ORIGINAL RESEARCH

Role of B-Type Natriuretic Peptide and N-Terminal Prohormone BNP as Predictors of Cardiovascular Morbidity and Mortality in Patients With a Recent Coronary Event and Type 2 Diabetes Mellitus
Emil Wolsk, MD, PhD; Brian Claggett, PhD; Marc A. Pfeffer, MD, PhD; Rafael Diaz, MD; Kenneth Dickstein, MD, PhD; Hertzel C. Gerstein, MD; Francesca C. Lawson, MD; Eldrin F. Lewis, MD, MPH; Aldo P. Maggioni, MD; John J. V. McMurray, MD, PhD; Jeffrey L. Probstfield, MD; Matthew C. Riddle, MD; Scott D. Solomon, MD; Jean-Claude Tardif, MD; Lars Køber, MD
Background-—Natriuretic peptides are recognized as important predictors of cardiovascular events in patients with heart failure, but less is known about their prognostic importance in patients with acute coronary syndrome. We sought to determine whether B-type natriuretic peptide (BNP) and N-terminal prohormone B-type natriuretic peptide (NT-proBNP) could enhance risk prediction of a broad range of cardiovascular outcomes in patients with acute coronary syndrome and type 2 diabetes mellitus.
Methods and Results-—Patients with a recent acute coronary syndrome and type 2 diabetes mellitus were prospectively enrolled in the ELIXA trial (n=5525, follow-up time 26 months). Best risk models were constructed from relevant baseline variables with and without BNP/NT-proBNP. C statistics, Net Reclassification Index, and Integrated Discrimination Index were analyzed to estimate the value of adding BNP or NT-proBNP to best risk models. Overall, BNP and NT-proBNP were the most important predictors of all outcomes examined, irrespective of history of heart failure or any prior cardiovascular disease. BNP significantly improved C statistics when added to risk models for each outcome examined, the strongest increments being in death (0.77–0.82, P<0.001), cardiovascular death (0.77–0.83, P<0.001), and heart failure (0.84–0.87, P<0.001). BNP or NT-proBNP alone predicted death as well as all other variables combined (0.77 versus 0.77).
Conclusions-—In patients with a recent acute coronary syndrome and type 2 diabetes mellitus, BNP and NT-proBNP were powerful predictors of cardiovascular outcomes beyond heart failure and death, ie, were also predictive of MI and stroke. Natriuretic peptides added as much predictive information about death as all other conventional variables combined.
Clinical Trial Registration-—URL: http://www.clinicaltrials.gov. Unique identifier: NCT01147250. ( J Am Heart Assoc. 2017;6: e004743. DOI: 10.1161/JAHA.116.004743.)
Key Words: acute coronary syndrome • biomarker • brain natriuretic peptide • cardiac outcomes • diabetes mellitus • Evaluation of Lixisenatide in Acute Coronary Syndrome trial • glucagon-like peptide-1 • natriuretic peptide • N-terminal prohormone B-type natriuretic peptide • risk model

P atients admitted with an acute coronary syndrome (ACS) are at increased risk of subsequent cardiovascular events, especially those with type 2 diabetes mellitus,1,2

who constitute %30% of all ACS patients.3 Determining the predictors of death, myocardial infarction (MI), heart failure (HF), and stroke among these patients is important as it

From the Division of Cardiovascular Medicine, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA (E.W., B.C., M.A.P., E.F.L., S.D.S.); Department of Cardiology, Rigshospitalet, Copenhagen, Denmark (E.W., L.K.); Estudios Clınicos Latinoamerica, Rosario, Argentina (R.D.); University of Bergen, Stavanger University Hospital, Stavanger, Norway (K.D.); Division of Endocrinology & Metabolism, McMaster University, Hamilton, Ontario, Canada (H.C.G.); Sanofi U.S., Bridgewater, NJ (F.C.L.); Research Center of the Italian Association of Hospital Cardiologists, Florence, Italy (A.P.M.); British Heart Foundation Cardiovascular Research Centre, University of Glasgow, United Kingdom (J.J.V.M.); Division of Cardiology, University of Washington Medical Center, Seattle, WA (J.L.P.); Division of Endocrinology, Oregon Health and Science University, Portland, OR (M.C.R.); Montreal Heart Institute, Universite de Montreal, Montreal, Canada (J.-C.T.).
Accompanying Tables S1 through S13 and Figure S1 are available at http://jaha.ahajournals.org/content/6/6/e004743/DC1/embed/inline-supplementary-mate rial-1.pdf
Correspondence to: Lars Køber, MD, Department of Cardiology, Heart Centre, Rigshospitalet, University of Copenhagen, Blegdamsvej 9, Copenhagen DK-2100, Denmark. E-mail: lars.koeber@regionh.dk
Received November 17, 2016; accepted March 2, 2017.
ª 2017 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley Blackwell. This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 1

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

ORIGINAL RESEARCH

assists in identifying the individuals at highest risk of these various outcomes. These individuals should be the focus of the most intensive secondary preventive strategies. In the same way, risk stratification may help motivate both patients and clinicians in secondary preventive efforts. Lastly, predictive models can be used to select the highest risk individuals for trials of new secondary preventive therapies.
B-type natriuretic peptide (BNP) and N-terminal prohormone B-type natriuretic peptide (NT-proBNP) levels are well established predictors of HF hospitalization and mortality in patients with HF.4–7 However, as the incidences of outcomes differ between patients with HF8 and patients with a recent ACS9 in a stable phase, it is less established how natriuretic peptides are associated with cardiovascular outcomes, especially subsequent MI and stroke, in the latter population.3,10–14 Concentrations of natriuretic peptides may be affected by both asymptomatic myocardial ischemia15,16 and atrial fibrillation,17,18 which could make BNP and NT-proBNP relevant as predictors of MI and stroke. However, trials investigating patients at high risk of atherosclerotic events have found conflicting results regarding both the predictive ability of natriuretic peptides and cardiovascular outcomes— including MI and stroke—as well as the predictive strength of BNP versus NT-proBNP.19–22
We wanted to expand knowledge of these natriuretic peptides as predictors of death, cardiovascular death, myocardial infarction, heart failure, and stroke in patients with a recent coronary event and type 2 diabetes mellitus enrolled in the Evaluation of Lixisenatide in Acute Coronary Syndrome trial (ELIXA, NCT01147250). Data from the ELIXA trial allowed us to compare the predictive strength of baseline BNP and NT-proBNP in a high risk ACS patient cohort with type 2 diabetes mellitus. In addition, prospective ascertainment during a reasonable follow-up period, and adjudication of a variety of cardiovascular events, ensured detailed and validated data for analyses.
Methods
The ELIXA trial (Evaluation of Lixisenatide in Acute Coronary Syndrome) trial included 6068 patients with type 2 diabetes mellitus and an acute coronary event within 180 days from randomization (index event).23 The study was approved by the appropriate national and institutional regulatory and ethics boards, and all subjects gave informed consent. The objective of the ELIXA trial was to assess the safety and efficacy of lixisenatide, a glucagon-like peptide-1 receptor agonist, on cardiovascular morbidity and mortality. Details of the trial design and the demographic and clinical characteristics of the included patients have been reported previously.24 In summary, patients were included in this randomized, double-blind, placebo-controlled, parallel-group

study, between 2010 and 2013, from 49 countries, and followed for a median of 25 months. Key exclusion criteria were percutaneous coronary intervention within 15 days of screening or planned percutaneous coronary intervention within 90 days after screening, coronary artery bypass graft treatment at the index event, an estimated glomerular filtration rate of less than 30 mL/min per 1.73 m2 of bodysurface area, a glycated hemoglobin level of less than 5.5% or more than 11.0%, or an inability to provide written informed consent. Patients were randomized to subcutaneous injections of either lixisenatide (maximum 20 lg daily) or placebo (volume matched) in addition to locally determined standards of care. The ELIXA trial showed that lixisenatide had a neutral effect with regard to the occurrence of the primary outcome (cardiovascular death, MI, stroke, or hospitalization for unstable angina) and HF hospitalization.23
Covariates and Outcomes
All data pertaining to baseline variables including demographics, anthropometrics, cardiovascular risk factors, and prior medical history were obtained at the time of randomization in the study. All events were reported to a centralized and independent adjudication committee who classified events according to prespecified definitions.24 Data on adjudicated time-to-event for outcomes of all-cause death (death), cardiovascular death (cardiovascular death), heart failure hospitalization (HF), fatal and nonfatal myocardial infarction (MI), and fatal and nonfatal stroke (stroke) were used for analyses.
BNP and NT-proBNP sampling was carried out at baseline. Samples were collected and analyzed at a core laboratory (Covance Central Laboratory Services, Meyrin, Switzerland). The Triage BNP assay was used to analyze BNP. The intraassay coefficient of variation was 1.1% to 3.1%. The interassay coefficient of variation was 1.8% to 6.6%. The Immulite NTproBNP assay was used to analyze NT-proBNP. The intraassay coefficient of variation was 2.3% to 5.4%. The interassay coefficient of variation was 4.0% to 6.4%. BNP and NT-proBNP samples from 5925 patients (98%) were obtained.
Statistical Analyses
Baseline characteristics shown in Table 1 were selected for best risk models. Patients without data on all these relevant variables, including BNP and NT-proBNP measurements, were excluded (n=543, 9%). The distributions of baseline BNP, NT-proBNP, and C-reactive protein were found to be rightskewed and were therefore log-transformed prior to analysis. Continuous variables were included in the models unless there was clear evidence of nonlinearity.

