Long-Term Outcomes of Coronary-Artery Bypass Grafting versus Stent Implantation - Long-Term Outcomes of Coronary-Artery Bypass Grafting versus Stent Implantation


The goal of the study was to evaluate long-term mortality comparing coronary artery bypass surgery (CABG) with percutaneous coronary intervention (PCI) with stenting among patients with multivessel disease.

Study Design

Study Design:

Patients Enrolled: 59,314
Mean Follow Up: Median 706 days in CABG patients and 585 days in PCI patients.
Mean Patient Age: Median 66 years
Female: 30

Patient Populations:

Multivessel coronary artery disease, defined as ≥70% stenosis in ≥2 of the 3 main arteries; New York state resident


Prior revascularization, left main disease >50%, or acute myocardial infarction within 24 hours prior to revascularization.

Primary Endpoints:


Secondary Endpoints:


Drug/Procedures Used:

Data were draw from two New York state cardiac registries (Cardiac Surgery Reporting System [CSRS] and the Percutaneous Coronary Intervention Reporting System [PCIRS]) from the period of January 1, 1997 through December 31, 2000. The study included 37,212 patients treated with CABG and 22,102 patients treated with PCI with stenting. Mortality and repeat revascularization were assessed through three years. Models were used to adjust for baseline risk and severity of illness prior to initial revascularization. Patients were grouped into one of five anatomical categories.

Principal Findings:

Compared with CABG, patients undergoing stenting were younger (median age 65 vs 67 years, p<0.001), and more often female (31.4% vs 29.1%, p<0.001). Ejection fraction was higher in the PCI group (median 53% for stenting vs 50% for CABG, p<0.001). Other risk factors were also lower in the stenting group, including stroke (4.4% vs 6.9%, p<0.001), carotid or cerebrovascular disease (3.5% vs 14.0%, p<0.001), and congestive heart failure (11.4% vs 19.5%, p<0.001). Double vessel disease was more frequent in stent patients (80.4% vs 30.7%), while triple vessel disease was more frequent in CABG patients (69.3% vs 19.6%, p<0.001).

Unadjusted in-hospital mortality was higher in the CABG group compared with the stent group (1.75% vs 0.68%, p<0.001). However, long-term risk-adjusted survival was higher in the CABG group compared with the PCI group in all anatomical subgroups, including triple-vessel disease with proximal left anterior descending artery (LAD) involvement (89.3% for CABG vs 84.4% for PCI, hazard ratio [HR] 0.64, 95% CI 0.56-0.74) or non-proximal LAD involvement (HR 0.74, 95% CI 0.62-0.90) and double-vessel disease without LAD involvement (93.3% for CABG vs 91.4% for PCI, HR 0.75, 95% CI 0.58-0.98), with non-proximal LAD involvement (HR 0.76, 95% CI 0.60-0.96), or with proximal LAD involvement (92.1% for CABG vs 89.8% for PCI, HR 0.75, 95% CI 0.66-0.86). The mortality benefit was similar when further stratified by diabetes and ejection fraction ≥40% or <40%. Among patients with three-vessel disease with proximal LAD involvement, even the univariate survival estimates at 3 years were significantly higher in the CABG group (89.2% for CABG vs 84.9% for stenting; p<0.001). Revascularization rates were also higher in the stent group compared with the CABG group (7.8% vs 0.3% for subsequent CABG; 27.3% vs 4.6% for subsequent PCI; p<0.001 for each).


Among patients with multivessel disease, long-term risk-adjusted survival rates were higher in patients treated with CABG compared with those treated with coronary stenting, both overall and in all anatomical subgroups. Likewise, the need for repeat revascularization was lower in the CABG group, despite a much higher pre-revascularization risk profile. Prior randomized trials comparing CABG with PCI have consistently shown lower rates of repeat revascularization with CABG. Additionally, mortality benefit for CABG has been shown in meta-analysis of randomized PCI vs CABG trials since individual trials lacked power to detect a difference. However, many of the randomized trials were conducted prior to widespread use of coronary stents, unlike the present study.

A limitation of the present study is the non-randomized treatment strategy, which was at the discretion of the physician. However, data were adjusted for identified confounders as well as the propensity to undergo either type of revascularization. Additionally, data for the trial were conducted prior to the use of drug-eluting stents, although no studies have ever shown a mortality effect associated with the use of drug-eluting stents.


Hannan EL, et al. Long-Term Outcomes of Coronary-Artery Bypass Grafting versus Stent Implantation. N Engl J Med 2005;352:2174-83.

Keywords: Registries, Stroke, Hospital Mortality, Drug-Eluting Stents, Survival Rate, Heart Failure, Risk Factors, Constriction, Pathologic, Coronary Artery Bypass, Cardiac Surgical Procedures, Angioplasty, Balloon, Coronary, Diabetes Mellitus

< Back to Listings