Author + information
- Received December 26, 2015
- Revision received March 22, 2016
- Accepted March 29, 2016
- Published online June 28, 2016.
- Jonathan C. Hsu, MD, MASa,∗ (, )
- Thomas M. Maddox, MD, MScb,
- Kevin Kennedy, MSc,
- David F. Katz, MDd,
- Lucas N. Marzec, MDd,
- Steven A. Lubitz, MD, MPHe,
- Anil K. Gehi, MDf,
- Mintu P. Turakhia, MD, MASg and
- Gregory M. Marcus, MD, MASh
- aCardiac Electrophysiology Section, Division of Cardiology, Department of Medicine, University of California, San Diego, La Jolla, California
- bVA Eastern Colorado Health Care System/University of Colorado School of Medicine, Denver, Colorado
- cMid America Heart Institute, Kansas City, Missouri
- dUniversity of Colorado School of Medicine, Denver, Colorado
- eCardiac Arrhythmia Service and Cardiovascular Research Center, Massachusetts General Hospital, Boston, Massachusetts
- fUniversity of North Carolina, Chapel Hill, North Carolina
- gVA Palo Alto Health Care System/Stanford University School of Medicine, Palo Alto, California
- hSection of Cardiac Electrophysiology, Division of Cardiology, Department of Medicine, University of California, San Francisco, San Francisco, California
- ↵∗Reprint requests and correspondence:
Dr. Jonathan C. Hsu, Cardiac Electrophysiology Section, Division of Cardiology, Department of Medicine, University of California-San Diego, 9444 Medical Center Drive, MC7411, La Jolla, California 92037.
Background Oral anticoagulation (OAC), rather than aspirin, is recommended in patients with atrial fibrillation (AF) at moderate to high risk of stroke.
Objectives This study sought to examine patient and practice-level factors associated with prescription of aspirin alone compared with OAC in AF patients at intermediate to high stroke risk in real-world cardiology practices.
Methods The authors identified 2 cohorts of outpatients with AF and intermediate to high thromboembolic risk (CHADS2 score ≥2 and CHA2DS2-VASc ≥2) enrolled in the American College of Cardiology PINNACLE (Practice Innovation and Clinical Excellence) registry between 2008 and 2012. Using hierarchical modified Poisson regression models adjusted for patient and practice characteristics, the authors examined the prevalence and predictors of aspirin alone versus OAC prescription in AF patients at risk for stroke.
Results Of 210,380 identified patients with CHADS2 score ≥2 on antithrombotic therapy, 80,371 (38.2%) were treated with aspirin alone, and 130,009 (61.8%) were treated with warfarin or non-vitamin K antagonist OACs. In the cohort of 294,642 patients with CHA2DS2-VASc ≥2, 118,398 (40.2%) were treated with aspirin alone, and 176,244 (59.8%) were treated with warfarin or non-vitamin K antagonist OACs. After multivariable adjustment, hypertension, dyslipidemia, coronary artery disease, prior myocardial infarction, unstable and stable angina, recent coronary artery bypass graft, and peripheral arterial disease were associated with prescription of aspirin only, whereas 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.
Conclusions In a large, real-world cardiac outpatient population of AF patients with a moderate to high risk of stroke, more than 1 in 3 were treated with aspirin alone without OAC. Specific patient characteristics predicted prescription of aspirin therapy over OAC.
Atrial fibrillation (AF) is the most common cardiac arrhythmia worldwide and imparts significant stroke risk (1). Risk stratification schemes, including the CHADS2 score (2) and the more recent CHA2DS2-VASc score (3), have been developed to estimate the risk of thromboembolism in AF patients on the basis of specific risk factors (2,3). In patients with AF determined to be at intermediate to high risk for thromboembolism, anticoagulation with warfarin (a vitamin K antagonist) or the newer non-vitamin K antagonist oral anticoagulant agents (OACs) reduces morbidity and mortality (4,5). Previous studies have demonstrated that aspirin therapy is not as beneficial as OAC for reduction of thromboembolism (6), yet it is well known that appropriate OAC prescription in AF patients at risk for stroke outside clinical trial settings falls short of guideline-based expectations (7). The extent to which AF patients at risk for stroke are prescribed only aspirin and the factors that determine such a practice are not well known. Additionally, the prevalence of concomitant antiplatelet agent prescription in contemporary patients with AF who are and are not prescribed OAC has not been well studied.
