Objective This study aimed to evaluate the efficacy of virtual reality (VR) in upper limb rehabilitation in subacute stroke, including the Fugl-Meyer Assessment Upper Extremity Scale (FMA-UE), Box and Block Test (BBT), and Action Research Arm Test (ARAT). Subgroup analysis was performed to explore the optimal intervention parameters for VR treatment. Methods PubMed, Embase, Cochrane Library, and Web of Science were thoroughly searched for randomized controlled trials (RCTs) on the use of VR for the rehabilitation of upper limb dysfunction in post-stroke hemiplegic patients until December 2025. Eligible studies were analyzed via RevMan 5.3 statistical software, with study quality assessed by the AOS scoring system, and data analysis performed using Stata 15. Results Fifteen studies were encompassed, comprising 612 cases overall, with 318 cases in the experimental group and 294 cases in the control group. VR rehabilitation was superior to traditional rehabilitation in FMA-UE score (SMD = 0.68, 95% CI 0.30–1.05, P = 0.000), but there was no significant difference in BBT (SMD = 0.44, 95% CI −0.02–0.90, P = 0.058), ARAT (SMD = 0.63, 95% CI −0.01–1.27, P = 0.052) scores ( P > 0.05). Subgroup analysis showed that FMA-UE scores were significantly improved when VR training exceeded 30 minutes/session, the frequency was 4–5 days/week, and the total intervention time was 3–4 weeks. Immersive VR training showed advantages in FMA-UE, BBT, and ARAT scores. According to the Cochrane 1.0 quality assessment system, the overall methodological quality of this study was rated as moderate. A total of eight follow-up studies were included: Four of the follow-up studies showed differences in outcome measures between the two groups. Conclusion VR may have advantages in promoting FMA-UE scores in the upper limbs of patients with subacute stroke. VR training with more than 30 minutes per session, more than 4 days per week, and a total intervention time of more than 3 weeks may be more conducive to the recovery of upper limb function (ULF) in subacute stroke. However, the efficacy of VR in improving fine motor skills and dexterity of the hand, assessed by the BBT and ARAT, remains to be elucidated. Although these findings are informative, the conclusions should be interpreted with caution in consideration of the study’s limitations. Registry number: CRD42025610757.
Wu et al. (2026) studied this question.