Abstract: Background　Available surveys show that about 55% to 75% of stroke survivors worldwide have limb dysfunction，which directly affects the motor function and daily living activities，and brings great mental stress and physical pain.Functional rehabilitation after stroke is one method to effectively reduce the disability rate.Virtual reality technology（VR） is one of the most widely used rehabilitation methods in recent years，which is safe，effective and timely responsive.However，clinical evaluation of its effectiveness and safety is not the same.Objective　To systematically overview the systematic reviews（SR）/meta-analyses of the effectiveness and safety of VR for stroke rehabilitation.Methods　Databases of PubMed，EMBase，the Cochrane Library（2018，Issue 6），CNKI，Wanfang Data，CBM and VIP were searched as of June 2018 and the SR/meta analyses based on VR for stroke rehabilitation were collected.Literature screening and data extraction were performed by two reviewers independently.AMSTAR was used to evaluate the methodological quality of the included articles.The GRADE system was used to evaluate the quality of evidence for main outcome indicators reported in the articles.Results　Finally，10 SR/meta analyses were included，involving 7 major outcome indicators.AMSTAR evaluation results showed that only two studies were of high quality and the other eight studies were of medium quality.The major problems of methodological quality assessment were lack of pre-design planning，omission of the status of publications of the included articles，lack of the list of excluded articles and the explaining of the conflicts of interest.Qualities of the evidence for indicators measuring 6 outcomes （the upper limb motor function，limb mobility，functional mobility，balance function，gait speed and the participation） were mainly graded as "medium"，and those for indicators related to stride length were mainly graded as "low".Three studies showed good VR safety and no adverse reactions，and one study showed that a small number of studies reported pain，headache or dizziness，but no serious adverse reactions were found.Conclusion　The methodological quality of included studies is unsatisfied，with a low grade of evidence.VR has good safety and can be used as a short-term clinical assistant rehabilitation for stroke patients.However，there are still some controversies about its long-term effectiveness.A large number of high-quality tests are needed to prove its long-term effectiveness and safety.Clinicians should be cautious when using the evidence for rehabilitation.

Key words: Stroke, Virtual reality technology, Rehabilitation, Overview of systematic reviews, Evidence-based medicine

摘要： 背景　调查显示全球有55%~75%的脑卒中幸存者会遗留肢体功能障碍，直接影响脑卒中患者运动功能和日常生活活动能力，给患者带来巨大的精神压力和躯体痛苦。脑卒中后的功能康复治疗是有效降低致残率的方法之一，虚拟现实技术（VR）是近年来被临床广泛应用的康复方法之一，具有安全有效、及时反馈等特点。然而临床对其有效性及安全性评价不一。目的　对VR应用于脑卒中康复的有效性及安全性的系统评价（SR）/Meta分析进行再评价。方法　计算机检索PubMed、EMBase、the Cochrane Library（2018年6期）、中国知网、万方数据知识服务平台、中国生物医学文献数据库和维普网，搜集基于VR干预脑卒中的SR/Meta分析，检索时限均为建库至2018年6月。由2位评价员独立筛选文献并提取资料后，采用AMSTAR工具评价纳入SR/Meta分析的方法学质量，并采用GRADE系统对SR/Meta分析的主要结局指标进行证据等级评价。结果　最终纳入10篇SR/Meta分析，包含7个主要结局指标。AMSTAR评价结果显示，8篇文献为中等质量，2篇为高质量，方法学质量存在的问题主要表现在：缺乏前期设计方案、未考虑纳入文献的发表状态、未列出排除文献清单及未说明相关利益冲突4个方面。GRADE评价结果显示：上肢运动功能、肢体活动度、功能性移动能力、平衡功能、步行速度以及参与度指标的证据质量整体以中等为主，行走步长指标整体以低为主。3篇SR/Meta分析显示VR安全性好，无不良反应；1篇SR/Meta分析显示少数研究报道患者出现疼痛、头痛或头晕，但未发现严重的不良反应。结论　纳入SR/Meta分析的方法学质量和结局指标证据等级评价均不高。VR安全性较好，短期可作为脑卒中患者临床辅助康复治疗措施，但在长期疗效方面尚存在争议，仍需要大量的高质量试验来证明其长期使用的有效性及安全性，临床在使用这些证据进行康复决策时尚需谨慎。

关键词: 卒中, 虚拟现实技术, 康复, 系统评价再评价, 循证医学