The cutting balloon is a new device for coronary angioplasty, which, by the combination of incision and dilatation of the plaque, is believed to minimize arterial wall trauma, the neoproliferative response, and subsequent restenosis. In this study, we sought to determine predictors of the restenosis using this technique. Seventy-seven patients underwent successful coronary angioplasty with cutting balloon alone. In 67 of these patients (87%), we performed a control angiogram at 6-month follow-up. Pre-, post-, and late angiographic results were evaluated by quantitative coronary analysis. Clinical and angiographic variables were correlated with restenosis as a binary variable and a continuous variable (late loss and late minimum luminal diameter). Univariate analysis showed that the immediate postprocedure minimum luminal diameter (MLD) was smaller in the restenotic group (defined as MLD > 50% by quantitative coronary angiography) than in the nonrestenotic group (1.90 +/- 0.47 mm vs. 2.19 +/- 0.56 mm, P < 0.05). In addition, the immediate percentage of stenosis was higher in the restenotic group than in the nonrestenotic group (37% +/- 10% vs. 27% +/- 11%, P < 0. 003). Multivariate analysis identified the immediate postcutting balloon percentage of stenosis as an independent determinant of binary restenosis (P < 0.008). When restenosis was defined as a continuous variable, the immediate postprocedure MLD was an independent predictor of late loss (P < 0.02) and of late MLD (P < 0. 0002). No clinical, preprocedure angiographic, or technical variables tested were associated with restenosis. The degree of postprocedural residual stenosis after cutting balloon angioplasty is predictive of late restenosis.