OBJECTIVES: Current drug-drug interaction (DDI) warnings lack details about the risk of harm and use general labels like “major” interaction. This is not useful to clinicians or patients. The purpose of this study was to develop a model that estimate risk of gastrointestinal (GI) bleeding for patients on oral anticoagulants (OAC) when taking non-steroid anti-inflammatory (NSAIDs) and/or antidepressants.

METHODS: A comprehensive literature research was conducted in PubMed to identify risk of GI bleeding based on age, previous history of GI bleeding, OAC (warfarin, apixaban, rivaroxaban, dabigatran and edoxaban), NSAIDs, antidepressants, glucocorticoids, antiplatelets, aspirin and proton pump inhibitors (PPI). Bleeding risk was estimated by incorporating evidence across multiple studies. The model was examined for face validity by estimating GI bleeding based on variety of patient scenarios.

RESULTS: The highest risk for GI bleeding among the OACs was edoxaban (Odds Ratio [OR] =3.3), with apixaban having the lowest GI bleeding risk among the OACs (OR=1.2). A history of previous GI bleeding was associated with the highest risk (OR=6.7). Among the NSAIDs, the highest risk was associated with ketorolac (OR=21.6). Among antidepressants, selective serotonin reuptake inhibitors had a higher GI bleeding risk (e.g. sertraline OR=1.4) than other antidepressants (e.g. mirtazapine OR=1.1). One example of face validity was that the estimated GI bleeding risk for a patient of 81 years of age with history of previous GI bleeding on dabigatran and mirtazapine had a risk of GI bleeding 33% that, when adding a PPI to his treatment plan the risk decreases to 25% and, if adding a diclofenac, it increases the GI bleeding risk up to 57%.

CONCLUSIONS: The OAC GI bleeding risk calculator incorporates patient-level risk factors to estimate bleeding risk. The model allows clinicians to enhance share decision making when visiting patients on OAC that are receiving NSAIDs and/or antidepressants.