THE EFFECT OF A MULTICOMPONENT INTERVENTION ON GASTROINTESTINAL BLEEDING RISK IN PATIENTS PRESCRIBED WARFARIN: THE ANTICOAGULATION WITH ENHANCED GASTROINTESTINAL SAFETY (AEGIS) RANDOMIZED TRIAL

Importance: Many patients who take anticoagulants are prescribed antiplatelet medications unnecessarily and/or fail to receive guideline-recommended gastroprotection with proton pump inhibitors (PPI), which increases the risk of gastrointestinal (GI) bleeding.
Objective: To evaluate the effectiveness of a multicomponent intervention – clinician notification with nurse facilitation (CNNF) – to reduce use of antiplatelet medications without a PPI in patients taking warfarin.
Methods: We conducted a cluster-randomized trial comparing CNNF to a wait-list control (usual care) in a single center. Patients were eligible if they were prescribed warfarin and an antiplatelet medication without a PPI and managed through the anticoagulation service. In CNNF, nurses sent electronic messages to clinicians of patients identified as high risk for upper GI bleeding, recommended consideration of either antiplatelet discontinuation or initiation of a PPI, provided a link to an evidence summary on appropriate antiplatelet prescribing, and offered to facilitate any recommended medication changes. Messages were preferentially directed to patients’ cardiologists and secondarily primary care providers. The primary outcome was proportion of patients who either discontinued antiplatelet therapy or initiated a PPI based on self-report after 7-10 weeks. The secondary outcome was proportion of patients with a documented recommendation by a clinician to make such a medication change by 7 weeks. Multiple imputation was used for patients with missing outcome data. A per-protocol analysis was done excluding patients lost to follow up or who were misclassified as high risk for upper GI bleeding at study entry.
Results: Equal numbers of patients were randomized to CNNF and wait-list control (110 each). The most common antithrombotic regimen was single-agent aspirin, in addition to warfarin (93%). For patients receiving CNNF, 50% had outreach to a cardiologist, 28% to a PCP, and 22% to other specialists. Intention-to-treat analysis demonstrated CNNF was associated with increased odds of discontinuing antiplatelet therapy or initiating PPI (adjusted odds ratio [aOR] 5.76, 95% CI 2.54, 13.0). The effect was stronger in per-protocol analysis (n=126, aOR 43.6, 95% CI 6.56, 290), in which 19/59 (33.2%) patients discontinued antiplatelet therapy and 13/59 (22.0%) initiated PPI with CNNF vs. 1/67 (1.5%) each in the wait-list control group. The intervention was also associated with increased odds of a recommendation to make a medication change (75/110 [68.2%] vs. 1/110 [0.9%], OR 19.86, 95% CI 10.63, 29.0). Surgeons and proceduralists were less likely to have recommended medication changes for their patients relative to other clinicians (OR <0.01).
Discussion: The multicomponent intervention effectively reduced gastrointestinal bleeding risk in patients prescribed warfarin.