Grant: Project Grant

Patients receiving care in Intensive Care (ICU) are at risk for developing stomach or duodenal (small intestine) "stress" ulcers. Stress ulcers occur commonly when life-support (a breathing machine) is required or when a patient develops a bleeding tendency due to their illness. Stress ulcers may cause serious bleeding. Given stress ulcers are common and associated with adverse outcome, most patients are routinely given one of two types of drugs, a 'proton pump inhibitor' (PPI) or a 'histamine-2 receptor blocker' (H2RB) to prevent them. While prevention is vital, these drugs can have adverse effects, including risk of certain infections. This study will establish which of these routinely provided drugs to prevent stress ulcers leads to lower risk of serious bleeding, prolonged ventilation, C. difficile infection and hospital death. The study is a 'randomized cluster, crossover registry' trial comparing two standard of care drugs used for the same purpose. In this study, all data is already routinely collected via existing care systems. There will be two phases. In phase one, half the ICUs will be randomly (like a coin toss) assigned to use PPIs to prevent ulcers while the other half will use H2RBs. In phase two, each ICU will switch to the opposite treatment. For any given patient, if there is a specific indication for either PPI or H2RB, the indicated treatment will be given irrespective of treatment assigned to that ICU at the time. In other words, patients will always receive the treatment that clinicians believe to be in their best interests. In addition to the direct benefit of comparing ulcer treatments, the study utilizes an innovative and pragmatic design that may transform how future trials are performed in ICU. Such a design using existing data sources, imposing no added health record access, could provide opportunity to efficiently detect small, but clinically important treatment effects at comparatively lower cost than conventional randomized trials.