OBJECTIVES: This study investigated the efficacy and safety of adding systemic (IV) bevacizumab (Bev) to hepatic arterial infusion (HAI) with floxuridine (FUDR)/dexamethasone (Dex) in unresectable primary liver cancer. METHODS: Patients with unresectable intrahepatic cholangiocarcinoma (ICC) or hepatocellular carcinoma (HCC) were treated with HAI FUDR/Dex plus IV Bev. Results were compared to a recent study of HAI without Bev in a similar patient population. RESULTS: Twenty-two patients (18 ICC, 4 HCC) were treated with HAI FUDR/Dex plus Bev; 7 (31.8%) had partial response and 15 (68.2%) had stable disease. Median survival was 31.1 months (CI 14.14-33.59), progression-free survival (PFS) 8.45 months (CI 5.53-11.05), and hepatic PFS 11.3 months (CI 7.93-15.69). In the previous trial with HAI alone (no Bev), the response was 50%; median survival, PFS, and hepatic PFS were 29.5, 7.3, and 10.1 months. In the present trial, bilirubin elevation (>2 mg/dl) was seen in 24% of patients and biliary stents were placed in 13.6%, versus 5.8 and 0%, respectively, in the HAI trial without Bev. Due to increased biliary toxicity, the trial was prematurely terminated. CONCLUSION: Adding Bev to HAI FUDR/Dex appeared to increase biliary toxicity without clear improvement in outcome (median PFS 8.45 vs. 7.3 months, and median survival 31.1 vs. 29.5 months, for HAI + Bev vs. HAI alone groups, respectively).