The aim of the present randomized, single center trial was to evaluate role of paclitaxel-eluting stent, Taxus compared with sirolimus-eluting stent, Cypher among diabetic patients with symptomatic coronary artery disease with de novo lesions. Patients were randomized to stent implantation with either the Taxus stent (n=125) or Cypher stent (n=125). Patients underwent angiographic follow-up at 6-8 months. Baseline clinical and angiographic characteristics were comparable between the treatment groups, with a lesion length of 12.4 mm in the Taxus group and 13.8 mm in the Cypher group. The primary endpoint of late in-stent lumen loss at angiographic follow-up at 6 to 8 months was 0.67 mm in the Taxus group and 0.43 mm in the Cypher (p=0.86 for non-inferiority; p=0.002 for superiority of the Cypher v. the Taxus group). Angiographic restenosis occurred more frequently in the Taxus group (16.5% v. 6.9%, p=0.03). Target lesion revascularization at 9 months was performed in 12.0% of the Taxus group and 6.4% of the Cypher group (p=0.13). There was no difference in 9-month mortality (4.8% for Taxus v. 3.2% for Cypher, p=0.52) or myocardial infarction (2.4% v. 4.0%, p=NS). Among diabetic patients with coronary artery disease, treatment with the sirolimuseluting stent was associated with a lesser late lumen loss at 6-8 months angiographic follow-up compared to treatment with the paclitaxel-eluting stent. Whether the angiographic benefits will also lead to clinical improvement needs to be further evaluated. The results of the present trial should be restricted to high-risk population of diabetic patients, and should not be extrapolated to other patient subgroups.