Author + information
- Received October 17, 2012
- Accepted November 27, 2012
- Published online June 4, 2013.
- Ziad Hijazi, MD*,†,* (, )
- Lars Wallentin, MD, PhD*,
- Agneta Siegbahn, MD, PhD*,‡,
- Ulrika Andersson, MSc*,
- Christina Christersson, MD, PhD*,†,
- Justin Ezekowitz, MBBCh, MSc§,
- Bernard J. Gersh, MD⋮,
- Michael Hanna, MD¶,
- Stefan Hohnloser, MD#,
- John Horowitz, MD**,
- Kurt Huber, MD††,
- Elaine M. Hylek, MD, MPH‡‡,
- Renato D. Lopes, MD, PhD§§,
- John J.V. McMurray, MD⋮⋮ and
- Christopher B. Granger, MD§§
- *Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden
- †Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden
- ‡Department of Medical Sciences, Clinical Chemistry, Uppsala University, Uppsala, Sweden
- §University of Alberta, Edmonton, Alberta, Canada
- ⋮Mayo Clinic, Rochester, Minnesota
- ¶Bristol-Myers Squibb, Princeton, New Jersey
- #J.W. Goethe-University, Frankfurt, Germany
- **University of Adelaide, Adelaide, South Australia, Australia
- ††Department of Cardiology and Emergency Medicine, Wilhelminen Hospital, Vienna, Austria
- ‡‡Boston University Medical Center, Boston, Massachusetts
- §§Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina
- ⋮⋮Western Infirmary, Glasgow, United Kingdom
- ↵*Reprint requests and correspondence:
Dr. Ziad Hijazi, Uppsala Clinical Research Center, Uppsala University, Uppsala Science Park, SE-752 37 Uppsala, Sweden.
Objectives This study sought to assess the prognostic value of N-terminal pro–B-type natriuretic peptide (NT-proBNP) in patients with atrial fibrillation (AF) enrolled in the ARISTOTLE (Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation) trial, and the treatment effect of apixaban according to NT-proBNP levels.
Background Natriuretic peptides are associated with mortality and cardiovascular events in several cardiac diseases.
Methods In the ARISTOTLE trial, 18,201 patients with AF were randomized to apixaban or warfarin. Plasma samples at randomization were available from 14,892 patients. The association between NT-proBNP concentrations and clinical outcomes was evaluated using Cox proportional hazard models, after adjusting for established cardiovascular risk factors.
Results Quartiles of NT-proBNP were: Q1, ≤363 ng/l; Q2, 364 to 713 ng/l; Q3, 714 to 1,250 ng/l; and Q4, >1,250 ng/l. During 1.9 years, the annual rates of stroke or systemic embolism ranged from 0.74% in the bottom NT-proBNP quartile to 2.21% in the top quartile, an adjusted hazard ratio of 2.35 (95% confidence interval [CI]: 1.62 to 3.40; p < 0.0001). Annual rates of cardiac death ranged from 0.86% in Q1 to 4.14% in Q4, with an adjusted hazard ratio of 2.50 (95% CI: 1.81 to 3.45; p < 0.0001). Adding NT-proBNP levels to the CHA2DS2VASc score improved C-statistics from 0.62 to 0.65 (p = 0.0009) for stroke or systemic embolism and from 0.59 to 0.69 for cardiac death (p < 0.0001). Apixaban reduced stroke, mortality, and bleeding regardless of the NT-proBNP level.
Conclusions NT-proBNP levels are often elevated in AF and independently associated with an increased risk of stroke and mortality. NT-proBNP improves risk stratification beyond the CHA2DS2VASc score and might be a novel tool for improved stroke prediction in AF. The efficacy of apixaban compared with warfarin is independent of the NT-proBNP level. (Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation [ARISTOTLE]; NCT00412984).
Atrial fibrillation (AF) increases the risk of stroke and death and consequently constitutes a significant societal health and economic problem (1,2). Prediction of a patient’s risk of stroke is most commonly made using a clinical score such as the CHADS2 or CHA2DS2VASc (heart failure, hypertension, age 75 years and older, diabetes, and previous stroke or transient ischemic attack, vascular disease, age 65 to 74 years, sex category [female sex], respectively) (3,4). B-type natriuretic peptides are secreted by myocytes in response to a number of stimuli including an increase in wall stress (5). Their secretion increases with aging and with left ventricular hypertrophy and in acute coronary syndromes, heart failure, chronic kidney disease, and AF (6,7). In these conditions, N-terminal pro-B-type natriuretic peptide (NT-proBNP) is a powerful predictor of prognosis (8–15). However, little is known about the predictive value of NT-proBNP in patients with AF in whom, at present, risk stratification based on clinical scores only offers a modest discriminating ability for the individual patients (4).