We sought to investigate the prevalence and predictors of treatment with aspirin only compared with OAC therapy by cardiovascular specialists in a cohort of outpatients with AF at a moderate to high risk for thromboembolism using data from the National Cardiovascular Data Registry (NCDR)’s Practice Innovation and Clinical Excellence (PINNACLE). Use of this prospective national registry of cardiovascular care in the United States provides a unique opportunity to examine patterns of both aspirin and OAC prescription in routine practice among outpatients with AF, as well as the use of other antiplatelet agents (i.e., thienopyridine prescription).
The NCDR PINNACLE registry was created in 2008 by the American College of Cardiology as the first national, prospective, office-based cardiac quality-improvement registry in the United States (8). Participating academic and private practices collect longitudinal, point-of-care data that includes patient demographics, symptoms, comorbidities, vital signs, medications, laboratory values, and recent hospitalizations with either paper forms, or modification of a practice’s electronic medical record using a standardized collection tool to comprehensively obtain and transmit uniform data. NCDR registry data quality assurance is maintained through standardized data collection and transmission protocols, rigorous data definitions, and periodic data quality audits, which have shown much >90% raw accuracy of data abstraction (9). Quality checks and analyses of the data have been performed at St. Luke’s Mid America Heart Institute (Kansas City, Missouri), the primary analytical center for the PINNACLE registry.
There were a total of 1,147,227 patients enrolled into the PINNACLE registry between January 1, 2008, and December 30, 2012. We excluded patients without AF (n = 689,722 excluded), patients deemed not able to be prescribed aspirin or OAC therapy (n = 17,627 excluded), as assessed by the treating clinician and specified on data collection forms (due to medical, patient, or system reason), and those not prescribed any antithrombotic therapy at all (n = 112,222 excluded). We further restricted the cohort to patients known to be at moderate to high risk for thromboembolism (CHADS2 score ≥2, with 1 point for congestive heart failure, hypertension, age ≥75 years, and diabetes, and 2 points for stroke/transient ischemic attack [TIA] ; n = 117,276 patients excluded). Exclusion criteria for our primary cohort (referred to as the CHADS2 score ≥2 cohort) were on the basis of the American College of Cardiology/American Heart Association/European Society of Cardiology guidelines in place during the study timeframe, which defined as a Class I indication (i.e., recommendation that the treatment is useful/effective) the use of OAC instead of aspirin alone for the primary prevention of stroke in at-risk AF patients with a CHADS2 score ≥2 (10). Therefore, our final primary study cohort was composed of 210,380 patients with AF from 123 practices in 38 states across the United States (Figure 1). Of the 123 practices included in the PINNACLE registry, 97 (78.9%) practices were located in an urban/suburban setting. The regional breakdown of practice locations included the following: 27 (22%) in the West; 17 (13.8%) in the Northeast; 33 (26.8%) in the Midwest; and 46 (37.4%) in the South. We also conducted a secondary analysis, using the CHA2DS2-VASc score, a more sensitive tool to risk-stratify AF patients who may be at risk for stroke and benefit from anticoagulant therapy (3); use of this risk score may have influenced cardiovascular specialist prescription of OAC during the study timeframe, as reflected in updated European Society of Cardiology guidelines from 2012 (11) and subsequently published updated guidelines after the study timeframe (12). These updated guidelines, published in 2014, advise the use of the CHA2DS2-VASc score for the assessment of stroke risk, and state as a Class I recommendation the use of OAC therapy for patients with nonvalvular AF and a CHA2DS2-VASc score ≥2 (12). In addition to carrying over all of the same exclusion criteria listed previously, this cohort was restricted to AF patients with a CHA2DS2-VASc score ≥2, with 1 point for congestive heart failure, hypertension, age ≥65 years [2 points if age ≥75 years], diabetes, female sex, and coronary or peripheral arterial disease, and 2 points for stroke/TIA. Selection of this CHA2DS2-VASc score ≥2 cohort led to inclusion of an additional 84,262 patients, for a total cohort of 