Traditional binary definitions of coronary restenosis based on 6-month continuous angiographic measurements (e.g., > 50% diameter stenosis) may give confusing results for lesions whose late percent stenosis falls near the arbitrary threshold. To determine the long-term clinical consequences of such lesions, the overall correlation between follow-up percent stenosis and the performance of subsequent ischemia-driven target vessel revascularization (triggered by significant angina or a positive exercise study result, or both) was examined in 443 consecutive lesions treated with directional coronary atherectomy or Palmaz-Schatz coronary stenting. Follow-up angiograms (available in 355 lesions, 82%) were stratified into 3 groups: severe late stenosis (> 70% stenosis, n = 59), moderate late stenosis (40% to 70% stenosis, n = 72), and minimal late stenosis (< 40% stenosis, n = 224). With an average clinical follow-up of 933 +/- 394 days, 92% of lesions in the "severe late stenosis" group were treated with ischemia-driven target vessel revascularization, compared with 0% of the lesions in the "minimal late stenosis" group. Ischemia-driven target vessel revascularization was performed in 38% of patients in the "moderate late stenosis" group. However, patients in this group who did not undergo revascularization (despite the fact that 43% of them had a late stenosis of > 50%) showed a similarly favorable long-term clinical outcome to patients with a minimal late stenosis. These results support a strategy of conservative management for the 20% of patients who have a moderate (40% to 70%) late stenosis after stenting or atherectomy, but do not have evidence of ischemia.