As the clinical significance of hemodynamic parameters remains controversial, the portohepatic gradient has been measured in 89 patients with cirrhosis by a transjugular approach. Relationships between portohepatic gradient and clinical, roentgenologic and prognostic patterns were studied with a maximum follow-up period of 36 months. Portohepatic gradient was not related to the rate of occurrence or rate of recurrence of digestive tract bleeding. Portohepatic gradient was related to the size of esophageal varices, p less than 0.01. Observation of large varices at endoscopy was associated with a higher risk of digestive tract bleeding. Portohepatic gradient was related to the number of portosystemic shunts opacified at portography, p less than 0.01. Portohepatic gradient was found to be related to these aspects of prognosis: the mean portohepatic gradient was higher in patients who died than it was in survivors, p less than 0.05. When patients were separated into two groups according to portohepatic gradient values--greater than or equal to 20.8, less than 20.8 millimeters of mercury--the actuarial survival rate was inversely related to portohepatic gradient at one, p less than 0.02, and at 12, p less than 0.02, months of follow-up study. This relationship could be demonstrated in the entire group of patients with cirrhosis and in the group of patients with digestive tract bleeding. This pattern seemed to be related to the risk of hemorrhage. In groups of patients and within a 24 month observation period, the survival rate was significantly related to the range of portohepatic gradient in each group.