294,642 patients. Because antiplatelet therapy may be indicated for reasons other than AF among those with cardiovascular disease, we also stratified analyses in both the CHADS2 score ≥2 and CHA2DS2-VASc score ≥2 cohorts by those patients with and without a coronary heart disease risk equivalent condition, which was considered a diagnosis of any of the following: coronary artery disease; unstable angina; stable angina; prior myocardial infarction; prior coronary artery bypass graft (CABG) surgery; or peripheral arterial disease. To minimize over-representation by patients with multiple visits, only data from each patient’s index visit during the study period were used. The index visit was considered the first encounter where a diagnosis of AF was specified. To ensure that misclassification of OAC prescription was not overlooked by examining only the index visit, we performed a sensitivity analysis to determine the number of patients that would be reclassified as being prescribed an OAC by increasing the window from baseline to within 1 year following the index visit.
Our primary study outcome was treatment with aspirin therapy (alone or in combination with another antiplatelet agent) versus any U.S. Food and Drug Administration–approved OAC for stroke prevention in patients with AF, which included warfarin, dabigatran or rivaroxaban (apixaban and edoxaban had not yet been approved by the U.S. Food and Drug Administration during the study timeframe). Aspirin therapy was defined as prescription of aspirin, either alone or in combination with clopidogrel, ticlopidine, prasugrel, and/or dipyridamole. Frequencies of OAC prescription with aspirin and/or other antiplatelet agents (i.e., thienopyridine or dipyridamole), as well as aspirin plus antiplatelet agent prescription were analyzed. Treatment with a thienopyridine was defined as prescription of clopidogrel, ticlopidine, or prasugrel. Patients treated with aspirin/dipyridamole combination only were included in the aspirin group.
Normally distributed continuous variables are expressed as mean ± SD, whereas categorical variables are expressed as proportions. Unadjusted differences were compared using the chi-square test for categorical variables, and Student t tests for continuous variables, as appropriate.
To investigate the independent associations of various characteristics with the outcome of aspirin alone versus OAC prescription, we constructed hierarchical modified Poisson regression models adjusted for patient and practice-level characteristics. These models included site as a random effect to account for patient clustering within sites. All covariates considered to be potential predictors were entered in the multivariable model and included patient-level characteristics (age, body mass index [BMI], sex, congestive heart failure, diabetes mellitus, stroke or TIA, systemic embolism, hypertension, dyslipidemia, coronary artery disease, unstable angina, stable angina, prior myocardial infarction, prior CABG surgery, and peripheral arterial disease) and practice-level characteristics (U.S. geographical region, urban location, and clinic volume). Candidate covariates selected for the multivariate analyses were chosen a priori on the basis of plausibility that they could be associated with differential prescription of aspirin alone versus OAC. Additional analyses adjusted for any thienopyridine use were performed to evaluate associations of these predictors with prescription of aspirin alone versus OAC, independent of the need for thienopyridine prescription. The highest missing rate for variables was BMI (28.7%); therefore, a missing indicator was included in models that contained this variable. Additionally, missing data were assumed to be missing at random and were imputed with 10 imputation sets, in which all patient variables were used to inform the imputation model (13). Because the rate of OAC prescription exceeded 10%, we used modified Poisson regression models at all steps to estimate relative prescription rates with 95% confidence intervals (CIs) directly (instead of odds ratios obtained from logistic regression, which may overestimate effect differences) (14). Statistical tests were 2-sided and considered significant if they yielded a p value <0.05. Analyses were performed using the SAS statistical package, version 9.3 (SAS Institute, Cary, North Carolina), R version 2.15.3 (Foundation for Statistical Computing, Vienna, Austria), and IVEWare (Institute for Social Research, University of Michigan, Ann Arbor, Michigan).
