The occurrence of balloon slippage (“watermelon seeding”; WMS) during treatment of patients with in-stent restenosis (ISR) has been described, but predisposing factors and the potential implications of this phenomenon remain unknown. In the Restenosis Intrastent: Balloon Angioplasty vs. Elective Stenting (RIBS) randomized study, 450 patients with ISR were included. Of these, 42 patients (9%) presented WMS during the procedure. WMS was detected in 26 patients (12%) in the balloon arm and 16 (7%) in the stent arm (P = 0.11). In the stent arm, WMS was only noticed during balloon predilation, never during stent implantation. As compared with 408 patients without WMS, patients with WMS had more severe (TIMI flow 1; 21% vs. 8%; P = 0.01) and diffuse (length > 15 mm: 45% vs. 28%; P = 0.02) ISR lesions. Patients with WMS required more balloon inflations, longer total inflation time, had more frequent crossover to stenting or ended the procedure with residual dissections, and eventually obtained poorer acute results (minimal lumen diameter, 2.35 ± 0.5 vs. 2.53 ± 0.5 mm; P = 0.03). In addition, at 6-month follow-up, patients with WMS had a smaller minimal lumen diameter (1.26 ± 0.7 vs. 1.61 ± 0.7 mm; P = 0.007) and a higher restenosis rate (56% vs. 37%; P = 0.017). On logistic regression analysis, the WMS phenomenon emerged as an independent predictor of recurrent restenosis (adjusted RR = 2.1; 95% CI = 1.1–4.1; P = 0.04). The WMS phenomenon may complicate treatment of patients with ISR. Long and severe lesions appear to predispose to this technical problem that never occurs during stent deployment. In patients with ISR, WMS is associated with cumbersome procedures and poorer acute and long-term angiographic results. © 2005 Wiley-Liss, Inc.