RESUMEN
OBJECTIVE: This study aimed to reveal the efficacy of physical therapy for patients with peripheral facial palsy. METHODS: A literature search was conducted using PubMed, Ichushi-Web, and Cochrane Central Register of Controlled Trials. Published randomized controlled trials comparing the physical therapy versus placebo/non-treatment for peripheral facial palsy such as Bell's palsy, Ramsay Hunt syndrome, and traumatic facial palsy were included for meta-analysis. The primary outcome was non-recovery at the end of the follow-up. Non-recovery was defined according to the authors' definition. The secondary outcomes were the composite score of the Sunnybrook facial grading system and sequelae (presence of synkinesis or hemifacial spasm) at the end of the follow-up. Data was analyzed using Review Manager software and pooled risk ratio (RR) or mean difference (MD) with 95% confidence intervals (CI) were calculated. RESULTS: Seven randomized controlled trials met the eligible criteria. The data on non-recovery from four studies was obtained and included 418 participants in the meta-analysis. Physical therapy might reduce non-recovery (RR = 0.51 [95% CI = 0.31-0.83], low quality). Pooling the data of composite score of the Sunnybrook facial grading system from three studies (166 participants) revealed that physical therapy might increase the composite scores (MD = 12.1 [95% CI = 3.11-21.0], low quality). Moreover, we obtained data on sequelae from two articles (179 participants). The evidence was very uncertain about the effect of physical therapy on reduction of sequelae (RR = 0.64 [95% CI = 0.07-5.95], very low quality). CONCLUSION: The evidence suggested that physical therapy reduces non-recovery in patients with peripheral facial palsy and improves the composite score of the Sunnybrook facial grading system, whereas the efficacy of physical therapy in reducing sequelae remained uncertain. The included studies had high risk of bias, imprecision, or inconsistency; therefore, the certainty of evidence was low or very low. Further well-designed randomized controlled trials are needed to confirm its efficacy.