RESUMO
BACKGROUND AND AIMS: Few studies have compared two or more cohorts of cirrhotic patients admitted for upper gastrointestinal bleeding (UGIB) several decades apart. Our aim was to compare epidemiological, clinical, therapeutic and prognostic characteristics of UGIB (whatever the source) in two cohorts of cirrhotic patients admitted to the emergency room of the same general hospital 2 decades apart. METHODS: One-hundred cases of UGIB in cirrhotic patients consecutively admitted between 1984 and 1990 (cohort A) were compared with 100 similar cases admitted between 2004 and 2009 (cohort B). RESULTS: The sex ratio (M/F: 2/1), mean age (approximately 55Y) and the proportion of patients with alcoholic cirrhosis (approximately 80%) did not change. Mean Child-Pugh score and the proportion of patients in Child-Pugh stage C increased from 7.6 and 19% in cohort A to 8.8 and 35% in cohort B (p < 0.001). Therapeutic intervention was performed during initial endoscopy in 13 cases from cohort A and 50 from cohort B (p < 0.001), respectively. The number of transfused patients (85 in cohort A, 58 in cohort B) and the number of red blood cell units administered on the first day (median: 4 in cohort A, 2 in cohort B) were significantly decreased in cohort B (p < 0.001). The rate of rebleeding (45 in cohort A, 11 in cohort B), the need for rescue surgery (8 in cohort A, 0 in cohort B) and the in-hospital mortality (24 in cohort A, 9 in cohort B) significantly decreased in the more recent cohort (p < 0.005). CONCLUSION: This study demonstrated that several characteristics of cirrhotic patients admitted with UGIB have changed over the past 2 decades. Above all, outcome has improved despite an increase in the severity of cirrhosis.