Bridging anticoagulation is commonly prescribed to patients with atrial fibrillation who are initiating therapy or require interruption of anticoagulation for procedures. Current guidelines recommend bridging for patients at high risk of stroke, but no data guide this recommendation. Among patients with atrial fibrillation and one or more risk factors for thromboembolic stroke, the recently published BRIDGE trial found forgoing bridging to be, on average, noninferior to bridging with respect to thromboembolic complications, with significantly fewer hemorrhagic complications. Depending on underlying risk, some subgroups may derive clinically meaningful benefit, but the cost-effectiveness of bridging among patients who derive theoretical benefit has not been examined.


We performed a Monte Carlo simulation to predict the cost-effectiveness of bridging anticoagulation among patients with non-valvular atrial fibrillation, across the spectrum of thromboembolic and hemorrhagic risk. We used literature-derived estimates of the underlying risks of ischemic stroke and intracranial hemorrhage, along with the changes in risk associated with warfarin alone and warfarin with bridging. Simulations were performed from a third-party payer perspective, with multi-way sensitivity analyses.


When stratified by CHADS2 and HAS-BLED scores, outpatient bridging anticoagulation reached commonly used cost-effectiveness thresholds in small proportions of simulated patients at high risk of stroke and low risk of hemorrhage. Hospital admission for bridging rarely meets commonly accepted cost-effectiveness thresholds. Results are sensitive to the costs of long-term care.


While carefully selected patients with non-valvular atrial fibrillation may benefit from bridging anticoagulation in initiation or interruption of warfarin, bridging that requires hospitalization is unlikely to be cost-effective from the perspective of a third-party payer. Outpatient bridging may be reasonable after stratification along thromboembolic and hemorrhagic risk and careful patient selection. From a third-party payer perspective, hospitalization specifically for bridging is inefficient.