
Boreham et al., 2000 [26]	United Kingdom	19.8 ± 0.3	–	22 (0/22)	Public stairway	Brisk pace (∼199 steps in 2 min)	Stair climbing, 2 min/session, 1−6 bouts/d, 5 d/wk, 7 wk	VO2max, PPO, BF%, TC, HDL-C
Boreham et al., 2005 [27]	United Kingdom	18.9 ± 0.6	20.9 ± 1.5	15 (0/15)	Public stairway	Brisk pace (90 steps/min)	Stair climbing, 2 min/session, 1−5 bouts/d, 5 d/wk, 8 wk	VO2max, TC, HDL-C, LDL-C, TG
Kennedy et al., 2007 [28]	United Kingdom	44.3 ± 7.4	25.3 ± 3.1	45 (22/23)	Workplace stairway	Moderate pace (75 steps/min)	Stair climbing, 2 min/session, 1−5 bouts/d, 5 d/wk, 8 wk	VO2max, BF%, TC, HDL-C, LDL-C, TG
Metcalfe et al., 2012 [29]	United Kingdom	26 ± 3	23.6 ± 1.1	29 (13/16)	Laboratory setting	Cycling at 60 W interspersed with 10−20-s all-out sprints	Sprint interval, 10 min/session, 3 bouts/d, 3 d/wk, 6 wk	VO2max
Andersen et al. 2013 [30]	Denmark	42 ± 10	23 ± 4	160 (35/125)	Workplace stairway	High intensity (∼90% HRR)	Stair climbing, 10 min/session, 1 bout/d, 5 d/wk, 10 wk	VO2max, BF%
Songsorn et al. 2016 [31]	United Kingdom	24 ± 6	22.9 ± 4.5	30 (10/20)	Laboratory setting	All-out intensity (20-s sprint)	Sprint interval, 0.5 min/session, 1 bout/d, 3 d/wk, 4 wk	VO2max, PPO
Ho et al., 2018 [32]	Singapore	36 ± 8	25.2 ± 4.0	16 (0/16)	Laboratory setting	All-out intensity (30-s Wingate sprint)	Sprint interval, 1.5 min/session, 3 bouts/d, 3 d/wk, 8 wk	VO2max, PPO
Perkin et al., 2019 [33]	United Kingdom	70.0 ± 4.0	25.0 ± 3.4	20 (6/14)	Home-based setting	AMRAP performed with correct technique	Resistance exercise, 5 min/session, 2 bouts/d, 7 d/wk, 4 wk	BF%
Jenkins et al., 2019 [34]	Canada	20.0 ± 1.8	21.9 ± 5.0	24 (5/19)	Supervised stairway setting	Vigorous intensity (fastest safe pace)	Stair climbing, 5 min/session, 3 bouts/d, 5 d/wk, 6 wk	VO2max, PPO
Wun et al. 2020 [35]	Singapore	34.1 ± 6.3	24.5 ± 5.9	33 (16/17)	Laboratory setting	All-out Wingate sprint	Sprint interval, 2 min/session, 3 bouts/d, 3 d/wk, 6 wk	VO2max, BF%, TC, HDL-C, LDL-C, TG
Stojanović et al. 2021 [36]	Serbia	75.7 ± 8.9	27.5 ± 5.1	168 (−/−)	Supervised group setting	Low-load resistance exercise (OMNI-RES 4−5/10)	Resistance exercise, 10 min/session, 2 bouts/d, 2 d/wk, 12 wk	TC, HDL-C, LDL-C, TG
Wanders et al. 2021 [37]	the Netherlands	59.6 ± 8.1	30.2 ± 2.5	24 (5/19)	Laboratory setting	Moderate intensity (50−70% HRmax)	Cycling, 5 min/session, 6 bouts/d, 1 d/wk, 4 wk	TC, HDL-C, LDL-C, TG
Michael et al. 2021 [38]	United Kingdom	31.76 ± 1.33	26.33 ± 1.18	60 (0/60)	Home or fitness-centre setting	Self-paced, vigorous intensity	Stair climbing, 5 min/session, 2−5 bouts/d, 5 d/wk, 8 wk	VO2max, BF%, TC, HDL-C, LDL-C, TG
Reljic et al. 2021 [39]	Germany	49.6 ± 12.3	39.6 ± 7.8	46 (20/26)	Supervised laboratory setting	High intensity (80−95% HRmax)	Cycling, 7 min/session, 2 bouts/d, 2 d/wk, 12 wk	VO2max, PPO, BF%, TC, HDL-C, LDL-C, TG
Yin et al., 2024 [40]	China	22.1 ± 2.1	22.1 ± 2.8	29 (14/15)	Campus stairway	All-out intensity (∼30-s stair sprint)	Stair climbing, 1 min/session, 3 bouts/d, 3 d/wk, 6 wk	VO2max, PPO, BF%
Wong et al. 2024 [41]	Singapore	24.0 ± 2.8	–	19 (8/11)	Laboratory setting	All-out intensity (30-s sprint)	Sprint interval, 0.5 min/session, 1 bout/d, 7 d/wk, 6 wk	VO2max, PPO, BF%, TC, HDL-C, LDL-C, TG
Zhou et al., 2024 [42]	China	22.1 ± 1.9	–	27 (13/14)	Stairway adjacent to the laboratory	Fastest safe pace	Stair climbing, 0.5 min/session, 6 bouts/d, 4 d/wk, 12 wk	VO2max, BF%
Brandt et al. 2024 [43]	United Kingdom	42.1 ± 11.1	26.9 ± 4.3	26 (0/26)	Worksite setting	AMRAP performed with correct technique	Resistance exercise, 10 min/session, 1 bout/d, 5 d/wk, 12 wk	BF%
Babir et al., 2025 [44]	Canada	54 ± 6	27.2 ± 2.7	77 (21/56)	Real-world, remotely delivered setting	Self-paced intensity (RPE 3.0 ± 0.7 on a 0−10 scale)	Resistance exercise, 1 min/session, 3 bouts/d, 3 d/wk, 12 wk	VO2max, PPO
Zhou et al., 2025 [45]	Korea	68.5 ± 3.1	24.8 ± 2.8	34 (−/−)	Home-based, remotely delivered setting	Vigorous intensity (≥70% HRmax; RPE 14−15)	Resistance exercise, 9 min/session, 3 bouts/d, 3 d/wk, 12 wk	VO2max, TC, HDL-C, LDL-C, TG

