INTRODUCTION: We conducted a randomized phase II study to evaluate the impact of adding bevacizumab (B) to 5-fluorouracil (5-FU), hydroxyurea (HU), and radiotherapy (FHX) for intermediate-stage and select T4 head and neck squamous cell cancers (HNSCC). PATIENTS AND METHODS: Eligible patients had newly diagnosed HNSCC. Randomization was 2:1 in favor of BFHX. All patients received 500 mg HU p.o. b.i.d., 600 mg/m(2)/day continuous infusion 5-FU, and b.i.d. radiotherapy with or without bevacizumab 10 mg/kg administered on day 1 of each 14-day cycle. Patients received five cycles consisting of chemoradiotherapy for 5 days followed by 9 days without therapy. RESULTS: Twenty-six patients were enrolled (19 BFHX and 7 FHX). The study was halted following unexpected locoregional progression. Two-year survival was 68%; 89% treated with FHX and 58% (95% confidence interval 33% to 78%) treated with BFHX. Two-year locoregional control was 80% after chemoradiotherapy and 85% after surgical salvage. All locoregional progression occurred in T4 tumors randomized to BFHX. Two patients receiving BFHX died during therapy, and one died shortly after therapy. No catastrophic bleeding events were seen. CONCLUSIONS: Locoregional progression seen in T4N0-1 tumors treated with BFHX was unexpected and led to study termination. The addition of bevacuzimab to chemoradiotherapy for HNSCC should be limited clinical trials.