摘要： 目的  系统评价左乙拉西坦添加治疗难治性部分性发作癫痫的疗效和药物安全性。方法  计算机检索1998 年1 月-2010 年12 月Cochrane 图书馆、MEDLINE、EMbase、社会科学引文索引、维普中文科技期刊、中国知网中国期刊全文数据库和中国生物医学文献数据库，并手工检索相关杂志，由两名研究者独立进行质量评价及数据分析，RevMan 5.0 统计软件进行Meta 分析。结果  根据Cochrane 5.0.2 版随机对照临床试验质量评价标准，纳入11 项随机对照临床试验共1981 例受试者（左乙拉西坦组1192 例、安慰剂对照组789 例）。Meta 分析结果显示，左乙拉西坦组每周部分性癫痫发作频率减少≥50%的病例数高于对照组（1000 mg/d：OR = 2.990，P = 0.000；2000 mg/d：OR = 3.870，P = 0.000；3000 mg/d：OR = 3.440，P = 0.000）；每周发作频率减少≥ 75%的病例明显高于对照组（1000 mg/d：OR = 3.130，P =0.000；2000 mg/d：OR = 5.060，P = 0.000；3000 mg/d：OR = 4.730，P = 0.000）；完全不发作病例明显高于对照组（1000 mg/d：OR = 5.080，P = 0.001；2000 mg/d：OR = 4.420，P = 0.050；3000 mg/d：OR = 4.150，P = 0.000）。左乙拉西坦组失访率与安慰剂对照组之间差异无统计学意义（P > 0.05）。治疗期间常见药物不良反应包括嗜睡、头晕、乏力、鼻咽炎、精神行为异常等，两组精神行为不良反应方面存在异质性（P = 0.360，I² =8.000%）。结论  现有证据显示，左乙拉西坦添加治疗难治性部分性发作癫痫的疗效与安慰剂组相比效果显著，保留率高；药物安全性应注意其所引起的精神行为异常。

关键词: 抗惊厥药； 癫痫, 复杂部分性； Meta分析； 随机对照试验

Abstract: Objective To evaluate the effects and tolerability of adjunctive levetiracetam (LEV) in refractory partial seizures. Methods Relevant research articles about randomized controlled trials of adjunctive LEV in refractory partial seizures from January 1998 to December 2010 were retrieved from Cochrane Library, MEDLINE, EMbase, Social Sciences Citation Index (SSCI), VIP, Chinese National Knowledge Infrastructure (CNKI) database, China Biology Medicine (CBM). Two reviewers independently evaluated the quality of the included articles and abstracted the data. A Meta-analysis was conducted by using RevMan 5.0 software. Results According to the enrollment criteria, eleven prospective, randomized controlled clinical trials with a total of 1192 in LEV group and 789 in placebo group were finally selected. The reduction in three endpoints (a 50% or greater reduction of partial seizure frequency per week, a 75% or greater reduction of partial seizure frequency per week and seizure free) was significant in LEV group than placebo group. There was no significance between LEV group and placebo group in the withdrawl rate (1000 mg/d: OR = 1.180, 95%CI: 0.690-2.010, P = 0.540; 2000 mg/d: OR = 1.530, 95%CI: 0.770-3.030, P = 0.230; 3000 mg/d: OR = 1.000, 95% CI: 0.620-1.600, P = 1.000). The following adverse events were associated with LEV: somnolence (OR = 1.720, 95%CI: 1.280-2.310, P = 0.000), dizziness (OR = 1.490, 95%CI: 1.000-2.220, P = 0.050), asthenia (OR = 1.670, 95%CI: 1.140-2.240, P = 0.008), nasopharyngitis (OR = 1.120, 95% CI: 0.710-1.760, P = 0.630), psychiatric and behavioral abnormalities (OR = 2.120, 95% CI: 1.370-3.280, P = 0.000). Conclusion LEV is effective and well tolerated when added to existing therapy in patients with refractory partial seizures compared with control drugs. Further studies are needed to identify the effects of monotherapy of LEV in partial seizures.

Key words: Anticonvulsants, Epilepsy, complex partial, Meta-analysis, Randomized controlled trials