You are on page 1of 12

Arrhythmia/Electrophysiology

Cardiac Biomarkers Are Associated With an Increased Risk of Stroke and Death in Patients With Atrial Fibrillation
A Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) Substudy
Ziad Hijazi, MD; Jonas Oldgren, MD, PhD; Ulrika Andersson, MSc; Stuart J. Connolly, MD; Michael D. Ezekowitz, MB, ChB; Stefan H. Hohnloser, MD; Paul A. Reilly, PhD; Dragos Vinereanu, MD, PhD; Agneta Siegbahn, MD, PhD; Salim Yusuf, MD, PhD; Lars Wallentin, MD, PhD
BackgroundCardiac biomarkers are strong predictors of adverse outcomes in several patient populations. We evaluated the prevalence of elevated troponin I and N-terminal pro-B-type natriuretic peptide (NT-proBNP) and their association to cardiovascular events in atrial fibrillation (AF) patients in the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) trial. Methods and ResultsBiomarkers at randomization were analyzed in 6189 patients. Outcomes were evaluated by Cox proportional hazards models adjusting for established cardiovascular risk factors and the CHADS2 and CHA2DS2-VASc risk scores. Patients were stratified based on troponin I concentrations: 0.010 g/L, n 2663; 0.010 to 0.019 g/L, n 2006; 0.020 to 0.039 g/L, n 1023; 0.040 g/L, n 497; and on NT-proBNP concentration quartiles: 387; 387 to 800; 801 to 1402; 1402 ng/L. Rates of stroke were independently related to levels of troponin I with 2.09%/year in the highest and 0.84%/year in the lowest troponin I group (hazard ratio [HR], 1.99 [95% CI, 1.173.39]; P 0.0040), and to NT-proBNP with 2.30%/year versus 0.92% in the highest versus lowest NT-proBNP quartile groups, (HR, 2.40 [95% CI, 1.41 4.07]; P 0.0014). Vascular mortality was also independently related to biomarker levels with 6.56%/year in the highest and 1.04%/year the lowest troponin I group (HR, 4.38 [95% CI, 3.05 6.29]; P 0.0001), and 5.00%/year in the highest and 0.61%/year in the lowest NT-proBNP quartile groups (HR, 6.73 [3.9511.49]; P 0.0001). Biomarkers increased the C-statistic from 0.68 to 0.72, P 0.0001, for a composite of thromboembolic events. ConclusionsElevations of troponin I and NT-proBNP are common in patients with AF and independently related to increased risks of stroke and mortality. Cardiac biomarkers seem useful for improving risk prediction in AF beyond currently used clinical variables. Clinical Trial RegistrationURL: http://www.clinicaltrials.gov. Unique identifier: NCT00262600. (Circulation. 2012;125:1605-1616.) Key Words: atrial fibrillation cardiac biomarkers natriuretic peptide risk prediction troponin

he prevalence of atrial fibrillation (AF) is increasing and is projected to reach epidemic proportions in coming decades.1 AF is associated with a 5-fold increase in the rate of ischemic stroke and doubled total mortality.2 Strategies for identifying patients at risk for thromboembolism are commonly based on clinical variables, eg, congestive heart

failure, hypertension, age, diabetes mellitus, and prior stroke or transient ischemic attack (TIA) in the widely used CHADS2 risk score.3 So far, no biochemical marker has been shown to provide incremental information. Cardiac troponin, an intracellular protein involved in heart muscle contraction, is an established biochemical marker of myocardial cell

Received April 20, 2011; accepted February 16, 2012. From the Uppsala Clinical Research Center and Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden (Z.H., J.O., L.W.); Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden (U.A.); Population Health Research Institute, Hamilton, Canada (S.J.C., S.Y.); Lankenau Institute for Medical Research and the Heart Center, Wynnewood, PA (M.D.E.); Department of Cardiology, J.W. Goethe University, Frankfurt, Germany (S.H.H.); Boehringer Ingelheim Pharmaceuticals, Ridgefield, CT (P.A.R.); Department of Cardiology, University Hospital of Bucharest, Bucharest, Romania (D.V.); and Uppsala Clinical Research Center and Department of Medical Sciences, Clinical Chemistry, Uppsala University, Uppsala, Sweden (A.S.). The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA. 111.038729/-/DC1. Correspondence to Ziad Hijazi, MD, Uppsala Clinical Research Center, Dag Hammarskjolds vag 14B, 1st floor, SE- 752 37 Uppsala, Sweden. E-mail Ziad.hijazi@ucr.uu.se 2012 American Heart Association, Inc. Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.111.038729

1605

1606

Circulation

April 3, 2012
of detection with this assay is 0.006 g/L with 0.014 g/L as the lowest concentration measurable with a coefficient of variation of 10% and 0.02 g/L as the 99th percentile upper reference limit (URL) for subjects aged 60 years14 and 0.04 g/L as the 99th percentile URL regardless of age.15 With present instrument calibration all troponin I concentrations 0.010 g/L are reported as 0.010 g/L and considered undetectable; levels 0.010 are regarded as detectable and reported with 2 significant figures; levels 0.020 g/L are considered elevated. NT-proBNP was analyzed by using the Sandwich Immunoassay, Elecsys, Roche Diagnostics. The analytic range extends from 20 to 35 000 ng/L according to the manufacturer. The upper reference level (97.5th percentile) in men and woman aged 40 to 65 years is 184 and 268 ng/L, respectively, and age 66 to 76 years, 269 and 391 ng/L, respectively.16 The lowest concentration measurable with a coefficient of variation 10% is 30 ng/L.17

damage. B-type natriuretic peptide (BNP), a neurohormone secreted from the cardiac ventricles, is a recognized marker of myocardial wall tension, and its inactive part, N-terminal fragment (N-terminal pro-B-type natriuretic peptide [NTproBNP]]), as well.4 Elevated levels of troponin and NTproBNP have repeatedly been demonstrated as important markers of increased mortality and morbidity in acute coronary syndromes,5,6 stable coronary artery disease,7,8, congestive heart failure,9,10 and even in general community-based populations.11,12 The prevalence and clinical significance of elevated cardiac troponin I and NT-proBNP in nonvalvular AF patients at risk for stroke is unknown.

Clinical Perspective on p 1616


The Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) trial recently demonstrated the superiority of dabigatran versus warfarin for stroke prevention in 18 113 AF patients.13 In this prospectively designed biomarker substudy, we investigated the prevalence of elevated troponin I and NT-proBNP and their association to cardiovascular events in a representative subgroup constituting one-third of the RE-LY cohort.

Statistical Analysis
The sample size for the substudy was prospectively estimated at 5744 patients, based on an expected event rate of 1.6%/year (corresponding to a total event rate of 3.2%), and the assumptions that the biomarker was related to outcome in the following manner: the total event rate for the lower quartile is 1.6%, for the middle quartiles 3.2%, and for the upper quartile 6.4%. A 2-sided test of the null hypothesis of no difference in event rate between the lower quartile and a middle quartile requires 1436 patients per group when the significance level is 5% and the power is 80%. The number of samples in total was thus calculated to 4 1436 5744, to account for missing information blood samples, was planned from 6200 patients of the 18 113 patients in the RE-LY study. The final number of samples in the present analysis was 6189, composed of 2561 patients participating in a comprehensive RE-LY biomarker substudy program and 3628 randomly selected patients with cardiac biomarkers (troponin I and NT-proBNP) obtained at randomization in the main RE-LY study. Patients were grouped according to quartiles of NT-proBNP levels, 387 ng/L (n 1547), 387 to 800 ng/L (n 1547), 801 to 1402 ng/L (n 1544), and 1402 ng/L (n 1551). Troponin I results were, as expected, extremely skewed, and patients were in the present study grouped according to cutoff levels based on results from previous studies, instead of 4 equally sized groups. Troponin I levels were undetectable ( 0.010 g/L) in 2663 (43.0%) patients, 2006 (32.4%) patients had detectable troponin I levels 0.010 to 0.019 g/L, ie, up to the 99th percentile for apparently healthy individuals 60 years age.18 One thousand twenty-three (16.5%) patients had slightly elevated troponin I levels 0.020 to 0.039 g/L, and 497 (8.0%) had clearly elevated troponin I levels 0.040 g/L, ie, above the 99th percentile URL for the troponin I assay regardless of age.15 Demographics and baseline characteristics were summarized for the troponin I and NT-proBNP level groups with the use of frequencies for categorical variables, and median and 25th and 75th quartiles for continuous variables. Because of the low numbers of patients in the CHADS2 scores, 0, 4, 5, and 6 patients were grouped in CHADS2 classes of 0 to 1, 2, and 3. For tests of differences among groups, the 2 test was used for categorical variables, and Kruskal-Wallis test was used for continuous variables. The risk of event is reported as percentage per year, which was calculated by dividing the total number of patients with events by the total number of patient-years of follow-up. Cumulative hazard plots were used to illustrate the timing of events. The relations between levels of troponin I and NT-proBNP at randomization and events were investigated by using Cox proportional hazards regression. Three different models were used. Model A was adjusted for study treatment, use of (prestudy) anticoagulant treatment at randomization and established risk factors for cardiovascular disease (age, sex, body mass index, smoking status, sitting systolic blood pressure, sitting heart rate, AF duration, AF type, creatinine clearance, diabetes, coronary artery disease, previous stroke/systemic embolism/TIA, heart failure, hypertension, treatment at randomization with aspirin, angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, and statins). Model B was only adjusted for study treatment

