By Cara Pellegrini, MD

Assistant Professor of Medicine, University of California, San Francisco, Cardiology Division; Electrophysiology Section, San Francisco VA Medical Center

Dr. Pellegrini reports no financial relationships relevant to this field of study.

SYNOPSIS: In a real-world population of atrial fibrillation patients at moderate to high risk of stroke, more than one-third were treated with aspirin as opposed to an oral anticoagulant. Those with conditions related to coronary heart disease were more likely to be treated with aspirin.

SOURCE: Hsu JC, Maddox TM, Kennedy K, et al. Aspirin instead of oral anticoagulant prescription in atrial fibrillation patients at risk for stroke. J Am Coll Cardiol 2016;67:2913-2923.

Built on multiple randomized, controlled trials, clinical registries, and meta-analyses, guidelines are consistent across societies in recommending oral anticoagulation (OAC) for patients suffering from atrial fibrillation (AF) at moderate to high risk of thromboembolism (i.e., CHA2DS2-VASc score ≥ 2). Yet, for various reasons, clinical practice falls short. Hsu et al examined patient and practice-based factors that may affect this decision in a real-world setting.

They abstracted data from the American College of Cardiology’s Practice Innovation and Clinical Excellence (PINNACLE) registry for patients presenting with AF and a CHADS2 score ≥ 2, or separately, a CHA2DS2-VASc score of ≥ 2. Patients with a documented medical, patient, or system reason that they could not receive aspirin or OAC were excluded from the study, as were patients who were not prescribed either agent. This resulted in cohorts of about 200,000-300,000 patients, each who were treated at 123 outpatient cardiovascular practices across the United States. Cohorts were stratified for the presence or absence of a coronary heart disease (CHD) equivalent condition (coronary artery disease, unstable angina, stable angina, prior myocardial infarction, prior coronary artery bypass graft surgery, or peripheral arterial disease) to further elucidate the effect of another potential indication for antiplatelet therapy. The primary outcome was treatment with aspirin, with or without another antiplatelet agent, vs. OAC therapy.

Remarkably, more than one-third of patients received aspirin instead of OAC: 38.2% in patients with a CHADS2 score ≥ 2 and 40.2% for those with a CHA2DS2-VASc score ≥ 2. After multivariable adjustment, CHD equivalent conditions or the presence of certain CHD risk factors (hypertension and dyslipidemia) were associated with the prescription of aspirin instead of OAC. Conversely, male sex, higher body mass index, prior stroke/transient ischemic attack, prior systemic embolism, and congestive heart failure were associated with more frequent prescription of OAC. There was a fair amount of variation in percentage of OAC prescriptions among sites, with an interquartile range of 56.4% to 69.9%. In addition to the appearance of underutilization of OAC therapy, researchers observed a potential overutilization of aspirin therapy, with 31.2% of patients without a CHD equivalent on OAC also prescribed aspirin. Usage of any thienopyridine (e.g., clopidogrel) was significantly lower in the OAC group. The authors concluded that despite the clear benefit of OAC use in this population, there appears to be a gap in care, most prominent in patients with or at risk for coronary artery disease.


These data are consistent with several older large-scale studies that showed aspirin prescription as high as 28-38% in AF patients at risk for stroke. It is disappointing to see that in the era of the CHADS2 and CHA2DS2-VASc risk stratification systems, and a setting of cardiovascular specific clinics with a focus on quality improvement, no improvement is evident. Non-U.S. data from the Global Anticoagulant Registry in the FIELD (GARFIELD) registry show slightly better performance internationally, with only 25.3% of patients scoring ≥ 2 on the CHADS2 receiving aspirin instead of OAC. Hsu et al extend previous work by defining patients most likely to miss out on the benefit of OAC treatment, and in so doing, suggesting targets for educational intervention.

Clearly there seems to be a concern among some providers to avoid the combination of aspirin and OAC, but the decision to give aspirin and withhold OAC is likely based on erroneous calculations of risk on the part of the providers, the patients, or both. Perhaps the message that aspirin has little if any benefit in risk reduction for stroke and systemic embolism also is not fully appreciated, and providers are hoping to get double the bang for their buck with a single prescription of aspirin. The lower utilization of OAC in women, who have been demonstrated to be at an increased risk of stroke compared to men, is harder to explain. The variance across practices in OAC usage suggests that focused intervention at the practice level may be important to improve OAC utilization. Perhaps greater clarity in the guidelines with regard to the need or lack thereof of aspirin therapy in addition to OAC use and the duration of antiplatelet following an acute coronary syndrome would help both increase OAC prescriptions and decrease unnecessary aspirin use. Such usage not only exposes patients to increased bleeding if used in conjunction with OAC, but seems to serve as a hindrance to the proper prescription of OAC in this population. It is undeniably difficult to convince patients to take on the nuisance, cost, and often visible risk of bleeding, for the abstract long-term decreased chance of a stroke (the stroke that doesn’t happen being invisible), and shortcomings are as much a failure of the system as that of any individual providers or practices.

The most important limitation of this study is that its information source is already somewhat dated. The agents with the lowest bleeding risk, apixaban and edoxaban, were not yet FDA-approved, and there was little penetrance of direct oral anticoagulants (DOACs) in the study population, wherein 90% of patients on OAC were treated with warfarin. How the greater availability and utilization of DOACs has affected OAC usage overall is unclear, but it likely has led to an increase. Conversely, PINNACLE clinic sites were staffed by cardiology specialists and were focused on quality improvement; true national utilization of OAC may be even lower. Bleeding risk (e.g., HAS-BLED scores) could not be captured in this study. However, as the net clinical benefit of warfarin actually increased with the HAS-BLED score, a higher bleeding risk alone should not justify the withholding of OAC. Hopefully these data have encouraged providers to closely examine their own practices, and enriched our understanding of treatment gaps.