In the ARISTOTLE (Apixaban for the Prevention of Stroke in Subjects With Atrial Fibrillation) trial (16), we obtained plasma samples for central biomarker analyses from 14,892 of 18,201 patients. In this pre-specified substudy, we examined whether NT-proBNP might predict the occurrence of stroke and other nonfatal and fatal cardiovascular events in anticoagulated patients with AF and an increased risk of stroke. We further tested the incremental value of measuring NT-proBNP levels in addition to established risk factors (including the CHA2DS2VASc score) to predict cardiovascular and bleeding events.
The ARISTOTLE trial
The details of the ARISTOTLE trial were published previously (16). Briefly, the ARISTOTLE trial was a double-blind, double-dummy, randomized clinical trial that enrolled 18,201 patients with AF and at least 1 CHADS2 risk factor for stroke or systemic embolism. Patients were randomized to warfarin (n = 9,081) or apixaban (n = 9,120). The primary endpoint was stroke or systemic embolism. Bleeding was classified according to the International Society on Thrombosis and Haemostasis (ISTH) criteria. The median length of follow-up was 1.9 years in the biomarker study.
Endpoints and clinical risk classification
The endpoints in the ARISTOTLE trial included stroke or systemic embolism; ischemic stroke and systemic embolism, hemorrhagic stroke, myocardial infarction; all-cause mortality; cardiac death (excluding bleeding and other noncardiac causes); ISTH major bleeding; and composites including stroke or systemic embolism, total death or cardiac death, and myocardial infarction. All endpoints were adjudicated by a blinded clinical events committee using pre-specified criteria (16). CHADS2 and CHA2DS2VASc scores were calculated for each patient based on the sum of the corresponding risk factors present at randomization. Patients were categorized by CHADS2 according to score (0 to 1, 2, or ≥3) and by CHA2DS2VASc scores (0 to 1, 2, 3, 4, and ≥5).
For the ARISTOTLE biomarker substudy, baseline blood samples were obtained from 14,892 patients. Plasma was frozen in aliquots and stored at −70°C until analyzed centrally. The NT-proBNP levels were determined with sandwich immunoassays on the Cobas Analytics e601 Immunoanalyzers (Roche Diagnostics, Mannheim, Germany) according to manufacturer instructions. The lower limit of detection for NT-proBNP with this assay is 5 ng/l. The analytical range extends from 20 to 35,000 ng/l according to the manufacturer. The upper reference level (97.5th percentile) in men and women 40 to 65 years of age is 184 and 268 ng/l, respectively, and 66 to 76 years of age, 269 and 391 ng/l, respectively (17). The lowest concentration with a coefficient of variation <10% is 30 ng/l (18).
These analyses included the 14,892 patients who provided blood samples for the biomarker study at randomization and also had available results of the evaluated biomarkers. Demographics and other baseline characteristics were summarized using frequencies for categorical variables and median and 25th and 75th percentiles for continuous variables. For tests of differences among groups, the chi-square test was used for categorical variables and the Kruskal-Wallis test was used for continuous variables.
Multivariable analysis of variance with natural logarithms of NT-proBNP as response variables and categorized baseline characteristics as explanatory variables was used to investigate the independent effect of each variable. Geometric means, calculated by antilogs of the model-adjusted means, were compared.
Efficacy analyses included all randomized patients and included all events from randomization until the efficacy cut-off date (pre-defined as January 30, 2011). Bleeding analyses were ‘on treatment’ including all randomized patients who received at least 1 dose of study drug and included all events from receipt of the study drug until 2 days after the last dose of the study drug.
