OBJECTIVE We classified the Doppler waveform seen in patients with portal hypertension and examined the associations of the waveform type with the diagnosis of Budd-Chiari syndrome and severity of the liver cirrhosis. METHODS The Doppler pattern of right and left hepatic veins in 100 consecutive Japanese patients with portal hypertension and esophagogastric varices was classified into six types: I, triphasic waveform; II, biphasic waveform without reversed flow; III, decreased amplitude of phasic oscillations; IV, flat waveform with fluttering; V, completely flat waveform with fluttering; VI, no waveform. All patients underwent computed tomography and magnetic resonance imaging. Patients in whom hepatic vein waveform showed type IV, type V, or type VI, positively underwent hepatic venography and inferior vena cavography. RESULTS Type I was seen in 31 of 100 patients, type II in 35, type III in 17, type IV in eight, type V in four, and type VI in five. Types I-IV waveform indicated no lesion in hepatic veins and inferior vena cava, type V indicated stenosis of hepatic veins or occlusion of inferior vena cava, and type VI, occlusion of hepatic veins. For one patient with type V hepatic veins, balloon angioplasty was done, and the waveform changed from type V to type II. Examining the relationship between hepatic vein waveform and the Child-Pugh score, liver function of type IV cases was worse than that of type I cases in 66 cirrhotic patients without hepatocellular carcinoma (p < 0.05). There was no clear relationship between hepatic vein waveform and portal venous perfusion, as based on Nordlinger's grade. CONCLUSIONS Our classification of hepatic vein waveform in Doppler ultrasonography is useful in diagnosing Budd-Chiari syndrome, in judging the efficiency of treatment for hepatic vein lesions, and in assessing severe liver function in cirrhotic patients.