目的 系統(tǒng)評價舒眠膠囊聯(lián)合化學藥治療失眠癥的臨床療效與安全性。方法 計算機檢索PubMed、中國生物醫(yī)學文獻數(shù)據(jù)庫（CBM）、中國知網（CNKI）、維普數(shù)據(jù)庫（VIP）和萬方數(shù)據(jù)庫，收集舒眠膠囊聯(lián)合化學藥（試驗組）對比單用化學藥（對照組）治療失眠癥的隨機對照試驗（RCT），并追溯納入研究的參考文獻。檢索時限為2000年1月-2017年7月。嚴格按照納入與排除標準篩選文獻、提取有效數(shù)據(jù)并對納入文獻進行方法學質量評價，采用RevMan 5.0軟件對各效應指標進行Meta-分析。結果 共納入10個RCT，共計1 051例失眠癥患者。Meta-分析結果顯示：試驗組臨床總有效率高于對照組[OR=2.23，95% CI（2.18，4.77），P<0.001]，兩組比較差異有統(tǒng)計學意義；試驗組治療周期末匹茲堡睡眠質量指數(shù)量表（PSQI）評分低于對照組[MD=-1.88，95% CI（-2.58，-1.17），P<0.001]，差異有統(tǒng)計學意義；試驗組與對照組不良反應發(fā)生率相當[OR=0.68，95% CI（0.33，1.40），P=0.29]，差異無統(tǒng)計學意義。結論 舒眠膠囊聯(lián)合化學藥治療失眠癥的臨床療效優(yōu)于單用化學藥治療。由于本系統(tǒng)評價納入的文獻質量較低、且樣本量小，尚需更多高質量RCT予以進一步的證實。

Objective To evaluate the efficacy and safety of Shumian Capsule combine with chemical drugs (experimental group) compared with chemical drugs (control group) in treatment of insomnia. Methods The randomized controlled trials (RCT) of Shumian Capsule combined with western medicine in treatment of insomnia were searched from the database including Pubmed, CBM, CNKI, VIP database, and Wanfang by computer from January 2000 to July 2017. References of included studies were also retrieved. Screened studies according to inclusion and exclusion criteria, extracted valid data, and assessed the methodological quality of including literatures. Meta-analysis was performed using RevMan 5.0 software. Results A total of ten RCTs were included, including 1051 patients with insomnia. Meta-analysis results showed that, compared with control group, experimental group improved total effective rate[OR=2.23, 95% CI(2.18, 4.77), P<0.001], and reduced the PSQI score[MD=-1.88, 95% CI(-2.58, -1.17), P<0.001], the difference was statistically; There was no signficant difference between the two group in the incidences of adverse effect[OR=0.68, 95% CI(0.33, 1.40), P=0.29]. Conclusion Shumian Capsule combined with chemical drugs showed better efficacy for the insomnia compared with chemical drugs alone. However, due to the low quality and small sample size, the RCTs with high quality were still needed to be further validated.