META- ANALYSIS OF CORONARY ARTERY BYPASS SURGERY COMPARED TO  PERCUTANEOUS TRANSLUMINAL ANGIOPLASTY WITH STENT IN DIABETIC PATIENTS Background: Diabetic patients are a group of primary interest in the study of myocardial revascularization. Aim: To compare coronary artery bypass grafting surgery (CABG) and percutaneous angioplasty with stents (PCI-S) in diabetic patients with coronary three-vessel or left main coronary artery disease. Material and methods: Meta-analysis of MEDLINE randomized controlled studies comparing CABG and PCI-S in diabetic patients. The primary outcome measure was major adverse cardiovascular events (MACCE), death, myocardial infarction, cerebrovascular accident (CVA) and coronary re-intervention. Secondary outcomes were the individual components of MACCE.  Results: Three studies comparing CABG and PCI-S met the inclusion criteria. One thousand sixty two patients were studied: 565 in the CABG group and 597 in the PCI-S group. At one year follow up MACCE occurred in 24.9 and 12.7% of patients in  PCI-S and CABG groups, respectively (Odds ratio (OR) 2.27; 95% confidence intervals (CI) 1.66 - 3.09). There were no differences in death or myocardial infarction. Strokes were less common in the PCI-S group (OR 0.25, 95% CI 0.09 -  0.68) and coronary re-intervention was required with higher frequency in the PCI-S group (OR 5.32, 95% CI 3.27 – 8.67). Conclusions: In diabetic patients with three-vessel coronary disease or left main coronary artery, revascularization with CABG had significantly less MACCE at one year than those treated with PCI-S. Stroke frequency was higher in CABG, coronary re-intervention was higher in PCI-S. These results must be interpreted cautiously.