The long-term follow-up of patients with the severe form of Manson's schistosomiasis who had had elective surgical treatment for portal hypertension, in a randomized trial, was clinically evaluated. Of 94 patients, proximal splenorenal shunting was performed in 32, esophagogastric devascularization with splenectomy in 32 and distal splenorenal shunting in 30. Patients were observed during a mean of 85.7 +/- 33.1 mo, excluding nine patients (9.6%) who were lost to follow-up. Recurrence of upper gastrointestinal tract bleeding occurred in 24.1% of the patients, without statistical differences among the three groups, but rebleeding because of varices was more frequent after esophagogastric devascularization with splenectomy. Hepatic encephalopathy was significantly higher after proximal splenorenal shunting (39.3%) when compared with distal splenorenal shunting (14.8%) and with esophagogastric devascularization with splenectomy (0%). Lethality was also significantly higher after proximal splenorenal shunting (42.9%) when compared with distal splenorenal shunting (14.8%) and with esophagogastric devascularization with splenectomy (7.1%). Indirect hyperbilirubinemia was absent after esophagogastric devascularization with splenectomy and more frequent after distal splenorenal shunting (52%) although also present after proximal splenorenal shunting (29.6%). Esophagogastric devascularization with splenectomy was demonstrated to be the best option because of the absence of encephalopathy and because of low mortality rates. Hepatic encephalopathy occurred after distal splenorenal shunting but in a lesser percentage than after proximal splenorenal shunting. The higher incidence of encephalopathy and lethality proscribes proximal splenorenal shunting in Manson'schistosomiasis.