Background The long-term effects of bone marrow-derived cells (BMC) transplantation in patients with acute myocardial infarction (AMI) have not been established.The present meta-analysis of randomized controlled trials with follow-up ≥2 years was performed to investigate the long-term effects of BMC therapy in patients after AMI.Methods Specific terms were used to conduct a systematic literature search of MEDLINE,EMBASE,the Cochrane Library and the Cochrane Central Register of Controlled Trials,and the China Biological Medicine Disk database from their inception to March 2012.A standardized protocol was used to extract information,and random effect model was used to analyze all data except major adverse events.Results Five trials comprising 510 patients were included.Compared with controls,BMC therapy significantly improved left ventricular ejection fraction (LVEF) (4.18％,95％ CI:2.02％ to 6.35％,P=0.0002),while mildly but not significantly reduced left ventricular end-systolic volume (-4.47 ml,95％ CI:-10.92 to 1.99,P=-0.17) and left ventricular end-diastolic volume (-2.29 ml,95％ CI:-9.96 to 5.39,P =0.56).Subgroup analysis revealed that significant improvement of LVEF induced by BMC therapy could be observed in patients with baseline LVEF ≤42％,but disappeared in those with baseline LVEF ＞42％.There were trends in favor of BMC therapy for most major clinical adverse events,though most differences were not significant.Conclusions Intracoronary BMC infusion in patients with AMI seems to be safe and may further improve LVEF on top of standard therapy; especially the beneficial effects could last for long term.The findings need to be validated in the future.