These experiments were designed to determine whether the limited cardiac output during open cardiac massage could be preferentially directed to the coronary and cerebral vessels by balloon occlusion of the descending thoracic aorta. Sixteen dogs were instrumented to monitor cardiac output and left atrial, right atrial, right ventricular, left ventricular, and arterial blood pressures. Measurements of myocardial and cerebral blood flow distribution during massage were made using the radioactive microsphere technique. Each animal underwent two episodes of fibrillation and resuscitation. In one episode the arrest was managed by open massage alone, and in the other, open massage was accompanied by balloon occlusion, with the order randomized. When compared to control, open cardiac massage was associated with a significant decrease in mean arterial pressure; however, the addition of balloon occlusion produced a 130% increase in the mean arterial pressure that was obtained during open CPR (control, 93 ± 5 mm Hg; massage alone, 35 ± 2 mm Hg; massage + balloon, 76 ± 2 mm Hg, P < 0.01). In a similar fashion, although the absolute blood flow was reduced by 50% when compared to control, the blood flow (ml/min/g) to the brain and heart during massage was 100% better when balloon occlusion was employed (brain: control, 0.41 ± 0.03; massage only, 0.05 ± 0.01; massage + balloon, 0.25 ± 0.02, P < 0.01; heart: control, 1.46 ± 0.11; massage alone, 0.35 ± 0.05; massage + balloon, 0.71 ± 0.05, P < 0.01). These results suggest that aortic occlusion significantly increased myocardial and cerebral perfusion patterns during ventricular fibrillation and open cardiac massage. Percutaneous transfemoral balloon aortic occlusion during CPR may be a useful adjunct to standard therapy.