Among the 210,380 patients with a CHADS2 score ≥2, there were 80,371 (38.2%) treated with aspirin alone and 130,009 (61.8%) treated with warfarin or non-vitamin K antagonist OACs. Of the 294,642 patients with CHA2DS2-VASc ≥2, 118,398 (40.2%) were treated with aspirin alone, and 176,244 (59.8%) were treated with warfarin or non-vitamin K antagonist OACs.
Patient and practice characteristics among treated AF patients in each cohort stratified by prescription of aspirin alone versus OAC are shown in Table 1. In both the CHADS2 score ≥2 cohort and the CHA2DS2-VASc score ≥2 cohort, AF patients prescribed aspirin alone instead of OAC were younger, had lower BMI, were more often female, and were more likely to have diabetes mellitus, hypertension, dyslipidemia, coronary artery disease, unstable angina, stable angina, prior myocardial infarction, prior CABG surgery, and peripheral artery disease. Compared with those who were prescribed OAC, patients in both cohorts who were prescribed aspirin alone were more often treated in the South and West, in nonurban practices, and at practices with a higher volume of patients. Patients with AF treated with aspirin or OAC therapy differed from patients with AF excluded from the analyses with respect to several baseline characteristics (Online Table 1).
In the CHADS2 score ≥2 cohort, the mean CHADS2 score was 2.8 ± 1.0. Of the 130,009 patients (61.8%) who were treated with OACs, warfarin was the most commonly used therapy (n = 118,178 [90.9%]), followed by dabigatran (n = 9,363 [7.2%]) and rivaroxaban (n = 2,468 [1.9%]). In the CHA2DS2-VASc score ≥2 cohort, the mean CHA2DS2-VASc score was 2.2 ± 1.2. Similarly, of the 176,244 patients (59.8%) who were treated with OACs, warfarin was the most commonly used therapy (n = 159,668 [90.6%]), followed by dabigatran (n = 13,122 [7.4%]), and rivaroxaban (n = 3,454 [2.0%]). In a sensitivity analysis evaluating reclassification of patients prescribed an OAC within 1 year of the index visit, a small proportion of patients (1.9%, n = 2,507 of 135,056) not prescribed an anticoagulant agent at baseline were prescribed OAC in follow-up. Evaluation of practice-level variation of OAC prescription revealed that the median practice rate for OAC prescription in the CHADS2 score ≥2 cohort was 64.9%, with a significant variation in OAC prescription, as evidenced by an interquartile range of 56.4% to 69.9%. The median practice rate for OAC prescription in the CHA2DS2-VASc score ≥2 cohort was 63.3%, with variation represented by an interquartile range of 55.2% to 68.7% (Figures 2A and 2B).
In both the CHADS2 score ≥2 cohort and the CHA2DS2-VASc score ≥2 cohort, concomitant use of any thienopyridine (mainly clopidogrel) was higher in those patients prescribed aspirin. Concomitant prescription of aspirin alone as well as aspirin + thienopyridine was not uncommon in patients who were prescribed an OAC (Table 2). Combination dual antiplatelet therapy with aspirin and any thienopyridine was much less common in patients prescribed OAC than aspirin, in the range of 5.7% to 6.0% in both the CHADS2 score ≥2 and CHA2DS2-VASc score ≥2 cohorts. Among patients treated with OAC, concomitant antiplatelet therapy was more often prescribed in patients prescribed a non-vitamin K antagonist OAC compared with warfarin (Online Table 2).