Abbreviations: BMI, body mass index; M, male; F, female; VO₂max, maximal oxygen uptake; PPO, peak power output; BF%, body fat percentage; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TG, triglycerides; HRR, heart rate reserve; HRmax, maximum heart rate; RPE, rating of perceived exertion; AMRAP, as many repetitions as possible; OMNI-RES, OMNI Resistance Exercise Scale; d/wk, days per week; wk, weeks.

Exercise intensity is presented according to the terminology and measures reported in the original studies.

Fig. 3.

Meta-regression analyses of age and effect sizes for each outcome: (a) VO₂max; (b) PPO; (c) BF%; (d) TC; (e) HDL-C; (f) LDL-C; and (g) TG.

3.5. Subgroup analyses

Further subgroup analyses were conducted according to session duration, number of exercise bouts per day, weekly intervention frequency, and intervention duration (Table 2). Session duration significantly moderated the subgroup differences for TC, LDL-C, and TG (P < 0.05). In the subgroup with a session duration ≤2 min, decreases were observed in TC (g = −0.53, 95% CI = −0.92 to −0.13), LDL-C (g = −0.47, 95% CI = −0.93 to −0.01), and TG (g = −0.53, 95% CI = −0.92 to −0.14). The number of exercise bouts per day significantly moderated the subgroup differences for VO₂max and HDL-C (P < 0.05). For VO₂max, the largest effect size was observed in the subgroup performing >5 bouts/day (g = 1.74, 95% CI = 0.35–3.14). For HDL-C, although the subgroup difference was statistically significant, no individual subgroup showed a statistically significant within-group effect. Weekly intervention frequency significantly moderated the subgroup differences for PPO, BF%, and HDL-C (P < 0.05). For PPO, the largest effect size was found in the 4−5 days/week subgroup (g = 1.43, 95% CI = 0.72–2.15). Intervention duration significantly moderated the subgroup differences for VO₂max and PPO (P < 0.05). For VO₂max, the largest effect size was observed in the 7−9 weeks subgroup (g = 1.33, 95% CI = 0.55–2.10).

