The aim of the present study was to evaluate angiographic late luminal loss after the implantation of sirolimus-eluting Cypher stents and paclitaxel-eluting Taxus stents in patients with diabetes. The study was a Danish multicenter, open-label, randomized trial. One hundred fifty-three patients with diabetes with coronary artery disease were randomized to Cypher (n = 76) or Taxus (n = 77) stent implantation. All patients were followed for 8 months. The primary end point was 8-month angiographic in-stent late luminal loss. This primary end point was reduced in the Cypher group compared with the Taxus group (0.23 +/- 0.54 vs 0.44 +/- 0.52 mm, p = 0.025). Angiographic in-segment restenosis at 8-month follow-up, a secondary end point, was present in 16 patients (Cypher, n = 6; Taxus, n = 10; p = 0.24). Target lesion revascularization was performed in 5 patients (6.5%) and 9 patients (11.8%) in the Cypher and Taxus groups, respectively (p = 0.25). Definite stent thrombosis was observed in 2 patients (in the Taxus group), no patients had probable stent thrombosis, and 1 patient in each group had possible stent thrombosis. Major adverse cardiac events (cardiac death, myocardial infarction, definite stent thrombosis, or target lesion revascularization) were observed in 17 patients (Cypher, n = 6; Taxus, n = 11; p = 0.19). In conclusion, angiographic in-stent late luminal loss is significantly reduced in patients with diabetes by use of the sirolimus-eluting Cypher stent compared with the paclitaxel-eluting Taxus stent.