The majority of patients in both the CHADS2 score ≥2 cohort (n = 127,188 [60.5%]) and CHA2DS2-VASc score ≥2 cohort (n = 171,401 [58.2%]) had a coronary heart disease equivalent condition. In both the CHADS2 score ≥2 and CHA2DS2-VASc score ≥2 cohorts, when stratified by the presence of a coronary heart disease risk equivalent diagnosis, aspirin use was more common in patients prescribed OAC, but did not top 55.2%. Even in patients without a coronary heart disease risk equivalent diagnosis, aspirin use was as high as 31.2% in patients also prescribed an OAC. Thienopyridine use was mostly reserved for patients with a coronary heart disease risk equivalent diagnosis (Table 2).
After multivariable adjustment, hypertension, dyslipidemia, coronary artery disease, prior myocardial infarction, unstable and stable angina, recent CABG, and peripheral arterial disease were associated with more frequent prescription of aspirin only, whereas male sex, higher BMI, prior stroke/TIA, prior systemic embolism, and congestive heart failure were associated with more frequent prescription of OAC in both the CHADS2 score ≥2 and CHA2DS2-VASc score ≥2 cohorts (Central Illustration). The associations of all covariates included in the adjusted multivariable models evaluating predictors of aspirin versus OAC prescription are summarized in Online Table 3. In subanalyses including any thienopyridine use in the multivariable model, thienopyridine prescription was associated with more frequent prescription of aspirin compared with OAC in both the CHADS2 score ≥2 (relative prescription rate: 1.61; 95% CI: 1.58 to 1.61; p < 0.001) and CHA2DS2-VASc score ≥2 cohorts (relative prescription rate: 1.55; 95% CI: 1.52 to 1.57; p < 0.001) (Online Figure 1).
In a large, quality-improvement registry of >200,000 outpatients with AF at intermediate to high stroke risk, as measured by both CHADS2 and CHA2DS2-VASc scores ≥2, approximately 40% were treated with aspirin alone instead of OAC prescription. Coronary atherosclerosis-related comorbidities, including hypertension, dyslipidemia, coronary artery disease, prior myocardial infarction, unstable and stable angina, recent CABG, and peripheral arterial disease were associated with more frequent prescription of aspirin only, whereas male sex, higher BMI, prior stroke/TIA, prior systemic embolism, and congestive heart failure were associated with more frequent prescription of OAC. Our findings have important implications for AF patients, particularly as the benefit of stroke risk reduction with OAC over aspirin prescription increases as annual stroke risk increases, as measured by both the CHADS2 and CHA2DS2-VASc scores (2,3), the latter of which includes coronary atherosclerosis and its equivalent disease processes. Because many of the predictors of aspirin use alone include conditions that may warrant aspirin therapy regardless of the presence of AF, much of the underutilization of appropriate anticoagulant agents may be driven by either the perception that aspirin by itself is sufficient or that the risk of aspirin plus anticoagulation is not worth the benefit.
