T he introduction of coronary stent implantation heralded a major advance in percutaneous coronary intervention (PCI), with dramatically improved in-hospital mortality and morbidity compared with plain old balloon angioplasty for the treatment of coronary artery disease. Several randomized controlled trials (RCTs) have established the superior long-term safety and efficacy of bare-metal stenting (BMS) compared with plain old balloon angioplasty (1-3). The introduction of drug-eluting stents (DES) resulted in a further significant reduction in restenosis, and has been associated with improved procedural outcomes in a more complex subset of patients and lesions (4-7). Thus, coronary stenting using bare-metal or drug-eluting platforms has become the cornerstone of PCI, and the indications for PCI have vastly expanded beyond those originally proposed.Pooled data from randomized DES trials have shown acceptable rates of all-cause mortality, myocardial infarction (MI) and revascularization in extended follow-up, up to four years (8). Although RCTs have excellent internal validity, direct extrapolation of their findings to the general population has been problematic due to their restrictive inclusion and exclusion criteria. Early data (9,10) regarding clinical outcomes in real-world patient reports from registries had limitations in both the follow-up period and the patient cohort size; the large New York State registry (11) comparing three-year outcomes in patients undergoing coronary stenting versus coronary artery bypass graft (CABG) surgery excluded all those with previous revascularization, left main stem (LMS) disease or MI within 24 h of the index procedure (11). More recently, several registries have compared two-(6,12-14) or three-year (15) outcomes in patients undergoing BMSonly versus DES-only PCI, and confirmed higher rates of subsequent revascularization in BMS-treated patients with either lower (12-14) or equivalent (6) rates of mortality or MI in DES-treated patients, or a HeaLtH outcomes/pubLic poLicy ©2010 Pulsus Group Inc. All rights reserved BACKgRounD: Randomized controlled trials report short-and medium-term outcomes following percutaneous coronary intervention (PCI), but their applicability to the general population is not known. Data regarding the long-term clinical outcomes of patients undergoing PCI are lacking. oBJECTIVE: To determine the long-term outcomes of 'all-comers' undergoing PCI at a large-volume tertiary cardiac referral centre. METhoDs: A total of 12,662 consecutive patients undergoing an index procedure and entered into the University Health Network's (Toronto, Ontario) prospective registry between April 2000 and September 2007 were identified. In-hospital outcomes were assessed. Follow-up data were obtained through linkage to a provincial registry. Kaplan-Meier analysis was performed to calculate unadjusted survival rates, and Cox multiple regression analysis identified independent predictors of late mortality, major adverse cardiac events and all cardiovascular events. REsuLTs: T...