A pragmatic assessment of proton pump inhibitors vs. histamine type 2 receptor antagonists on clinically important gastrointestinal bleeding and mortality when used for stress ulcer prophylaxis in the ICU.

Abstract

Approximately 1%-5% of critically ill patients experience clinically important gastrointestinal bleeding (CIGB). This study assessed the effectiveness and safety of proton pump inhibitors (PPIs) compared to histamine type 2 receptor antagonists (H2RAs) for prevention of CIGB in mechanically ventilated patients. This is a retrospective, single-center, pharmacoepidemiologic study. This study was carried out in six intensive care units (ICUs). Critically ill adults admitted between 9/1/14 and 9/1/19 who received PPIs or H2RAs within 24h of intubation and for ≥48h were included in this study. PPIs or H2RAs for stress ulcer prophylaxis. Primary outcomes were CIGB occurring 48h after ICU admission and hospital mortality. Secondary outcomes were pneumonia, Clostridioides difficile infection (CDI), acute kidney injury, myocardial infarction/ischemia, thrombocytopenia, and delirium. Outcomes were defined using International Classification of Diseases, Ninth Revision (ICD-9) and Tenth Revision (ICD-10)-codes with manual cross-reference for a hemoglobin drop, transfusion, or hemodynamic compromise to further define CIGB. Of 3873 patients, 2061 (53.2%) received PPIs. CIGB was rare but higher in the PPI group (0.34% vs. 0%, RR=1, 95% CI, 1-1; p=0.013); however, substantial group differences existed possibly predisposing the PPI group to CIGB. Hospital mortality was higher in the PPI group (42.1% vs. 29.1%, RR=1.23, 95% CI, 1.17-1.29; p<0.0001). PPIs remained an independent risk factor for mortality after multivariate adjustment (RR=1.61, 95% CI, 1.39-1.88; p<0.0001). Rates of secondary outcomes were similar between groups except thrombocytopenia (4.3% vs. 2.2%, RR=1.02, 95% CI, 1.01-1.03; p=0.0003) and delirium (83.7% vs. 78.1%, RR=1.34, 95% CI, 1.18-1.53; p<0.0001) that were higher in the PPI group. Proton pump inhibitors were associated with CIGB; however, the overall rate of CIGB was low. Compared to H2RAs, PPIs were associated with hospital mortality. Further identification of appropriate selection criteria for ulcer prophylaxis and comparisons of pharmacologic prevention strategies are warranted.