The choice of drug-eluting versus bare metal stents is based on costs and expectations of restenosis and thrombosis risk. Approaches to stent placement vary from covering just the zone of maximal obstruction to stenting well beyond the lesion boundaries (normal-to-normal vessel). The independent effects of stented lesion length, nonstented lesion length, and excess stent length, on coronary restenosis have not been evaluated for bare metal or drug-eluting stents. We analyzed the angiographic follow-up cohort (1,181 patients) from 6 recent bare metal stent trials of de novo lesions in native coronary arteries. Stent length exceeded lesion length in 87% of lesions (mean lesion length 12.4 +/- 6.3 mm, mean stent length 20.0 +/- 7.9 mm, mean difference 7.6 +/- 7.9 mm). At 6- to 9-month follow-up, the mean percent diameter stenosis was 39.1 +/- 20.1%. In an adjusted multivariable model of percent diameter stenosis, each 10 mm of stented lesion length was associated with an absolute increase in percent diameter stenosis of 7.7% (p <0.0001), whereas each 10 mm of excess stent length independently increased percent diameter stenosis by 4.0% (p <0.0001) and increased target lesion revascularization at 9 months (odds ratio 1.12, 95% confidence interval 1.02 to 1.24). Significant nonstented lesion length was uncommon (12.5% of cases). In summary, stent length exceeded lesion length in most stented lesions, and the amount of excess stent length increased the risk of restenosis independent of the stented lesion length. This analysis supports a conservative approach of matching stent length to lesion length to reduce the risk of restenosis with bare metal stents.