Despite a well-established association of AF with stroke, significant lack of OAC prescription to reduce thromboembolism in at-risk candidates has been demonstrated in several large-scale studies, with aspirin prescription prevalence as high as 28% to 38% in patients who were prescribed an antithrombotic agent at all (15,16). These previous studies described primarily U.S. patients in the era of warfarin therapy, and were before the promulgation of the importance of stroke risk scores to aid in risk stratification. Clinical risk scores, including the CHADS2 and CHA2DS2-VASc scores, have been developed to elucidate and quantify stroke risk in AF patients to aid in the decision to prescribe antithrombotic therapies (2,3). In our study, we used the CHADS2 score, because this risk scheme was the predominant scoring system contemporary with the study period. In order to expand upon the robustness of our findings, we also studied the same cohort of AF patients as assessed by the CHA2DS2-VASc score, because this risk scheme may improve discrimination of AF patients at risk for stroke and thromboembolism and is supported by updated guidelines published near the end of and after the study timeframe (11,12). In the GARFIELD registry (Global Anticoagulant Registry in the FIELD), a multinational, observational study of 10,614 AF patients enrolled between 2009 and 2011 at 540 sites in 19 non-U.S. countries and cared for by general practitioners, cardiologists, and neurologists, 25.3% of patients with a CHADS2 score ≥2 were prescribed aspirin instead of OAC, but the study did not evaluate predictors of aspirin prescription (17). Because 38.2% of AF patients with a CHADS2 score ≥2 were prescribed aspirin alone in our study, there is clearly a continued lack of a guideline-adhering prescription of OAC in U.S. cardiovascular specialist practices. The assertion that aspirin does not adequately protect against thromboembolism in AF patients has been extensively studied, highlighted by a meta-analysis of 4,052 patients with AF who experienced nearly one-half the stroke risk when treated with warfarin rather than aspirin (hazard ratio: 0.55; 95% CI: 0.43 to 0.71) (18). Although inadequate knowledge regarding these data may explain the guideline nonadherence among cardiology specialists, other reasons, including justifiable clinical circumstances, concerns for bleeding in specific patients, or patient refusal of this therapy, may be responsible.
Patients with risk factors for coronary atherosclerosis and myocardial infarction are often treated with aspirin as primary prevention therapy, and those with previous acute coronary syndrome are treated with aspirin as secondary prevention therapy, as per clinical guidelines (19). In our large cohort of AF patients at risk for stroke, risk factors for coronary atherosclerosis, including hypertension and dyslipidemia, predicted aspirin over OAC therapy prescription. Additionally, characteristics associated with coronary atherosclerosis, including coronary artery disease itself, prior myocardial infarction, unstable and stable angina, recent CABG, and peripheral arterial disease, also were more often predictive of aspirin prescription. It is possible that cardiovascular specialists may be more inclined to treatment with aspirin therapy in such patients, despite concomitant indications for anticoagulation due to AF. This may reflect the misperception that aspirin exhibits adequate efficacy compared with OAC, or may be driven by concerns for bleeding risk when using combination therapy (20). That prior stroke/TIA, prior systemic embolism, and congestive heart failure were associated with more frequent prescription of OAC suggests that these specific components of both the CHADS2 and CHA2DS2-VASc scores may be weighed more by the prescribing clinician in a decision to choose appropriate OAC over aspirin-only therapy. Interestingly, after multivariable adjustment, men with AF in our study had a 6% greater odds of OAC prescription versus aspirin therapy, despite previous studies that demonstrate an increased stroke risk in women (3). This paradoxical practice pattern suggests that female sex as an important stroke risk factor in AF remains largely unappreciated. Previous data from the PINNACLE registry suggest that a sex bias may exist on the basis of lower prescription rates of OAC in women treated in real-world U.S. cardiology practices (21). Significant variability of OAC prescription across individual practices was observed (as shown in Figure 2); this suggests that focusing on factors at the practice level, and not simply at the individual physician level, may prove important in future efforts to rectify proper prescription of OAC in AF patients at risk for stroke.