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 2

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

Table 1. Characteristics of All Included Patients

Randomized to lixisenatide Age, y Male (%) BMI, kg/m2 Race
Asian Black Other White Ethnicity—Hispanic Region Africa/Near East Asia Pacific Eastern Europe North America South and Centr. America Western Europe Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Heart rate, bpm Current smoker Former smoker Medical history MI HF Atrial fibrillation/flutter PAD TIA Ventricular tachycardia Stroke CABG Implanted pacemaker Carotid disease Hypertension Index event STEMI NSTEMI Unstable angina pectoris PCI at index event Insulin-treated Duration of diabetes mellitus, y Retinopathy

No Cardiovascular Events (n=4626)
2327 (50.3%) 59.7Æ9.5 3238 (70.0%) 30.0Æ5.6
648 (14.0%) 14 (3.1%) 385 (8.3%) 3451 (74.6%) 1396 (30.2%)
215 (4.6%) 597 (12.9%) 1172 (25.3%) 563 (12.2%) 1551 (33.5%) 528 (11.4%) 129Æ17 77Æ10 70Æ10 511 (11.0%) 2113 (45.7%)
918 (19.8%) 905 (19.5%) 240 (5.2%) 271 (5.9%) 83 (1.8%) 57 (1.2%) 201 (4.3%) 309 (6.7%) 102 (2.2%) 87 (1.9%) 3449 (74.6%)
2146 (46.4%) 1702 (36.8%) 778 (16.8%) 2943 (63.6%) 1699 (36.7%) 8.8Æ7.9 452 (9.8%)

DOI: 10.1161/JAHA.116.004743

Cardiovascular Events (n=899)
449 (49.9%) 63.3Æ9.7 627 (69.7%) 30.3Æ6.1
82 (9.1%) 47 (5.2%) 86 (9.6%) 684 (76.1%) 250 (27.8%)
54 (6.0%) 72 (8.0%) 241 (26.8%) 153 (17.0%) 273 (30.4%) 106 (11.8%) 131Æ19 77Æ11 71Æ11 117 (13.0%) 409 (45.5%)

P Value
0.84 <0.001 0.88 0.24 <0.001
0.16 <0.001
0.86 0.18 0.001 0.09 0.92

340 (37.7%) 330 (36.7%) 121 (13.5%) 142 (15.8%) 44 (4.9%) 17 (1.9%) 91 (10.1%) 151 (16.8%) 41 (4.6%) 52 (5.8%) 761 (84.6%)
297 (33.0%) 432 (48.1%) 170 (18.9%) 463 (51.5%) 460 (51.2%) 11.9Æ9.5 139 (15.5%)

<0.001 <0.001 <0.001 <0.001 <0.001 0.12 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001

<0.001 <0.001 <0.001 <0.001

Continued

Journal of the American Heart Association 3

ORIGINAL RESEARCH

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

ORIGINAL RESEARCH

Table 1. Continued

Neuropathy Asthma COPD HbA1c, % HDL, mg/dL LDL, mg/dL eGFR, mL/min per 1.73 m2 Albuminuria
Normoalbuminuria Microalbuminuria Macroalbuminuria Hemoglobin, g/dL Na, mmol/L Albumin, g/dL CRP, mg/dL BNP, pg/mL NT-proBNP, pg/mL

No Cardiovascular Events (n=4626)
714 (15.4%) 114 (2.5%) 173 (3.7%) 7.6Æ1.3 43Æ11 77Æ34 77.5Æ21.1
3558 (76.9%) 829 (17.9%) 239 (5.2%) 13.8Æ1.4 140.4Æ2.9 4.1Æ0.3 2.0 (1.9–2.0) 95 (92–98) 285 (274–295)

Cardiovascular Events (n=899)
205 (22.8%) 40 (4.4%) 76 (8.5%) 7.9Æ1.3 43Æ11 83Æ39 68.1Æ20.6
544 (60.5%) 234 (26.0%) 121 (13.5%) 13.5Æ1.5 140.3Æ3.1 3.9Æ0.4 2.7 (2.4–2.9) 198 (184–213) 703 (644–766)

P Value
<0.001 <0.001 <0.001 <0.001 0.65 <0.001 <0.001 <0.001
<0.001 0.75 <0.001 <0.001 <0.001 <0.001

BMI indicates body mass index; bpm, beats per minute; BNP, B-type natriuretic peptide; CABG, coronary artery bypass grafting; COPD, chronic obstructive pulmonary disease; CRP, Creactive protein; eGFR, estimated glomerular filtration rate; HbA1c, glycated hemoglobin; HDL, high-density lipoproteins; HF, heart failure; LDL, low-density lipoproteins; MI, myocardial infarction; NSTEMI, non-ST elevation myocardial infarction; NT-proBNP, N-terminal prohormone B-type natriuretic peptide; PAD, peripheral artery disease; PCI, percutaneous coronary intervention; STEMI, ST elevation myocardial infarction; TIA, transient ischemic attack.

Cox proportional hazard modeling was used to create best risk models without BNP or NT-proBNP using forward selection with a cut-off value of 0.05. Separate base risk models were created for the following outcomes; death, cardiovascular death, fatal or nonfatal MI, fatal or nonfatal HF hospitalization, as well as fatal or nonfatal stroke. The variables selected were ordered according to their v2 value and sorted in descending order for each outcome. The predictive ability of base risk models were assessed using Harrell’s C statistics. Selected 30-day models were made for comparison with previous studies. Using the selected variables from the base model, comparison between the predictive ability of the base model compared to the base model with log2BNP/log2NT-proBNP was assessed for all outcomes. Changes in C statistics, Net Reclassification Index (NRI), and Integrated Discrimination Index (IDI) were estimated to evaluate the incremental value of adding BNP or NT-proBNP to best risk models using a set time of 2 years comparable to the average follow-up time (somersd package, STATA 13. survIDINRI package, R 2.3.2).
Identification of baseline variables independently associated with BNP/NT-proBNP were obtained using forward selection regression models with P<0.001 as a cut-off. The 5 variables with the highest v2 value were listed along with the r2 values.

To identify the most significant predictive threshold of BNP/NT-proBNP values, we divided the continuous BNP/ NT-proBNP concentrations into arbitrary threshold concentrations (ie, 35, 100, 125, 200, 300. . .1000, 2000, 5000). Then using HF hospitalization as an outcome, a fully adjusted Cox model with forward selection identified the BNP threshold concentration that most significantly separated patients into a lower versus a higher risk group. A univariate approach was also carried out using receiver operating characteristic analysis (SENSPEC package, STATA 13) with binary outcome of HF hospitalization to determine the optimal cut-off value with respect to Youden index (ie, sensitivity+specificityÀ1). Interaction analyses between natriuretic peptides and timing of the baseline sample in temporal relation to the index ACS event for outcome of death was also performed, as was interaction between natriuretic peptides and type of ACS index event (STelevation MI [STEMI], Non-ST elevation MI [NSTEMI], Unstable Angina Pectoris [UAP]).
To assess whether the relationship between baseline BNP/ NT-proBNP and hazard was linear, fully adjusted Cox spline models for each outcome with transformed BNP/NT-proBNP were analyzed. Concentrations below 35 pg/mL (BNP) or 125 pg/mL (NT-proBNP) were considered normal, as these concentrations are commonly referenced as diagnostic

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 4

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al
thresholds for excluding HF in patients presenting in a nonacute manner.25 Hence, the risk of death in patients with these levels were used as references. Both unadjusted and adjusted models were used to confirm findings. Risk modeling, including interaction analyses (history of HF and log2BNP/log2NT-proBNP), discriminatory statistics, and Cox spline models were also analyzed in patients stratified according to medical history of HF for BNP.
All BNP/NT-proBNP values were summarized as geometric meanÆ95% CI. A significance level of 0.05 was considered statistically significant.
Results
Baseline Characteristics
Our study included 5525 patients comprising 91% of the included patients in the ELIXA trial. The median follow-up time was 26 months. In our population, 4626 (84%) patients did not experience any cardiovascular event confirmed by adjudication. Baseline characteristics of patients with or without a cardiovascular event are listed in Table 1 (Baseline characteristics according BNP quartiles and as linear covariates are listed in Tables S1 and S2). Compared with patients not experiencing a cardiovascular event, those that did were in general older and more burdened with comorbidity, and were more likely to have micro- or macroalbuminuria, and a lower estimated glomerular filtration rate. Blood pressure was similar in both groups. Baseline BNP and NT-proBNP were elevated in those subsequently experiencing any cardiovascular event.
Predictive Variables
In separate models, BNP and NT-proBNP were the most significant predictors for each of death from any cause, death from a cardiovascular cause, HF, and stroke among the studied variables. The natriuretic peptides were the second most significant predictors for MI (Tables 2 and 3). Apart from BNP/NT-proBNP, the 14 other variables that conferred the greatest information were the following: Prior MI, body mass index, NSTEMI (index event), heart rate (HR), glycated hemoglobin, percutaneous coronary intervention at the index event (percutaneous coronary intervention), cerebrovascular disease (prior stroke/transient ischemic attack), atrial fibrillation, prior HF, sodium concentration, macroalbuminuria, peripheral artery disease (PAD), age, and LDL concentration. Fifteen and 16 variables were independently associated with concentrations of BNP/NT-proBNP at the a=0.001 level and accounted for 26% and 34% of patient-level variability, respectively. The 5 strongest associated variables are listed in Tables S3 through S5.
DOI: 10.1161/JAHA.116.004743

Table 2. Predictors of Outcomes Ranked 1 to 5 According to v2 Value Using Base Variables and BNP (n=5525)

Outcome
Death (397 events) Cardiovascular death (286 events) MI (473 events) HF (221 events) Stroke (115 events)

1st
log2BNP (v2:203, HR 1.67) log2BNP (v2:201, HR 1.82) Prior MI (v2:46, HR 1.96) log2BNP (v2:135, HR 1.80) log2BNP (v2:23, HR 1.35)

2nd
AF (v2:11, HR 1.60) HbA1c (v2:13, HR 1.18) log2BNP (v2:44, HR 1.23) BMI per 5 (v2:34, HR 1.33) Prior TIA (v2:13, HR 3.12)

3rd
NSTEMI (v2:11, HR 1.41) AF (v2:13, HR 1.79) NSTEMI (v2:29, HR 1.66) HR per 10 (v2:17, HR 1.30) Macroalbuminuria (v2:11, HR 2.35)

4th
Na* (v2:10, HR 1.08) NSTEMI (v2:9, HR 1.46) Prior stroke (v2:13, HR 1.69) Prior HF (v2:13, HR 1.82) Age per 10 (v2:9, HR 1.36)

5th
HR per 10 (v2:10, HR 1.17) HR per 10 (v2:9, HR 1.19) PAD (v2:10, HR 1.52) Prior MI (v2:12, HR 1.67) LDL per 10 (v2:4, HR 1.06)

Hazard ratios reflect 1 unit changes if nothing else is stated. For log2BNP that translates into a doubling of the untransformed BNP concentrations. Macroalbuminuria: >300 mg albumin excretion/24 hours. All variables are significant, P<0.05. AF indicates atrial fibrillation/flutter; BNP, B-type natriuretic peptide; HbA1c, glycated hemoglobin; HR, heart rate; LDL, low-density lipoproteins; MI, myocardial infarction; NSTEMI, non-ST elevation myocardial infarction at index event; PAD, peripheral artery disease; TIA, transient ischemic attack. *1 mmol/L decreases below 140 mmol/L.