Methods
Study Population and Trial Design
The study organization, trial design, patient characteristics and outcomes of the RE-LY study have been published previously.13 In brief, RE-LY was a prospective, multicenter, randomized trial comparing 2 blinded doses of dabigatran with open-label warfarin for a minimum of 12 months in 18 113 patients. Inclusion criteria were documented AF and at least one of the following risk factors for stroke: previous stroke or TIA; congestive heart failure or reduced left ventricular ejection fraction ( 40%); at least 75 years of age; or at least 65 years of age with diabetes mellitus, hypertension, or coronary artery disease. Exclusion criteria included severe heart valve disorder, recent stroke, increased risk of hemorrhage, creatinine clearance 30 mL/min, or active liver disease. The 6189 (of 18113) patients in the present study represented 446 of 951 sites in 38 of 44 countries in the RE-LY trial. The primary efficacy outcome in the study was fatal and nonfatal stroke (ischemic, hemorrhagic, or unspecified) or systemic embolism, and the secondary outcomes were total mortality, vascular (including hemorrhagic) mortality, nonvascular mortality, and a composite thromboembolic end point consisting of ischemic stroke, systemic embolism, myocardial infarction, pulmonary embolism, and vascular mortality (excluding hemorrhagic death). The primary safety outcome was major bleeding. Median follow-up was 2.0 years for the main trial and 2.2 years for the substudy population. Outcomes were assessed by study visits scheduled at 3-month intervals during the first year and 4-month intervals thereafter. Each event was classified by 2 independent adjudicators from an international team, blinded to treatment assignments. Definition of the outcomes has been described previously.13

Blood Sampling
Venous blood was drawn at randomization, before initiation of study treatment, with the use of a 21/22 gauge needle into Vacutainer tubes containing EDTA. The blood was centrifuged within 30 minutes at 2000g for 10 minutes. The tubes were thereafter immediately frozen at 20 C or colder. Aliquots were stored at 70C to allow batch analysis.

Laboratory Methods
All plasma samples were centrally analyzed in Uppsala Clinical Research Center laboratory, Sweden. Troponin I was analyzed with the Access AccuTnI assay (Beckman Coulter, Inc, Fullerton, CA), a 2-site immunoenzymatic (sandwich) immunoassay. The lower limit

Hijazi et al
and the CHADS2 score (1 point for each of congestive heart failure, hypertension, age 75 years, diabetes mellitus, and 2 points for previous stroke/TIA) by using 6 CHADS2 classes (CHADS2 score of 5 and 6 combined). Model C was only adjusted for study treatment and the CHA2DS2-VASc score (1 point for each of congestive heart failure, hypertension, age 65 but 75 years, diabetes mellitus, prior stroke or TIA, vascular disease, and sex category; and 2 points for each of age of 75 years and prior stroke or TIA) by using 7 classes (scores of 0 and 1 combined, and scores of 7, 8, and 9 combined). The 2 cardiac biomarkers were included separately and jointly, as well, in the models. The results of hazard ratios for model A are presented in Results and tables. Models B and C yielded similar hazard ratios (see online-only Data Supplement Tables I and II). The effects of randomized treatment on outcome in relation to level of biomarker was evaluated with Cox proportional hazards model with treatment group, biomarker, and their interaction as dependent variables. The proportional hazards assumption with respect to the cardiac biomarkers was assessed by visual inspection of log-cumulative hazard plots and by extending the Cox model with a time-by-biomarker interaction factor. In addition, sensitivity analyses using logistic regression were performed (results not shown). The increased discriminative values of troponin I and NT-proBNP were investigated by estimating the difference in C-statistics between models with and without respective cardiac biomarker19 and also the integrated discrimination improvement measure (IDI) as described by Pencina et al.20 In these analyses, the occurrence/nonoccurrence of stroke or systemic embolism and composite thromboembolic events, respectively, during the follow-up period was used as a binary response, and the C value will be the same as the area under the ROC curve. The relative IDI was calculated to facilitate interpretation of the IDI.21 A probability value of 0.05 from 2-sided tests was considered to indicate statistical significance. The statistical software package SAS, version 9.2 for Windows (SAS institute, Cary, NC) was used for all analyses.

Cardiac Biomarkers in Atrial Fibrillation

1607

undetectable troponin I ( 0.010 g/L) and low NT-proBNP 387 ng/L, and highest (37.6% and 38.6%, respectively) in the groups with highest troponin I ( 0.040) g/L and highest NT-proBNP ( 1402 ng/L), respectively (Table 1).

Stroke or Systemic Embolism


During the median of 2.2 years follow-up, there were 183 events of stroke or systemic embolism. The annual rates of stroke or systemic embolism were lowest, 0.84%, in the group with undetectable troponin I, in comparison with 2.09% (HR, 1.99; 95% CI, 1.173.39) in the highest troponin I group (P 0.0040) (Figure 1A, Table 2). In relation to NT-proBNP, the annual rates of stroke or systemic embolism were lowest, 0.92%, in the quartile group with low NT-proBNP, in comparison with 2.30% (HR, 2.40; 95% CI, 1.41 4.07) in the highest NT-proBNP (P 0.0014) (Figure 1B, Table 2). The effects of troponin I remained significant (P 0.0232) after adjustment for NTproBNP in the multivariable model, and for NT-proBNP (P 0.0112) adjusted for troponin I, as well.

Mortality
During the follow up 450 patients died, 297 of which died of vascular causes. In the group with undetectable troponin I, annual vascular death rate was 1.04% in comparison with 6.56% (HR, 4.38; 95% CI, 3.05 6.29) in the highest troponin I group (P 0.0001) (Figure 2A, Table 2). In relation to NT-proBNP, the annual rates of vascular death were 0.61% in the lowest NT-proBNP group, in comparison with 5.00% (HR, 6.73; 95% CI, 3.9511.49) in the highest NT-proBNP group (P 0.0001) (Figure 2B, Table 2). The effects of troponin I remained significant (P 0.0001) after adding NT-proBNP to the multivariable model, and for NT-proBNP (P 0.0001) adjusted for troponin I, as well. One hundred fifty-three patients died of nonvascular causes. The rate of nonvascular mortality was higher among patients with elevated troponin I levels, P 0.0127, but without a gradual increase by troponin I levels (Table 2). There was no significant association between NT-proBNP levels and nonvascular death.