Table 2.

Subgroup analyses based on intervention characteristics (g, 95% CI).

Variable	VO2max	PPO	BF%	TC	HDL-C	LDL-C	TG
Session time (min)	P = 0.263	P = 0.397	P = 0.633	P = 0.030	P = 0.052	P = 0.008	P = 0.002
≤2	0.96 (0.49–1.46)	0.48 (−0.15 to 1.10)	0.17 (−0.11 to 0.45)	−0.53 (−0.92 to −0.13)	0.36 (−0.91 to 1.63)	−0.47 (−0.93 to −0.01)	−0.53 (−0.92 to −0.14)
3−7	1.06 (0.64–1.48)	0.60 (−0.07 to 1.27)	0.70 (−1.12 to 2.52)	0.17 (−0.23 to 0.57)	−0.62 (−1.68 to 0.44)	0.39 (−0.15 to 0.94)	0.50 (−0.22 to 1.22)
>7	1.28 (0.17–2.40)	1.20 (0.37–2.02)	0.02 (−0.28 to 0.32)	−0.35 (−0.63 to 0.07)	−0.22 (−0.50 to 0.06)	−0.25 (−0.52 to 0.03)	0.27 (−0.55 to 1.08)
Times (sessions/day)	P = 0.041	P = 0.126	P = 0.135	P = 0.223	P = 0.017	P = 0.304	P = 0.283
≤2	0.57 (0.22 to 0.92)	0.52 (−0.08 to 1.11)	−0.05 (−0.29 to 0.20)	−0.47 (−0.95 to 0.01)	−0.22 (−0.62 to 0.18)	−0.20 (−0.51 to 0.06)	−0.17 (−0.63 to 0.29)
3−5	1.06 (0.57–1.55)	0.54 (−0.16 to 1.24)	−0.04 (−0.39 to 0.31)	−0.18 (−0.56 to 0.20)	1.41 (−1.02–3.85)	−0.05 (−0.49 to 0.39)	0.10 (−0.51 to 0.71)
>5	1.74 (0.35–3.14)	1.42 (0.73–2.11)	1.28 (−0.01–2.56)	−0.01 (−0.70 to 0.68)	−0.97 (−2.72 to 0.78)	−0.03 (−0.50 to 0.44)	0.17 (−1.06 to 1.40)
Frequency (days/week)	P = 0.287	P = 0.015	P = 0.003	P = 0.494	P = 0.028	P = 0.379	P = 0.181
≤3	0.98 (0.45–1.51)	0.62 (0.16–1.07)	−0.20 (−0.56 to 0.16)	−0.24 (−0.46 to −0.01)	−0.21 (−0.46 to 0.04)	−0.20 (−0.51 to 0.10)	0.08 (−0.22 to 0.37)
4−5	0.73 (0.48 to 0.99)	1.43 (0.72–2.15)	0.14 (−0.15 to 0.43)	−0.50 (−0.95 to −0.04)	1.02 (−0.78 to 2.83)	−0.05 (−0.49 to 0.39)	−0.50 (−1.41 to 0.40)
>5	1.59 (−0.06 to 3.24)	−0.21 (−1.12 to 0.71)	1.02 (−0.56 to 2.60)	−0.35 (−2.32 to 1.62)	−0.85 (−2.39 to 0.69)	−0.01 (−0.37 to 0.35)	0.21 (−2.06 to 2.48)
Duration (weeks)	P = 0.008	P = 0.046	P = 0.081	P = 0.852	P = 0.690	P = 0.436	P = 0.659
≤6	1.19 (0.52–1.86)	1.04 (0.24–1.85)	0.06 (−0.31 to 0.43)	−0.44 (−1.13 to 0.24)	−0.40 (−1.29 to 0.50)	−0.32 (−0.81 to 0.18)	−0.36 (−1.42 to 0.70)
7−9	1.33 (0.55–2.10)	0.35 (−0.28 to 0.99)	0.94 (−0.61 to 2.49)	−0.40 (−0.85 to 0.38)	1.16 (−0.19 to 2.50)	−0.02 (−1.05 to 1.05)	−0.02 (−0.88 to 0.87)
>9	0.58 (0.18 to 0.97)	0.48 (−0.17 to 1.12)	0.05 (−0.44 to 0.54)	−0.22 (−0.55 to 0.10)	−0.18 (−0.43 to 0.07)	−0.10 (−0.44 to 0.24)	0.19 (−0.27 to 0.65)

