Intracoronary stenting has been proposed as an adjunct to balloon angioplasty to improve the immediate and long-term results. However, late luminal narrowing has been reported following the implantation of a variety of stents. One of the studies conducted with the Wiktor stent is a prospective registry designed to evaluate the feasibility, safety and efficacy of elective stent implantation in patients with documented restenosis of a native coronary artery. To identify angiographic variables predicting recurrence of restenosis, the angiograms of the first 91 patients with successful stent implantation and without clinical evidence of (sub)acute thrombotic stent occlusion were analyzed with the Computer Assisted Angiographic Analysis System using automated edge detection. The incidence of restenosis was 44% by patient and 45% by stent according to the 0.72 mm criterion, and 30% by patient and 29% by stent according to the 50% diameter stenosis criterion. The risk for restenosis for several angiographic variables was determined using an univariate analysis and is expressed as odds ratio with corresponding confidence interval. The only statistically significant predictor of restenosis was the relative gain when it exceeded 0.48 using the 0.72 mm criterion (odds ratio 2.7, 95% confidence interval 1.1-6.4). Furthermore, the relation between the relative gain (increase in minimal luminal diameter normalized to vessel size) as angiographic index of vessel wall injury and relative loss (decrease in minimal luminal diameter normalized to vessel size) as index of neointimal thickening was analyzed using a linear regression analysis. When using the categorical approach to address restenosis, there is an increased risk for recurrent restenosis when the relative gain exceeds 0.48. The continuous approach underscores this concept by indicating a weak but positive relation between the relative gain and relative loss.