Results
Baseline Characteristics in Relation to Levels of Cardiac Biomarkers
Patient characteristics are summarized in Table 1. Troponin I levels ranged from 0.010 to 7.1 g/L, with a median of 0.011, 25th and 75th percentile values of 0.010 and 0.019 g/L. Detectable levels of troponin I ( 0.010 g/L) were found in 3526 (57.0%) and elevated levels of troponin I ( 0.020 g/L) in 1520 (24.6%) patients (Table 1). Several patient characteristics, eg, higher age, previous myocardial infarction, history of congestive heart failure, AF rhythm at randomization (at the time of blood sampling), and lower creatinine clearance were significantly associated with higher troponin I levels, all P 0.0001 (Table 1). NT-proBNP levels ranged from 5 to 44959 ng/L, with a median of 801, 25th and 75th percentile values of 387 and 1403 ng/L. Several patient characteristics, most prominently age, AF rhythm at randomization, history of congestive heart failure, and lower creatinine clearance, were significantly associated with higher NT-proBNP levels, all P 0.0001 (Table 1). The proportion of patients with CHADS2 risk scores 0 to 1 was highest (37.3% and 38.6%, respectively) in groups with the lowest levels of cardiac biomarkers (troponin I 0.010 g/L and NT-proBNP 387 ng/L) and lowest (25.4% and 25.6%, respectively) in groups with the highest levels of cardiac biomarkers (troponin I 0.040 g/L and NT-proBNP 1402 ng/L, respectively). The opposite applied to the proportion of patients with a CHADS2 scores 3, which was lowest (27.8% and 25.5%, respectively) in groups with

Myocardial Infarction
There were 103 myocardial infarctions during follow-up. The annual rates were 0.49% in the group with undetectable troponin I and 1.69% (HR 3.04; 95% CI 1.64 5.64) in the highest troponin I group (P 0.0052) (Table 2). The association of troponin I and myocardial infarctions remained significant after adding NT-proBNP to the model. There was no significant relation between NT-proBNP levels and myocardial infarctions.

Composite Thromboembolic End Point


There were 482 composite thromboembolic events (ischemic stroke, systemic embolism, pulmonary embolism, myocardial infarction, and vascular death excluding hemorrhagic death). In the group with undetectable troponin, the annual rates of the composite thromboembolic end point were 2.00% in comparison with 8.85% (HR, 3.43; 95% CI, 2.57 4.56) in the highest troponin I group (P 0.0001) (Table 2). Concerning NT-proBNP, the annual rates of the composite thromboem-

1608
Table 1.

Circulation

April 3, 2012

Baseline Characteristics According to Troponin I and NT-proBNP group


Troponin I

Groups No. in substudy Follow-up time, median (25th, 75th pct), y Region West Europe North America Age, median (25th, 75th pct), y Body mass index, median (25th, 75th pct), kg/m2 Male sex, n (%) Current smoker, n (%) Systolic blood pressure, median (25th, 75th pct), mm Hg Heart rate median (25th, 75th pct) AF rhythm at baseline (%) Type of AF, n (%) Paroxysmal Persistent Permanent AF duration, n (%) 3 mo 3 mo2 y 2y Heart failure, n (%) Hypertension, n (%) Age 75 y, n (%) Diabetes mellitus, n (%) Previous stroke/TIA, n (%) CHADS2 score, n (%) 01 2 3 Prior myocardial infarction, n (%) Coronary artery disease, n (%) CrCL at baseline, mL/min (25th-75th pct) Medications at baseline, n (%) Aspirin -blocker ACE inhibitor and/or ARB Statin Amiodarone

All 6189 2.20 (1.78, 2.52) 3512 (56.7) 72.0 (67.0, 77.0) 28.1 (25.3, 31.6) 3944 (63.7) 483 (7.8) 130.0 (120.0, 144.0) 72.0 (62.0, 82.0) 4519 (73.2) 1848 (29.9) 1609 (26.0) 2731 (44.1) 1794 (29.0) 1441 (23.3) 2954 (47.7) 1859 (30.0) 4852 (78.4) 2356 (38.1) 1322 (21.4) 1216 (19.6) 2025 (32.7) 2197 (35.5) 1967 (31.8) 1078 (17.4) 1540 (24.9) 69.0 (54.2, 87.1) 2214 (35.8) 4102 (66.3) 4313 (69.7) 2670 (43.1) 700 (11.3)

0.010 g/L 2663 2.22 (1.88, 2.54) 1553 (58.3) 71.0 (66.0, 76.0) 28.2 (25.6, 31.8) 1631 (61.2) 202 (7.6) 130.0 (120.0, 142.0) 72.0 (63.0, 82.0) 1867 (70.3) 871 (32.7) 726 (27.3) 1066 (40.0) 756 (28.4) 646 (24.3) 1261 (47.4) 618 (23.2) 2112 (79.3) 883 (33.2) 519 (19.5) 538 (20.2) 992 (37.3) 932 (35.0) 739 (27.8) 330 (12.4) 578 (21.7) 73.4 (58.1, 90.3) 900 (33.8) 1762 (66.2) 1761 (66.1) 1198 (45.0) 296 (11.1)

0.010 0.019 g/L 2006 2.25 (1.74, 2.56) 1175 (58.6) 73.0 (67.0, 78.0) 28.1 (25.2, 31.5) 1288 (64.2) 157 (7.8) 133.0 (120.0, 145.0) 72.0 (62.0, 82.0) 1490 (74.5) 580 (28.9) 474 (23.6) 951 (47.4) 562 (28.0) 461 (23.0) 983 (49.0) 588 (29.3) 1575 (78.5) 838 (41.8) 447 (22.3) 387 (19.3) 636 (31.7) 710 (35.4) 660 (32.9) 375 (18.7) 516 (25.7) 68.1 (53.6, 85.4) 722 (36.0) 1352 (67.4) 1447 (72.1) 851(42.4) 214 (10.7)

0.020 0.039 g/L 1023 2.08 (1.67, 2.40) 529 (51.7) 73.0 (67.0, 78.0) 28.0 (25.1, 31.9) 685 (67.0) 79 (7.7) 130.0 (120.0, 145.0) 72.0 (63.0, 82.0) 783 (76.8) 259 (25.3) 285 (27.9) 479 (46.8) 316 (30.9) 220 (21.5) 487 (47.6) 436 (42.6) 791 (77.3) 426 (41.6) 239 (23.4) 183 (17.9) 271 (26.5) 371 (36.3) 381(37.2) 241 (23.6) 293 (28.6) 64.3 (50.1, 82.0) 394 (38.5) 656 (64.1) 737 (72.0) 409 (40.0) 118 (11.5)

0.040 g/L 497 2.10 (1.71, 2.46) 255 (51.3) 72.0 (65.0, 78.0) 27.6 (24.8, 31.2) 340 (68.4) 45 (9.1) 130.0 (120.0, 140.0) 72.0 (62.5, 81.0) 379 (76.4) 138 (27.8) 124 (24.9) 235 (47.3) 160 (32.2) 114 (22.9) 223 (44.9) 217 (43.7) 374 (75.3) 209 (42.1) 117 (23.5) 108 (21.7) 126 (25.4) 184 (37.0) 187 (37.6) 132 (26.6) 153 (30.8) 63.1 (49.3, 81.1) 198 (39.8) 332 (66.8) 368 (74.0) 212 (42.7) 72 (14.5)

P Value*

0.0001 0.0190 0.0008 0.7369 0.0027 0.9794 0.0001 0.0001

0.2226

0.0001 0.1773 0.0001 0.0153 0.2560 0.0001

0.0001 0.0001 0.0001 0.0093 0.3445 0.0001 0.0396 0.1127

NT-proBNP indicates N-terminal pro-B-type natriuretic peptide; ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; CrCL, creatine clearance; AF, atrial fibrillation; pct, percentile; and TIA, transient ischemic attack. *The P value is based on Kruskal-Wallis test for continuous variables and the 2 test for categorical variables.

bolic end point were 1.78% in the lowest quartile group of NT-proBNP, in comparison with 6.76% (HR, 3.55; 95% CI, 2.515.02) in the highest NT-proBNP group (P 0.0001) (Table 2). The effects of troponin I remained significant (P 0.0001) after adding NT-proBNP to the multivariable model, and for NT-proBNP (P 0.0001) adjusted for effects of troponin I, as well.

Cardiac Biomarkers Levels in Relation to CHADS2 Score


Figure 3A illustrates the annual rates of the composite thromboembolic end point according to troponin I levels and CHADS2 score. The pattern of gradually higher rates of the composite thromboembolic end point concomitant with higher troponin I levels was consistent in all CHADS2 scores, including the group with CHADS2 scores 0 to 1. The highest annual rates of the composite end point, 11.4%, were found in the group with CHADS2 3 and highest troponin I ( 0.040 g/L) compared with the lowest annual risk of 1.48%, in the group with CHADS2 0 to 1 and undetectable troponin I ( 0.010 g/L). Annual rates of the composite thromboembolic end point according to NT-proBNP levels and CHADS2 score are

Major Bleeding
There were 334 major bleeds. In the group with undetectable troponin I, the annual rate of major bleed was 1.72% in comparison with 4.38% (HR, 2.01; 95% CI, 1.39 2.90) in the highest troponin I group (P 0.0009) (Table 2). There was no significant association between NT-proBNP levels and major bleeding.

