Sixty consecutive coronary patients operated on by the same team in 1992 were divided into two groups: group 1 (30 patients) using intermittent oxygenated cold Fresenius solution antegrade and retrograde (FR), group 2 (30 patients) using warm retrograde blood cardioplegia (WRC) with the Fremes solution initially antegrade and retrograde (high potassium solution) then continuous retrograde low potassium solution. All patients were submitted to only arterial grafts (3 to 4) using both internal mammary arteries and the the right gastroepiploic artery. There were no differences in mean preoperative data between the 2 groups. The times of aortic cross-clamping (P < 0.05) and bypass after release of the aortic clamp (P < 0.01) were significantly higher in the WRC group. No significant difference was observed in the number of postoperative supraventricular arrhythmias or electrocardiographic infarctions. A significant difference was observed with higher values of the enzymes (aspartate amino transferase, creatine kinase) for the WRC group on the first (P < 0.05) and the second postoperative days (P < 0.01). More patients in the WRC group received vasoactive or inotropic drugs in the intensive care unit, where they stayed a longer time because of hemodynamic instability or enzyme elevation (P < 0.05). In conclusion, for coronary arterial revascularization, WRC is technically more demanding and does not appear to afford optimal myocardial protection.