A significant number of patients prescribed OAC for AF were also prescribed antiplatelet therapy. Dual antiplatelet therapy was prescribed in 19.2% to 20.6% of the patients who were prescribed aspirin alone. Among those receiving OAC, aspirin was prescribed in just over one-half of all patients with coronary heart disease equivalent conditions and one-third of patients without any coronary heart disease equivalent conditions. It is also important to emphasize that a substantial proportion of AF patients without a cardiovascular risk equivalent (approximately one-third) were prescribed both an OAC and aspirin, placing those patients at higher risk for bleeding, without any evidence of benefit. This category of patients likely represents “low-hanging fruit” for changing practice (presumably via prescriber education) in a fashion that would mitigate risk for bleeding, without reducing the efficacy of thromboembolism prevention (22). Triple therapy, including an anticoagulant agent and 2 antiplatelet drugs, was prescribed in approximately 6%. Although guidelines recommend aspirin therapy in all acute coronary syndrome patients who are able to take it (23), consensus guidelines for anticoagulant-indicated AF patients with stable coronary heart disease and its equivalents are lacking, with either little mention (24), or recommendations to only continue antiplatelet therapy for 1 year after an acute coronary syndrome. Our finding that a large proportion of patients prescribed OAC are also prescribed aspirin therapy suggests that further studies evaluating cardiovascular outcomes in AF patients with stable coronary heart disease prescribed various antithrombotic strategies are needed. Additionally, clarification and subsequent promulgation of consensus guidelines on concomitant antiplatelet and anticoagulation use appears warranted, particularly because coronary heart disease risk equivalents have now been incorporated into the CHA2DS2-VASc score as part of consensus AF guidelines (12).
First, although the PINNACLE registry does ascertain whether an anticoagulant agent is contraindicated, the data are not sufficiently granular to calculate the HAS-BLED score (i.e., the PINNACLE registry does not capture data on medication use predisposing to bleeding, labile international normalized ratios, or alcohol or drug use) (25), nor does it contain data on renal insufficiency or aspirin dose prescribed. This additional information may have been helpful to quantify the bleeding risk among those who received aspirin versus OAC prescription to determine whether bleeding risk was associated with prescription preference. However, because the net clinical benefit of warfarin actually increases with the HAS-BLED score (given that the score generally mirrors the stroke-risk scores), a higher bleeding risk alone should not justify aspirin, rather than an anticoagulant agent. In fact, data support a net clinical benefit for OAC over aspirin therapy in AF patients at risk for stroke, despite increased bleeding risks (26). Second, the PINNACLE program enrolled patients from motivated cardiology practices dedicated to quality improvement. Therefore, antithrombotic therapy prescription patterns in other U.S. practices may differ from those reported in this study, potentially reducing the generalizability of our results. Third, specific data are unavailable regarding previous bleeding complications or exact reasons for contraindications to anticoagulant therapy, and therefore we cannot determine the validity of a reported contraindication. Fourth, the main data analyzed consisted of OAC prescription at the index AF visit. Although the index visit may not capture OAC prescription at any time in patient follow-up, a sensitivity analysis that included follow-up visits out to 1 year after the index visit demonstrated that only a small proportion of additional patients (<2%) were prescribed OAC. Fifth, the PINNACLE registry does not contain follow-up data to ascertain outcomes, such as subsequent stroke or bleeding, and does not contain follow-up laboratory data to assess therapeutic anticoagulation status with warfarin. Future studies should focus on these outcomes as they relate to aspirin versus OAC prescription, as well as effects of prescription of additional antiplatelet agents. Finally, some may argue that the PINNACLE data collection form may not reflect actual prescription of the drug, much less what patients actual receive or consume; however, that distinction is arguably minimally relevant for the purposes of this study, because the data recorded on the form likely more purely reflects the intent or perceived “correct” prescription of medications. Nevertheless, we cannot rule out the possibility that, despite best efforts for accurate data collection, underreporting of OAC prescription occurred by those recording this information.
Despite a lack of evidence for aspirin's use for thromboembolic prophylaxis in patients with AF at moderate to high risk of stroke, and clear benefit with the use of OACs, nearly 1 in 3 of these patients in this large, quality-improvement registry of cardiology outpatients with AF at intermediate to high risk of stroke were prescribed aspirin alone. Several specific patient characteristics predicted prescription of aspirin therapy over OAC therapy, particularly comorbidities related to coronary atherosclerosis and its risk equivalent diseases. These data indicate a gap in care, most prominent in patients with or at risk for coronary artery disease, and should draw attention to a high rate of prescription of aspirin therapy in AF patients at risk for stroke, despite previous data that show aspirin to be inferior to OAC in this population. The specific patient characteristics associated with this practice, including those related to coronary artery disease, highlight opportunities to improve appropriate prescription of OAC in AF, including identifying knowledge gaps that might be informed by future studies.
