Improved survival after variceal hemorrhage over an 11-year period in the Department of Veterans Affairs

OBJECTIVES: Over the past two decades, several modalities have become widely used in the management of esophageal variceal hemorrhage. The effectiveness of these measures on the outcome of patients with this type of hemorrhage remains unknown. METHODS: Using the Department of Veterans Affairs (VA) Patient Treatment File, we identified two cohorts of patients diagnosed with an initial variceal hemorrhage: An early cohort during 1981-1982 (1339 patients), and a late cohort during 1988-1991 (3636 patients). Each cohort was followed for 6 yr for rebleeding and death. Analyses were performed with proportional hazards survival analysis controlling for confounding factors. RESULTS: On presentation, patients in the late cohort were older (57 yr vs 55 yr, p < 0.0001) and had more ascites (25% vs 13%, p < 0.0001), more peritonitis (4% vs 2%, p < 0.0001), and more encephalopathy (14% vs 9%, p = 0.0003). The late cohort experienced a significant decline in mortality at 30 days (20.8% vs 29.6%, p = 0.0001) and at 6 yr (69.7% vs 74.5%, p = 0.0001). This improvement was accentuated in multivariate survival analysis when controlling for the more severe illness in the late cohort. For patients who survived the first 30 days, no significant difference in 6-yr mortality was found on univariate analysis between the early cohort (63.7%) and late cohort (61.8%) (p = 0.25), but survival was slightly better in the late cohort on multivariate analysis (p = 0.01). In the late cohort, patients with sclerotherapy during the initial hospitalization had better 30-day (17%) and 6-yr mortality (68%) than did the rest of the late cohort. CONCLUSIONS: Between the years 1981-1982 and 1988-1991, improvements in long-term survival after an initial episode of esophageal variceal hemorrhage resulted primarily from better short-term mortality. Sclerotherapy offers a partial explanation for improved survival. (C) 2000 by Am. Coll. of Gastroenterology.