The relationship between NT-proBNP and outcomes was evaluated both in a simple and multivariable Cox regression analysis. The multivariable analyses included established risk factors (age, sex, body mass index, smoking status, systolic blood pressure, heart rate, AF type, diabetes, heart failure, previous stroke/systemic embolism/transient ischemic attack, hypertension, previous myocardial infarction, previous peripheral arterial disease, coronary artery bypass grafting/percutaneous coronary intervention, treatment at randomization with aspirin, angiotensin-converting enzyme inhibitor or angiotensin II receptor blocker, and amiodarone), randomized treatment, region, use of warfarin within 7 days before randomization, use of statin medication within 30 days before randomization score, and biomarkers (cystatin C and high-sensitivity troponin T). For the major bleeding endpoint, a history of anemia and of spontaneous or clinical relevant bleeding was also included in the established risk factors and the multivariable analysis. The hazard ratios (HRs) and 95% confidence intervals (CIs), using the group with the lowest biomarker levels as reference, were reported. The assumption of proportional hazards for the factors included in the Cox regression analyses was assessed visually using log-cumulative hazard plots (not shown).
Treatment effects were compared according to the NT-proBNP group using a Cox proportional hazards model including treatment group, NT-proBNP quartile group, and treatment by NT-proBNP interaction as covariates. The treatment HRs are reported at each level of NT-proBNP, regardless of the significance of interaction.
The incidences of the different endpoints were summarized in relation to randomized treatment, quartiles of the NT-proBNP level, and CHA2DS2VASc scores.
The increased discriminative values of adding NT-proBNP level to models with only a CHA2DS2VASc score and study treatment were investigated by estimating the difference in C-statistics between models with and without NT-proBNP levels and also the integrated discrimination improvement measure (IDI), as described by Pencina et al. (19,20). In these analyses, the occurrence/nonoccurrence of an event during the follow-up period was used as a binary response, and the C value will be the same as the area under the receiver-operating characteristic curve. The relative IDI was calculated to facilitate interpretation of the IDI (21). In addition, continuous (category-free) net reclassification improvement (NRI) was analyzed as a measure of probability of better reclassification minus the probability of worse reclassification with the new model. The NRI among events and nonevents as well as the total NRI were analyzed.
Kaplan-Meier estimates of the cumulative risk to the first occurrence of an event were calculated and plotted. Event rates per 100 patient-years of follow-up were reported. All statistical tests were 2 tailed and performed at the 0.05 significance level. There were no adjustments for multiple comparisons. The Clinical Trials section at Uppsala Clinical Research Center conducted the statistical analyses, using the statistical software package SAS, version 9.3 for Windows (SAS Institute, Cary, North Carolina) for all analyses.
Baseline characteristics and distribution levels of NT-proBNP
The median NT-proBNP concentration was 714 ng/l (25th and 75th percentiles = 363 and 1,250 ng/l), and there was no difference between the warfarin and apixaban groups. The 25th percentile value was close to the upper reference level of NT-proBNP in healthy subjects (i.e., 75% of patients had an elevated level) (17).
Baseline characteristics according to NT-proBNP quartiles are shown in Table 1. Many characteristics were associated with NT-proBNP in univariate analysis, with higher levels in older subjects and women and in patients with other cardiovascular disease, diabetes, and renal dysfunction. NT-proBNP was also related to AF type. These baseline characteristics remained associated with the NT-proBNP level in a multivariable analysis (p < 0.05 for all). The strongest relationship was with AF type, with a more than 3-fold higher geometric mean NT-proBNP in patients with persistent/permanent AF compared with paroxysmal AF (ratio of geometric means: 3.35; 95% CI: 3.22 to 3.50; p < 0.0001), followed by reduced creatinine clearance (Cockcroft-Gault), heart failure, and age as independent predictors of NT-proBNP levels (p < 0.0001 for all). Patients with higher CHADS2 and CHA2DS2VASc scores had higher NT-proBNP levels. Among patients in the highest quartile of NT-proBNP level >1,250 ng/l, 81.5% had a CHA2DS2VASc score >2 in contrast to only 59% in the lowest quartile of patients with NT-proBNP level ≤363 ng/l.
