BACKGROUND: Clinical studies suggest that granulocyte colony-stimulating factor (G-CSF)-mobilized stem cells are recruited to ischemic myocardium and differentiate into specialized cells such as cardiomyocytes, endothelial cells, and smooth muscle cells, and may improve left ventricular function. OBJECTIVES: The aim of this study was to investigate the effectiveness and tolerability of G-CSF treatment with regard to global left ventricular function in patients with myocardial infarction (MI). METHODS: A literature search was conducted of MEDLINE, Cochrane Controlled Trials Register, EMBASE, Science Citation Index, and PubMed (all from their inception to March 2007). Reference lists of papers and reviews on the topic were also searched. We selected the following criteria for trials included in this study: (1) randomized controlled trial (RCT) design of MI routine therapy comparing G-CSF with placebo or blank control in patients with MI; (2) > or =3 to < or =12 months' follow-up after G-CSF treatment; (3) diagnosis of acute MI (AMI) (< or =14 days from onset of new ST-segment elevation infarction) or old MI (OMI) (>14 days from onset); (4) complete left ventricular ejection fraction (LVEF) data and major adverse cardiovascular event (MACE) reports; and (5) the availability of demographic characteristics of patients and the duration and dose of G-CSF treatment. This information was independently extracted by 2 of the investigators using a standardized protocol. RESULTS: Of the 14 RCTs meeting the inclusion criteria, 7 RCTs were deemed eligible for further analysis. The remaining studies included 364 patients (G-CSF groups, 179; control groups, 185; mean age range, 49.8-63.0 years). A significant increase in follow-up LVEF (LVEF(follow-up)) was observed in the G-CSF groups compared with the control groups (2.96%; 95% CI, 0.98-4.94; P = 0.003), and the LVEF change from baseline to follow-up (LVEF(Delta)) also significantly increased (3.46%; 95% CI, 0.60-6.32; P = 0.018). The heterogeneity was significant across the studies with regard to LVEF(follow-up) (P = 0.068) and the LVEFA (P = 0.001). The relative risk (RR) for the prevalence of MACEs, including ventricular arrhythmia (RR, 0.65; 95% CI, 0.29-1.49), rehospitalization for heart failure (RR, 2.00; 95% CI, 0.36-11.17), and the composite of other cardiovascular events (ie, cardiac death, recurrent MI, infarct-vessel revascularization procedure, and stroke) (RR, 1.07; 95% CI, 0.71-1.60), was not significantly different in the G-CSF treatment groups compared with the control groups. The overall risk for MACE was also not significantly different between the 2 groups (RR, 0.93; 95% CI, 0.57-1.28). CONCLUSION: Based on the studies included in this meta-analysis, G-CSF treatment improved the LVEF in AMI (but not OMI) at 3 to 12 months follow-up. Treatment with G-CSF was generally well tolerated.