Abstract 4: Variation in Published Stroke Rates Results in Wide Variation in the Net Clinical Benefit of Anticoagulation for Atrial Fibrillation

Abstract

Listen

Translate article

Background: The risk of ischemic stroke without anticoagulation for a given CHA 2 DS 2 -VASc score varies significantly across published cohorts. The resulting impact on the net clinical benefit of anticoagulation in atrial fibrillation (AF) is not known. We aimed to determine the effect of variation in published stroke rates on the net clinical benefit of anticoagulation. Methods: A decision analysis study. We used ischemic strokes rates corresponding to CHA 2 DS 2 -VASc scores from four studies - ATRIA study, SPORTIF, Swedish AF study, and the Danish National Patient Registry - in a 28-state Markov state transition model that compared the benefit of oral anticoagulation vs. no anticoagulation. During each monthly cycle, patients face a chance of stroke and hemorrhage, either of which may lead to death, neurologic sequelae, or symptom resolution. The simulation ran for the entire life expectancy of the patient. We applied this decision analytic model to a cohort of 33434 patients with incident AF (Jan 06 to Jun 09) in the Kaiser Permanente Northern and Southern California integrated health care delivery systems. Results: Of the 33434 patients, 27122 had a CHA 2 DS 2 -VASc score ≥2. Among those with CHA 2 DS 2 -VASc score ≥2, use of ATRIA Study stroke rates yielded the lowest population benefit (2576 QALYs, 0.06 expected median per person, IQR -0.02 to 0.18), while the Danish National Patient Registry stroke rates yielded the greatest population benefit (28934 QALYs, 0.70 expected median per person, IQR 0.38 to 1.47). The CHA 2 DS 2 -VASc threshold at or above which the population derives the greatest net clinical benefit from anticoagulation varies based on the stroke rates used: CHA 2 DS 2 -VASc ≥3 using ATRIA Study stroke rates, ≥2 using Swedish AF study stroke rates, ≥1 using SPORTIF and ≥0 using the Danish National Patient Registry stroke rates. Conclusion: Using current guidelines, variation in published off-anticoagulation stroke rates results in tenfold variation in the estimated net clinical benefit of anticoagulation. Guidelines should better reflect the uncertainty of the current approach that uses a CHA 2 DS 2 -VASc threshold to recommend anticoagulation.