Hepatocellular carcinoma (HCC) is one of the more common malignant diseases in the world. Here we have investigated role of hepatic venous outflow obstruction in the development of HCC. During a 10-year period from November 1986 to December 1996, 1,748 patients with clinical evidence of either portal hypertension, hepatic venous outflow obstruction, or inferior vena cava obstruction without Behçet's disease (BD) and 512 patients with Behçet's disease were examined at Hacettepe University Hospital. The presence of and the effect of hepatic venous obstruction on the subsequent development of HCC was assessed. In each case, hepatic vein thrombosis was assessed by hepatic venography and by digital subtraction angiography (DSA), computed tomography (CT), ultrasonography (US), and liver biopsy. Coagulation factors, including protein C, protein S, anti-thrombin III, and routine laboratory studies assessing the coagulability of blood were also investigated. The role of hepatic venous outflow obstruction on the subsequent development of HCC was determined by periodic laboratory investigations that included alpha-fetoprotein (AFP), ultrasonography, and when indicated liver biopsy. During the same time period all patients diagnosed as having HCC were investigated to identify all potential etiologic factors responsible for the HCC. Fifty-five (10.7%) of the 512 patients with BD were found to have one or more large vein thromboses. Sixteen of these 55 (29%) patients had hepatic vein thrombosis. During the follow-up period HCC developed in 2 of these 16 patients (12.5%), 34 and 21 months after a diagnosis of hepatic vein thrombosis was established. Forty patients from a total of 1,748 patients with clinical evidence of portal hypertension and cirrhosis, but without BD, were found to have evidence of hepatic vein thrombosis. Twenty-one of these 40 patients had an identifiable underlying disorder responsible for their hepatic vein thrombosis. Despite a full clinical and laboratory investigation in the other 19 patients, the etiologic factor responsible for the hepatic vein thrombosis remained obscure. Only one of these 19 patients, who also had portal vein thrombosis, developed HCC during a 9-year follow-up. Thus, a total of three of the 56 (5.36%) of cases of hepatic vein thrombosis developed an HCC. All of the hepatic tumors were of the multicentric, nodular, rapidly growing type. Despite the presence of hepatic vein thrombosis, there was no clear-cut histologic evidence for cirrhosis. Our experience suggests that hepatic vein thrombosis may be a contributing factor responsible for HCC development. Moreover, we advise that individuals with hepatic vein thrombosis should be assessed periodically for the development of HCC.