Non-variceal upper GI bleeding in patients already hospitalized for another condition.

Objectives: To compare outpatients (OPs) presenting with non-variceal upper gastrointestinal bleeding (NVUGIB) to those who started hemorrhaging while in a hospital (inpatients, IPs) in a contemporary setting and to better identify predictors of outcome.

Methods: Retrospective data from the Canadian Registry of Patients With Upper Gastrointestinal Bleeding Undergoing Endoscopy (RUGBE). Descriptive, inferential, and multivariate logistic regression models were carried out in 469 IPs (68.5+/-14 years, 36% women) and 1,395 OPs (65.5+/-18 years, 39% women) in 18 Canadian community and tertiary care centers.

Results: Main outcomes were rebleeding, mortality, and their predictors. IPs differed from OPs in disease acuity (P=0.02) and comorbidities (3.1+/-1.7 vs. 2.3+/-1.5, P<0.001), and were admitted longer (7.2+/-7.4 vs. 5+/-5.4 days, P<0.001) and more often to intensive care unit (ICU; 40.5% vs. 16%, P<0.001). Ulcers or erosions predominated (83% vs. 85%, P=0.28), treated by endotherapy (38% vs. 36%, P=0.46). More IPs received proton pump inhibitors (PPIs; 88% vs. 83%, P=0.009). Mortality was greater for IPs (11% vs. 3.5%, P<0.001), but rebleeding (15.7% vs. 13.4%, P=0.23) and surgery (6.9% vs. 6.4%, P=0.72) were not. Among IPs, comorbidity (odds ratio, OR=1.15; 95% confidence interval, CI: 1.01-1.32) and endoscopic high-risk stigmata increased (OR=3.86, 95% CI:2.05-7.26), whereas PPI decreased (OR=0.20, 95% CI:0.10-0.42) rebleeding; high-risk stigmata (OR=3.13, 95% CI:1.23-7.99) and rebleeding (OR=4.19, 95% CI:2.06-8.55) increased mortality, whereas low disease acuity was protective (OR=0.20; 95% CI:0.46-0.90).

Conclusions: IPs are sicker than OPs. Endoscopic hemostasis and PPI therapy favorably affect rebleeding in IPs, whereas patient characteristics principally determine the threefold greater IPs mortality.