Estimating the Stroke Risk Threshold for Initiating Non-Vitamin K Antagonist Oral Anticoagulation in Atrial Fibrillation: Markov Decision Model Analysis.

backgroundRandomized trials have clearly demonstrated the benefits of anticoagulant therapy in patients with atrial fibrillation who are at high risk of ischemic stroke. However, less is known about the benefit of anticoagulation in low-risk patients, and exactly how low baseline stroke risk justifies further attempts to reduce it with direct oral anticoagulants (DOACs) remains unclear.

methodsWe developed a Markov decision model to estimate the impact of initiating DOACs on quality-adjusted life years (QALYs) on a 20-year time horizon in patients with atrial fibrillation across a range of nonanticoagulated ischemic stroke risk. The model incorporated data from randomized controlled trials on the effects of DOACs on the severity and risk of ischemic stroke, major bleeding, and mortality, as well as previous evidence on their impact on quality of life. Nonanticoagulated event rates were averaged from previous observational studies.

resultsThe tipping point in the annual nonanticoagulated ischemic stroke rate, at which DOAC treatment resulted in equal cumulative QALYs as withholding therapy, was 0.65%. Below this risk threshold, DOAC therapy yielded slightly fewer QALYs, while, above it, DOAC therapy resulted in increasingly higher QALYs. At nonanticoagulated stroke risk levels of 1%, 2%, and 3%, the mean QALY gains with DOACs per patient during a 20-year simulation were 0.13, 0.53, and 1.00, respectively, whereas, at the stroke risk level of 0.4%, DOAC therapy resulted in 0.01 lower QALYs per patient.

conclusionsIn this simulation, DOAC therapy versus no anticoagulation was associated with a net benefit on QALYs in patients with atrial fibrillation with an annual nonanticoagulated stroke risk >0.65%, with the magnitude of benefit increasing with higher stroke risk.