Abstract Rationale Critically ill patients at risk of gastrointestinal bleeding receive either a proton pump inhibitor (PPI) or histamine-2 receptor antagonist (H2RA) for stress ulcer prophylaxis (SUP). Venovenous extracorporeal membrane oxygenation (VV-ECMO) is associated with an increased risk of bleeding events, including gastrointestinal bleeding. However, there are no guidelines for SUP in patients supported with VV-ECMO. We examined the association between the use of either a PPI or H2RA as initial stress ulcer prophylaxis in patients receiving VV-ECMO support and several clinical outcomes including in-hospital mortality, upper gastrointestinal bleed (UGIB), and Clostridium difficile infection (CDI). Methods This was a single-center retrospective cohort study of patients supported with VV-ECMO in the Duke University Hospital (DUH) medical intensive care unit (MICU) from 2013-2023. The study was approved by Duke Health Institutional Review Board. Patients that received SUP therapy with either a PPI or H2RA within 48 hours of initiating VV-ECMO in the study ICU were included. Patients that were exposed to both PPI and H2RA on more than 1 day within 96 hours of cannulation, transferred to the study ICU 24 hours after cannulation, died within 48 hours of cannulation, or had a confirmed UGIB at time of cannulation were excluded. Patient characteristics and clinical outcomes were obtained through manual review of the electronic health record. The primary outcome was in-hospital mortality. Secondary outcomes were confirmed UGIB and CDI. Outcomes were compared using an odds ratio (OR). Results During the study period, 367 patients received VV-ECMO support in the study ICU, of whom 302 met inclusion criteria. Of these 302 patients, 175 (57.9%) received a PPI and 127 (42.1%) received a H2RA for SUP. In-hospital mortality was 37.7% in the PPI group and 41.7% in the H2RA group (OR 0.85, 95% confidence interval CI 0.54-1.33). Clinically important UGIB occurred in 5.1% of patients in the PPI group and 4.7% in the H2RA group (OR 1.09, 95% CI 0.40, 3.15). Acquired CDI during hospitalization occurred in 4.0% in the PPI group and 2.4% in the H2RA group (OR 1.72, 95% CI 0.47, 6.22). Figure 1 Conclusion There were no significant differences in clinical outcomes of in-hospital mortality, clinically important UGIB, or acquired CDI between the PPI and H2RA groups. These findings are consistent with current guidelines that recommend either PPI or H2RA for stress ulcer prophylaxis for critically ill patients at risk of gastrointestinal bleeding, including those supported with VV-ECMO. This abstract is funded by: Duke Department of Medicine Faculty Resident Research Grant
Watson et al. (Fri,) studied this question.