HOME
Return to Table of Contents
151 RISK FACTORS FOR GASTROINTESTINAL BLEEDING IN THE HOSPITALIZED PATIENT J Wyse, A Barkun, C Baldry, S Dial BACKGROUND: The incidence and risk factors for hospital acquired upper gastrointestinal bleeding (HUGIB) are unknown. Most of the research in this area has been limited to the critical care setting.
McGill University Health Centre, McGill University, Montreal, Quebec
OBJECTIVE: To determine the incidence and risk factors for significant HUGIB.
METHODS: We assembled a cohort of all adult patients admitted for at least two days from January 1st - December 31st, 2003, to any of the four McGill University Health Centre hospitals using a merged administrative and pharmacy database. Cases were identified based on having been prescribed an intravenous pantoprazole bolus and infusion on day two or later of their hospital admission. Patients admitted with an UGIB within 30 days of a hospital discharge were also defined as HUGIB. Random charts (10% of the cases) were reviewed to confirm correct patient selection. Multivariate analyses were carried out to identify independent predictors of HUGIB.
RESULTS: The incidence of HUGIB was less than 1% (209 of 21714). The mortality rate amongst these patients was 27%. On univariate analysis: older age, longer hospital stay, and an ICU admission, were strongly associated with an increased risk of HUGIB. Certain co-morbidities (diabetes, cirrhosis, congestive heart failure, chronic renal failure, HIV, undergoing cancer therapy and being a transplant recipient) were also predictive of HUGIB. Medication exposures that were significant risk factors included vasopressors, ASA, corticosteroids, coumadin, and anti-platelet agents, but not NSAIDS. On multivariate analysis the following risk factors remained statistically significant: older age, ICU admission, diabetes, cirrhosis, chronic renal failure, undergoing cancer therapy and HIV. The only medication exposures that remained significant were vasopressors. There did not appear to be a protective effect of proton pump inhibitors or H2 receptor antagonists on multivariate analysis.
CONCLUSION: HUGIB is rare and strongly associated with certain clinical co-morbidities. The only medication exposure associated with an increased risk of HUGIB was vasopressor use. The lack of an apparent protective effect of PPI exposure was surprising, and may reflect a lack of power or that the underlying pathobiology of HUGIB is not strictly related to acid suppression.