Restenosis after coronary angioplasty is the single complication that most limits this revascularization procedure in clinical practice. The process is largely unpredictable and the lesion-related factors predisposing to restenosis are poorly understood, with little consensus in published reports. In this study using detailed quantitative angiographic measurements to assess 490 lesions, the simple lesion characteristics associated with restenosis were defined and the relation to the restenosis process documented. Restenosis was defined as an absolute deterioration in the minimal lumen diameter by greater than or equal to 0.72 mm, a criterion based on the 95% confidence intervals for repeat angiographic measurements. This was chosen in an attempt to separate spurious changes due to a poor angiographic result and the variability of angiographic measurements from significant changes due to the restenosis process. The principal determinants of restenosis were found to be a large improvement in the minimal lumen diameter at the time of dilation (1.13 mm for the restenosis group compared with 0.86 mm for the no restenosis group [p less than 0.0001]) and an optimal postangioplasty result (minimal lumen diameter 2.28 mm in the restenosis group compared with 2.05 mm [p less than 0.001] in the no restenosis group, corresponding to a 25% and a 30% diameter stenosis, respectively [p less than 0.0001]). These observations reported for the first time suggest that the distinction needs to be made between a "clinical restenosis" of greater than or equal to 50% diameter stenosis and the "restenosis process" as measured by the absolute changes occurring during and after angioplasty.(ABSTRACT TRUNCATED AT 250 WORDS)