方法：计算机检索Cochrane Library、EMbase、PubMed、中国生物医学数据库、中国期刊全文数据库、维普数据库和万方数据库，收集所有更昔洛韦对比阿昔洛韦治疗单纯疱疹病毒性角膜炎的随机对照试验，由两名评价员独立筛选合格文献并提取相关信息进行分析。根据Cochrane协作网推荐的工具进行偏倚风险评估，用Revman 5.0软件进行统计学分析。

结果：最终纳入14个随机对照试验，共4 820例单纯疱疹病毒性角膜炎患者。按以患者人数为观察对象和以患眼数为观察对象分组分析，并对纳入研究根据随访时间进行亚组分析。更昔洛韦组在有效率、复发率等方面优于阿昔洛韦组，差异具有统计学意义\〖RR=1.22， 95%CI(1.10～1.36)； OR=4.50， 95%CI(2.02～ 10.04)； RR=0.23， 95%CI(0.10～0.52)\〗。更昔洛韦组不良反应发生率比阿昔洛韦组少，差异具有统计学意义\〖RR=0.12，95%CI(0.03～ 0.46)\〗。两组的不良反应均可自行缓解。

结论：目前证据表明，更昔洛韦相比阿昔洛韦治疗单纯疱疹病毒性角膜炎疗效好、安全性高。

METHODS: All of the randomized controlled trials for the study of ganciclovir versus acyclovir in the treatment of herpes simplex virus keratitis were collected from Cochrane Library, EMbase, PubMed, Chinese Bio-medicine Database, China Journal Full-text Database, VIP Database and WanFang Database. Then the data were extracted and evaluated by two reviewers independently. Risk of bias assessment was evaluated by a tool recommended by Cochrane Library. Revman 5.0 software was used for statistical analysis.

RESULTS: Finally, 14 randomized controlled trials were included, 4 820 patients totally. Subgroups were used according to the number of patients and diseased eyes as well as the difference of follow-up time. For and relapse rate, ganciclovir group was overmatch acyclovir group. There were statistical differences between the two groups \〖RR=1.22, 95%CI(1.10-1.36); OR=4.50, 95%CI(2.02-10.04); RR=0.23, 95%CI(0.10-0.52)\〗. Compared with acyclovir, ganciclovir had less side-effect. There were statistical differences between the two groups \〖RR=0.12, 95%CI(0.03-0.46)\〗. All of the side effects of the two groups can be relieved by themselves.

CONCLUSION: Current evidence suggests that the ganciclovir is more efficient and safe than acyclovir in the treatment of herpes simplex virus keratitis.