Anticoagulation thresholds based on CHA2DS2-VASc risk score varied from population to population, researchers reported in the Annals of Internal Medicine.
After accounting for differing rates of stroke in published studies, the benefit of warfarin anticoagulation varied nearly fourfold, said Sachin J. Shah, MD, of the University of California San Francisco and his associates. They called for guidelines that “better reflect the uncertainty in current thresholds of stroke risk score for recommending anticoagulation.”
Oral anticoagulation markedly reduces risk of ischemic stroke in patients with atrial fibrillation but increases the risk of major bleeding, including intracranial hemorrhage, which often is fatal. Therefore, when deciding whether to recommend oral anticoagulation, physicians must estimate clinical net benefit by quantifying the difference between reduction in stroke risk and increase in major bleeding risk, weighted by the severity of each outcome.
Guidelines on nonvalvular atrial fibrillation from the European Society of Cardiology and joint guidelines from the American Heart Association, American College of Cardiology, and Heart Rhythm Society (AHA/ACC/HRS) recommend oral anticoagulation when CHA2DS2-VASc (congestive heart failure, hypertension, age, diabetes, stroke, and vascular disease) risk score is 2 or greater. These guidelines implicitly assume that a particular CHA2DS2-VASc score denotes the same amount of risk across populations, even though a recent meta-analysis found otherwise, as the researchers noted.
To further test this assumption, they applied an existing Markov model to data from more than 33,000 members of the ATRIA-CVRN cohort. All patients had nonvalvular atrial fibrillation, were members of Kaiser Permanente Northern California, and were diagnosed during 1996-1997. About 81% had a CHA2DS2-VASc score of at least 2. For each patient, the researchers produced four estimates of the net clinical benefit of oral anticoagulation based on ischemic stroke rates from ATRIA, the Swedish AF cohort study, the SPORTIF study, and the Danish National Patient Registry.
Optimal anticoagulation thresholds were a CHA2DS2-VASc score of 3 or more using stroke rates from ATRIA, 2 or more based on Swedish AF rates, 1 or more based on SPORTIF rates, and 0 or more using rates from the Danish National Patient Registry. Oral anticoagulation thresholds were lower but still varied widely after accounting for the lower rates of intracranial hemorrhage associated with non–vitamin K antagonist therapy.
Therefore, current guidelines based on CHA2DS2-VASc score may need revising “in favor of more accurate, individualized assessments of risk for both ischemic stroke and major bleeding,” the investigators 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.”
The study had no primary funding source. Dr. Shah reported having no conflicts of interest. Three coinvestigators disclosed research support from relevant pharmaceutical or device companies.
SOURCE: Shah SJ et al. Ann Intern Med. 2018 Sep 25. doi: 10.7326/M17-2762