RESUMO
OBJECTIVE: To assess the occurrence, clinical predictors, and associated mortality of all-cause emergency readmissions after acute upper gastrointestinal bleeding (AUGIB). PATIENTS AND METHODS: All patients with AUGIB from an area of 600 000 inhabitants in Sweden admitted in a single institution in 2009-2011 were retrospectively identified. All medical records were scrutinized and relevant data (such as comorbid illness and medications, endoscopy, rebleeding, inhospital mortality, and 30-day emergency readmission) were extracted. The Charlson comorbidity index was calculated. RESULTS: A total of 174 out of 1056 patients discharged alive following AUGIB (16.5%) had an emergency readmission within 30 days. Nineteen percent of readmissions were because of rebleeding, whereas the rest were because of other reasons, mainly bacterial infections (9.8%) and cardiovascular events (8%). Inhospital mortality did not differ significantly between index admissions and readmissions (13.7 vs. 9.8%, P=0.181). In logistic regression analysis, only a higher Charlson comorbidity index [odds ratio (OR): 1.154, 95% confidence interval (CI): 1.056-1.261] was related to emergency readmission. Bisphosphonate use (OR: 3.933, 95% CI: 1.264-12.233), previous AUGIB (OR: 2.407, 95% CI: 1.157-5.009), and length of stay at index admission (>5 days; OR: 0.246, 95% CI: 0.093-0.649) were found to be independent predictors of postdischarge rebleeding. CONCLUSION: All-cause emergency readmission following AUGIB is frequent. It is related to rebleeding in one-fifth of cases and mortality is similar to that in index admissions. The presence of comorbid illness appears to predict readmissions. Reduced length of stay and bisphosphonate use appear to be important, potentially modifiable, predictors of postdischarge rebleeding.