NT-proBNP in relation to outcomes and study treatment
In this substudy cohort, a total of 397 patients (1.40%/year) experienced stroke or systemic embolism, and 1,075 (3.69%/year) died; 547 (1.88%/year) died of a cardiac cause and 674 (2.61%/year) experienced a major bleeding episode during a median follow-up of 1.9 years. Higher baseline NT-proBNP concentration was strongly associated with each of the major clinical outcomes examined except major bleeding, even after adjustment for the multivariable model (Fig. 1). For example, the annualized rate of stroke or systemic embolism ranged from 0.74% in the bottom NT-proBNP quartile (Q1: ≤363 ng/l) to 2.21% in the top quartile (Q4: >1,250 ng/l); adjusted HR (Q4 vs. Q1): 2.35; 95% CI: 1.62 to 3.40; p < 0.0001) (Fig. 1). The adjusted HR (Q4 vs. Q1) for all-cause mortality was 2.25 (95% CI: 1.80 to 2.81, p < 0.0001), and for cardiac mortality, it was 2.50 (95% CI: 1.81 to 3.45, p < 0.0001). Higher NT-proBNP was not associated with an increased risk of major bleeding (adjusted HR [Q4 vs. Q1]: 1.07; 95% CI: 0.82 to 1.40; p = 0.0667).
There was no significant interaction between baseline NT-proBNP levels and the effect of randomized treatment in relation to any of these or the other study outcomes (Fig. 2). Kaplan-Meier plots illustrating the associations between NT-proBNP (by quartile) and these outcomes are shown in Figure 3.
NT-proBNP levels in relation to CHA2DS2VASc score for risk assessment and prognostic discrimination
Annual rates of stroke or systemic embolism according to NT-proBNP levels and CHA2DS2VASc score are illustrated in Figure 4A. The rate increased with both increasing CHA2DS2VASc score and higher NT-proBNP level. The highest annual rate of stroke and systemic embolism (2.45%) was found in the group with a CHA2DS2VASc score ≥3 and an NT-proBNP concentration >1,250 ng/l, compared with an average annual rate of 0.56% in patients with a CHA2DS2VASc score ≤2 and NT-proBNP levels ≤363 ng/l. Adding NT-proBNP to the CHA2DS2VASc score improved the predictive model, with an increase in the C-statistic from 0.620 (95% CI: 0.592 to 0.647) to 0.646 (95% CI: 0.619 to 0.673) (p = 0.0009) and a relative IDI improvement of 47% (IDI: 0.0023; 95% CI: 0.0015 to 0.0030; p < 0.0001). For cardiac deaths, increasing NT-proBNP levels had a distinctively greater impact on outcomes than the CHA2DS2VASc score groups (Fig. 4B). The C-statistic increased from 0.592 (95% CI: 0.568 to 0.617) to 0.691 (95% CI: 0.669 to 0.714) (p < 0.0001), with a relative IDI improvement of 270% (IDI: 0.0142; 95% CI: 0.0119 to 0.0165; p < 0.0001) compared with the CHA2DS2VASc alone. The findings concerning the composite ischemic event endpoint (ischemic stroke, systemic embolism, myocardial infarction, and cardiac death) were very similar. The C-statistic increased from 0.598 (95% CI: 0.579 to 0.617) to 0.660 (95% CI: 0.641 to 0.678) (p < 0.0001), with a relative IDI improvement of 162% (IDI: 0.0119; 95% CI: 0.0099 to 0.0139; p < 0.0001). In addition to IDI, a category-free (continuous) NRI analysis was performed (Table 2). The continuous NRI was 0.289 (95% CI: 0.195 to 0.384, p < 0.0001) for stroke or systemic embolism with events contributing 0.315 and nonevents −0.025. For vascular death, the continuous NRI was 0.508 (95% CI: 0.423 to 0.593, p < 0.0001) with events contributing 0.112 and nonevents 0.396. For stroke or systemic embolism, the improved reclassification was among patients experiencing events, whereas for cardiovascular death, the improvement was highest for nonevent patients.
The major findings of this study were that NT-proBNP levels are elevated in three-fourths of patients with AF and at least 1 risk factor for stroke. The NT-proBNP concentration was related to age, sex, and comorbidity, as expected, and most strongly to the type of AF, with higher levels in patients with persistent/permanent AF compared with paroxysmal AF. The NT-proBNP level was independently and gradually related to the risk of stroke or systemic embolism and all-cause and cardiac mortality. This predictive value of NT-proBNP was incremental to the CHA2DS2VASc score and persisted after adjustment for established risk factors including the CHA2DS2VASc score and indicators of renal function and troponin. There was no interaction between NT-proBNP and treatment (i.e., apixaban was superior to warfarin in terms of both efficacy and safety across NT-proBNP quartiles).