Hijazi et al
Table 1. Continued
NT-proBNP 387 ng/L 1547 2.17 (1.78, 2.42) 872 (56.4) 70.0 (65.0, 75.0) 28.6 (25.6, 32.0) 965 (62.4) 121 (7.8) 134.0 (122.0, 145.0) 65.0 (60.0, 76.0) 502 (32.6) 955 (61.7) 325 (21.0) 267 (17.3) 529 (34.2) 393 (25.4) 625 (40.4) 292 (18.9) 1273 (82.3) 422 (27.3) 359 (23.2) 296 (19.1) 597 (38.6) 556 (35.9) 394 (25.5) 190 (12.3) 315 (20.4) 76.4 (61.1, 93.9) 587 (37.9) 922 (59.6) 1053 (68.1) 704 (45.5) 278 (18.0) 387 800 ng/L 1547 2.23 (1.79, 2.54) 895 (57.9) 71.0 (66.0, 76.0) 28.4 (25.7, 32.0) 1056 (68.3) 132 (8.5) 130.0 (120.0, 144.0) 73.0 (64.0, 82.0) 1233 (80.0) 373 (24.1) 401 (25.9) 773 (50.0) 396 (25.6) 331 (21.4) 820 (53.0) 422 (27.3) 1207 (78.0) 538 (34.8) 335 (21.7) 299 (19.3) 546 (35.3) 548 (35.4) 453 (29.3) 214 (13.8) 354 (22.9) 73.2 (57.8, 91.1) 519 (33.5) 968 (62.6) 1067 (69.0) 667 (43.1) 173 (11.2)

Cardiac Biomarkers in Atrial Fibrillation

1609

8011402 ng/L 1544 2.23 (1.84, 2.57) 887 (57.4) 73.0 (67.0, 78.0) 27.9 (25.2, 31.6) 977 (63.3) 119 (7.7) 131.0 (120.0, 144.0) 75.0 (65.0, 84.0) 1379 (89.4) 257 (16.7) 447 (29.0) 839 (54.4) 392 (25.4) 349 (22.6) 803 (52.0) 473 (30.6) 1193 (77.3) 660 (42.7) 332 (21.5) 293 (19.0) 485 (31.4) 537 (34.8) 522 (33.8) 296 (19.2) 414 (26.8) 68.3 (54.6, 85.2) 531 (34.4) 1070 (69.3) 1059 (68.6) 668 (43.3) 122 (7.9)

1402 ng/L 1551 2.19 (1.71, 2.50) 858 (55.3) 74.0 (68.0, 79.0) 27.5 (24.7, 30.8) 946 (61.0) 111 (7.2) 130.0 (120.0, 140.0) 75.0 (66.0, 86.0) 1405 (90.8) 263 (17.0) 436 (28.1) 852 (54.9) 477 (30.8) 368 (23.7) 706 (45.5) 672 (43.3) 1179 (76.0) 736 (47.5) 296 (19.1) 328 (21.1) 397 (25.6) 556 (35.8) 598 (38.6) 378 (24.4) 457 (29.5) 59.0 (46.1, 74.8) 577 (37.2) 1142 (73.6) 1134 (73.1) 631 (40.7) 127 (8.2)

P Value*

0.0001 0.0001 0.0002 0.5589 0.0001 0.0001 0.0001 0.0001

0.0001

0.0001 0.0001 0.0001 0.0456 0.3901 0.0001

0.0001 0.0001 0.0001 0.0272 0.0001 0.0081 0.0613 0.0001

illustrated in Figure 3B. The general trend of increasing rates of the composite thromboembolic end point with increasing NT-proBNP levels was consistent in all CHADS2 scores, including the group with CHADS2 scores 0 to 1. The highest annual rate of the composite end point, 8.99%, was found in the group with CHADS2 3 and highest NTproBNP ( 1402 ng/L) compared with the lowest yearly risk of 1.55%, in the group with CHADS2 0 to 1 and lowest NT-proBNP ( 387 ng/L). Figure 3C illustrates the annual rates of the composite thromboembolic end point in relation to cardiac biomarkers only. The pattern of gradually higher rates of the composite thromboembolic end point is apparent in all biomarker groups, with the highest annual rate of the composite end

point, 12.0%, found in patients with highest levels of troponin I ( 0.040 g/L) in combination with highest levels of NT-proBNP ( 1402 ng/L).

Study Treatment and Subgroups


For all the above outcomes there were no significant interactions between troponin I or NT-proBNP groups and effects of study treatment with warfarin or dabigatran 110 mg or 150 mg. Exploratory analyses of relations between troponin or NT-proBNP levels and outcomes within subgroups defined by age ( 75; 75 years), sex, regions (North America and Western Europe versus others, Asian versus non-Asian countries), CHADS2 score (0 1; 2; 3 6), prevalent coronary artery disease versus no coronary artery disease, or prevalent

1610

Circulation

April 3, 2012

0.07

0.06

0.040 g/L 0.020-0.039 g/L 0.010-0.019 g/L <0.010 g/L

Cumulative Hazard Rate

0.05

0.04

0.03

0.02

0.01

0.00 0
Numbers at Risk
<0.010 ug/L 2663 0.010-0.019 ug/L 2006 0.020-0.039 ug/L 1023 0.040 ug/L 497 2639 1969 989 476 2598 1925 958 459

12 Months

18
2382 1711 835 411

24
1715 1269 593 286

30
701 543 212 117

0.07

0.06

>1402 ng/L 801-1402 ng/L 387-800 ng/L <387 ng/L

Figure 1. Cumulative hazard rates for stroke or systemic embolism, according to troponin I levels at randomization (A) and NT-proBNP levels at randomization (B).

Cumulative Hazard Rate

0.05

0.04

0.03

0.02

0.01

0.00 0
Numbers at Risk
<387 ng/L 387-800 ng/L 801-1402 ng/L >1402 ng/L 1547 1547 1544 1551 1539 1528 1516 1490 1522 1500 1486 1432

12 Months

18
1371 1342 1352 1274

24
959 975 996 933

30
347 411 430 385

cardiovascular disease (prior stroke, prevalent congestive heart failure, and/or coronary artery disease) versus no cardiovascular disease, yielded similar results as in the total material with no significant interactions (data not shown).

Predictive Ability of Cardiac Biomarkers in AF for Thromboembolic Events


In this cohort the C-statistic for stroke and systemic embolism was 0.605 for troponin I, 0.598 for NT-proBNP, and 0.631 when combining the cardiac biomarkers (Table 3). A model based on the CHADS2 score (model B) yielded a C-statistic of 0.614. The separate addition of each cardiac biomarker improved the predictive model significantly, and even further when adding both cardiac biomarkers to a model simultaneously (Table 3). The IDI was significant for all models when adding cardiac biomarkers, separately or combined, with relative IDI ranging from 24% to 124% depending on the model. For the composite thromboembolic end point, the C-statistic was 0.639 for troponin I, 0.636 for NT-proBNP,

and 0.676 when combining troponin I and NT-proBNP (Table 3). A model based on the CHADS2 score yielded a C-statistic of 0.590. The separate addition of a cardiac biomarker improved the predictive model significantly, and even further when adding both cardiac biomarkers simultaneously to the model (Table 3). The IDI was significant for all models when adding cardiac biomarkers, separately or combined, with relative IDI ranging from 41% to 425% depending on the model. The improvements in C-statistic and IDI in relation to the CHA2DS2-VASc score (model C) were similar to the improvements with the CHADS2 score (model B) (Table 3).