Abbreviations: g, Hedges' g; 95% CI, 95% confidence interval.

Overall, the effects of exercise snacks on cardiorespiratory fitness outcomes appeared to be more strongly influenced by the number of exercise bouts per day, weekly intervention frequency, and intervention duration, whereas some body composition and blood lipid outcomes were mainly influenced by session duration.

3.6. Risk of bias and sensitivity analyses

The results of the risk-of-bias assessment are shown in Figs. 4a and 4b. Few studies were judged to be at high risk of bias, and most were rated as having low risk of bias or some concerns. For the randomization process (D1), 5 studies were rated as low risk, 14 studies as some concerns, and 2 studies as high risk. For deviations from intended interventions (D2), all 21 studies were rated as having some concerns. For missing outcome data (D3), all studies were rated as low risk of bias. For measurement of the outcome (D4), 14 studies were rated as low risk and 7 studies as some concerns. For selection of the reported result (D5), 9 studies were rated as low risk and 12 studies as some concerns. Overall, the main sources of bias were concentrated in the domains of the randomization process, deviations from intended interventions, and selection of the reported result, whereas the risk of bias due to missing outcome data was low.

Fig. 4.

Risk-of-bias assessment and publication bias analyses of the included studies.

The results of the publication bias assessment are shown in Fig. 4c. For VO₂max, both Egger’s test and Begg’s test were statistically significant (Egger’s P < 0.001; Begg’s P = 0.006). For BF%, statistical significance was observed in Begg’s test (Egger’s P = 0.061; Begg’s P = 0.005). By contrast, neither Egger’s test nor Begg’s test reached statistical significance for PPO, TC, HDL-C, LDL-C, or TG (P > 0.05). The leave-one-out sensitivity analyses (Supplementary Figures S1–S28) showed that the direction and statistical significance of the pooled effects for the main outcomes did not materially change after excluding each study in turn.

In the additional sensitivity analysis restricted to studies with a reported mean participant age of ≥60 years, three studies met the age criterion (Supplementary Table S2). VO₂max and BF% were each reported by only one eligible study and were therefore not pooled, while no eligible study reported PPO data. The single-study estimate suggested a favorable effect on VO₂max (g = 0.71, 95% CI = 0.02–1.41, P = 0.04) but not for BF% (g = 0.05, 95% CI = −0.85 to 0.95, P = 0.91). TC, HDL-C, LDL-C, and TG were each reported by two studies involving 202 participants. Exercise snacks significantly reduced TC (g = −0.35, 95% CI = −0.63 to −0.07, P = 0.01, I² = 0%), whereas no significant pooled effects were observed for HDL-C, LDL-C, or TG.

