Characteristics of included studies
Figure 1 presents the PRISMA flow diagram outlining the systematic search and selection process. The initial electronic screening identified 5030 potentially relevant articles. After title and abstract screening, 103 studies were deemed eligible and underwent full-text review. Subsequently, 14 RCTs met the inclusion criteria and were retained for meta-analysis.
Flow chart of the study selection procedure
The included RCTs were published between 2005 and 2023 (median year, 2014), comprising a total of 3376 participants (1683 boys), aged 3–7 years. Sample sizes ranged from 30 to 1434, with a mean participant age of 4.98 years (SD = 0.51). Intervention durations spanned 8–48 weeks (mean 20.62 weeks, SD 13.50), frequencies ranged from 1 to 5 sessions per week (mean 3.28, SD 1.17), and session lengths varied from 15 to 90 min (mean 36.40 min, SD 17.12).
The number of independent data points differed by physical fitness outcome: upper muscle strength (10 studies) [26,27,28,29,30,31], lower muscle strength (14 studies) [6, 26,27,28,29,30,31,32,33,34], flexibility quality (10 studies) [26, 28,29,30,31, 34, 35], coordination (6 studies) [26,27,28,29,30, 33, 36][, cardiorespiratory fitness (CRF; 11 studies) [6, 26, 28,29,30,31,32], dynamic balance (11 studies) [26, 28, 29, 31, 36,37,38], and static balance (10 studies) [6, 27, 28, 30, 33, 37, 38]. Detailed study characteristics are presented in Supplementary 3.
The results of pairwise meta-analysis
Summary of overall physical fitness outcomes
Seventy-three independent data points from 14 studies contributed to the analysis of overall physical fitness, which encompassed all measured sub-components. As depicted in Fig. 2, chronic PA interventions exerted a small but significant positive effect on overall physical fitness in preschool children (SMD = 0.30, 95% CI: 0.22 to 0.38, I2 = 72.8%, p < 0.001). Subgroup analyses demonstrated that intervention duration (≥ 16 weeks vs. < 16 weeks), session duration (≥ 40 min vs. < 40 min), and frequency (> 3 times/week vs. ≤ 3 times/week) each influenced the pooled effect size.
Effects of physical activity on preschool children’s physical fitness with different moderators
Muscle strength
PA interventions significantly improved upper muscle strength (SMD = 0.30, 95% CI: 0.14 to 0.46, I2 = 33.2%, p < 0.001). In subgroup analyses, interventions ≥ 16 weeks in duration (SMD = 0.35, 95% CI: 0.15 to 0.54, I2 = 41.0%, p = 0.001), session lengths < 40 min (SMD = 0.41, 95% CI: 0.19 to 0.62, I2 = 35.1%, p < 0.001), and frequencies ≤ 3 times/week (SMD = 0.35, 95% CI: 0.16 to 0.53, I2 = 20.0%, p < 0.001) all yielded significant gains in upper muscle strength.
PA interventions produced a significant improvement in lower muscle strength (SMD = 0.29, 95% CI: 0.12 to 0.46, I2 = 75.5%, p = 0.001). Subgroup analyses indicated meaningful effects in both < 16 weeks (SMD = 0.51, 95% CI: 0.23 to 0.80, I2 = 44%, p < 0.001) and ≥ 16 weeks of intervention (SMD = 0.20, 95% CI: 0.02 to 0.38, I2 = 74.4%, p = 0.031). Session length (≥ 40 min: SMD = 0.22, 95% CI: 0.01 to 0.43, I2 = 61.9%, p = 0.045; < 40 min: SMD = 0.36, 95% CI: 0.07 to 0.65, I2 = 80.8%, p = 0.014) also influenced lower muscle strength. Of note, only the > 3 times/week subgroup displayed significant improvement (SMD = 0.36, 95% CI: 0.11 to 0.61, I2 = 83.3%, p = 0.005).
Flexibility quality
PA interventions significantly enhanced flexibility (SMD = 0.27, 95% CI: 0.03 to 0.52, I2 = 72.4%, p = 0.028). Subgroup analyses revealed that session lengths ≥ 40 min (SMD = 0.33, 95% CI: 0.06 to 0.59, I2 = 54.9%, p = 0.016) and frequencies > 3 times/week (SMD = 0.65, 95% CI: 0.18 to 1.13, I2 = 79.6%, p = 0.007) contributed substantially to the observed improvement.
CRF
PA interventions led to a significant increase in CRF (SMD = 0.29, 95% CI: 0.10 to 0.48, I2 = 72.4%, p = 0.003). Subgroup analyses indicated significant improvements with both < 16 weeks (SMD = 0.32, 95% CI: 0.10 to 0.54, I2 = 0%, p = 0.005) and ≥ 16 weeks of intervention (SMD = 0.28, 95% CI: 0.04 to 0.52, I2 = 77.0%, p = 0.022), as well as with PA frequencies of > 3 times/week (SMD = 0.42, 95% CI: 0.04 to 0.80, I2 = 85.5%, p = 0.031) and ≤ 3 times/week (SMD = 0.20, 95% CI: 0.02 to 0.39, I2 = 27.0%, p = 0.031). Interestingly, the < 40-min session subgroup (SMD = 0.27, 95% CI: 0.05 to 0.49, I2 = 43.2%, p = 0.015) also showed notable CRF improvements.
Coordination
PA interventions significantly enhanced coordination (SMD = 0.54, 95% CI: 0.31 to 0.76, I2 = 57.5%, p < 0.001). All examined subgroups—longer (≥ 16 weeks: SMD = 0.49, 95% CI: 0.20 to 0.78, I2 = 65.7%, p = 0.001) versus shorter (< 16 weeks: SMD = 0.63, 95% CI: 0.34 to 0.91, I2 = 0.0%, p < 0.001) durations, longer (≥ 40 min: SMD = 0.50, 95% CI: 0.21 to 0.78, I2 = 68.4%, p = 0.001) versus shorter (< 40 min: SMD = 0.63, 95% CI: 0.31 to 0.96, I2 = 0.2%, p < 0.001) session lengths, and higher (> 3 times/week: SMD = 0.50, 95% CI: 0.26 to 0.75, I2 = 60.4%, p < 0.001) versus lower (≤ 3 times/week: SMD = 0.70, 95% CI: 0.13 to 1.26, I2 = 46.7%, p = 0.016) frequencies—exhibited statistically significant effects.
Balance
PA interventions significantly improved static balance (SMD = 0.38, 95% CI: 0.10 to 0.65, I2 = 80.7%, p = 0.007). Within subgroup analyses, frequencies of ≤ 3 times/week yielded a significant benefit (SMD = 0.33, 95% CI: 0.11 to 0.55, I2 = 43.8%, p = 0.003). All studies included in this analysis of static balance used ≥ 40-min PA sessions. No significant improvement was noted in dynamic balance, and subgroup analyses did not alter the null findings.
The results of dose–response relationship
Overall physical fitness
A non-linear dose–response curve was identified between weekly PA dose (0–380 min/week) and overall physical fitness (Fig. 3). The optimal dose was approximately 270 min/week (SMD = 0.557, 95% CrI: 0.347 to 0.825), indicating that moderate-to-high volumes of structured PA may maximize improvements in preschool children’s global fitness levels.
Dose–response relationship between overall exercise dose and change in the physical fitness
Muscle strength
An inverted U-shaped association emerged between weekly PA dose (0–350 min/week) and upper muscle strength (Fig. 4a). The optimal dose converged at approximately 260 min/week (SMD = 0.337, 95% CrI: 0.068 to 0.735), suggesting that interventions beyond this range may yield diminishing returns.
Dose–response relationship between different exercise doses and change in the different specific fitness components
Lower muscle strength displayed a similarly inverted U-shaped pattern (Fig. 4b), with effective doses spanning 0–350 min/week and an estimated optimum of around 260 min/week (SMD = 0.424, 95% CrI: 0.235 to 0.746). This finding aligns closely with that of upper muscle strength, emphasizing the importance of balancing intensity and volume.
Flexibility quality
Flexibility quality improved with PA volumes up to 340 min/week (Fig. 4c). The peak effect occurred at roughly 190 min/week (SMD = 0.345, 95% CrI: 0.078 to 0.739), suggesting that, compared with muscle strength outcomes, lower cumulative exercise doses may suffice to enhance flexibility quality in preschool populations.
CRF
Dose–response modeling for CRF (Figure 4d) demonstrated continued gains up to 380 min/week, with the maximum benefit at the highest dose examined (SMD = 0.255, 95% CrI: 0.106 to 0.496). This suggests that higher weekly PA volumes may be particularly advantageous for improving aerobic capacity in young children.
Coordination
Coordination displayed a wide range of effective doses, from 0 to 380 min/week (Fig. 4e). The apparent optimal volume reached approximately 300 min/week (SMD = 0.419, 95% CrI: 0.075 to 1.082). This finding highlights that motor skill development—particularly coordination—benefits from relatively high PA volumes.
Static balance
Improvements in static balance were most pronounced at doses up to 180 min/week (Fig. 4f), peaking around that threshold (SMD = 0.431, 95% CrI: 0.224 to 0.821). Compared with other outcomes (e.g., CRF and coordination), static balance may not require as high a weekly PA volume to achieve meaningful gains.
Connectivity assessment
A connectivity assessment was undertaken to verify adequate linkages among all dose levels. Insufficient connectivity may compromise statistical power and yield biased findings when direct comparisons are infeasible. The assessment confirmed no connectivity deficits among the examined PA doses, thereby supporting the robustness of the subsequent dose–response modeling (Supplementary 11).
Quality assessment of evidence and risk of bias
Overall, 10 studies (40%) exhibited a low risk of bias, 10 (40%) had an unclear risk, and 5 (20%) demonstrated a high risk of bias. Study-level details of risk of bias are provided in Supplementary 8. Except for upper muscle strength, several outcomes displayed relatively high heterogeneity (Supplementary 9). Publication bias was assessed via funnel plots and Egger’s tests. Potential publication bias was noted for flexibility, lower muscle strength, coordination, CRF, and static balance (p < 0.05), warranting cautious interpretation. The p-values for all remaining outcomes exceeded 0.05, indicating no significant publication bias in those measures (Supplementary 10).