BACKGROUND The proximal left anterior descending artery (LAD) is a challenging area for percutaneous interventions; therefore, coronary artery bypass grafting is often considered and sometimes performed even in patients with single-vessel disease involving the proximal LAD. This study compares mid-term results of LAD revascularization using drug-eluting stents (Cypher) with minimally invasive direct coronary artery bypass grafting (MIDCAB). METHODS From May 2002 to December 2003, 376 consecutive patients underwent myocardial revascularization of the LAD, 272 by Cypher and 104 by MIDCAB. After matching for age, sex, and extent of coronary artery disease, two groups of 83 patients each were used to compare the two revascularization modalities. The groups were similar; however, ejection fraction of less than 0.35 was more prevalent in the MIDCAB group and prior percutaneous coronary intervention in the Cypher group. RESULTS Thirty-day mortality was 1.1% in the MIDCAB and 0% in the Cypher group. Mean follow-up was 22.5 months. Two late cardiac deaths occurred in the MIDCAB group and one in the Cypher group (p = NS). Angina returned in 35% of the Cypher group and in 8.4% of the MIDCAB group (p < 0.001). There were 14 (16.8%) reinterventions in the Cypher compared with three (3.6%) in the surgical group (p = 0.005). Cox proportional hazard model showed that assignment to the Cypher group was the only independent predictor of reangina (hazard ratio [HR], 6.17, 95% confidence interval [CI], 2.46 to 15.4). Treatment with Cypher was also an independent predictor of reintervention (HR 8.26, 95% CI, 1.68 to 40). CONCLUSIONS Despite improved results of percutaneous interventions with Cypher to the LAD, mid-term clinical outcome of patients treated with MIDCAB was better.