Benefit of anticoagulation for stroke prevention may vary in patients with afib, study finds

An accompanying editorial said the CHA2DS2-VASc cutoff should be considered a starting point for individualized assessment rather than a threshold that applies the same way to everyone.


Guidelines for prescribing anticoagulants for stroke prevention in atrial fibrillation may not adequately reflect the uncertainty of stroke risk scores, a recent study found.

Researchers performed a Markov model decision analysis in 33,434 community-dwelling older adults to determine how variation in published stroke rates affects the net expected clinical benefit of anticoagulation in patients with incident atrial fibrillation. The current guideline-recommended threshold for oral anticoagulation is a CHA2DS2-VASc score of 2 of higher. Warfarin was used for the base case, and a secondary analysis modeled the use of non-vitamin K antagonist oral anticoagulants (NOACs). Differing stroke rate estimates were taken from the ATRIA (AnTicoagulation and Risk Factors in Atrial Fibrillation) study, the Danish National Patient Registry, the Swedish AF cohort study, and the SPORTIF (Stroke Prevention using Oral Thrombin Inhibitor in atrial Fibrillation study). Estimated gains in quality-adjusted life-years (QALYs) were compared. The study results were published Sept. 25 by Annals of Internal Medicine.

The study cohort was made up of 33,434 adults from the ATRIA-CVRN (AnTicoagulation and Risk Factors in Atrial Fibrillation-Cardiovascular Research Network), 45% of whom were ages 75 years or older at cohort entry and 45% of whom were women. Overall, 81% had a CHA2DS2-VASc score of 2 of higher when they were diagnosed. The 27,179 patients who had a CHA2DS2-VASc score of 2 or higher had the least population benefit with warfarin anticoagulation when stroke rates from the ATRIA (AnTicoagulation and Risk Factors in Atrial Fibrillation) study were used and the most benefit when those from the Danish National Patient Registry were used (6,290 QALYs vs. 24,1110 QALYs; P<0.001). Population benefit was 13,230 QALYs and 16,710 QALYs when stroke rates came from the SPORTIF and Swedish AF studies, respectively. Median benefit per person with oral anticoagulation also varied according to the stroke rate used.

Optimal CHA2DS2-VASc scores for anticoagulation yielding the greatest population benefit were 3 or higher with stroke rates from ATRIA (7,355 QALYs gained), 2 or higher with stroke rates from the Swedish AF cohort study (16,710 QALYs gained), 1 or higher with stroke rates from the SPORTIF study (17,740 QALYs gained), and 0 or higher with stroke rates from the Danish National Patient Registry (34,770 QALYs gained). The researchers calculated the chance of individual benefit among the 27,179 patients with a CHA2DS2-VASc score of 2 or higher and found that 16,182 (60%) would not benefit from oral anticoagulation based on ATRIA stroke rates, 5,319 (20%) would not benefit with SPORTIF stroke rates, 4,498 (17%) would not benefit with Swedish AF rates, and 776 (3%) would not benefit with Danish National Patient Registry rates. Optimal score thresholds decreased when the authors accounted for lower rates of intracranial hemorrhage linked to NOAC use versus warfarin but continued to vary widely.

The authors noted that the benefits measured in their study may not be generalizable to other populations, among other limitations, but said that their findings highlight the importance of accurate, precise estimates of ischemic stroke risk in patients with atrial fibrillation, as well as the need for standardized methods to obtain such estimates. “Our findings also indicate that the current guidelines based on CHA2DS2-VASc score may need to be revised in favor of more accurate, individualized assessments of risk for both ischemic stroke and major bleeding,” the authors wrote. “Until such time, guidelines should better reflect the uncertainty of the current approach in which a patient's CHA2DS2-VASc score is used as the primary basis for recommending [oral anticoagulation].”

An accompanying editorial agreed that the study calls the optimal CHA2DS2-VASc cutoff into question for given populations or given patients and said that it should be considered a starting point for individualized assessment rather than a threshold that applies the same way to everyone. The editorial noted that the CHA2DS2-VASc score is still the leading approach to assessment of atrial fibrillation risk and anticoagulation benefit but suggested it can be improved. “When it comes to the conversation about the risks and benefits of anticoagulation for our patients with [atrial fibrillation], we must remember that each patient is an individual and has his or her own ‘score,’” the editorialists wrote.