3.7. Certainty of evidence

The certainty of evidence was rated as moderate for VO₂max and PPO, and as low for BF%, TC, HDL-C, LDL-C, and TG. The certainty ratings for VO₂max and PPO were downgraded mainly because of some risk of bias in the included studies and between-study heterogeneity. The certainty of evidence for BF% and the lipid-related outcomes was rated as low, mainly due to substantial heterogeneity, imprecision in the effect estimates, and possible publication bias for some outcomes. Overall, the current evidence supports the beneficial effects of exercise snacks on cardiorespiratory fitness-related outcomes, whereas the evidence for body composition and blood lipid outcomes remains relatively limited.

4. Discussion

The beneficial effects of exercise snacks were more consistent for cardiorespiratory fitness than for body composition and blood lipid outcomes. Overall, VO₂max and PPO improved, suggesting that multiple brief exercise bouts distributed throughout the day can provide a sufficient cumulative stimulus to the cardiorespiratory system. Repeated moderate-to-vigorous activity may promote adaptations in oxygen delivery and utilization, peripheral blood flow, skeletal muscle oxidative function, and neuromuscular recruitment, even when the total exercise volume is relatively low [[46], [47], [48], [49]].

In contrast, the findings for BF%, TC, HDL-C, LDL-C, and TG were less consistent, with substantial heterogeneity for several outcomes. Body fat and blood lipid markers generally require longer intervention periods, greater cumulative energy expenditure, or concurrent dietary changes to produce measurable improvements [50,51]. Their responses may also be influenced by baseline health status, exercise modality, intervention dose, supervision, and measurement methods. In addition, some metabolic benefits of brief activity bouts may be more readily reflected in postprandial responses than in short-term changes in BF% or lipid concentrations [52,53].

Age-related analyses suggested that intervention effects may vary across study populations. Improvements in VO₂max were more evident in the lower- and intermediate-age strata than in the higher-age stratum, and meta-regression indicated that the effect size tended to decline as study-level mean age increased. These patterns may reflect age-related changes in cardiovascular reserve, skeletal muscle oxidative capacity, recovery, and adaptive responsiveness [[54], [55], [56]].

Intervention structure may also contribute to variation in effects. Improvements in cardiorespiratory fitness appeared more evident with a greater number of daily bouts, higher weekly frequency, and longer intervention duration, whereas very brief sessions showed more favorable effects for some blood lipid outcomes [57,58]. Nevertheless, these subgroup findings were based on observational comparisons between relatively small numbers of studies and should not be interpreted as definitive evidence of an optimal exercise-snack prescription [59]. Overall, exercise snacks appear to be a time-efficient strategy for improving cardiorespiratory fitness, whereas their effects on body fat and blood lipid profiles remain uncertain and require further high-quality trials.

Previous evidence has emphasized that the health effects of physical activity depend not only on total exercise volume but also on how activity is accumulated throughout the day. Prolonged sedentary behavior is associated with adverse cardiovascular and metabolic outcomes, while brief, manageable bouts of activity may provide a practical means of increasing daily physical activity and interrupting extended sitting [60,61].

Systematic reviews and intervention studies have suggested that exercise snacks are a feasible and time-efficient strategy for improving physical activity participation and selected health outcomes, particularly in physically inactive adults [62]. More recently, Zhang et al. reported improvements in VO₂max and peak power-related outcomes and identified potential variation in intervention effects across age groups [12]. Building on this evidence, the present study examined age-related variation using both categorical study-level age strata and continuous-age meta-regression. VO₂max and PPO improved overall; the improvement in VO₂max was more evident in the lower- and intermediate-age strata, whereas PPO improved across all three strata. Meta-regression further suggested that the effect on VO₂max tended to decrease as study-level mean age increased.

The present review also broadened the scope of previous evidence by evaluating BF% and blood lipid outcomes. Unlike the relatively consistent improvements in cardiorespiratory fitness, no significant overall effects were observed for TC, HDL-C, LDL-C, or TG, and the unfavorable pooled direction for BF% was accompanied by substantial heterogeneity. Brief repeated exercise bouts may be sufficient to promote adaptations in oxygen delivery, utilization, and exercise performance, whereas changes in body composition and blood lipids may require greater cumulative exercise volume, longer intervention duration, or concurrent changes in energy balance [63].