Discussion
The present RE-LY substudy demonstrated a high prevalence of detectable and elevated troponin I and considerably elevated NT-proBNP in patients with nonvalvular AF and a raised risk of stroke. The degree of troponin I and NTproBNP elevations were both independently related to a

Hijazi et al
Table 2.
Outcome Stroke and systemic embolism

Cardiac Biomarkers in Atrial Fibrillation

1611

Cox Proportional Hazards Model With Troponin I and NT-proBNP at Randomization in Relation to Outcomes
Biomarkers Troponin I group ( g/L) Group 0.010 0.0100.019 0.0200.039 0.040 NT-proBNP quartile group (ng/L) 387 387800 8011402 1402 Total Number of Events, n (%/y) 48 (0.84) 73 (1.72) 41 (1.99) 21 (2.09) 30 (0.92) 40 (1.22) 40 (1.22) 73 (2.30) 59 (1.04) 102 (2.40) 70 (3.39) 66 (6.56) 20 (0.61) 46 (1.40) 72 (2.19) 159 (5.00) 49 (0.86) 52 (1.22) 42 (2.04) 10 (0.99) 27 (0.83) 42 (1.28) 27 (0.82) 57 (1.79) 28 (0.49) 37 (0.87) 21 (1.02) 17 (1.69) 21 (0.64) 28 (0.85) 24 (0.73) 30 (0.94) 114 (2.00) 167 (3.93) 112 (5.43) 89 (8.85) 58 (1.78) 97 (2.95) 112 (3.40) 215 (6.76) 98 (1.72) 125 (2.94) 67 (3.25) 44 (4.38) 66 (2.03) 74 (2.25) 80 (2.43) 114 (3.58) 1.12 (0.781.62) 1.16 (0.791.70) 1.47 (1.012.15) 1.48 (1.131.93) 1.50 (1.092.07) 2.01 (1.392.90) 0.1503 1.07 (0.741.54) 1.07 (0.731.57) 1.28 (0.871.88) 1.78 (1.262.53) 2.06 (1.442.94) 3.55 (2.515.02) 0.0009 1.45 (1.111.90) 1.45 (1.052.01) 1.89 (1.302.75) 0.5272 1.73 (1.362.21) 2.11 (1.612.76) 3.43 (2.574.56) 0.0001 1.66 (1.172.36) 1.81 (1.262.58) 2.79 (1.963.98) 0.0040 1.72 (0.933.18) 1.49 (0.762.91) 1.93 (0.983.82) 0.0001 1.63 (1.282.08) 1.83 (1.392.41) 2.77 (2.063.72) 0.0001 1.59 (0.972.62) 1.77 (0.993.17) 3.04 (1.645.64) 0.2383 1.62 (0.883.00) 1.31 (0.662.56) 1.53 (0.763.08) 0.0001 1.31 (0.762.24) 0.79 (0.431.45) 1.42 (0.802.50) 0.0052 1.56 (0.952.57) 1.72 (0.953.09) 2.88 (1.535.42) 0.4499 1.19 (0.801.77) 1.92 (1.252.93) 0.89 (0.451.77) 0.0767 1.25 (0.732.14) 0.73 (0.401.34) 1.30 (0.732.30) 0.0122 2.44 (1.404.27) 3.71 (2.156.41) 6.73 (3.9511.49) 0.0127 1.18 (0.791.75) 1.87 (1.212.88) 0.83 (0.411.67) 0.0862 1.94 (1.402.69) 2.31 (1.613.30) 4.38 (3.056.29) 0.0001 2.23 (1.273.89) 3.19 (1.855.52) 5.07 (2.958.71) 0.0150 1.35 (0.802.27) 1.31 (0.762.27) 2.40 (1.414.07) 0.0001 1.75 (1.262.42) 1.88 (1.312.70) 3.20 (2.204.65) 0.0001 1.79 (1.232.59) 1.97 (1.283.04) 1.99 (1.173.39) 0.0014 1.29 (0.762.17) 1.21 (0.702.09) 2.09 (1.223.58) 0.0001 HR (95% CI) P Value Effect of Biomarker Level 0.0040 1.71 (1.182.48) 1.76 (1.132.73) 1.68 (0.972.89) 0.0112 Addition of Cardiac Biomarker* HR (95% CI) P Value Effect of Biomarker Level 0.0232

Vascular death

Troponin I group ( g/L)

0.010 0.0100.019 0.0200.039 0.040

NT-proBNP quartile group (ng/L)

387 387800 8011402 1402

Nonvascular death

Troponin I group ( g/L)

0.010 0.0100.019 0.0200.039 0.040

NT-proBNP quartile group (ng/L)

387 387800 8011402 1402

Myocardial infarction

Troponin I group ( g/L)

0.010 0.0100.019 0.0200.039 0.040

NT-proBNP quartile group (ng/L)

387 387800 8011402 1402

Composite thromboembolic outcome

Troponin I group ( g/L)

0.010 0.0100.019 0.0200.039 0.040

NT-proBNP quartile group (ng/L)

387 387800 8011402 1402

Major bleed

Troponin I group ( g/L)

0.010 0.0100.019 0.0200.039 0.040

NT-proBNP quartile group (ng/L)

387 387800 8011402 1402

Six thousand ninety-two patients had values for all covariates and were included in the Cox regressions. Numbers of patients in each group were as follows: troponin I group 1 2663, group 2 2006, group 3 1023, and group 4 497. For NT-proBNP; Q1 1547, Q2 1547, Q3 1544, and Q4 1551. NT-proBNP indicates N-terminal pro-B-type natriuretic peptide; and HR, hazard ratio. *Adding NT-proBNP to multivariable analysis of troponin I, and troponin I to multivariable analysis of NT-proBNP, respectively. Stroke, systemic embolism, pulmonary embolism, myocardial infarction, vascular death (excluding hemorrhagic death).

1612

Circulation

April 3, 2012

0.20

0.040 g/L 0.020-0.039 g/L 0.010-0.019 g/L <0.010 g/L

0.15 Cumulative Hazard Rate

0.10

0.05

0.00 0
Numbers at Risk
<0.010 ug/L 2663 0.010-0.019 ug/L 2006 0.020-0.039 ug/L 1023 0.040 ug/L 497 2648 1985 994 480 2618 1947 970 464

12 Months

18
2407 1736 853 420

24
1738 1292 612 291

30
714 558 217 120

0.20

>1402 ng/L 801-1402 ng/L 387-800 ng/L <387 ng/L

Figure 2. Cumulative hazard rates for vascular death, according to troponin I levels at randomization (A) and NT-proBNP levels at randomization (B).

0.15 Cumulative Hazard Rate

0.10

0.05

0.00 0
Numbers at Risk
<387 ng/L 387-800 ng/L 801-1402 ng/L >1402 ng/L 1547 1547 1544 1551 1543 1535 1525 1504 1533 1513 1502 1451

12 Months

18

24

30

1385 1363 1367 1301

970 995 1010 958

352 423 436 398

raised risk of stroke or systemic embolism, mortality, and other cardiovascular events. The levels of these biomarkers added prognostic information beyond currently used clinical risk scores, eg, the CHADS2 or CHA2DS2-VASc risk scores. Elevation of troponin was initially identified as a sensitive indicator of myocardial damage and myocardial infarction and also as an indicator of raised risk of reinfarction and mortality in patients with acute coronary syndromes.5,6,22 At a later stage, slight elevations of troponin were observed in a proportion of patients with stable coronary artery disease and also associated with a worse outcome.8 In advanced heart failure, elevated troponin is associated with progressive left ventricular dysfunction and increased mortality.9 Recently, more sensitive assays have identified detectable troponin levels also in elderly healthy men, predicting coronary heart disease events and mortality independent of conventional major coronary risk factors.12 The prognostic value of natriuretic peptides has previously been established for a variety of cardiovascular diseases.

