PURPOSE To evaluate the safety and efficacy of polyvinyl alcohol (PVA) terminal chemoembolization and to identify the prognostic factors associated with survival in hepatocellular carcinoma (HCC) patients with hepatic arteriovenous shunts (HAVS). MATERIALS AND METHODS Of 133 patients' managements were retrospectively analyzed. HAVS was classified into three types: slow-flow, intermediate-flow and high-flow. The size of the PVA used was determined following the scheme: slow-flow HAVS: 300-500μm PVA; intermediate-flow HAVS: 500-710μm PVA; high-flow HAVS: 710-1000μm PVA. The HCCs with slow-flow and intermediate-flow HAVS were embolized by PVA plus chemotherapeutic agents lipiodol emulsion, while the high-flow HAVS were treated by PVA with chemotherapeutic agents. Survival curves were calculated by Kaplan-Meier method and compared by log-rank test. The influence of possible prognostic factors on survival were analyzed by multivariate Cox proportional-hazards method. RESULTS The median overall survival (OS) of 133 patients was 9.1 months. The median OS of the slow-flow type, intermediate-flow type and high-flow type patients were 10.8, 9.1 and 7.3 months, respectively. There was no statistically significant difference among different HAVS types (P=0.239). The 30-day mortality was 3.8%. Cox multivariate survival analysis revealed that initial preoperative AFP value≥400ng/ml (HR=2.105, P=0.006) was an independent risk factor. While multiple embolization (HR=0.482, P=0.011), tumor remission (HR=0.431, P=0.041) and multimodality therapy (HR=0.416, P=0.004) were independent protection factors. CONCLUSION It is safe and effective for HCCs with HAVS treated by terminal chemoembolization therapy with PVA plus chemotherapeutic agents lipiodol emulsion (or PVA plus chemotherapeutic agents). The HCCs with HAVS achieves good prognosis with multiple embolization, tumor remission and multimodality therapy, while achieves poor prognosis with inital preoperative high AFP value (≥400ng/ml).