Sirolimus-eluting stents have been developed to prevent restenosis in the treatment of coronary artery disease. We investigated the risk of restenosis with use of sirolimus-eluting stents compared with bare-metal stents to assess possible differences.We enrolled 352 patients in whom one coronary artery required treatment, with diameter 2·5–3·0 mm and lesion length 15–32 mm. We randomly assigned patients sirolimus-eluting stents (n=175) or bare-metal stents (control, n=177). At 8 months we assessed differences in minimum lumen diameter and binary restenosis within the lesion (restenosis of ≤50% diameter, including 5 mm vessel segments proximal and distal to stented segment). Patients were also followed up for 9 months for major adverse cardiac events. Analysis was by intention to treat.Stent implantation was successful in 100% of sirolimus-stent patients and 99·4% of controls. The mean diameter of treated coronary arteries was 2·55 mm (SD 0·37) and mean lesion length was 15·0 mm (6·0). Multiple stents were implanted in 170 (48%) patients. At 8 months, minimum lumen diameter was significantly higher with sirolimus-eluting stents than with control stents (2·22 vs 1·33 mm, p<0·0001). The rate of binary restenosis was significantly reduced with sirolimus-eluting stents compared with control stents (5·9 vs 42·3%, p=0·0001). Significantly fewer patients with sirolimus-eluting stents had major adverse cardiac events at 9 months than did controls (8·0 vs 22·6%, p=0·0002), due mainly to a lower need for target-lesion revascularisations (4·0 vs 20·9%, p<0·0001).Sirolimus-eluting stents are better than bare-metal stents for treatment of single long atherosclerotic lesions in a coronary vessel smaller than 3 mm in diameter.Published online Sept 30, 2003. http://image.thelancet.com/extras/03art9099web.pdf