Even in community-based populations without heart failure, natriuretic peptide levels predict risk of death and cardiovascular events.11,23 BNP and NT-proBNP are also prognostic of short- and long-term mortality in patients with acute coronary syndromes24,25 and stable coronary artery disease.7 Elevated levels of BNP also provide prognostic information in acute decompensated heart failure26 and in chronic heart failure.10 The simultaneous use of both cardiac biomarkers has displayed even further improvements in risk stratification in patients with acute coronary syndromes.27 Our findings in the present RE-LY substudy extend these observations to a novel population by demonstrating the prognostic importance of troponin I and NT-proBNP in patients with AF. At present, risk stratification in AF is based on clinical variables, with the CHADS2 risk score being the most widely used. Although easy to apply, all clinical risk scores including the novel CHA2DS2-VASc score have, at best, only a modest discriminating ability for the individual patients, in this respect not very different from CHADS2 risk score, with C-statistics in

Hijazi et al

Cardiac Biomarkers in Atrial Fibrillation

1613

Figure 3. Stroke, systemic embolism, pulmonary embolism, myocardial infarction, vascular death (excluding hemorrhagic death) in relation to troponin I levels and CHADS2 scores (A), NT-proBNP levels and CHADS2-scores (B), and troponin I and NT-proBNP levels (C). Total number of patients given in each bar.

the range from 0.54 to 0.65.28 We demonstrated the additive value of troponin I and NT-proBNP to CHADS2 score by stratifying the RE-LY biomarker cohort based on cardiac biomarkers and CHADS2 score (Figure 3A). Within every CHADS2 score stratum, there was also a further gradation of thromboembolic risk in relation to the troponin I and NTproBNP levels. Even in low-risk patients with a CHADS2 score of 0 to 1, any increase in troponin I ( 0.020 g/L) doubled the risk, and highly elevated troponin I levels ( 0.040 g/L) raised the risk 5-fold, surpassing the annual risk of patients with a CHADS2 score of 2 and undetectable troponin I levels. The incremental information from NT-proBNP levels was similar, with a 2.5-fold increase in risk when comparing patients with the highest and lowest quartiles of NT-proBNP within the group of patients with CHADS2 score of 0 to 1 (Figure 3B). The significantly improved C-statistics and IDI support the improvement of risk prediction of thromboembolic events when adding cardiac biomarkers to the CHADS2 and the CHA2DS2-VASc risk scores. A CHADS2 score 2 is currently the generally accepted indication for treatment with oral anticoagulants in AF patients.29 Therefore, it is noteworthy that, in the present population, a group with low CHADS2 score of 0 to 1 and elevated levels of any or both of these cardiac biomarkers had

a higher annual rate of a composite of thromboembolic events than patients with higher CHADS2 scores and undetectable troponin I and/or low NT-proBNP levels. Conversely, patients without elevated cardiac biomarkers might have substantially lower risk than perceived by the CHADS2 score. Similar findings were shown in relation to the novel CHA2DS2-VASc risk score. Patients with high CHADS2 or CHA2DS2-VASc risk score and elevated cardiac biomarkers remain at high risk for thromboembolic events despite preventive treatment with effective oral anticoagulants. Such patients might be considered for intensified pharmacological treatment with angiotensin-converting enzyme inhibitors, angiotensin receptor blockers or statins, ablation therapy, left atrial appendage closure devices, left atrial volume reduction, and, perhaps, also myocardial perfusion stress test or coronary angiogram for further risk stratification and potential percutaneous coronary interventions. The results of this study also document an association between elevated troponin I levels and risk of major bleeding. In AF, several of the variables in scores for estimating stroke risk are the same components used for assessing bleeding risk.30 There are some previous studies in acute coronary syndrome populations linking peak troponin I levels to subsequent increase of bleeding rate.31 The causality is

1614

Circulation

April 3, 2012

Table 3. Receiver Operating Characteristics for Stroke or Systemic Embolism, and the Composite of Stroke, Systemic Embolism, Pulmonary Embolism, Myocardial Infarction, Vascular Death (Excluding Hemorrhagic Death)
Stroke and Systemic Embolism Model Troponin I NT-proBNP Troponin I NT-proBNP CHADS2 Troponin I NT-proBNP Troponin I NT-proBNP CHA2DS2-VASc Troponin I NT-proBNP Troponin I NT-proBNP CV risk factors Troponin I NT-proBNP Troponin I NT-proBNP C-statistic 0.605 0.598 0.631 0.614 0.646 0.637 0.658 0.618 0.647 0.633 0.654 0.679 0.690 0.692 0.700 0.0228 Referent 0.0398 0.1157 0.0141 Referent 0.0492 0.2393 0.0279 Referent 0.2620 0.1486 0.0641 0.0032 0.0041 0.0064 24 30 47 0.0033 0.0035 0.0058 59 63 103 0.0033 0.0036 0.0059 70 76 124 0.0027 67 P Value Referent IDI Relative IDI, % C-statistic 0.640 0.640 0.676 0.596 0.667 0.662 0.691 0.612 0.674 0.668 0.696 0.679 0.703 0.707 0.720 0.0001 Referent 0.0001 0.0001 0.0001 Referent 0.0001 0.0001 0.0001 Referent 0.0007 0.0001 0.0001 0.0165 0.0136 0.0247 49 40 73 0.0195 0.0160 0.0295 152 125 230 0.0200 0.0177 0.0355 239 211 425 0.0126 57 Composite Thromboembolic Outcome* P Value Referent IDI Relative IDI, %

NT-proBNP indicates N-terminal pro-B-type natriuretic peptide; IDI, integrated discrimination improvement; CV, cardiovascular; AF, atrial fibrillation; TIA, transient ischemic attack. *Stroke, systemic embolism, pulmonary embolism, myocardial infarction, vascular death (excl hemorrhagic death). All P 0.01. CVs risk factors: Adjusted for study treatment, use of (prestudy) anticoagulant treatment at randomization and established risk factors for cardiovascular disease, ie, age, sex, body mass index, smoking status, sitting systolic blood pressure, sitting heart rate, AF duration, AF type, creatinine clearance, diabetes, coronary artery disease, previous stroke/systemic embolism/TIA, heart failure, hypertension, treatment at randomization with aspirin, angiotensin-converting enzyme inhibitors or angiotensin II receptor blockers, and statins.

unknown, but elevated troponin I levels might contribute to the identification of a more fragile AF subpopulation more likely to bleed during anticoagulation. The NT-proBNP levels were not related to major bleeding risk in the present analysis. The mechanism for the prognostic value of elevated troponin I levels in patients with AF can currently only be an area for speculation. In acute coronary syndrome, troponin levels reflect necrosis of myocytes as a consequence of myocardial ischemia. In patients with chronic heart failure, stable coronary disease, apparently healthy elderly individuals, or as in this cohort of stable AF patients, there might be alternative explanations such as the increased ventricular rate that might lead to oxygen demand/mismatch and myocardial ischemia, volume and pressure overload, changes in microvascular blood flow, atrial calcium overload, oxidative stress, or alterations in tissue structure.29,32,33 The pathogenesis of thrombi in AF involves poorly contractile atrium,34 hypercoagulable state,35 and endothelial dysfunction.36 Troponin release in AF patients may be connected to several of these mechanisms associated with myocardial dysfunction, apoptosis, inflammation, and fibrosis in the atrial and ventricular musculature, as well. The higher proportion of patients with troponin elevation in permanent AF as an indicator of increased AF burden and a more advanced cardiac disease supports this hypothesis and also the relation to major bleeding events. The mechanism for the prognostic value of elevated NT-proBNP levels in AF seems easier to understand. Elevated natriuretic peptides reflect the myocyte response to increased wall tension. This is usually seen in settings of left

ventricular systolic or diastolic dysfunction, ventricular hypertrophy,37 increasing age, and female sex.38 An acute coronary syndrome, renal dysfunction, and high-output states may cause elevation of NT-proBNP as well.39 Concerning natriuretic peptides in AF, NT-proBNP level is a predictor of future development of AF, independent of other risk factors, including echocardiographic parameters in older adults.40 In accordance with the present results, the levels of NT-proBNP have previously been shown to be elevated in patients with AF, either with or without structural heart disease in comparison with matched controls in sinus rhythm.41 After restoration of sinus rhythm, either by cardioversion or ablation therapy, the level of natriuretic peptides falls rapidly.42,43 Therefore, there are arguments for NT-proBNP being of atrial origin in AF,44 in contrast to the pathophysiology of heart failure, where it is derived mainly from the ventricles.45 Some studies support this by indicating that atrial stretch is a source of BNP in patients with AF.46 The level of natriuretic peptides in AF may therefore, to some extent reflect, atrial dysfunction, which is an established marker of atrial thrombus formation,34 and thereby provide a plausible mechanism for the prognostic importance of elevated NT-proBNP levels and thromboembolic events as shown in the present study. There are several limitations of this study. The findings concern a population with nonvalvular AF with at least 1 risk factor for stroke. To extend and apply the results to AF patients without any clinical stroke risk factors, further studies in other AF populations are warranted. Furthermore, the study design, with all study participants receiving oral

Hijazi et al
anticoagulants, does not allow final conclusions concerning the optimal cutoff value of cardiac biomarkers as a decisionsupport tool for improved selections of AF patients for oral anticoagulation. The mechanisms behind the release and origin of NT-proBNP in AF patients without congestive heart failure need further clarification, and the exact mechanisms behind troponin elevations and their relations to events such as stroke and bleeding in AF, as well. Although the current troponin I assay has shown robust analytic performance in patients who have coronary artery disease and are older,12,14 newly developed assays with even higher sensitivity and precision might provide even better identification of patients at increased risk and with a potential benefit of oral anticoagulation.

