OBJECTIVE--Active compression-decompression (ACD) cardiopulmonary resuscitation (CPR) appears to improve ventilation and coronary perfusion when compared with standard CPR. The objective was to evaluate potential benefits of this new CPR technique in patients with out-of-hospital cardiac arrest in St Paul, Minn. DESIGN--Ten-month, prospective, randomized parallel-group design. SETTING--St Paul, Minn, population 270,000. PATIENTS--All normothermic victims of nontraumatic cardiac arrest older than 8 years who received CPR. MAIN OUTCOME MEASURES--Return of spontaneous circulation, admission to the intensive care unit (ICU), return of baseline neurological function (alert and oriented to person, place, and time), survival to hospital discharge, survival to hospital discharge with return of baseline neurological function, and complications. RESULTS--Seventy-seven patients received standard CPR and 53 patients received ACD CPR. The mean emergency medical services call response interval was less than 3.5 minutes. When all patients were considered, a higher percentage of ACD CPR patients had a return of spontaneous circulation and were admitted to the ICU vs standard CPR (45% vs 31%, and 40% vs 26%, respectively), but these trends were not statistically significant (P < .10 and P < .10). No statistically significant differences were found between hospital discharge rates (12 [23%] of 53 for ACD CPR vs 13 [17%] of 77 for standard CPR), return to baseline neurological function (10 [19%] of 53 for ACD CPR vs 13 [17%] of 77 for standard CPR), or return to baseline neurological function at hospital discharge (nine [17%] of 53 for ACD CPR vs 12 [16%] of 77 for standard CPR). Return of spontaneous circulation, ICU admission, and neurological recovery in both CPR groups were highly correlated with downtime (time from collapse to emergency medical system personnel arrival to the scene in witnessed arrests). With less than 10 minutes' downtime, survival to the ICU was 59% (19/32) with ACD CPR and 33% (16/49) with standard CPR (P < .02), return to baseline neurological function was 31% (10/32) with ACD CPR and 20% (10/49) with standard CPR (P = .27), and hospital discharge rate was 38% (12/32) with ACD CPR and 20% (10/49) with standard CPR (P = .17). Complication rates in patients admitted to the hospital were similar in both groups. CONCLUSIONS--This study demonstrates that ACD CPR appears to be more effective than standard CPR in a well-defined subset of victims of out-of-hospital cardiac arrest during the critical early phases of resuscitation. Based on this study, a larger study should be performed to evaluate the potential long-term benefits of ACD CPR.