A prospective randomized controlled trial was performed to determine optimal flow rates and hemoglobin concentrations for continuous normothermic blood cardioplegia and to compare warm heart surgery with standard intermittent cold blood cardioplegia. Thirty-five patients received intermittent cold blood cardioplegia, low hemoglobin low flow, low hemoglobin high flow, high hemoglobin low flow, or high hemoglobin high flow warm blood cardioplegia (seven patients per group: low hemoglobin, 50 g/l; high hemoglobin, 80 g/l; low flow, less than 80 ml/min; high flow, greater than 80 ml/min). Hypothermia resulted in a significantly greater accumulation of ADP and AMP during cross clamp, consistent with impaired mitochondrial function. Low hemoglobin low flow warm blood cardioplegia increased myocardial oxygen consumption and coronary sinus blood flow after cross clamp release, and also decreased lactate consumption. Postoperative myocardial performance and diastolic compliance were reduced in low hemoglobin low flow warm patients, and diastolic compliance was increased with high hemoglobin high flow warm blood cardioplegia when compared with cold patients. In this study, continuous normothermic cardioplegia was safe when delivered at 80 ml/min or greater, with a hemoglobin concentration of at least 80 g/l, affording myocardial metabolic and functional recovery comparable to that found after intermittent cold blood cardioplegia.