Objective To evaluate the effectiveness and safety of Lipo-prostaglandin E1 injection in treating viral hepatitis.Methods We searched MEDLINE, EMBASE, The Cochrane Library and CNKI from 1978 to June 2007. We identified randomized control led trials of Kai Shi injection versus other medicines or blank controlin treating viral hepatitis. The quality of included trials was evaluated independently by two reviewers. Meta-analyses were performed with The Cochrane Collaboration’s RevMan 4.2.7 software. Results Fourteen studies involving 1 218 patients were included, one of these compared lipo-prostaglandin E1 injection versus Mai Anding injection, one compared lipo-prostaglandin E1 injection versus potassium-magnesium aspartate injection, and the other 12 compared Lipo-prostaglandin E1 injection versus blank control. Allincluded studies were assessed in terms of randomization, allocation concealment and blinding; and all were graded C(poor quality). Meta-analyses showed that, the total effective rate was significantly higher in the lipo-prostaglandin E1 injection group[RR 1.45, 95%CI (1.29, 1.63)] and the mortality was lower[RR 0.66, 95%CI (0.53, 0.83)] compared with the blank control group, but the incidence of phlebitis was significantlyhigher [RR 7.70, 95%CI (2.57, 23.07)]. There was no significant di f ference between Mai Anding inject ion and lipo-prostaglandin E1 injection in the total effective rate, but Lipo-prostaglandin E1 injection was more effective in improving patients’ liver functions. Compared with potassium-magnesium aspartate injection, the total effective rate was significantly higher in the lipo-prostaglandin E1 injection group[RR1.54, 95%CI (1.14, 2.08)].Conclusion The evidence currently available shows that the effectiveness and safety of lipo-prostaglandin E1 injection are not significantly different from those of Mai Anding injection for patients with viral hepatitis. Compared with potassium-magnesium aspartate injection, Lipo-prostaglandin E1 injection could significantly improve the total effective rate, but since we only include 1 relevant randomized trials, the strength of this evidence is weak. When compared with the blank control, Lipo-prostaglandin E1 injection significantly improved the total effective rate, decreased mortality but increased the incidence of side effects and the existing evidence is insufficiant to show whether Lipo-prostaglandin E1 injection improves patients’ liver functions.