Cardiac Biomarkers in Atrial Fibrillation

1615

Conclusion
Elevated levels of troponin I and NT-proBNP are common in patients with AF and at least 1 risk factor for stroke and associated with an increase in the risk for stroke or systemic embolism and vascular events. The prognostic information from the troponin I and NT-proBNP levels are independent of and additive to established clinical risk factors such as the CHADS2 and CHA2DS2-VASc risk scores.

Source of Funding
The RE-LY trial was funded by Boehringer Ingelheim Pharmaceuticals.

Disclosures
All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Hijazi reports receiving lecture fees and an institutional research grant from BoehringerIngelheim. Dr Oldgren reports receiving consulting and lecture fees, and grant support from Boehringer Ingelheim, and consultant and lecture fees from Bayer and Bristol-Myers Squibb. Ms Andersson reports no conflicts of interest. Dr Connolly reports receiving consulting fees, lecture fees, and grant support from Boehringer Ingelheim. Dr Ezekowitz reports receiving consulting fees, lecture fees, and grant support from Boehringer Ingelheim and Aryx Therapeutics; consulting fees from Sanofi-aventis; and lecture fees and grant support from Portola Pharmaceuticals. Dr Hohnloser reports receiving consulting fees and lecture fees from Boehringer Ingelheim, St. Jude Medical, and Sanofi-aventis, and lecture fees from Cardiome. Dr Reilly is an employee of Boehringer Ingelheim. Dr Vinereanu reports receiving lecture fees and grant support from Boehringer Ingelheim. Dr Siegbahn reports consulting fees, lecture fees, and grant support from Boehringer Ingelheim; lecture fees and grants from Eli Lilly; and grant support from AstraZeneca. Dr Yusuf reports receiving consulting fees, lecture fees, and grant support from Boehringer Ingelheim; and consulting fees from AstraZeneca, BristolMyers Squibb, and Sanofi-aventis. Dr Wallentin reports receiving consulting and lecture fees, honoraria, and research grants from Boehringer Ingelheim; research grants from AstraZeneca, Bristol-Myers Squibb, GlaxoSmithKline, and Schering-Plough; honoraria from AstraZeneca, Bristol-Myers Squibb, Eli Lilly, GlaxoSmithKline, and Schering-Plough; consultant fees from Athera Biotechnologies, AstraZeneca, Eli Lilly, GlaxoSmithKline, and Regado Biotechnologies; and lecture fees from AstraZeneca and Eli Lilly.

References
1. Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, Singer DE. Prevalence of diagnosed atrial fibrillation in adults: National implications for rhythm management and stroke prevention: the anticoagulation and risk factors in atrial fibrillation (atria) study. JAMA. 2001;285: 2370 2375. 2. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: The Framingham Study. Stroke. 1991;22:983988.

3. Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the national registry of atrial fibrillation. JAMA. 2001;285: 2864 2870. 4. Boomsma F, van den Meiracker AH. Plasma a- and b-type natriuretic peptides: physiology, methodology and clinical use. Cardiovasc Res. 2001;51:442 449. 5. James SK, Armstrong P, Barnathan E, Califf R, Lindahl B, Siegbahn A, Simoons ML, Topol EJ, Venge P, Wallentin L. Troponin and c-reactive protein have different relations to subsequent mortality and myocardial infarction after acute coronary syndrome: a GUSTO-IV substudy. J Am Coll Cardiol. 2003;41:916 924. 6. Lindahl B, Toss H, Siegbahn A, Venge P, Wallentin L. Markers of myocardial damage and inflammation in relation to long-term mortality in unstable coronary artery disease. FRISC study group. Fragmin during instability in coronary artery disease. N Engl J Med. 2000;343: 1139 1147. 7. Kragelund C, Grnning B, Kber L, Hildebrandt P, Steffensen R. N-terminal pro-B-type natriuretic peptide and long-term mortality in stable coronary heart disease. N Engl J Med. 2005;352:666 675. 8. Omland T, de Lemos JA, Sabatine MS, Christophi CA, Rice MM, Jablonski KA, Tjora S, Domanski MJ, Gersh BJ, Rouleau JL, Pfeffer MA, Braunwald E. A sensitive cardiac troponin T assay in stable coronary artery disease. N Engl J Med. 2009;361:2538 2547. 9. Horwich TB, Patel J, MacLellan WR, Fonarow GC. Cardiac troponin I is associated with impaired hemodynamics, progressive left ventricular dysfunction, and increased mortality rates in advanced heart failure. Circulation. 2003;108:833 838. 10. Tsutamoto T, Wada A, Maeda K, Hisanaga T, Maeda Y, Fukai D, Ohnishi M, Sugimoto Y, Kinoshita M. Attenuation of compensation of endogenous cardiac natriuretic peptide system in chronic heart failure: prognostic role of plasma brain natriuretic peptide concentration in patients with chronic symptomatic left ventricular dysfunction. Circulation. 1997;96:509 516. 11. Wang TJ, Larson MG, Levy D, Benjamin EJ, Leip EP, Omland T, Wolf PA, Vasan RS. Plasma natriuretic peptide levels and the risk of cardiovascular events and death. N Engl J Med. 2004;350:655 663. 12. Zethelius B, Johnston N, Venge P. Troponin I as a predictor of coronary heart disease and mortality in 70-year-old men: a community-based cohort study. Circulation. 2006;113:10711078. 13. Connolly SJ, Ezekowitz MD, Yusuf S, Eikelboom J, Oldgren J, Parekh A, Pogue J, Reilly PA, Themeles E, Varrone J, Wang S, Alings M, Xavier D, Zhu J, Diaz R, Lewis BS, Darius H, Diener HC, Joyner CD, Wallentin L. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139 1151. 14. Eggers KM, Lagerqvist B, Venge P, Wallentin L, Lindahl B. Persistent cardiac troponin I elevation in stabilized patients after an episode of acute coronary syndrome predicts long-term mortality. Circulation. 2007;116: 19071914. 15. Uettwiller-Geiger D, Wu AH, Apple FS, Jevans AW, Venge P, Olson MD, Darte C, Woodrum DL, Roberts S, Chan S. Multicenter evaluation of an automated assay for troponin I. Clini Chem. 2002;48:869 876. 16. Johnston N, Jernberg T, Lindahl B, Lindback J, Stridsberg M, Larsson A, Venge P, Wallentin L. Biochemical indicators of cardiac and renal function in a healthy elderly population. Clin Biochem. 2004;37: 210 216. 17. Yeo KT, Wu AH, Apple FS, Kroll MH, Christenson RH, Lewandrowski KB, Sedor FA, Butch AW. Multicenter evaluation of the Roche NT-proBNP assay and comparison to the Biosite Triage BNP assay. Clin Chim Acta. 2003;338:107115. 18. Venge P, Lagerqvist B, Diderholm E, Lindahl B, Wallentin L. Clinical performance of three cardiac troponin assays in patients with unstable coronary artery disease (a FRISC II substudy). Am J Cardiol. 2002;89: 10351041. 19. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics. 1988;44:837 845. 20. Pencina MJ, DAgostino RB Sr, DAgostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Stat Med. 2008;27:157172. 21. Pencina MJ, DAgostino RB Sr, DAgostino RB Jr, Vasan RS. Comments on integrated discrimination and net reclassification improvements practical advice. Stat Med. 2008;27:207212. 22. Antman EM, Tanasijevic MJ, Thompson B, Schactman M, McCabe CH, Cannon CP, Fischer GA, Fung AY, Thompson C, Wybenga D,

