One hundred seventeen consecutive patients undergoing repeat percutaneous transluminal coronary angioplasty (PTCA) were studied to assess procedural success and recurrent restenosis rates. Clinical, anatomic and procedural variables were examined as predictors of recurrent restenosis using stepwise logistic regression analysis. Primary success was achieved in 114 patients (97.5%). One patient (0.8%) died after acute occlusion. No other in-hospital complications were encountered. After a mean follow-up interval of 218 +/- 160 days, 72 of 114 successfully dilated patients (63%) remained angina free. There were no late deaths. Three patients (2.6%) experienced a late myocardial infarction. Follow-up arteriography was performed in 100 patients (88%), of whom 32% had recurrent restenosis (greater than 50% luminal diameter narrowing). On univariate analysis, the presence of 3 clinical variables at repeat PTCA was associated with significantly higher recurrent restenosis rates compared with their absence, that is, unstable angina (48 vs 20%, p = 0.003), diabetes (61 vs 26%, p = 0.003) and hypertension (46 vs 18%, p = 0.003). Patients with recurrent restenosis had a shorter interval between first and second PTCA compared with those who remained patent (136 +/- 116 vs 214 +/- 163 days, p = 0.018). Multivariate analysis confirmed unstable angina, diabetes and hypertension as independent predictors of recurrent restenosis. Repeat PTCA may be performed for restenosis with a high likelihood of success and low incidence of complications. The rate of recurrent restenosis is similar to that reported for initial angioplasty. Patients with unstable angina, diabetes and hypertension appear to be at higher risk for recurrent restenosis.