Journal of the American Heart Association 5

ORIGINAL RESEARCH

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al
Enhanced Prediction With BNP and NT-proBNP
To estimate the predictive strength of BNP alone, C statistics were compared between base models without BNP versus BNP alone. This showed that the discriminatory ability of base models without BNP versus BNP was similar in outcomes of death (Harrell’s C statistics: 0.77 both models [0–30 days: 0.82 versus 0.88, P=0.26]) and in cardiovascular death (Harrell’s C statistics: 0.77 versus 0.79, P=0.17).
In contrast, BNP was significantly less discriminatory compared with best risk models without BNP for the outcomes of MI (Harrell’s C statistics: 0.71 versus 0.62), HF (Harrell’s C statistics: 0.84 versus 0.77) and stroke (Harrell’s C statistics: 0.75 versus 0.67) (all P≤0.01). Similar estimates and trends were evident for NT-proBNP.
The strength of BNP and NT-proBNP as contributors to risk prediction translated into augmented predictive ability of risk models for all outcomes, as summarized in the increases in C statistics, NRI, and Integrated Discrimination Index (Table 4). Although BNP and NT-proBNP did not increase the C statistics for the risk model for stroke, both peptides improved NRI significantly.
BNP and NT-proBNP were the most predictive variables in risk models of type of cardiovascular death; however, when added to best risk models, BNP and NT-proBNP only significantly improved the predictive ability (Harrel’s C statistics) in outcomes of fatal HF and sudden death; fatal HF (n=39) (Base model: 0.850, BNP: +0.085, NT-proBNP: +0.086, both P<0.001), and sudden death (n=116) (Base model: 0.773, BNP: +0.024, NT-proBNP: +0.037, P=0.20 and 0.046, respectively), but not in outcome of fatal MI (n=52) (Base model: 0.827, BNP: +0.025, NT-proBNP: +0.016, P=0.24 and 0.37, respectively).
In univariate analysis, a BNP concentration of 228 pg/mL best separated patients into a lower versus higher group at risk of subsequent HF (sensitivity: 0.62, specificity: 0.77, Youden index: 0.39), with a corresponding threshold for NTproBNP of 751 pg/mL (sensitivity: 0.67, specificity: 0.74, Youden index: 0.41). In adjusted Cox models, a BNP concentration of 500 pg/mL provided the most significant threshold by which to further identify patients at lower versus higher risk of subsequent HF (HR 3.0 [2.1–4.1], P<0.0001), with a corresponding threshold for NT-proBNP of 700 pg/mL (HR 2.5 [1.7–3.5], P<0.0001).
Predictive Strength of BNP Compared to NTproBNP
There was no significant increase in C statistics when BNP was included in the best risk models compared to NT-proBNP in outcomes of death +0.002 (P=0.55), cardiovascular death +0.0002 (P=0.97), MI +0.002 (P=0.50), HF +0.0001 (P=0.98),
DOI: 10.1161/JAHA.116.004743

Table 3. Predictors of Outcomes Ranked 1 to 5 According to v2 Value Using Base Variables and NT-proBNP (n=5525)

Outcome
Death (397 events) Cardiovascular death
(286 events) MI (473 events) HF (221 events) Stroke (115 events)

1st
log2NT-proBNP (v2:215, HR 1.52) log2NT-proBNP (v2:226, HR 1.65) Prior MI (v2:49, HR 2.00) log2NT-proBNP (v2:132, HR 1.61) log2NT-proBNP (v2:19, HR 1.25)

2nd
NSTEMI (v2:10, HR 1.40) HbA1c (v2:14, HR 1.20) log2NT-proBNP (v2:37, HR 1.17) BMI per 5 (v2:31, HR 1.31) Prior TIA (v2:13, HR 3.09)

3rd
PCI (v2:10, HR 0.71) NSTEMI (v2:10, HR 1.47) NSTEMI (v2:29, HR 1.66) NSTEMI (v2:14, HR 1.90) Macroalbuminuria (v2:10, HR 2.34)

4th
DBP* (v2:8, HR 1.02) AF (v2:9, HR 1.62) Prior stroke (v2:13, HR 1.71) Prior HF (v2:14, HR 1.73) Age per 10 (v2:7, HR 1.32)

5th
Na† (v2:8, HR 1.04) Prior HF (v2:8, HR 1.44) PAD (v2:9, HR 1.48) Prior MI (v2:12, HR 1.69) LDL per 10 (v2:4, HR 1.05)

Hazard ratios reflect 1 unit changes if nothing else is stated. For log2NT-proBNP that translates into a doubling of the untransformed NT-proBNP concentrations. Macroalbuminuria: >300 mg albumin excretion/24 hours. All variables are significant, P<0.05. AF indicates atrial fibrillation/flutter; BMI, body mass index; HbA1c, glycated hemoglobin; HF, heart failure; HR, heart rate; LDL, low-density lipoproteins; MI, myocardial infarction; NT-proBNP, N-terminal prohormone B-type natriuretic peptide; NSTEMI, non-ST elevation myocardial infarction at index event; PAD, peripheral artery disease; PCI, percutaneous coronary intervention; TIA, transient ischemic attack. *1 mm Hg decreases below 70 mm Hg. †1 mmol/L decreases below 140 mmol/L.

Journal of the American Heart Association 6

ORIGINAL RESEARCH

ORIGINAL RESEARCH

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

Table 4. Discriminatory Changes in Best Risk Models With and Without BNP and NT-proBNP

C Statistics in Each Model (n=5525)

Death (397 events) BNP NT-proBNP

Base Model
0.77 (74–0.79)

BNP/NT-proBNP in Model
0.82 (0.80–0.84)* 0.81 (0.79–0.83)*

Cardiovascular death (286 events) BNP 0.77 (0.74–0.80) NT-proBNP
MI (473 events)

0.83 (0.81–0.86)* 0.83 (0.80–0.86)*

BNP NT-proBNP

0.71 (0.68–0.73)

0.72 (0.70–0.75)* 0.72 (0.67–0.74)*

HF (221 events) BNP NT-proBNP

0.84 (0.81–0.86)

0.87 (0.85–0.89)* 0.87 (0.84–0.89)*

Stroke (115 events)

BNP NT-proBNP

0.74 (0.70–0.79)

0.76 (0.72–0.80) 0.76 (0.72–0.80)

NRI
30.6% (25.2–36.8)* 24.3% (17.4–29.3)*
36.0% (27.4–41.4)* 30.9% (21.7–36.9)*
14.3% (9.3–19.5)* 10.6% (5.7–16.6)*
35.4% (24.7–40.6)* 29.9% (21.8–36.6)*
19.3% (8.8–29.9)* 17.2% (6.3–28.1)*

IDI
5.0% (3.5–7.2)* 3.3% (2.1–5.0)*
5.6% (3.4–8.6)* 4.0% (2.4–6.3)*
1.2% (0.6–2.1)* 0.8% (0.3–1.6)*
5.0% (3.0–7.6)* 3.8% (2.2–5.8)*
0.4% (0–1.2) 0.2% (0–0.8)

NRI and IDI summarized as mean percent improvement Æ95% CI. BNP indicates B-type natriuretic peptide; HF, heart failure; IDI, Integrated Discrimination Index; MI, myocardial infarction; NRI, Net Reclassification Index; NT-proBNP, N-terminal prohormone B-type natriuretic peptide. *P<0.05, comparison between base model and BNP/NT-proBNP model.

stroke +0.002 (P=0.88). The comparable estimates were confounded by a significant correlation between BNP and NTproBNP (Spearman’s rho 0.86, P<0.0001).
Subgroup and Sensitivity Analyses
In our population, 22% had a history of HF. As sensitivity analysis, patients were stratified according to history of HF at baseline to provide ranking and estimates of important risk factors, although there was no significant interaction with natriuretic peptides and history of HF (death; BNP, P=0.57. NT-proBNP, P=0.21, Figure S1). BNP was the strongest prognostic variable for all outcomes examined in both groups (Æprior HF; Tables S6 through S8). Sensitivity analysis was also done in the subset of patients (52%) without any prior cardiovascular disease at baseline (HF, atrial fibrillation, peripheral artery disease, transient ischemic attack/stroke, ventricular tachycardia, coronary artery bypass graft or MI apart from index event). The same trends were also evident in this subset (Tables S9 and S10).
Analysis in the subset of patients that had information on left ventricular ejection fraction present at their index ACS was performed (n=3390). Left ventricular ejection fraction was not among the 3 most significant predictors across all outcomes when added to the variable list. The predictive ability of BNP and NT-proBNP was comparable in

this subset compared to the entire cohort (Tables S11 through S13).
The timing of the sampling in relation to the index ACS event did not affect the risk estimates for death for BNP (P=0.63) or NT-proBNP (P=0.46), nor did the type of ACS (BNP, P=0.30; NT-proBNP, P=0.32).
There was no significant interaction between sex and concentrations of BNP (P=0.17) or NT-proBNP (P=0.58) when tested in a fully adjusted model.
Discussion
Our goal was to examine the strength of BNP and NT-proBNP in predicting a range of cardiovascular outcomes in high risk ACS patients with type 2 diabetes mellitus. We found that baseline levels of these natriuretic peptides were elevated in patients with a subsequent cardiovascular event during follow-up compared with those not having an event. The levels of natriuretic peptides most likely reflect that all patients recently suffered a coronary event and on average had a high comorbid burden.
The significance of elevated natriuretic peptides was reiterated when BNP and NT-proBNP were added to risk models of major cardiovascular outcomes. Ranked according to the strength of prediction, BNP and NT-proBNP were the primary predictive variables in all outcomes examined, except

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 7

ORIGINAL RESEARCH

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

MI, where they were the second most predictive variable. The predictive strength of BNP and NT-proBNP was also evident when viewing the v2 values, which were magnitudes higher than other significant variables in death, cardiovascular death and HF, moderately higher in stroke, and comparable to having had a prior MI in outcome of MI. This was also mirrored in the ability of BNP and NT-proBNP to predict causes of death, where a significant predictive contribution of BNP and NT-proBNP was present when added to risk models in outcome of sudden death and HF death, whereas none was found for fatal MI. The cut-off point that most significantly divided patients into a higher versus lower risk group of subsequent HF hospitalization was 500 and 700 pg/mL for BNP and NT-proBNP, in multivariate analysis.
Natriuretic peptides are primarily recognized as predictors of mortality and HF hospitalization, whereas our finding of a strong predictive ability of natriuretic peptide levels in outcomes of MI and stroke is less validated, especially in ACS patients. This ability could be attributed to higher levels of natriuretic peptides in patients with asymptomatic myocardial ischemia15,16 and paroxysmal atrial fibrillation,17,18 which could predispose to both MI and stroke.
Our risk models also identified other important predictors apart from BNP and NT-proBNP. Risk models in other diabetic populations (TREAT [Trial to Reduce Cardiovascular Events

with Aranesp (darbepoetin-alfa) Therapy]),26 UKPDS [UK Prospective Diabetes Study]27) have yielded results that also highlight the importance of the risk factors we identified, such as prior HF, glycated hemoglobin, age, heart rate, albuminuria, and cardiac arrhythmias.28,29 Important differences were that these studies either used composite end points28 or only single outcomes29 when examining predictors. Furthermore, the diabetic patients on which these risk models were based were either at higher risk (TREAT: 81.1 deaths per 1000 PY) or lower risk of death (UKPDS: 18.9 deaths per 1000 PY) than in the present study (32.3 deaths per 1000 PY). Nonetheless, traditional cardiovascular risk factors combined with markers of chronic dysglycemia seem to persist as predictors of adverse cardiovascular events despite the differences in the diabetic populations studied.
When BNP or NT-proBNP was used to predict death or cardiovascular death alone compared to using all variables available excluding these natriuretic peptides, estimates of C statistics were comparable, but the stand-alone natriuretic peptide receiver operating characteristic area under the curve values exceeded those reported earlier in patients with HF or coronary artery disease.19,30 Thus, a single measurement of BNP or NT-proBNP contains the same predictive information about risk of death as all other variables listed in Table 1. Of note, the reverse was seen in cardiovascular morbidity

Figure. The association of BNP and NT-proBNP concentrations and risk of all-cause death. The hazard of death is depicted with 95% CIs. The reference of hazard ratio=1.0 corresponds to a BNP concentration of 35 pg/mL, and a NT-proBNP concentration of 125 pg/mL. BNP indicates B-type natriuretic peptide; NTproBNP, N-terminal prohormone BNP.

