Surgical procedures for bleeding esophagogastric varices when sclerotherapy fails: a prospective study.

In a 7-year period, 692 patients were admitted to the Heinz-Kalk Hospital with bleeding esophageal varices. All patients were first treated with endoscopic sclerotherapy (ES). Fourteen patients exsanguinated during the first 2 hours of admission or refused treatment. Long-term injection sclerotherapy was performed in 311 Child-Pugh C patients. Among the remaining 367 patients, 182 were in class A and 185, class B. In 194 patients, ES was successful. In 173 patients with at least two rebleeding episodes despite long-term sclerotherapy, specific selection criteria were used to assess suitability for a shunt. Eighty-eight patients received a shunt: 54, a narrow-lumen mesocaval (NLMS) shunt; 32, a distal splenorenal shunt; 1, a portacaval shunt; and 1, a proximal splenorenal Linton shunt. There was no significant difference in mortality at 30 days; however, late mortality in the ES group was 36% and in the shunt group, 17%. This difference was statistically significant in favor of the shunt operation (p less than 0.01). Thus, using specific selection criteria, shunt procedures performed in Child-Pugh class A and B patients are the best form of treatment for bleeding esophagogastric varices resistant to long-term ES. Furthermore, the narrow-lumen mesocaval shunt is a good alternative to the distal splenorenal shunt if the latter is technically impossible to perform or hemodynamically not advisable.