BACKGROUND: The aim of this study was to investigate the efficacy and safety of paclitaxel-coated balloon (PCB) for the treatment of the bare-metal stent restenosis (BMS-ISR) and drug-eluting stent restenosis (DES-ISR). METHODS: This study was a prospective, multicenter, randomized (2:1) trial conducted in 208 patients with 213 in-stent restenosis lesions (BMS-ISR: 123 lesions, DES-ISR: 90 lesions) at 13 centers in Japan. Patients were randomly assigned to a PCB group (137 patients with 142 lesions) or a conventional balloon angioplasty (BA) group (71 patients with 71 lesions). The primary end point was target vessel failure at 6-month follow-up. RESULTS: Clinical and angiographic follow-up 6 months after intervention was performed in 207 patients (99.5%) with 208 lesions (97.7%). Target vessel failure was noted in 6.6% of the PCB group and 31.0% of the BA group (P < .001). Recurrent restenosis occurred in 4.3% of the PCB group and 31.9% of the BA group (P < .001). Late lumen loss was lower in the PCB group than in the BA group (0.11 ± 0.33 mm vs 0.49 ± 0.50 mm, P < .001). In PCB-treated lesions, recurrent restenosis occurred in 1.1% of patients with BMS-ISR and in 9.1% of patients with DES-ISR (P = .04). Late lumen loss was lower in patients with BMS-ISR than in patients with DES-ISR (0.05 ± 0.28 mm vs 0.18 ± 0.38 mm, P = .03). CONCLUSIONS: This randomized clinical study suggested that PCB provided much better clinical and angiographic outcomes than did conventional BA in patients with BMS-ISR and DES-ISR. Drug-eluting stent restenosis was associated with poorer outcomes compared with BMS-ISR after treatment with PCB.