Abstract
Background Upper gastrointestinal bleeding (UGIB) is a common medical emergency with significant
mortality. Despite developments in endoscopic and clinical management, only minor
improvements in outcomes have been reported.
Methods This was a retrospective cohort study of patients with non-malignant UGIB emergency
admissions in England between 2003 and 2015, using Hospital Episode Statistics. Multilevel
logistic regression analysis examined the associations with mortality.
Results 242 796 patients with an UGIB admission were identified (58.8 % men; median age 70
[interquartile range (IQR) 53 – 81]). Between 2003 and 2015, falls occurred in both
30-day mortality (7.5 % to 7.0 %; P < 0.001) and age-standardized mortality (odds ratio (OR) 0.74, 95 % confidence interval
[CI] 0.69 – 0.80; P < 0.001), including from variceal bleeding (OR 0.63, 95 %CI 0.45 – 0.87; P < 0.005). Increasing co-morbidity (Charlson score > 5, OR 2.94, 95 %CI 2.85 – 3.04;
P < 0.001), older age (> 83 years, OR 6.50, 95 %CI 6.09 – 6.94; P < 0.001), variceal bleeding (OR 2.03, 95 %CI 1.89 – 2.18; P < 0.001), and a weekend admission (Sunday, OR 1.18, 95 %CI 1.12 – 1.23; P < 0.001) were associated with 30-day mortality. Of deaths at 30 days, 8.9 % were
from ischemic heart disease (IHD) and the cardiovascular age-standardized mortality
rate following UGIB was high (IHD deaths within 1 year, 1188.4 [95 %CI 1036.8 – 1353.8]
per 100 000 men in 2003).
Conclusions Between 2003 and 2015, 30-day mortality among emergency admissions with non-malignant
UGIB fell by 0.5 % to 7.0 %. Mortality was higher among UGIB admissions at the weekend,
with important implications for service provision. Patients with UGIB had a much greater
risk of subsequently dying from cardiovascular disease and addressing this risk is
a key management step in UGIB.
