Six of 85 patients (7%) with alcoholic liver disease undergoing transhepatic portal pressure measurement had either stagnant (3 patients) or reversed (3 patients) portal blood flow documented by gentle hand injection of 1 to 2 ml of angiographic contrast. Portal blood flow was uniformly hepatopetal in 24 patients with nonalcoholic liver disease. Recurrent spontaneous hepatic encephalopathy and sodium retention occurred in 4 of 6 patients with stagnant or reversed portal flow; gastrointestinal bleeding was not seen. Standard laboratory tests of liver function were widely variable. Net portal pressure was lower in this group than in patients with alcoholic liver disease and forward portal flow (9.2 f 2.6 vs. 15.6 2 4.1 mm Hg, p c 0.001). Wedged hepatic vein pressure was 1 to 7 mm Hg higher than portal vein pressure in patients with reversed portal flow. The ar- terioportal extraction of bile acid was calculated from the difference in concentration between artery and portal vein, and total functional hepatic blood flow was calculated from the hepatic extraction and systemic clearance of indocyanine green. Extraction was 070, and hepatic blood flow was 0.469 liter per min in a patient with hepatofugal portal flow and recurrent encephalopathy. Extraction was 2070, and hepatic blood flow was 4.014 liters per min in a patient who had never had encephalopathy. These data indicate that arterioportal communications may be sinusoidal or presinusoidal in patients who lose forward portal flow and that the amount of flow in the arterioportal circuit, together with its efficiency, largely determine the clinical outcome.
The occurrence of stagnant (to-and-fro) and reversed portal vein blood flow is recorded in patients with cirrhosis (1-4). However, the pathophysiology of this condition is not well defined. We report six patients with cirrhosis and absence of portal inflow whose clinical, hemodynamic and metabolic features shed further light on this extreme derangement of portal hemodynamics.