The prognostic value of natriuretic peptides is established in elderly subjects living in the community and in patients with stable coronary artery disease, acute coronary syndromes, and heart failure (10,12–15). Our study extends these findings to subjects with AF and is consistent with the recent findings of the RE-LY trial (11). However, because there were considerably more patients in the present ARISTOTLE biomarker cohort (N = 14,892) compared with RE-LY (N = 6,189), we were able to demonstrate the independent incremental value of adding NT-proBNP to a recommended risk score (CHA2DS2VASc score) in relation to individual endpoints (e.g., subtypes of stroke including intracranial bleeding). Despite adjustment for established risk factors including CHA2DS2VASc score and biomarkers reflecting different pathways of the renal function (cystatin-C) and myocardial state (high-sensitivity troponin T), the improved risk stratification obtained with NT-proBNP remained highly significant, including more than a doubled risk of stroke and cardiac death (Fig. 1). Further, a consistent risk over the duration of the trial was seen in the NT-proBNP quartiles, allowing the identification of a greater absolute benefit with apixaban compared with warfarin treatment at higher NT-proBNP levels (Fig. 2).
In clinical practice, risk stratification in AF patients is commonly based on clinical risk scores, such as the established CHA2DS2VASc risk score, which is derived from clinical variables such as the presence of congestive heart failure, hypertension, age 75 years and older, diabetes, previous stroke or transient ischemic attack, vascular disease, age 65 to 74 years, and sex category. The present study demonstrated the independent predictive value of NT-proBNP in addition to clinical risk stratification with the CHA2DS2VASc score for improving the prognostic discrimination of risk of stroke or systemic embolism and cardiac death. The C-statistics improved substantially from 0.62 to 0.65 and 0.59 to 0.69 regarding stroke and cardiac death events, respectively. For stroke or systemic embolism events, the NT-proBNP levels and the CHA2DS2VASc score distinctively provided complementary information, as illustrated in Figure 4A. The lowest risk was seen in patients with no or minimal elevated levels of NT-proBNP (≤363 ng/l) and a low CHA2DS2VASc score (≤2). Despite a CHA2DS2VASc score of 0 to 1, if the NT-proBNP level was >1,250 ng/l, the annual risk of stroke and systemic embolism equaled those with a score of ≥3 and levels ≤363 ng/l (Fig. 4A). Concerning cardiac death, mainly NT-proBNP level, but not the CHA2DS2VASc score, provided prognostic information, as visually illustrated in Figure 4B, and with 270% improvement according to relative IDI. For both outcomes, a substantial proportion of patients were appropriately reclassified to higher or lower risks as displayed with the category free NRI analysis.
The effect of study treatment in relation to NT-proBNP levels was also analyzed in this study. NT-proBNP is an established sensitive marker for heart failure (6). Heart failure is a well-known risk factor for thromboembolism in AF as well as an independent marker for increased bleeding risk (4,22). We therefore analyzed the effect of the study treatment in relation to NT-proBNP levels and cardiovascular events including major bleeding. Despite the significant increased risk and predictive effect concerning stroke or systemic embolism and mortality with increasing NT-proBNP levels, there was no study treatment interaction with regard to outcomes, thus displaying the superior effect of apixaban over warfarin in terms of both efficacy and safety across NT-proBNP quartiles.
The clinical characteristics included in the CHA2DS2VASc risk score have been identified primarily to provide information on stroke but not on other adverse events in AF patients. Accordingly, also in relation to the composite ischemic event endpoint, increasing NT-proBNP levels were more strongly associated with outcomes than the CHA2DS2VASc score, and the C-statistic increased from 0.60 to 0.66 (p < 0.0001). AF is a well-described risk factor of increased mortality (1). Despite this evidence, in the clinical setting, risk stratification and treatment are mainly focused on primary or secondary stroke prevention. Although in an anticoagulated AF population such as in the RE-LY and ARISTOTLE trial, the total numbers and annual rates of mortality are more than doubled compared with stroke or systemic embolism events (23,24). This study clearly displays the powerful risk prediction gained with NT-proBNP measurements in an AF population concerning mortality, both independently and compared with CHA2DS2VASc score. This novel prognostic information may be usable for treatment selection to further improve survival in AF patients. The consistent results from 2 very large, systematic and prospective AF trials such as RE-LY (11) and ARISTOTLE, together including 21,081 patients, enable definite conclusions of the incremental prognostic value of NT-proBNP level for stroke, cardiac, total death, and other ischemic events in patients with AF.