Variations in intervention structure may also contribute to differences across studies. The subgroup analyses suggested that improvements in cardiorespiratory fitness may be related to daily bout frequency, weekly intervention frequency, and intervention duration, while very brief sessions showed more favorable effects for some blood lipid outcomes. These findings are compatible with previous evidence suggesting that repeated interruptions to sedentary behavior and the accumulation of brief activity bouts may influence physiological responses [64].

The key advantages of exercise snacks lie in their feasibility, low time requirement, and ease of integration into daily routines [65]. For adults who are sedentary, physically inactive, or unable to allocate uninterrupted time for conventional exercise, exercise snacks may provide a practical means of increasing physical activity participation [66]. However, time efficiency alone does not guarantee long-term adherence or scalability in real-world settings. The feasibility of exercise snacks is likely to depend on the context in which they are performed and on how much control individuals have over their daily routines. In occupational settings, implementation may be constrained by workload, limited autonomy over breaks, and a lack of suitable spaces or organizational support. In home, transport, and community settings, sustainability may instead be shaped by safety, mobility, environmental accessibility, family responsibilities, and the availability of behavioral support. Exercise snacks may therefore be easier to adopt among individuals with flexible routines and supportive environments, but less practical for those facing rigid schedules, physical limitations, or competing daily demands. Based on the present findings, exercise snacks should be viewed as a complement to conventional exercise rather than a replacement. Their main value may lie in lowering barriers to participation, encouraging more active daily behavior, and improving cardiorespiratory fitness through an accessible and adaptable approach.

Although this study provides several valuable insights, some limitations should be acknowledged. First, although the number of included studies was greater than in earlier reviews, the certainty of evidence for BF% and the blood lipid outcomes remained low, and several outcomes showed substantial heterogeneity; therefore, these findings should still be interpreted cautiously. Second, current exercise snack interventions vary considerably in exercise modality, intensity, session duration, within-day distribution, and degree of supervision. Third, although studies combining exercise snacks with dietary interventions were excluded, the extent of dietary control varied across the included trials, which may have contributed to the inconsistencies in body composition and blood lipid outcomes. Future studies should therefore monitor or standardize dietary intake throughout the intervention. Fourth, the analyses related to age were based on study-level mean age rather than individual participant data, and thus may be subject to ecological bias. Fifth, potential publication bias was observed for VO₂max and BF%.

5. Conclusions

Current evidence suggests that exercise snacks, as a time-saving, flexible, and easily implementable form of exercise intervention, show relatively consistent potential benefits for improving cardiorespiratory fitness in adults, whereas their effects on body composition and blood lipid outcomes remain inconclusive. The number of exercise bouts per day, weekly intervention frequency, and intervention duration may be related to improvements in cardiorespiratory fitness, while session duration may be associated with selected blood lipid outcomes. Overall, exercise snacks may serve as a promising complement to traditional exercise. However, further high-quality studies are needed to determine their suitability across age groups, identify optimal implementation protocols, and clarify their long-term effects on body composition and blood lipid profiles.

CRediT authorship contribution statement

L.M.: Conceptualization, Data curation, Formal analysis, Methodology, Project administration, Visualization, Writing – original draft, Writing – review & editing. Y.L.: Investigation. J.C.: Investigation. H.Z.: Conceptualization, Funding acquisition, Methodology, Project administration, Supervision, Validation, Writing – review & editing.

Ethics approval and consent to participate

Not applicable.

Declaration of Generative AI and AI-assisted technologies in the writing process

No generative AI tools were used in manuscript writing, data analysis, figure creation or artwork preparation in this study.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

Data will be made available on request.

Declaration of competing interest

The authors declare that they have no competing interests.

Appendix A. Supplementary data

The following is Supplementary data to this article: