Direction-specific effects of balance training on dynamic balance in athletes with chronic ankle instability: a multilevel meta-analysis of randomized controlled trials

Chronic ankle instability (CAI) is a prevalent consequence of ankle sprains among athletes and is often linked to compromised dynamic balance. This meta-analysis aims to evaluate the effects of balance training (BT) on dynamic balance in athletes with CAI, while also exploring whether the pooled effects differ based on participant characteristics, types of comparators, intervention modalities, and assessment tools. A systematic search was conducted in PubMed, Web of Science, PsycINFO, and the Cochrane Library until April 7, 2025. Thirteen randomized controlled trials (RCTs) involving 468 athletes diagnosed with CAI were included in this review. A multilevel (three-level) random-effects meta-analysis was performed to synthesize both composite and direction-specific dynamic balance outcomes. Additionally, exploratory subgroup analyses, Egger’s regression test, trim-and-fill analysis, CR2-adjusted cluster-robust sensitivity analyses, and GRADE certainty assessments were conducted to evaluate the robustness of the findings. BT statistically significantly improved dynamic balance across all pooled outcomes, including the composite score (SMD = 1.32, 95% CI: 0.08 to 2.55), anterior direction (SMD = 0.67, 95% CI: 0.29 to 1.06), posteromedial direction (SMD = 0.59, 95% CI: 0.16 to 1.02), and posterolateral direction (SMD = 0.74, 95% CI: 0.12 to 1.36). Substantial between-study heterogeneity was observed (I² = 76.6%–94.1%), and the approximate prediction intervals for all four outcomes crossed the null value, indicating uncertainty in the robustness of the findings. Exploratory subgroup analyses suggested possible variation across regions, participant types, comparator types, intervention modalities, and assessment tools, although these findings were inconsistent and based on small subgroup sample sizes. CR2-adjusted sensitivity analyses showed that the anterior, posteromedial, and posterolateral effects remained statistically significant, whereas the composite effect was attenuated and no longer statistically significant. BT may improve direction-specific dynamic balance outcomes in athletes with CAI, particularly in the anterior, posteromedial, and posterolateral directions. Although the composite outcome was statistically significant in the primary three-level meta-analysis, it exhibited less robustness in the CR2-adjusted sensitivity analysis. Given the substantial heterogeneity, wide prediction intervals, and limited evidence base, the pooled effects should be interpreted with caution, and subgroup findings should be considered exploratory.