ABSTRACT
Background
Patients with atrial fibrillation (AF) have a high rate of all-cause mortality that is only partially attributable to vascular outcomes. While the competing risk of death may affect expected anticoagulant benefit, guidelines do not account for it. We sought to determine if using a competing risks framework materially affects the guideline-endorsed estimate of absolute risk reduction attributable to anticoagulants.
Methods
We conducted a secondary analysis of 12 RCTs that randomized patients with AF to oral anticoagulants or either placebo or antiplatelets. For each participant, we estimated the absolute risk reduction (ARR) of anticoagulants to prevent stroke or systemic embolism using two methods. First, we estimated the ARR using a guideline-endorsed model (CHA 2 DS 2 -VASc) and then again using a Competing Risk Model that uses the same inputs as CHA 2 DS 2 -VASc but accounts for the competing risk of death and allows for non-linear growth in benefit over time. We compared the absolute and relative differences in estimated benefit and whether the differences in estimated benefit varied by life expectancy.
Results
7933 participants had a median life expectancy of 8 years (IQR 6, 12), determined by comorbidity-adjusted life tables. 43% were randomized to oral anticoagulation (median age 73 years, 36% women). The guideline-endorsed CHA 2 DS 2 -VASc model estimated a larger ARR than the Competing Risk Model (median ARR at 3 years, 6.9% vs. 5.2%). ARR differences varied by life expectancies: for those with life expectancies in the highest decile, 3-year ARR difference (CHA 2 DS 2 -VASc model – Competing Risk Model 3-year risk) was −1.2% (42% relative underestimation); for those with life expectancies in the lowest decile, 3-year ARR difference was 5.9% (91% relative overestimation).
Conclusion
Anticoagulants were exceptionally effective at reduced stroke risk. However, anticoagulant benefits were misestimated with CHA 2 DS 2 -VASc, which does not account for the competing risk of death nor decelerating treatment benefit over time. Overestimation was most pronounced in patients with the lowest life expectancy and when benefit was estimated over a multi-year horizon.
Full Text Availability
The license terms selected by the author(s) for this preprint version do not permit archiving in PMC. The full text is available from the preprint server.