A systematic review and meta-analysis of the effects of exercise training on dysfunction in acute, subacute, and chronic stroke patients

Objective: This systematic review and meta-analysis aimed to compare the effectiveness of exercise training initiated in the acute, subacute, and chronic phases after stroke. Methods: The Wanfang, CNKI, VIP, PubMed, Web of Science, Cochrane Library, and EMBASE databases were comprehensively searched for relevant articles from their inception to December 2025. Studies comparing outcomes among the acute stage exercise training group (ASETG), subacute stage exercise training group (SSETG), and chronic stage exercise training group (CSETG) for stroke patients were included. The outcomes assessed were improvements in the Fugl-Meyer Assessment (FMA), Berg Balance Scale (BBS), Action Research Arm Test (ARAT), Modified Barthel Index (MBI), and Modified Ashworth Scale (MAS) after treatment. Meta-analysis of comparable data was performed using Review Manager 5.3 software provided by the Cochrane Collaboration. Results: A total of 16 randomized controlled trials (RCTs) and 22 observational studies involving 5,254 patients were included. The meta-analysis indicated that the ASETG showed significantly greater improvements in FMA scores compared to the SSETG [WMD = 7.95, 95% CI (6.73, 9.16)]. This trend was also observed in the FMA-UE [WMD = 4.26, 95% CI (3.09, 5.44)] and FMA-LE [WMD = 4.57, 95% CI (3.73, 5.41)] subscales when compared to the SSETG. Furthermore, the ASETG outperformed the CSETG in FMA scores [WMD = 5.31, 95% CI (3.89, 6.72)]. In addition, the ASETG exhibited significantly greater improvements in BBS scores compared to the SSETG [WMD = 3.64, 95% CI (1.14, 6.15)] and in MBI scores [WMD = 10.66, 95% CI (9.55, 11.77)]. Moreover, the SSETG showed a greater improvement in ARAT scores compared to the CSETG [WMD = 2.70, 95% CI (1.81, 3.59)]. Conclusion: This review suggests earlier post-stroke exercise may improve functional outcomes (FMA, BBS, MBI, ARAT). However, due to heterogeneity and inclusion of observational studies, findings should be interpreted cautiously. High-quality RCTs are needed to confirm the optimal timing. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD42024519257, identifier PROSPERO (CRD42024519257).