Several of the clinical variables used to assess ischemic stroke risk in AF patients are also used to stratify the risk of major bleeding events, of which the most devastating and feared are the intracerebral hemorrhages (25). Due to the large cohort size in the present study, we are, for the first time, able to discriminate and analyze subtypes of ischemic or hemorrhagic stroke with regard to NT-proBNP levels. The results clearly display a significant elevated risk of ischemic stroke with increasing NT-proBNP level (Fig. 1). These novel findings add important information for the use of NT-proBNP measurements for risk stratification in stable AF patients with regard to competing risks and thromboembolic events.
In the present study, more than three-fourths of the patients had NT-proBNP levels above the upper reference limit for elderly individuals (17). AF rhythm rather than sinus rhythm at the time of blood sampling was most strongly associated with higher NT-proBNP levels, with a >3-fold higher geometric mean. Persistent or permanent compared with paroxysmal AF, older age, lower creatinine clearance, and clinical history of coronary artery disease and heart failure were also associated with higher NT-proBNP levels in this AF cohort, which affirms the results of some previously published small studies (26). Elevated natriuretic peptides reflect the myocyte response to increased wall tension. In AF, levels of natriuretic peptides are elevated compared with matched controls in sinus rhythm in various settings (26–28) and decrease rapidly after restoration of sinus rhythm (29–32). In contrast to the pathophysiology of heart failure in which NT-proBNP is derived mainly from the ventricles, there is support for NT-proBNP release from the atrium in patients with AF as a response to atrial stretch (33–35). Accordingly, elevated NT-proBNP levels in AF may partly originate from atrial dysfunction, an established marker associated with the formation of atrial thrombi (36), and thereby provide a plausible mechanism for the prognostic importance of elevated NT-proBNP levels and thromboembolic events as presented in this ARISTOTLE biomarker study.
The present findings are derived from a clinical trial population with AF and at least 1 risk factor for stroke and may therefore not be immediately extrapolated to the general AF population. The study design does not permit final conclusions about the optimal cutoff value of NT-proBNP as a decisive tool to select patients for different antithrombotic strategies because all study participants received oral anticoagulants.
The NT-proBNP level is elevated in the majority of patients with persistent or permanent AF. The degree of NT-proBNP elevation has a strong independent association with the risk of stroke and mortality. The relative benefits of apixaban compared with warfarin are consistent, regardless of the NT-proBNP levels, and accordingly the absolute benefits of apixaban will be greater in patients with AF and higher NT-proBNP levels. NT-proBNP might therefore be a novel tool for the selection of patients with AF both for anticoagulation and other treatments.
Editorial assistance was provided by Ebba Bergman, PhD, at Uppsala Clinical Research Center, Sweden.
The trial was funded by Bristol-Myers Squibb, Co., Princeton, New Jersey and Pfizer Inc., New York, New York, and coordinated by the Duke Clinical Research Institute, Durham, North Carolina, and Uppsala Clinical Research Center, Uppsala, Sweden. Both academic research institutes and the sponsors had full access to the database. The Uppsala Clinical Research Center received funding from Bristol-Myers Squibb Co. for the statistical analysis performed by Ms. Andersson. Dr. Hijazi has received lecture fees and an institutional research grant from Boehringer-Ingelheim. Dr. Wallentin has received research grants from AstraZeneca, Merck/Schering-Plough, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline; is a consultant for Merck/Schering-Plough, Regado