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 8

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

ORIGINAL RESEARCH

(MI, HF, stroke) where best risk models outperformed a single BNP or NT-proBNP measurement. As visually depicted in the figure, differences in natriuretic peptides conferred %20-fold changes in risk of death. This large risk gradient enables easier identification of patients at higher versus lower risk and reinforces the prognostic information contained in BNP and NT-proBNP levels.
The addition of either natriuretic peptide to best risk models improved the discriminatory ability significantly. This was evident in the changes in C statistics, NRI, and Integrated Discrimination Index. C statistics increased 0.01 to 0.06 depending on outcome, which was reflected in NRI increases of 10.6% to 35.4%. These changes make both BNP and NTproBNP valuable prognostic determinants that should be considered in future risk determination in a comparable population. Furthermore, history of HF did not influence the predictive ability of BNP or NT-proBNP across all outcomes, including MI and stroke.
Having diabetes mellitus seems to influence levels of natriuretic peptides in both the absence and presence of cardiovascular disease,31,32 which potentially could change the relationship between levels of natriuretic peptides and risk of cardiovascular events. Whether natriuretic peptides predict outcomes differently in ACS patients with type 2 diabetes mellitus compared to similar patients without diabetes mellitus is not known. The stand-alone discriminatory strength of BNP and NT-proBNP in cardiovascular death was somewhat lower (receiver operating characteristic area under the curve: BNP 0.58; NT-proBNP 0.68) in patients with coronary artery disease from the Prevention of Events With Angiotensin Converting Enzyme (PEACE) trial (%16% patients had diabetes mellitus). In comparison, when BNP was sampled 2 to 4 days after the infarct in STEMI patients included in the Enoxaparin Tenecteplase-Tissue-Type Plasminogen Activator With or Without Glycoprotein IIb/IIIa Inhibitor as Reperfusion Strategy in ST-Segment Elevation Myocardial Infarction-Thrombolysis In Myocardial Infarction-23 (ENTIRE-TIMI-23) trial (%13% patients had diabetes mellitus), the receiver operating characteristic area under the curve for death after 30 days was 0.81,11 which is comparable to our results (area under the curve30 days: 0.88). This could suggest that the severity of the coronary pathology influences the predictive strength of natriuretic peptides and/or the timing of the sample used for risk determination is important, as shown by Lindahl et al.33 Earlier studies have shown that levels of natriuretic peptides are dynamic in the subacute phase following an MI,34 and smaller studies suggest that patients are at higher risk of death and left ventricular remodeling if natriuretic peptides continue to be elevated after the MI, compared with those with decreasing levels.35,36 In our study, patients were randomized within 180 days from their ACS. The timing of the baseline sample in relation to their index ACS did not affect the risk

estimates of BNP or NT-proBNP, nor did the type of index event (STEMI, NSTEMI, or UAP). This suggests that the predictive value of natriuretic peptides is retained from shortly after the event until at least 6 months after the ACS in patients with type 2 diabetes mellitus, irrespective of coronary pathology.
As the availability of BNP and NT-proBNP analyses differs between institutions and regions, we also assessed the predictive value of BNP compared to NT-proBNP. In the PEACE trial, BNP was only a predictor of HF, while NT-proBNP was a predictor of cardiovascular death, HF, and stroke in coronary artery disease patients.19 Neither biomarker predicted MI. In the Valsartan Heart Failure Trial (Val-Heft), NT-proBNP proved superior to BNP in predicting a composite of morbidity and mortality and HF hospitalization in chronic HF patients; however, the incremental discriminatory value of NT-proBNP versus BNP was small.30 In our study of ACS patients with type 2 diabetes mellitus, both natriuretic peptides had comparable predictive strength in all outcomes studied (death, cardiovascular death, MI, HF, and stroke), albeit NRI and Integrated Discrimination Index values increased slightly more with BNP across all outcomes.
Our results expand knowledge of earlier findings of enhanced risk prediction using natriuretic peptides in patients with a recent ACS3,10–12 to also include patients with a recent ACS and type 2 diabetes mellitus. This prevalent population has only marginally been studied in this context.37–40 The use of BNP or NT-proBNP for risk prediction in all ACS patients, irrespective of diabetes mellitus status, is now further substantiated. Whether drugs that directly influence natriuretic peptide concentrations (eg, sacubitril41) can modify the incidence of the cardiovascular outcomes examined in this study remains speculative.
Limitations
To learn more about how diabetes mellitus per se affects the predictive ability of natriuretic peptides, a similar study design with ACS patients with and without type 2 diabetes mellitus would have been optimal. Furthermore, our average follow-up time of %2 years precludes any estimates on the long-term predictive ability of BNP and NT-proBNP. Data on the severity of the index ACS were not obtained (eg, troponin), which as a marker of myocardial damage could have attenuated the prognostic ability of BNP/NT-proBNP. Extrapolating from the present study to other ACS patients with type 2 diabetes mellitus should be done cautiously, as inclusion criteria may have led to selection bias compared to patients not included in this trial. Importantly, patients with estimated glomerular filtration rate <30 mL/min per 1.73 m2 were excluded from this study, and renal function is shown to affect levels of natriuretic peptides.42

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 9

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

ORIGINAL RESEARCH

Conclusion
In a population of patients with a recent ACS and type 2 diabetes mellitus, BNP and NT-proBNP are important and comparable predictors of death, cardiovascular death, MI, HF, as well as stroke. Both natriuretic peptides improve the discriminatory ability significantly when added to best risk models with known predictors of adverse cardiovascular outcomes, irrespective of prior history of HF or prior cardiovascular disease.
Sources of Funding
Sanofi funded the ELIXA trial. EW was supported by an unrestricted grant from The Danish Council for Independent Research (DFF—4183-00550).
Disclosures
Wolsk: supported by unrestricted grants from The Danish Council for Independent Research (DFF – 4183-00550), Fonden af 17-12-1981, Eva & Henry Frænkels Mindefond, Kong Christian den Tiendes Fond, Knud Højgaards Fond, and Direktør Ib Henriksens Fond. Claggett: None. Pfeffer: Received research Grant Support from Novartis, Sanofi. Consultant: AstraZeneca, Bayer, Boehringer Ingelheim, DalCor, Genzyme, Gilead, GlaxoSmithKline, Janssen, Lilly, Medicines Company, Merck, Novartis, Novo Nordisk, Relypsa, Sanofi, Teva and Thrasos. Stock Options: DalCor. Other: Patent awarded to BWH regarding the use of inhibitors of the renin-angiotensin system in MI. Licensed by Novartis, Dr. Pfeffer’s share irrevocably assigned to charity. Diaz: Received grants from Sanofi. Dickstein: Member of the ELIXA Executive Steering Committee. Gerstein: is supported by the McMasterSanofi Population Health Institute Chair in Diabetes Research and Care. Gerstein has received research grant support from Sanofi, Lilly, AstraZeneca and Merck, honoraria for speaking from Sanofi, Novo Nordisk, and AstraZeneca, and Berlin Chemie, and consulting fees from Sanofi, Lilly, AstraZeneca, Merck, Novo Nordisk, Abbot, Amgen, Boehringer Ingelheim, and Kaneq Bioscience. Lawson: Sanofi employee. Lewis: received research support from Sanofi. Maggioni: Honoraria for participation in the ELIXA Executive/Steering Committee sponsored by Sanofi. McMurray: None. Probstfield: None. Riddle: AstraZeneca: Research support through my institution and honoraria for consulting. Elcelyx: Honoraria for consulting. Eli Lilly: Research support through my institution and Honoraria for consulting. GlaxoSmithKline: Honoraria for consulting. NovoNordisk: Research support through my institution. Sanofi: Research support through my institution and honoraria for consulting and for speaking at professional meetings. Theracos: Honoraria for consulting. Valeritas:

