Objectives: Early angioplasty after thrombolysis is now recommended for ST-elevation myocardial infarction, but the current guidelines propose a wide time-window ranging between 3 and 24 h after lytic administration. To identify the optimal timing for PCI after thrombolysis, we analyzed frequency and time course of the adverse events in patients randomized in the multicenter CARESS-in-AMI trial. Methods: 598 high-risk patients with STEMI recruited in the CARESS-in-AMI study, were divided into the Immediate PCI group (IMM, n=298), Rescue PCI group (RES, n=107) and Standard Treatment Arm without rescue PCI (STA, n=193). Results: RES patients had worse pre-procedural TIMI flow and post-procedural blush grade. At 30 days, there were 23 deaths: 11 (10.3%) in RES, 9 (3%) in IMM and 3 (1.6%) in STA (pb0.001). There were 22 episodes of refractory ischemia or re-infarction: 17 (8.8%) in the STA group, 4 (1.6%) in IMM and 1 (0.9%) in RES (pb0.001). In the RES group 10/11 (90.9%) deaths occurred before day 5. In the STA group, all deaths and the majority of ischemic events occurred after day 3. A reduction of risk of de Conclusions: The mortality benefit of immediate referral to PCI after pharmacological treatment for STEMI derives from a reduction in the time to reperfusion of patients with failed thrombolysis in need of rescue PCI. In patients with evidence of successful reperfusion, "elective" PCI within 3 days may be sufficient to reduce the recurrent ischemic events. (C) 2010 Elsevier Ireland Ltd. All rights reserved.