To determine if clinically accessible hemodynamic and blood gas measurements are of value in predicting outcome of countershock after prolonged ventricular fibrillation (VF) and artificial cardiopulmonary support, 14 dogs were studied during 30 minutes of VF using two randomly assigned closed-chest techniques. Seven dogs underwent conventional CPR; the other seven were supported with a pneumatic thoracic vest and abdominal binder, which were inflated synchronously with the airway. Ascending aortic (Ao), right atrial (RA), and instantaneous coronary perfusion pressures (Ao - RA) were measured at five-minute intervals. Ao and RA blood samples were analyzed at 10, 20, 25 and 30 minutes for PO2, PCO2, and pH. After 25 minutes, 1 mg epinephrine was given intravenously, and five minutes later defibrillation was attempted. If unsuccessful, repeated countershocks, conventional pharmacologic therapy, and artificial support were continued. If a perfusing spontaneous cardiac rhythm did not result within an additional 30 minutes, the experiment was terminated. Six animals developed a perfusing cardiac rhythm after one or more countershocks (Group 1); eight failed to develop a perfusing rhythm after repeated countershocks and an additional 30 minutes of resuscitative effort (Group 2). Five Group 1 dogs received vest/binder artificial support. When measured values were averaged over the study period, Group 1 was found to have a significantly greater Ao end-diastolic pressure (AoEDP) and peak diastolic coronary perfusion pressure (CPP) when compared to Group 2 (23 +/- 6 vs 14 +/- 8 mm Hg, P less than .05; and 22 +/- 6 vs 5 +/- 10 mm Hg, P less than .01, respectively).(ABSTRACT TRUNCATED AT 250 WORDS)