Honoraria for consulting. Solomon: Received research support from Sanofi. Tardif: Has recevied research grants from Amarin, AstraZeneca, DalCor, EliLilly, Esperion, Merck, Pfizer, Sanofi and Servier; honoraria from DalCor, Pfizer, Sanofi and
Servier; and holds a minor equity interest in DalCor. Køber: Personal fees from Sanofi and Novartis as speaker, outside
the submitted work.
References
1. Haffner SM, Lehto S, R€onnemaa T, Py€or€al€a K, Laakso M. Mortality from coronary heart disease in subjects with type 2 diabetes and in nondiabetic subjects with and without prior myocardial infarction. N Engl J Med. 1998;339:229–234.
2. Beckman JA, Creager MA, Libby P. Diabetes and atherosclerosis: epidemiology, pathophysiology, and management. JAMA. 2002;287:2570–2581.
3. Scirica BM, Kadakia MB, de Lemos JA, Roe MT, Morrow DA, Li S, Wiviott SD, Kontos MC. Association between natriuretic peptides and mortality among patients admitted with myocardial infarction: a report from the ACTION Registryâ–GWTGTM. Clin Chem. 2013;59:1205–1214.
4. Jhund PS, Anand IS, Komajda M, Claggett BL, McKelvie RS, Zile MR, Carson PE, McMurray JJV. Changes in N-terminal pro-B-type natriuretic peptide levels and outcomes in heart failure with preserved ejection fraction: an analysis of the IPreserve study. Eur J Heart Fail. 2015;17:809–817.
5. Kristensen SL, Jhund PS, Køber L, McKelvie RS, Zile MR, Anand IS, Komajda M, Cleland JGF, Carson PE, McMurray JJV. Relative importance of history of heart failure hospitalization and N-terminal pro-B-type natriuretic peptide level as predictors of outcomes in patients with heart failure and preserved ejection fraction. JACC Heart Fail. 2015;3:478–486.
6. Lourencßo P, Ribeiro A, Pintalh~ao M, Silva S, Bettencourt P. Predictors of sixmonth mortality in BNP-matched acute heart failure patients. Am J Cardiol. 2015;116:744–748.
7. Berger R, Huelsman M, Strecker K, Bojic A, Moser P, Stanek B, Pacher R. Btype natriuretic peptide predicts sudden death in patients with chronic heart failure. Circulation. 2002;105:2392–2397.
8. Pfeffer MA, Swedberg K, Granger CB, Held P, McMurray JJV, Michelson EL, Olofsson B, Ostergren J, Yusuf S, Pocock S; CHARM Investigators and Committees. Effects of candesartan on mortality and morbidity in patients with chronic heart failure: the CHARM-Overall programme. Lancet. 2003;362:759–766.
9. Pfeffer MA, McMurray JJV, Velazquez EJ, Rouleau J-L, Køber L, Maggioni AP, Solomon SD, Swedberg K, Van de Werf F, White H, Leimberger JD, Henis M, Edwards S, Zelenkofske S, Sellers MA, Califf RM; Valsartan in Acute Myocardial Infarction Trial Investigators. Valsartan, captopril, or both in myocardial infarction complicated by heart failure, left ventricular dysfunction, or both. N Engl J Med. 2003;349:1893–1906.
10. Omland T, Aakvaag A, Bonarjee VV, Caidahl K, Lie RT, Nilsen DW, Sundsfjord JA, Dickstein K. Plasma brain natriuretic peptide as an indicator of left ventricular systolic function and long-term survival after acute myocardial infarction. Comparison with plasma atrial natriuretic peptide and N-terminal proatrial natriuretic peptide. Circulation. 1996;93:1963–1969.
11. Mega JL, Morrow DA, De Lemos JA, Sabatine MS, Murphy SA, Rifai N, Gibson CM, Antman EM, Braunwald E. B-type natriuretic peptide at presentation and prognosis in patients with ST-segment elevation myocardial infarction: an ENTIRE-TIMI-23 substudy. J Am Coll Cardiol. 2004;44:335–339.
12. Morrow DA, de Lemos JA, Sabatine MS, Murphy SA, Demopoulos LA, DiBattiste PM, McCabe CH, Gibson CM, Cannon CP, Braunwald E. Evaluation of B-type natriuretic peptide for risk assessment in unstable angina/non-STelevation myocardial infarction: B-type natriuretic peptide and prognosis in TACTICS-TIMI 18. J Am Coll Cardiol. 2003;41:1264–1272.
13. Scirica BM, Bhatt DL, Braunwald E, Raz I, Cavender MA, Im K, Mosenzon O, Udell JA, Hirshberg B, Pollack PS, Steg PG, Jarolim P, Morrow DA. Prognostic implications of biomarker assessments in patients with type 2 diabetes at high cardiovascular risk: a secondary analysis of a randomized clinical trial. JAMA Cardiol. 2016;1:989–998.
14. 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

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 10

ORIGINAL RESEARCH

Natriuretic Peptides as Predictors of Cardiovascular Outcomes Wolsk et al

coronary artery disease: a Global Utilization of Strategies To Open occluded arteries (GUSTO)-IV substudy. Circulation. 2003;108:275–281.
15. Sabatine MS, Morrow DA, de Lemos JA, Omland T, Desai MY, Tanasijevic M, Hall C, McCabe CH, Braunwald E. Acute changes in circulating natriuretic peptide levels in relation to myocardial ischemia. J Am Coll Cardiol. 2004;44:1988–1995.
16. Wiersma JJ, van der Zee PM, van Straalen JP, Fischer JC, van Eck-Smit BLF, Tijssen JGP, Trip MD, Piek JJ, Verberne HJ. NT-pro-BNP is associated with inducible myocardial ischemia in mildly symptomatic type 2 diabetic patients. Int J Cardiol. 2010;145:295–296.
17. Wachter R, Lahno R, Haase B, Weber-Kr€uger M, Seegers J, Edelmann F, Wohlfahrt J, Gelbrich G, G€orlitz A, Kermer P, Vollmann D, Hasenfuß G, Gr€oschel K, Stahrenberg R. Natriuretic peptides for the detection of paroxysmal atrial fibrillation in patients with cerebral ischemia—the Find-AF Study. PLoS One. 2012;7:e34351.
18. Sinner MF, Stepas KA, Moser CB, Krijthe BP, Aspelund T, Sotoodehnia N, Fontes JD, Janssens ACJW, Kronmal RA, Magnani JW, Witteman JC, Chamberlain AM, Lubitz SA, Schnabel RB, Vasan RS, Wang TJ, Agarwal SK, McManus DD, Franco OH, Yin X, Larson MG, Burke GL, Launer LJ, Hofman A, Levy D, Gottdiener JS, K€a€ab S, Couper D, Harris TB, Astor BC, Ballantyne CM, Hoogeveen RC, Arai AE, Soliman EZ, Ellinor PT, Stricker BHC, Gudnason V, Heckbert SR, Pencina MJ, Benjamin EJ, Alonso A. B-type natriuretic peptide and C-reactive protein in the prediction of atrial fibrillation risk: the CHARGE-AF Consortium of communitybased cohort studies. Europace. 2014;16:1426–1433.
19. Omland T, Sabatine MS, Jablonski KA, Rice MM, Hsia J, Wergeland R, Landaas S, Rouleau JL, Domanski MJ, Hall C, Pfeffer MA, Braunwald E; PEACE Investigators. Prognostic value of B-type natriuretic peptides in patients with stable coronary artery disease: the PEACE Trial. J Am Coll Cardiol. 2007;50:205–214.
20. Blankenberg S. Comparative impact of multiple biomarkers and N-terminal pro-brain natriuretic peptide in the context of conventional risk factors for the prediction of recurrent cardiovascular events in the Heart Outcomes Prevention Evaluation (HOPE) Study. Circulation. 2006;114:201–208.
21. Mishra RK, Beatty AL, Jaganath R, Regan M, Wu AHB, Whooley MA. B-type natriuretic peptides for the prediction of cardiovascular events in patients with stable coronary heart disease: the Heart and Soul Study. J Am Heart Assoc. 2014;3:e000907. DOI: 10.1161/JAHA.114.000907.
22. Eggers KM, Kempf T, Venge P, Wallentin L, Wollert KC, Lindahl B. Improving long-term risk prediction in patients with acute chest pain: the Global Registry of Acute Coronary Events (GRACE) risk score is enhanced by selected nonnecrosis biomarkers. Am Heart J. 2010;160:88–94.
23. Pfeffer MA, Claggett B, Diaz R, Dickstein K, Gerstein HC, Køber LV, Lawson FC, Ping L, Wei X, Lewis EF, Maggioni AP, McMurray JJV, Probstfield JL, Riddle MC, Solomon SD, Tardif J-C; ELIXA Investigators. Lixisenatide in patients with type 2 diabetes and acute coronary syndrome. N Engl J Med. 2015;373:2247– 2257.
24. Bentley-Lewis R, Aguilar D, Riddle MC, Claggett B, Diaz R, Dickstein K, Gerstein HC, Johnston P, Køber LV, Lawson F, Lewis EF, Maggioni AP, McMurray JJV, Ping L, Probstfield JL, Solomon SD, Tardif J-C, Wu Y, Pfeffer MA; ELIXA Investigators. Rationale, design, and baseline characteristics in Evaluation of LIXisenatide in Acute Coronary Syndrome, a long-term cardiovascular end point trial of lixisenatide versus placebo. Am Heart J. 2015;169:631–638.e7.
25. McMurray JJV, Adamopoulos S, Anker SD, Auricchio A, B€ohm M, Dickstein K, Falk V, Filippatos G, Fonseca C, Gomez-Sanchez MA, Jaarsma T, Køber L, Lip GYH, Maggioni AP, Parkhomenko A, Pieske BM, Popescu BA, Rønnevik PK, Rutten FH, Schwitter J, Seferovic P, Stepinska J, Trindade PT, Voors AA, Zannad F, Zeiher A; Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology, Bax JJ, Baumgartner H, Ceconi C, Dean V, Deaton C, Fagard R, Funck-Brentano C, Hasdai D, Hoes A, Kirchhof P, Knuuti J, Kolh P, McDonagh T, Moulin C, Popescu BA, Reiner Z, Sechtem U, Sirnes PA, Tendera M, Torbicki A, Vahanian A, Windecker S, McDonagh T, Sechtem U, Bonet LA, Avraamides P, Ben Lamin HA, Brignole M, Coca A, Cowburn P, Dargie H, Elliott P, Flachskampf FA, Guida GF, Hardman S, Iung B, Merkely B, Mueller C, Nanas JN, Nielsen OW, Orn S, Parissis JT, Ponikowski P; ESC Committee for Practice Guidelines. ESC guidelines for the diagnosis and treatment of acute and chronic heart failure 2012: the Task Force for the Diagnosis and Treatment of Acute and Chronic Heart Failure 2012 of the European Society of Cardiology. Developed in collaboration with the Heart Failure Association (HFA) of the ESC. Eur J Heart Fail. 2012;14:803– 869.
26. Pfeffer MA, Burdmann EA, Chen C-Y, Cooper ME, de Zeeuw D, Eckardt K-U, Feyzi JM, Ivanovich P, Kewalramani R, Levey AS, Lewis EF, McGill JB, McMurray JJV, Parfrey P, Parving H-H, Remuzzi G, Singh AK, Solomon SD, Toto R; TREAT Investigators. A trial of darbepoetin alfa in type 2 diabetes and chronic kidney disease. N Engl J Med. 2009;361:2019–2032.

