Number needed to treat for net effect of anticoagulation in atrial fibrillation: Real-world vs. clinical-trial evidence

Aims The net benefit of oral anticoagulants (OACs) in atrial fibrillation (AF) is poorly understood. We aimed to determine the NNT for net effect (NNTnet) using calculator of absolute stroke risk (CARS) in anticoagulated patients with AF in real-world and clinical trial cohorts. Methods Post-hoc analysis of patient-level data from the real-world Murcia AF Project and AMADEUS clinical trial. Baseline risk of stroke was determined using CARS. The risk of stroke and major bleeding events with OAC were determined using the number of respective events at 1-year. NNTnet was calculated as a reciprocal of the net effect of absolute risk reduction with OAC (NNTnet = 1/(absolute risk reduction of stroke[ARR(stroke)] - absolute risk increase of major bleeding[ARI(bleeding)])). Results In total, 3511 patients were included (1306 [37.2%] real-world patients and 2205 [62.8%] clinical trial participants). The absolute 1-year stroke risk was similar across both cohorts. In the real-world cohort, OAC was associated with a 4.0% ARR(stroke), 25 NNTbenefit, 1.0% ARI(bleeding), 100 NNTharm and 34 NNTnet. In the clinical trial cohort, OAC was associated with a 3.8% ARR(stroke), 27 NNTbenefit, 1.6% ARI(bleeding), 63 NNTharm and 46 NNTnet. In both cohorts, the NNTnet was significantly lower in patients with an excess stroke risk of >= 2% by CARS. Conclusion Overall, the NNTnet approach in AF incorporates information regarding baseline risk of stroke and major bleeding, and relative effects of OAC with the potential to include multiple additional outcomes and weighting of events based on their perceived effects by individual patients.