We compared the effectiveness of antegrade cardioplegia (A group, 100 cases) and combined antegrade-retrograde cardioplegia (A + R group, 66 cases) on myocardial protection in patients undergoing coronary artery bypass grafting for more than 3 vessels. The two groups were similar in patients' age, the extent of coronary artery diseases and preoperative risk factors, cardiac output and ejection fractions. The mean number of grafts was 3.3 in A group and 3.6 in A + R group (p < 0.05). The mean number of ITA grafts used was 1.2 in A group and 1.4 in A + R group (p < 0.05). Aortic cross-clamp time was 114 minutes in A group and 138 minutes in A + group (p < 0.01). The serum CK-MB concentration on postoperative day 1 was 38 IU/L in A group and 25 IU/L in A + R group (p < 0.05). The left ventricular stroke work index 3 hours after reperfusion was 33 gm min/m2 in A group and 43 gm min/m2 in A + R group (p < 0.05). The maximum dose of catecholamine required was similar in two groups. Intra-aortic balloon pump was used in 11 patients (11%) in A group and in 2 patients (3.6%) in A + R group (p < 0.05). We conclude that the combined method of intermittent antegrade and continuous retrograde perfusion of cardioplegic solution provides better myocardial protection than intermittent antegrade cardioplegia alone for multivessel coronary artery bypass surgery.