27. Intensive blood-glucose control with sulphonylureas or insulin compared with conventional treatment and risk of complications in patients with type 2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group. Lancet. 1998;352:837–853.
28. McMurray JJV, Uno H, Jarolim P, Desai AS, de Zeeuw D, Eckardt K-U, Ivanovich P, Levey AS, Lewis EF, McGill JB, Parfrey P, Parving H-H, Toto RM, Solomon SD, Pfeffer MA. Predictors of fatal and nonfatal cardiovascular events in patients with type 2 diabetes mellitus, chronic kidney disease, and anemia: an analysis of the Trial to Reduce cardiovascular Events with Aranesp (darbepoetin-alfa) Therapy (TREAT). Am Heart J. 2011;162:748–755.e3.
29. Stevens RJ, Kothari V, Adler AI, Stratton IM; United Kingdom Prospective Diabetes Study (UKPDS) Group. The UKPDS risk engine: a model for the risk of coronary heart disease in Type II diabetes (UKPDS 56). Clin Sci (Lond). 2001;101:671–679.
30. Masson S, Latini R, Anand IS, Vago T, Angelici L, Barlera S, Missov ED, Clerico A, Tognoni G, Cohn JN; Val-HeFT Investigators. Direct comparison of B-type natriuretic peptide (BNP) and amino-terminal proBNP in a large population of patients with chronic and symptomatic heart failure: the Valsartan Heart Failure (Val-HeFT) data. Clin Chem. 2006;52:1528–1538.
31. Luchner A, Behrens G, Stritzke J, Markus M, Stark K, Peters A, Meisinger C, Leitzmann M, Hense H-W, Schunkert H, Heid IM. Long-term pattern of brain natriuretic peptide and N-terminal pro brain natriuretic peptide and its determinants in the general population: contribution of age, gender, and cardiac and extra-cardiac factors. Eur J Heart Fail. 2013;15:859–867.
32. Peng Q, Hu W, Su H, Yang Q, Cheng X. Levels of B-type natriuretic peptide in chronic heart failure patients with and without diabetes mellitus. Exp Ther Med. 2013;5:229–232.
33. Lindahl B, Lindb€ack J, Jernberg T, Johnston N, Stridsberg M, Venge P, Wallentin L. Serial analyses of N-terminal pro-B-type natriuretic peptide in patients with non-ST-segment elevation acute coronary syndromes: a Fragmin and fast Revascularisation during In Stability in Coronary artery disease (FRISC)-II substudy. J Am Coll Cardiol. 2005;45:533–541.
34. White M, Rouleau JL, Hall C, Arnold M, Harel F, Sirois P, Greaves S, Solomon S, Ajani U, Glynn R, Hennekens C, Pfeffer M. Changes in vasoconstrictive hormones, natriuretic peptides, and left ventricular remodeling soon after anterior myocardial infarction. Am Heart J. 2001;142:1056–1064.
35. Sandhu S, Har BJ, Aggarwal SG, Kavanagh KM, Ramadan D, Exner DV; REFINE Investigators. Predictive value of repeated versus single N-terminal pro B-type natriuretic peptide measurements early after-myocardial infarction. Pacing Clin Electrophysiol. 2009;32(suppl 1):S86–S89.
36. Fertin M, Hennache B, Hamon M, Ennezat PV, Biausque F, Elkohen M, Nugue O, Tricot O, Lamblin N, Pinet F, Bauters C. Usefulness of serial assessment of B-type natriuretic peptide, troponin I, and C-reactive protein to predict left ventricular remodeling after acute myocardial infarction (from the REVE-2 study). Am J Cardiol. 2010;106:1410–1416.
37. Wannamethee SG, Welsh P, Lowe GD, Gudnason V, Di Angelantonio E, Lennon L, Rumley A, Whincup PH, Sattar N. N-terminal pro-brain natriuretic peptide is a more useful predictor of cardiovascular disease risk than C-reactive protein in older men with and without pre-existing cardiovascular disease. J Am Coll Cardiol. 2011;58:56–64.
38. Pfister R, Tan D, Thekkanal J, Hellmich M, Erdmann E, Schneider CA. NT-proBNP measured at discharge predicts outcome in multimorbid diabetic inpatients with a broad spectrum of cardiovascular disease. Acta Diabetol. 2007;44:91–97.
39. James SK, Lindahl B, Timmer JR, Ottervanger JP, Siegbahn A, Stridsberg M, Armstrong P, Califf R, Wallentin L, Simoons ML. Usefulness of biomarkers for predicting long-term mortality in patients with diabetes mellitus and non-STelevation acute coronary syndromes (a GUSTO IV substudy). Am J Cardiol. 2006;97:167–172.
40. Verges B, Zeller M, Desgres J, Dentan G, Laurent Y, Janin-Manificat L, L’Huillier I, Rioufol G, Beer J-C, Makki H, Rochette L, Gambert P, Cottin Y. High plasma N-terminal pro-brain natriuretic peptide level found in diabetic patients after myocardial infarction is associated with an increased risk of in-hospital mortality and cardiogenic shock. Eur Heart J. 2005;26:1734– 1741.
41. Vardeny O, Miller R, Solomon SD. Combined neprilysin and renin-angiotensin system inhibition for the treatment of heart failure. JACC Heart Fail. 2014;2:663–670.
42. Vickery S, Price CP, John RI, Abbas NA, Webb MC, Kempson ME, Lamb EJ. Btype natriuretic peptide (BNP) and amino-terminal proBNP in patients with CKD: relationship to renal function and left ventricular hypertrophy. Am J Kidney Dis. 2005;46:610–620.

DOI: 10.1161/JAHA.116.004743

Journal of the American Heart Association 11

Supplemental material

Table S1. Characteristics of all included patients grouped according to quartiles of BNP

BNP (5-49 pg/ml) n=1381

randomized to lixisenatide age (yrs) male (%) BMI (kg/m2) Race
Asian Black Other White ethnicity – hispanic Region Africa/Near East Asia Pacific Eastern Europe North America South and Centr. America Western Europe Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) heart rate (bpm) current smoker former smoker Medical history MI HF atrial fibrillation/flutter PAD TIA ventricular tachycardia stroke CABG implanted pacemaker carotid disease hypertension Index event STEMI NSTEMI UAP PCI treatment at index ACS insulin-treated duration of diabetes (yrs) retinopathy neuropathy asthma COPD HbA1c (%) HDL(mg/dl) LDL (mg/dl) eGFR (ml/min/1.73m2) Albuminuria normoalbuminuria microalbuminura macroalbuminuria Hemoglobin (g/dL) Na (mmol/l) albumin (g/dl) CRP (mg/dl) BNP (pg/ml) NT-proBNP (pg/ml)

666 (48.2%) 56.7 ± 9.3
1019 (73.8%) 30.8 ± 5.7
204 (14.8%) 57 (4.1%) 102 (7.4%) 1018 (73.7%) 375 (27.2%) 86 (6.2%) 184 (13.3%) 290 (21.0%) 219 (15.9%) 421 (30.5%) 181 (13.1%) 128.4 ± 15.1 78.0 ± 9.2 70.5 ± 9.9 194 (14.0%) 656 (47.5%) 252 (18.2%) 172 (12.5%) 30 (2.2%) 65 (4.7%) 29 (2.1%) 13 (0.9%) 45 (3.3%) 51 (3.7%) 17 (1.2%) 28 (2.0%) 1001 (72.5%) 458 (33.2%) 591 (42.8%) 332 (24.0%) 903 (65.4%) 463 (33.5%)
7.8 ± 7.2 122 (8.8%) 215 (15.6%) 33 (2.4%) 42 (3.0%)
7.7 ± 1.3 1.1 ± 0.3 2.1 ± 0.9 82.6 ± 20.4 1134 (82.1%) 200 (14.5%) 47 (3.4%) 14.1 ± 1.4 140.2 ± 2.7 41.6 ± 3.2 1.9 (1.8-2.0) 28 (27-28) 78 (74-81)

BNP (50-105 pg/ml) n=1383
745 (53.9%) 59.5 ± 9.1
968 (70.0%) 30.3 ± 5.8
186 (13.4%) 44 (3.2%) 115 (8.3%) 1038 (75.1%) 407 (29.4%) 71 (5.1%) 175 (12.7%) 332 (24.0%) 181 (13.1%) 438 (31.7%) 186 (13.4%) 129.9 ± 17.0 77.0 ± 9.6 69.0 ± 10.3 174 (12.6%) 614 (44.4%) 299 (21.6%) 245 (17.7%) 63 (4.6%) 86 (6.2%) 33 (2.4%) 14 (1.0%) 68 (4.9%) 101 (7.3%) 23 (1.7%) 28 (2.0%) 1068 (77.2%) 591 (42.7%) 559 (40.4%) 233 (16.8%) 906 (65.5%) 539 (39.0%)
9.2 ± 8.2 134 (9.7%) 224 (16.2%) 53 (3.8%) 59 (4.3%)
7.7 ± 1.3 1.1 ± 0.3 2.0 ± 1.0 78.6 ± 20.8 1074 (77.7%) 249 (18.0%) 60 (4.3%) 13.9 ± 1.3 140.3 ± 2.8 40.9 ± 3.1 1.9 (1.8-2.0) 73 (72-74) 208 (200-215)

BNP (106-218 pg/ml) n=1378
696 (50.5%) 61.3 ± 9.2
982 (71.3%) 29.9 ± 5.6
183 (13.3%) 33 (2.4%) 117 (8.5%) 1045 (75.8%) 413 (30.0%) 58 (4.2%) 172 (12.5%) 377 (27.4%) 181 (13.1%) 449 (32.6%) 141 (10.2%) 130.0 ± 17.8 77.2 ± 10.5 69.5 ± 10.5 134 (9.7%) 640 (46.4%) 314 (22.8%) 304 (22.1%) 96 (7.0%) 111 (8.1%) 23 (1.7%) 21 (1.5%) 67 (4.9%) 157 (11.4%) 43 (3.1%) 39 (2.8%) 1062 (77.1%) 684 (49.6%) 505 (36.6%) 189 (13.7%) 870 (63.1%) 555 (40.3%)
9.4 ± 8.3 141 (10.2%) 209 (15.2%) 36 (2.6%) 79 (5.7%)
7.7 ± 1.3 1.1 ± 0.3 2.0 ± 0.9 74.4 ± 21.0 1019 (73.9%) 269 (19.5%) 90 (6.5%) 13.8 ± 1.4 140.5 ± 3.0 40.4 ± 3.5 2.1 (2.0-2.3) 150 (148-152) 470 (453-487)

BNP (218-4231 pg/ml) n=1383
669 (48.4%) 63.5 ± 9.7
896 (64.8%) 29.3 ± 5.6
157 (11.4%) 55 (4.0%) 137 (9.9%) 1034 (74.8%) 451 (32.6%) 54 (3.9%) 138 (10.0%) 414 (29.9%) 135 (9.8%) 516 (37.3%) 126 (9.1%) 129.9 ± 19.3 76.5 ± 10.8 71.3 ± 10.9 125 (9.0%) 612 (44.3%) 393 (28.4%) 514 (37.2%) 172 (12.4%) 151 (10.9%) 42 (3.0%) 26 (1.9%) 112 (8.1%) 151 (10.9%) 60 (4.3%) 44 (3.2%) 1079 (78.0%) 710 (51.3%) 479 (34.6%) 194 (14.0%) 727 (52.6%) 602 (43.5%) 10.7 ± 9.0 194 (14.0%) 271 (19.6%) 32 (2.3%) 69 (5.0%)
7.7 ± 1.2 1.1 ± 0.3 2.0 ± 0.9 68.1 ± 20.2 875 (63.3%) 345 (24.9%) 163 (11.8%) 13.3 ± 1.5 140.5 ± 3.2 39.3 ± 4.0 2.4 (2.3-2.6) 431 (419-443) 1541 (1468-1618)