1616

Circulation

April 3, 2012
of elevated cardiac troponin I levels in a population-based sample of elderly subjects. Eur Heart J. 2008;29:22522258. Jeremias A, Gibson CM. Narrative review: alternative causes for elevated cardiac troponin levels when acute coronary syndromes are excluded. Ann Intern Med. 2005;142:786 791. The Stroke Prevention in Atrial Fibrillation Investigators Committee on Echocardiography. Transesophageal echocardiographic correlates of thromboembolism in high-risk patients with nonvalvular atrial fibrillation. The stroke prevention in atrial fibrillation investigators committee on echocardiography. Ann Intern Med. 1998;128:639 647. Asakura H, Hifumi S, Jokaji H, Saito M, Kumabashiri I, Uotani C, Morishita E, Yamazaki M, Shibata K, Mizuhashi K. Prothrombin fragment F1 2 and thrombin-antithrombin III complex are useful markers of the hypercoagulable state in atrial fibrillation. Blood Coagul Fibrinolysis. 1992;3:469 473. Conway DS, Pearce LA, Chin BS, Hart RG, Lip GY. Plasma von Willebrand factor and soluble p-selectin as indices of endothelial damage and platelet activation in 1321 patients with nonvalvular atrial fibrillation: relationship to stroke risk factors. Circulation. 2002;106: 19621967. Braunwald E. Biomarkers in heart failure. N Engl J Med. 2008;358: 2148 2159. Redfield MM, Rodeheffer RJ, Jacobsen SJ, Mahoney DW, Bailey KR, Burnett JC Jr. Plasma brain natriuretic peptide concentration: impact of age and gender. J Am Coll Cardiol. 2002;40:976 982. Daniels LB, Maisel AS. Natriuretic peptides. J Am Coll Cardiol. 2007; 50:23572368. Patton KK, Ellinor PT, Heckbert SR, Christenson RH, DeFilippi C, Gottdiener JS, Kronmal RA. N-terminal pro-B-type natriuretic peptide is a major predictor of the development of atrial fibrillation: the Cardiovascular Health Study. Circulation. 2009;120:1768 1774. Shelton RJ, Clark AL, Goode K, Rigby AS, Cleland JG. The diagnostic utility of N-terminal pro-B-type natriuretic peptide for the detection of major structural heart disease in patients with atrial fibrillation. Eur Heart J. 2006;27:23532361. Wozakowska-Kaplon B. Effect of sinus rhythm restoration on plasma brain natriuretic peptide in patients with atrial fibrillation. Am J Cardiol. 2004;93:15551558. Yamada T, Murakami Y, Okada T, Okamoto M, Shimizu T, Toyama J, Yoshida Y, Tsuboi N, Ito T, Muto M, Kondo T, Inden Y, Hirai M, Murohara T. Plasma atrial natriuretic peptide and brain natriuretic peptide levels after radiofrequency catheter ablation of atrial fibrillation. Am J Cardiol. 2006;97:17411744. Goetze JP, Friis-Hansen L, Rehfeld JF, Nilsson B, Svendsen JH. Atrial secretion of B-type natriuretic peptide. Eur Heart J. 2006;27:1648 1650. Yasue H, Yoshimura M, Sumida H, Kikuta K, Kugiyama K, Jougasaki M, Ogawa H, Okumura K, Mukoyama M, Nakao K. Localization and mechanism of secretion of B-type natriuretic peptide in comparison with those of A-type natriuretic peptide in normal subjects and patients with heart failure. Circulation. 1994;90:195203. Inoue S, Murakami Y, Sano K, Katoh H, Shimada T. Atrium as a source of brain natriuretic polypeptide in patients with atrial fibrillation. J Card Fail. 2000;6:9296.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

Braunwald E. Cardiac-specific troponin I levels to predict the risk of mortality in patients with acute coronary syndromes. N Engl J Med. 1996;335:13421349. Zethelius B, Berglund L, Sundstrom J, Ingelsson E, Basu S, Larsson A, Venge P, Arnlov J. Use of multiple biomarkers to improve the prediction of death from cardiovascular causes. N Engl J Med. 2008;358:21072116. de Lemos JA, Morrow DA, Bentley JH, Omland T, Sabatine MS, McCabe CH, Hall C, Cannon CP, Braunwald E. The prognostic value of B-type natriuretic peptide in patients with acute coronary syndromes. N Engl J Med. 2001;345:1014 1021. James SK, Lindahl B, Siegbahn A, Stridsberg M, Venge P, Armstrong P, Barnathan ES, Califf R, Topol EJ, Simoons ML, Wallentin L. N-terminal pro-brain natriuretic peptide and other risk markers for the separate prediction of mortality and subsequent myocardial infarction in patients with unstable coronary artery disease: a Global Utilization of strategies To Open occluded arteries (GUSTO)-IV substudy. Circulation. 2003; 108:275281. Fonarow GC, Peacock WF, Phillips CO, Givertz MM, Lopatin M. Admission B-type natriuretic peptide levels and in-hospital mortality in acute decompensated heart failure. J Am Coll Cardiol. 2007;49: 19431950. Sabatine MS, Morrow DA, de Lemos JA, Gibson CM, Murphy SA, Rifai N, McCabe C, Antman EM, Cannon CP, Braunwald E. Multimarker approach to risk stratification in non-ST elevation acute coronary syndromes: simultaneous assessment of troponin I, C-reactive protein, and B-type natriuretic peptide. Circulation. 2002;105:1760 1763. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263272. Camm AJ, Kirchhof P, Lip GY, Schotten U, Savelieva I, Ernst S, Van Gelder IC, Al-Attar N, Hindricks G, Prendergast B, Heidbuchel H, Alfieri O, Angelini A, Atar D, Colonna P, De Caterina R, De Sutter J, Goette A, Gorenek B, Heldal M, Hohloser SH, Kolh P, Le Heuzey JY, Ponikowski P, Rutten FH, Vahanian A, Auricchio A, Bax J, Ceconi C, Dean V, Filippatos G, Funck-Brentano C, Hobbs R, Kearney P, McDonagh T, Popescu BA, Reiner Z, Sechtem U, Sirnes PA, Tendera M, Vardas PE, Widimsky P, Agladze V, Aliot E, Balabanski T, Blomstrom-Lundqvist C, Capucci A, Crijns H, Dahlof B, Folliguet T, Glikson M, Goethals M, Gulba DC, Ho SY, Klautz RJ, Kose S, McMurray J, Perrone Filardi P, Raatikainen P, Salvador MJ, Schalij MJ, Shpektor A, Sousa J, Stepinska J, Uuetoa H, Zamorano JL, Zupan I. Guidelines for the management of atrial fibrillation: the task force for the management of atrial fibrillation of the European Society of Cardiology (ESC). Eur Heart J. 2010;31: 2369 2429. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJ, Lip GY. A novel user-friendly score (HAS-BLED) to assess one-year risk of major bleeding in atrial fibrillation patients: The Euro Heart Survey. Chest. 2010:10931100. Al-Mallah M, Bazari RN, Jankowski M, Hudson MP. Predictors and outcomes associated with gastrointestinal bleeding in patients with acute coronary syndromes. J Thromb Thrombolysis. 2007;23:5155. Eggers KM, Lind L, Ahlstrom H, Bjerner T, Ebeling Barbier C, Larsson A, Venge P, Lindahl B. Prevalence and pathophysiological mechanisms

33.

34.

35.

36.

37. 38.

39. 40.

41.

42.

43.

44. 45.

46.

CLINICAL PERSPECTIVE
Troponin I and N-terminal pro-B-type natriuretic peptide (NT-proBNP) are widely available biomarker analyses for diagnosis and prognostication of cardiac diseases. We demonstrated that elevated levels of troponin I and NT-proBNP were common in patients with atrial fibrillation, with troponin elevations observed in 25% of the patients and NT-proBNP elevations in almost 75% of the patients. Elevated levels of any or both of these cardiac biomarkers provided incremental prognostic information to the currently used CHADS2 and CHA2DS2-VASc risk scores. Beyond clinical risk factors, an elevated level of troponin I or NT-proBNP identifies patients with a doubled risk for stroke or systemic embolism, and up to 6-fold increased risk of vascular death. Determination of troponin I or NT-proBNP may be used to improve risk prediction in patients with atrial fibrillation beyond currently used clinical risk scores.