Abstract: Objective  To analyze and evaluate the effect of levosimendan in the treatment of severe sepsis and septic shock through metaanalysis, and then to provide evidentbased evidence for its application in the treatment of sepsis. Methods  The CNKI, VIP, Wanfang database, PubMed, EMBASE, Ovidand Cochrane library database were searched for the randomized controlled trials of levosimendan in the treatment of severe sepsis and septic shock from the database inception to October 2015. Metaanalysis was conducted on the mortality, the left ventricular ejection fraction (LVEF), the cardiac index (CI),  troponin I (TnI), lactate (Lac)  and the amount of norepinephrine  by using RevMan5.3 software.Results  A total of 13 articles involving 491 cases were included, of which 251 cases were in levosimendan group, while the other 240 cases were in the control group. Compared with the control group, the mortality of the patients in the levosimendan group decreased significantly (RR=0.78，95%CI=0.630.96，P=0.02). Cardiac index (WMD=0.79, 95%CI=0.561.03, P＜0.001), LVEF (WMD=9.03%，95%CI=6.74%11.31%, P＜0.00001)   of  the levosimendan  group increased significantly. And the TnI (WMD=-1.22，95%CI=-1.52-0.77，P＜0.00001), lactate (WMD=-1.41，95%CI=-1.56-1.27，P＜0.00001)   decreased significantly compared with control group. There was no significant difference in the amount of norepinephrine between the two groups (WMD=-0.01, 95%CI=-0.030.01，P=0.45).Conclusion  The evidence of meta analysis supports the application  of levosimendan in paitents with severe sepsis or septic shock.

Key words: sepsis, shock, septic, levosimendan, meta analysis