Background and aims: Recent progress in the treatment of variceal bleeding might have reduced the impact of variceal bleeding on survival in patients with esophageal varices. We conducted a retrospective cohort study in an attempt to re-evaluate the clinical significance of variceal bleeding.
Methods: A cohort of 304 patients with liver cirrhosis and esophageal varices, who had no previous history of variceal bleeding and no prophylactic therapy, was studied.
Results: During a median follow-up period of 32 months, 55 patients (18%) bled from varices and 111 (37%) died. Variceal hemorrhages accounted for 15% of total deaths. The mortality of first variceal bleeding was 25% in the whole group, but was remarkably different depending on liver function at the time of bleeding (0% in grade Child A vs 55% in grade C; P < 0.05). Among the survivors of first bleeding, 30% experienced rebleeding. Form of varix, red color sign and heavy drinking were the independent risk factors for first variceal bleeding. Multivariate analysis revealed that variceal bleeding still had a significant (P < 0.001) impact on death in the whole cohort, when other independent prognostic factors such as age, ascites, encephalopathy, platelet count, serum albumin level and hepatocellular carcinoma were adjusted. Furthermore, in subgroup analyses, variceal bleeding was more strongly (P < 0.001) linked to death in patients with alcoholic cirrhosis than in those with non-alcoholic cirrhosis, and showed a significant association with survival only for the patients in Child grade B.
Conclusions: Variceal bleeding has various prognostic impacts depending on the etiology of cirrhosis or on the degree of liver dysfunction, and this needs to be taken into account in the prophylaxis against first variceal bleeding.