P value 0.009 <0.001 <0.001 <0.001 0.020
0.019 <0.001
0.035 0.003 <0.001 <0.001 0.24 <0.001 <0.001 <0.001 <0.001 0.11 0.10 <0.001 <0.001 <0.001 0.12 0.002 <0.001 <0.001 <0.001 <0.001 <0.001 0.007 0.05 0.005 0.64 0.58 0.17 <0.001 <0.001 <0.001 0.033 <0.001 <0.001

Table S2. Linear regression of BNP and NT-proBNP with all variables listed in the model

randomized to lixisenatide age (yrs) male (%)
BMI (kg/m2) Race
Black vs. white
Asian vs. white other vs. white ethnicity – hispanic Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg)
heart rate (bpm) current smoker former smoker Medical history
MI HF atrial fibrillation/flutter
PAD TIA ventricular tachycardia stroke
CABG implanted pacemaker carotid disease
hypertension Index event STEMI NSTEMI
PCI treatment at index ACS insulin-treated duration of diabetes (yrs)
Retinopathy neuropathy asthma COPD
HbA1c (%) HDL(mg/dl) LDL (mg/dl) eGFR (ml/min/1.73m2)
micro vs. Normoalbuminuria macro vs. Normoalbuminuria Hemoglobin (g/dL)
Na (mmol/l) albumin (g/dl)

BNP P value

0.0 (-14.0, 14.0) 1.9 (1.0, 2.9)
15.0 (-3.6, 33.6)
-7.0 (-8.4, -5.6)

0.996 <0.001 0.113
<0.001

19.4 (-20.4, 59.2)
-37.9 (-64.6, -11.2) -1.0 (-27.3, 25.4) 19.8 (0.7, 38.9) -1.0 (-1.6, -0.5)
1.3 (0.4, 2.2) 2.1 (1.4, 2.8) -5.0 (-29.9, 19.8) 3.0 (-13.0, 18.9)

0.340
0.005 0.943 0.042 <0.001
0.007 <0.001 0.692 0.717

28.5 (10.4, 46.5) 114.2 (95.7, 132.7)
34.2 (4.4., 64.0) 27.7 (-0.5, 56.0) 2.5(-44.8, 49.9) 82.3 (20.9, 143.6)
42.7 (10.6, 74.7) -5.0 (-32.3, 22.4) 6.5 (-39.1, 52.1)
-28.1 (-74.8, 18.6) -20.6 (-38.8, -2.3)

0.002 <0.001
0.024 0.054 0.916 0.009
0.009 0.722 0.781
0.239 0.027

103.0 (81.2, 124.8)
50.0 (28.8, 71.2) -41.0 (-56.7, -25.3) -12.7 (-29.1, 3.7)
-0.1 (-1.1, 1.0) -12.1 (-37.1, 12.9) -19.3 (-40.0, 1.4) 0.8 (-42.7, 44.4)
-2.1 (-32.2, 32.9) -3.6 (-9.7, 2.6) -10.5 (-37.7, 16.7) 1.4 (7.0, 9.7)
-1.8 (-2.1, -1.4) 67.3 (48.7, 85.9) 89.6 (57.6, 121.6)
-21.4 (-27.4, -15.4) 2.4 (-0.2, 4.9)
-13.8 (-16.1, -11.5)

<0.001
<0.001 <0.001 0.130
0.900 0.343 0.068 0.970
0.905 0.254 0.449 0.750
<0.001 <0.001 <0.001
<0.001 0.072 <0.001

NT-proBNP
16.2 (-66.4, 98.8) 7.2 (1.6, 12.8) 8.2 (-101.4, 117.9) -46.8 (-55.2, -38.4)
27.7 (-207.3, 262.8) -225.1 (-382.7, -67.4) 127.4 (-28.3, 283.2) 169.9 (57.0, 282.8) -5.3 (-8.6, -2.0) 6.6 (1.1, 12.2) 13.8 (9.7, 18.0) -78.0 (-224.8, 68.8) -31.9 (-126.2, 62.5)
35.3 (-71.3, 141.9) 648.8 (539.4, 758.2) 232.8 (57.0, 408.7) 290.0 (123.2, 456.7) -48.3 (-327.7, 231.2) 182.8 (-179.4, 545.0) 105.2 (-84.0, 294.4) -166.9 (-328.3, -5.5) -20.6 (-289.9, 248.8) -120.8 (-396.6, 154.9) -95.5 (-203.4, 12.4)
579.6 (451.0, 708.3) 294.5 (169.3, 419.6) -179.6 (-272.2, -86.9) -17.6 (-114.5, 79.3) -4.6 (-10.7, 1.5) -68.3 (-215.9, 79.2) -12.6 (-135.0, 109.8) -46.0 (-303.1, 211.1) -67.4 (-274.5, 139.7) -51.8 (-88.1, -15.5) -91.0 (-251.5, 69.5) 2.0 (-47.3, 51.2) -14.4 (-16.7, -12.2) 443.7 (334, 553.4) 661.0 (472.2, 849.9) -98.3 (-133.6, -63.1) -5.1 (-20.3, 10.0) -85.4 (-99.1, -71.8)

P value
0.701 0.012 0.883 <0.001
0.817 0.005 0.109 0.003 0.002 0.019 <0.001 0.298 0.508
0.516 <0.001 0.009 0.001 0.735 0.323 0.276 0.043 0.881 0.390 0.083
<0.001 <0.001 <0.001
0.722 0.137 0.364 0.840 0.726 0.523 0.005 0.267 0.938 <0.001 <0.001 <0.001 <0.001
0.506 <0.001

Table S3. Variables independently associated with BNP/NT-proBNP concentrations, and corresponding r2 values for regression models with the 5 listed variables
1st 2nd 3rd 4th

5th

log2BNP
(r2=0.20)
log2NT-proBNP
(r2=0.28)

Albumin

Age (per 10 years)

(χ2:318, coeff:- 0.09) (χ2:316, coeff: -0.35)

Prior HF (χ2: 312, coeff: 0.78)

S TEMI (χ2:227, coeff: 0.57)

BMI (per 5 kg/m2) (χ2:80, coeff: -0.15)

eGFR (χ2:672, coeff: 0.27)

Albumin (χ2:527, coeff: -0.14)

S TEMI (χ2: 381, coeff: 0.87)

BMI (per 5 kg/m2) (χ2:305, coeff: -0.34)

prior HF (χ2:262, coeff: 0.86)

Table S4. Unadjusted estimates of predictors of outcomes found significant in multivariate models using base variables and BNP (n=5525)

Outcome

1st 2nd

3rd

4th 5th

Death
(397 events )
CV death
(286 events )
MI
(473 events )
HF
(221 events )
Stroke
(115 events )

log2BNP HR 1.90 (1.78-2.03)
log2BNP HR 2.01 (1.87-2.17)
prior MI HR 2.63 (2.19-3.16)
log2BNP HR 1.96 (1.80-2.14)
log2BNP HR 1.49 (1.33-1.68)

AF HR 2.90 (2.22-3.79)
HbA1c HR 1.23 (1.13-1.34)
log2BNP HR 1.33 (1.25-1.41)
BMI per 5 HR 1.31 (1.19-1.43)
prior TIA HR 4.90 (2.69-8.91)

NSTEMI HR 1.53 (1.26-1.86)
AF HR 3.14 (2.31-4.25)
NSTEMI HR 1.90 (1.59-2.28)
HR per 10 HR 1.34 (1.18-1.51)
macroalbuminuria HR 3.05 (1.86-4.99)

Na* HR 1.10 (1.05-1.16)
NSTEMI HR 1.61 (1.28-2.04)
prior stroke HR 2.50 (1.89-3.33)
prior HF HR 4.16 (3.19-5.42)
Age per 10 HR 1.61 (1.32-1.95)

HR per 10 HR 1.21 (1.11-1.34)
HR per 10 HR 1.25 (1.12-1.40)
PAD HR 2.52 (1.94-3.29)
prior MI HR 3.09 (2.37-4.03)
LDL per 10 HR 1.07 (1.02-1.12)

Table S5. Unadjusted estimated of predictors of outcomes found significant in multivariate models using base variables and NT-proBNP (n=5525)

Outcome
Death
(397 events )
CV death
(286 events )
MI
(473 events )
HF
(221 events )
Stroke
(115 events )

1st 2nd 3rd 4th

5th

log2NT-proBNP HR 1.64 (1.56-1.73)
log2NT-proBNP HR 1.72 (1.62-1.83)
prior MI HR 2.63 (2.19-3.16)
log2NT-proBNP HR 1.68 (1.57-1.80)
log2NT-proBNP HR 1.34 (1.22-1.48)

NSTEMI HR 1.53 (1.26-1.86)
HbA1c HR 1.23 (1.13-1.34)
log2NT-proBNP HR 1.23 (1.18-1.29)
BMI per 5 HR 1.31 (1.19-1.43)
prior TIA HR 4.90 (2.69-8.91)

PCI HR 0.52 (0.42-0.63)
NSTEMI HR 1.61 (1.28-2.04)
NSTEMI HR 1.90 (1.59-2.28)
NSTEMI HR 1.91 (1.47-2.49)
macroalbuminuria HR 3.05 (1.86-4.99)

DBP# HR 1.04 (1.02-1.06)
AF HR 3.14 (2.31-4.25)
prior stroke HR 2.50 (1.89-3.33)
prior HF HR 4.16 (3.19-5.42)
Age per 10 HR 1.61 (1.32-1.95)

Na* HR 1.10 (1.05-1.16)
prior HF HR 2.87 (2.27-3.62)
PAD HR 2.52 (1.94-3.29)
prior MI HR 3.09 (2.37-4.03)
LDL per 10 HR 1.07 (1.02-1.12)

Table S6. Predictors of outcomes ranked according to χ2 value using base variables and BNP and stratified according to history of heart failure (No prior HF, n=4290; Prior HF, n=1235)

Outcome
Death
(397 events )

1st 2nd

3rd

No prior HF

log2BNP (χ2:108, HR 1.68)

HR per 10 (χ2:11, HR 1.24)

Prior HF

log2BNP (χ2:108, HR 1.72)

AF (χ2:13, HR 1.93)