Biosciences, Evolva, Portola, C.S.L. Behring, Athera Biotechnologies, Boehringer-Ingelheim, AstraZeneca, GlaxoSmithKline, and Bristol-Myers Squibb/Pfizer; has received lecture fees from AstraZeneca, Boehringer Ingelheim, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, Schering-Plough; and honoraria from Boehringer-Ingelheim, AstraZeneca, Bristol-Myers Squibb/Pfizer, GlaxoSmithKline, and Schering-Plough/Merck. Dr. Siegbahn has received institutional grants from AstraZeneca and Boehringer-Ingelheim. Dr. Ezekowitz has received consulting fees and travel support from Bristol-Myers Squibb/Pfizer and research support, honoraria, or payments for continuing medical education events from Alere, AstraZeneca, Amgen, Abbott, Servier, Johnson & Johnson, Pfizer, and Bristol-Myers Squibb. Dr. Gersh is a consultant for Ortho-McNeil Janssen, GE Healthcare, St. Jude Medical, Abbott Laboratories, Medispec, Merck, and Boston Scientific. Dr. Hanna is an employee of Bristol-Myers Squibb and receives stock as a part of compensation. Dr. Hohnloser has received consulting fees from sanofi-aventis, St. Jude Medical, Boehringer-Ingelheim, Cardiome, Pfizer, Bayer, and Bristol-Myers Squibb; and lecture fees from sanofi-aventis, Bristol-Myers Squibb, Pfizer, Boehringer-Ingelheim, Medtronic, and St. Jude Medical. Dr. Horowitz has received consulting fees, travel support from and is an Executive Committee Member for Bristol-Myers Squibb; and lecture fees from Bristol-Myers Squibb and Pfizer. Dr. Huber has received lecture fees from AstraZeneca, Bristol-Myers Squibb/Pfizer, Boehringer-Ingelheim, Bayer, Daiichi Sankyo, and sanofi-aventis. Dr. Hylek has received consultant fees from Johnson & Johnson, Boehringer-Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo, Pfizer, and Ortho-McNeil-Janssen; and lecture fees from Boehringer Ingelheim and Bayer. Dr. Lopes has received grants from Bristol-Myers Squibb, AstraZeneca, Boehringer-Ingelheim, and Daiichi Sankyo; and consulting fees from Bristol-Myers Squibb and Boehringer-Ingelheim. Dr. McMurray’s employer (Glasgow University) was paid by Bristol-Myers Squibb/Pfizer for his participation as an Executive Committee member in the ARISTOTLE trial. Dr. Granger has received grants from Bristol-Myers Squibb, Boehringer-Ingelheim, Astellas, AstraZeneca, GlaxoSmithKline, Medtronic Foundation, Merck, sanofi-aventis, and the Medicines Company; consulting fees from Bristol-Myers Squibb, Pfizer, Boehringer-Ingelheim, AstraZeneca, Eli Lily, GlaxoSmithKline, Hoffman-La Roche, Novartis, Otsuka, Sanofi-Aventis, and the Medicines Company; and travel support from Pfizer, Novartis, Otsuka, and Hoffmann-La Roche. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- atrial fibrillation
- confidence interval
- hazard ratio
- integrated discrimination improvement
- International Society on Thrombosis and Haemostasis
- net reclassification improvement
- N-terminal pro-B-type natriuretic peptide
- Received October 17, 2012.
- Accepted November 27, 2012.
- American College of Cardiology Foundation
- Benjamin E.J.,
- Wolf P.A.,
- D'Agostino R.B.,
- Silbershatz H.,
- Kannel W.B.,
- Levy D.
- Wolf P.A.,
- Abbott R.D.,
- Kannel W.B.
- Camm A.J.,
- Kirchhof P.,
- Lip G.Y.,
- et al.
- Boomsma F.,
- van den Meiracker A.H.
- Daniels L.B.,
- Maisel A.S.
- Redfield M.M.,
- Rodeheffer R.J.,
- Jacobsen S.J.,
- Mahoney D.W.,
- Bailey K.R.,
- Burnett J.C. Jr..
- Arakawa N.,
- Nakamura M.,
- Aoki H.,
- Hiramori K.
- Fonarow G.C.,
- Peacock W.F.,
- Phillips C.O.,
- Givertz M.M.,
- Lopatin M.
- Hijazi Z.,
- Oldgren J.,
- Andersson U.,
- et al.
- James S.K.,
- Lindahl B.,
- Siegbahn A.,
- et al.
- Tsutamoto T.,
- Wada A.,
- Maeda K.,
- et al.
- Friberg L.,
- Rosenqvist M.,
- Lip G.Y.
- Ellinor P.T.,
- Low A.F.,
- Patton K.K.,
- Shea M.A.,
- Macrae C.A.
- Shelton R.J.,
- Clark A.L.,
- Goode K.,
- Rigby A.S.,
- Cleland J.G.
- Goetze J.P.,
- Friis-Hansen L.,
- Rehfeld J.F.,
- Nilsson B.,
- Svendsen J.H.
- Yasue H.,
- Yoshimura M.,
- Sumida H.,
- et al.