COMPETENCY IN PATIENT CARE: In a cardiac outpatient population of patients with AF at moderate to high risk of stroke, a relatively high proportion of patients were treated with aspirin rather than anticoagulant agents, and this was associated with conditions related to coronary heart disease.
TRANSLATIONAL OUTLOOK: Future studies should investigate the impact of the availability of non-vitamin K antagonist OACs on the choice of antithrombotic agent for patients with AF and the role of performance improvement programs on adherence to guideline-directed management.
For a supplemental figure and tables, please see the online version of this article.
This research was supported by the American College of Cardiology Foundation’s National Cardiovascular Data Registry (NCDR). The views expressed in this paper represent those of the authors, and do not necessarily represent the official views of the NCDR or its associated professional societies identified at www.ncdr.com. PINNACLE Registry is an initiative of the American College of Cardiology Foundation. Bristol-Myers Squibb and Pfizer Inc. are Founding Sponsors of the PINNACLE Registry. Dr. Hsu has received honoraria from St. Jude Medical, Medtronic, Biotronik, Janssen Pharmaceutical, and Bristol-Myers Squibb; has received research support from Biotronik and Biosense Webster. Dr. Lubitz has received grants from the National Institutes of Health (NIH) and the Doris Duke Charitable Foundation. Dr. Gehi has received speaker honoraria from Medtronic, Zoll Medical, St. Jude Medical, and Biotronik. Dr. Turakhia has received research support from the Veterans Affairs, Gilead Sciences, iRhythm, Medtronic, Janssen Pharmaceuticals, and SentreHeart; is a consultant to Janssen Pharmaceuticals, Medtronic, St. Jude Medical, and Precision Health Economics; and has equity ownership in thryva and Zipline. Dr. Marcus has received research support from the NIH, PCORI, Medtronic, Pfizer, Rhythm Diagnostic Systems, and SentreHeart; and is a consultant for and has equity ownership in InCarda. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- body mass index
- coronary artery bypass graft
- confidence interval
- National Cardiovascular Data Registry
- oral anticoagulation/anticoagulant
- transient ischemic attack
- Received December 26, 2015.
- Revision received March 22, 2016.
- Accepted March 29, 2016.
- American College of Cardiology Foundation
- Aguilar M.I.,
- Hart R.,
- Pearce L.A.
- Chan P.S.,
- Oetgen W.J.,
- Buchanan D.,
- et al.
- Messenger J.C.,
- Ho K.K.L.,
- Young C.H.,
- et al.,
- NCDR Science and Quality Oversight Committee Data Quality Workgroup
- Fuster V.,
- Rydén L.E.,
- Cannom D.S.,
- et al.
- Camm A.J.,
- Lip G.Y.H.,
- De Caterina R.,
- et al.,
- ESC Committee for Practice Guidelines (CPG)
- January C.T.,
- Wann L.S.,
- Alpert J.S.,
- et al.
- Raghunathan T.E.,
- Solenberger P.W.,
- Van Hoewyk J.
- Zou G.
- Kakkar A.K.,
- Mueller I.,
- Bassand J.P.,
- et al.,
- GARFIELD Registry Investigators
- Vandvik P.O.,
- Lincoff A.M.,
- Gore J.M.,
- et al.
- Thompson L.E.,
- Maddox T.M.,
- Bradley S.M.,
- et al.
- Steinberg B.A.,
- Kim S.,
- Piccini J.P.,
- et al.,
- ORBIT-AF Investigators and Patients
- O'Gara P.T.,
- Kushner F.G.,
- Ascheim D.D.,
- et al.