Aims Lumen enlargement during repeat percutaneous coronary intervention for in-stent restenosis has been shown to be the result of both stent over-expansion and decrease in neointimal tissue. How these two different mechanisms of action may influence outcome and target lesion revascularization after repeat intervention for in-stent restenosis is unclear.Methods Intravascular ultrasound guided repeat intervention for in-stent restenosis was carried out either with balloon angioplasty, or with a combination of rotational atherectomy plus balloon angioplasty. Clinical follow-up at 1 year, including death, myocardial infarction, or need for revascularization, was obtained.Results Seventy patients were included in this study; 40 were treated by balloon alone, and 30 by combination of rotational atherectomy plus balloon. Event-free survival probability was 76 5%. The mechanism of lumen enlargement, be it stent over-expansion or tissue removal, had no influence on long-term clinical evolution. The only independent predictor was the minimal lumen crosssectional area at the end of the procedure, the larger the lumen cross-sectional area, the higher the event-free probability. The cut-off point of the lumen cross-sectional area was set at 4·7 mm 2 by discriminant analysis. Event-free survival was 69 15% in patients with <4·7 mm 2 lumen cross-sectional area and 91 8% in patients with >4·7 mm 2 lumen cross-sectional area (P=0·008).Conclusions This study showed that the only independent predictor of late clinical outcome after percutaneous re-intervention for in-stent restenosis was final lumen size, no matter which means were used to achieve it.