Resuscitability and outcome after prolonged cardiac arrest were compared in dogs with standard external cardiopulmonary resuscitation (CPR) vs. closed-chest emergency cardiopulmonary bypass (CPB). Ventricular fibrillation (VF) was with no blood flow from VF 0 min to VF 10 min. Subsequent CPR basic life support (BLS) was from 10 min to VF 15 min. Then, group I (n = 13) received CPR advanced life support (ALS) from VF 15 min until restoration of spontaneous circulation to occur not later than VF 40 min. Group II (n = 14) received CPR-ALS from VF 15 min to VF 20 min without defibrillation, and then total CPB to defibrillation attempts started at VF 20 min, followed by assisted CPB to 2 h. Total ischemia time (no-flow time plus CPR time of MAP less than 50 mmHg) was unexpectedly shorter in group I (14.3 +/- 2.5 min) than in group II (18.6 +/- 2.3 min) (P less than 0.01). During CPR-BLS, coronary perfusion pressures were 25 +/- 9 mmHg in group I and 18 +/- 8 mmHg in group II (NS). Epinephrine during CPR-ALS, before countershock, raised coronary perfusion pressure to 40 +/- 10 mmHg in group I and 27 +/- 10 mmHg in group II (NS). In group II, coronary perfusion pressure increased during total CPB to 58 +/- 16 mmHg (P less than 0.01 vs. group I). Spontaneous normotension was restored in 11/13 dogs of group I and all 14 dogs of group II (NS). Ten dogs in each group followed protocol and survived to 96 h. Five of ten in group I and six of ten in group II were neurologically normal (NS). We conclude that: (1) Reperfusion with CPB yields higher coronary perfusion pressures than reperfusion with CPR-ALS; and (2) even after no blood flow for 10 min, optimized CPR can result in cardiovascular resuscitability and neurologic recovery, similar to those achieved by CPB.