Age per 10 (χ2:9, HR 1.28)
Race (χ2:13, HR 1.42)

CV death
(286 events )

No prior HF

log2BNP (χ2:101, HR 1.80)

AF (χ2:12, HR 2.45)

Prior HF

log2BNP (χ2:95, HR 1.84)

CABG (χ2:10, HR 2.03)

NSTEMI (χ2:9, HR 1.64)
PAD (χ2:9, HR 1.90)

MI
(473 events )

No prior HF Prior HF

log2BNP (χ2:29, HR 1.23)
log2BNP (χ2:26, HR 1.34)

prior MI (χ2:27, HR 1.86)
prior MI (χ2:17, HR 2.10)

PAD (χ2:21, HR 2.03)
NSTEMI (χ2:15, HR 2.00)

HF
(221 events )
Stroke
(115 events )

No prior HF

log2BNP (χ2:52, HR 1.70)

HR per 10 (χ2:22, HR 1.53)

macroalbuminuria (χ2:13, HR 2.78)

Prior HF

log2BNP (χ2:100, HR 1.94)

BMI per 5 (χ2:28, HR 1.39)

DBP* (χ2:12, HR 1.06)

No prior HF Prior HF

log2BNP

macroalbuminuria

(χ2:15, HR 1.40) (χ2:12, HR 2.98)

log2BNP (χ2:8, HR 1.31)

TIA (χ2:6, HR 2.46)

TIA (χ2:10, HR 3.98)
MI (χ2:5, HR 2.20)

Hazard rat io’s reflect 1 unit changes if nothing else is stated. For log2BNP that translates into a doubling of the untransformed BNP concentrations. Macroalbuminuria: >300 mg albu min excret ion/24 h. *per 1 mmHg decrease below ≥75 mmHg.. AF- atrial fibrillation/flutter. NSTEMI – non-ST elevation myocardial infarction at index event. HR – heart rate. PAD – peripheral artery disease. TIA – transient ischemic attack. Du ration of T2D is per year since diagnosis. All variables are significant, p<0.05.

Table S7. Discriminatory changes in best risk models with and without BNP stratified according to history of heart failure (No prior HF, n=4290; Prior HF, n=1235)

C statistics in each model (n=5525)

Death (397 events)

Base model

BNP in model

No prior HF 0.77 (0.73-0.80) 0.80 (0.77-0.84)* Prior HF 0.69 (0.64-0.73) 0.78 (0.75-0.81)*

NRI
25.7% (17.3-34.7)* 38.5% (28.7-45.4)*

CV death No prior HF 0.76 (0.72-0.80) 0.81 (0.77-0.85)* (286 events) Prior HF 0.73 (0.68-0.77) 0.82 (0.79-0.85)*

27.4% (17.7-35.6)* 39.6% (28.6-51.1)*

MI No prior HF 0.72 (0.69-0.75) 0.73 (0.70-0.75)# (473 events) Prior HF 0.71 (0.66-0.76) 0.73 (0.68-0.77)

13.7% (5.6-19.4)* 21.7% (8.4-30.9)*

HF No prior HF 0.83 (0.79-0.87) 0.86 (0.83-0.90)* (221 events) Prior HF 0.75 (0.71-0.80) 0.81 (0.77-0.85)*

29.4% (15.3-39.4)* 38.8% (26.6-46.6)*

Stroke

No prior HF 0.76 (0.71-0.82) 0.79 (0.73-0.84)

(115 events) Prior HF 0.67 (0.58-0.75) 0.69 (0.61-0.76)

17.6% (1.8-31.1)* 21.3% (0.0-35.2)*

IDI
4.0% (2.4-6.8)* 7.8% (4.4-11.4)* 3.7% (1.9-6.8)* 6.9% (3.6-11.6)* 0.8% (0.2-1.9)* 2.4% (0.9-4.7)* 3.4% (1.4-6.7)* 9.0% (5.1-13.3)* 0.1% (-0.6-1.6) 1.0% (0.1-3.9)*

*p<0.05, co mparison between base model and BNP model. #p=0.053, co mparison between base model and BNP model. NRI – Net Reclassification Index. IDI – Integrated Discrimination Index. NRI and IDI summarized as mean percent improvement ±95%CI..

Table S8. Discriminatory changes in best risk models without BNP compared to BNP alone, in all patients (n=5525) and stratified according to history of heart failure (No prior HF, n=4290; Prior HF, n=1235).

All patients

No prior HF

Prior HF

Outcome Base model

BNP

Base model

BNP

Base model

BNP

Death
(397 events )
CV death
(286 events )
MI
(473 events )
HF
(221 events )
Stroke
(115 events )

0.77 (0.74-0.79) 0.77 (0.75-0.80) 0.77 (0.74-80)

0.75 (0.71-0.78)

0.71 (0.67-0.74)

0.76 (0.73-0.80)*

0.77 (0.74-0.80) 0.79 (0.76-0.82) 0.76 (0.72-0.80) 0.76 (0.72-0.81) 0.73 (0.68-0.77)

0.78 (0.74-0.82)

0.71 (0.68-0.73) 0.63 (0.61-0.66)* 0.72 (0.69-0.75) 0.62 (0.59-0.65)* 0.71 (0.66-0.76)

0.66 (0.61-0.70)

0.84 (0.81-0.86) 0.78 (0.75-0.81)* 0.83 (0.79-0.87) 0.76 (0.72-0.80)* 0.75 (0.71-0.80)

0.74 (0.69-0.78)

0.75 (0.70-0.79) 0.67 (0.62-0.72)* 0.76 (0.71-0.82) 0.67 (0.60-0.74)* 0.67 (0.58-0.75)

0.61 (0.52-0.70)

*p<0.05, significant difference between base model co mpared to BNP model.

Table S9. Predictors of outcomes ranked according to χ2 value using base variables and BNP in patients without prior CV disease (n=2899)

Outcome

1st 2nd 3rd

Death
(112 events )
CV death
(74 events )
MI
(153 events )
HF
(39 events )
Stroke
(31 events )

log2BNP (χ2:81, HR 1.78)

DBP* (χ2:14, HR 1.05)

log2BNP (χ2:81, HR 2.04)

DBP* (χ2:16, HR 1.06)

log2BNP (χ2:18, HR 1.27)

macroalbuminuria (χ2:12, HR 2.34)

log2BNP (χ2:31, HR 1.87)

duration of T2D (χ2:13, HR 1.06)

carotid disease (χ2:41, HR 34.79)

microalbuminuria (χ2:16, HR 5.08)

Na# (χ2:7, HR 0.84) duration of T2D (χ2:10, HR 1.04) hyperte nsion (χ2:10, HR 1.89)
albumin (χ2:6, HR 0.91)
DBP* (χ2:16, HR 1.09)

Hazard rat io’s reflect 1 unit changes if nothing else is stated. For log2BNP that translates into a doubling of the untransformed BNP concentrations. Microalbu minuria ≥30-300 mg albumin excretion/24h. Macroalbuminuria: >300 mg albu min excret ion/24h.. *per 1 mmHg decrease below ≥75 mmHg .. T2D – Type 2 diabetes (years). All variables are significant, p<0.05.

Table S10. Predictors of outcomes ranked according to χ2 value using base variables and NTproBNP in patients without prior CV disease (n=2899)

Outcome

1st 2nd 3rd

Death
(112 events )
CV death
(74 events )
MI
(153 events )
HF
(39 events )
Stroke
(31 events )

log2proBNP (χ2:81, HR 1.61)

DBP* (χ2:16, HR 1.05)

log2proBNP (χ2:81, HR 1.81)

DBP* (χ2:16, HR 1.06)

log2proBNP (χ2:18, HR 1.21)

macroalbuminuria (χ2:11, HR 2.23)

log2proBNP (χ2:53, HR 1.91)

duration of T2D (χ2:14, HR 1.06)

carotid disease (χ2:37, HR 29.09)

microalbuminuria (χ2:16, HR 5.06)

age per 10 (χ2:8, HR 1.36) duration of T2D (χ2:10, HR 1.04) hyperte nsion (χ2:10, HR 1.92)
NSTEMI (χ2:5, HR 2.06)
DBP* (χ2:15, HR 1.08)

Hazard rat io’s reflect 1 unit changes if nothing else is stated. For log2BNP that translates into a doubling of the untransformed BNP concentrations. Microalbu minuria ≥30-300 mg albumin excretion/24h. Macroalbuminuria: >300 mg albu min excret ion/24h. DBP* d iastolic blood pressure per 1 mmHg decrease below ≥75 mmHg . NSTEM I – non-ST
elevation myocardial infarction at index event. T2D – Type 2 diabetes (years). All variables are significant, p<0.05.

Table S11. Predictors of outcomes ranked according to χ2 value using base variables and BNP with and without adding information on LVEF (n=3390)

Outcome Death
(236 events )
CV death
(166 events )
MI
(290 events )
HF
(148 events )
Stroke
(70 events )

1st 2nd

3rd

Variables Variables
+LVEF

log2BNP (χ2:99, HR 1.60)

Duration of T2D (χ2:17, HR 1.03)

log2BNP (χ2:108, HR 1.60)

Duration of T2D (χ2:17, HR 1.03)

male (χ2:9, HR 1.59)
male (χ2:9, HR 1.59)

Variables Variables
+LVEF

log2BNP (χ2:131, HR 1.86)

HbA1c (χ2:10, HR 1.22)

log2BNP (χ2:95, HR 1.79)

HbA1c (χ2:11, HR 1.23)

Duration of T2D (χ2:10, HR 1.03) Duration of T2D (χ2:10, HR 1.03)

Variables Variables
+LVEF

prior MI (χ2:43, HR 2.21)
prior MI (χ2:43, HR 2.21)

NSTEMI (χ2:24, HR 1.81)
NSTEMI (χ2:24, HR 1.81)

log2BNP (χ2:23, HR 1.21)
log2BNP (χ2:23, HR 1.21)

Variables Variables
+LVEF

log2BNP (χ2:97, HR 1.80)
log2BNP (χ2:97, HR 1.80)

BMI per 5 (χ2:25, HR 1.33)
BMI per 5 (χ2:25, HR 1.33)

HR per 10 (χ2:20, HR 1.43)
HR per 10 (χ2:20, HR 1.43)

Variables Variables
+LVEF

log2BNP (χ2:21, HR 1.42)
log2BNP (χ2:21, HR 1.42)

TIA (χ2:11, HR 3.59)
TIA (χ2:11, HR 3.59)

LDL per 10 (χ2:7, HR 1.08)
LDL per 10 (χ2:7, HR 1.08)

Table S12. Predictors of outcomes ranked according to χ2 value using base variables and NTproBNP with and without adding information on LVEF at index ACS (n=3390)

Outcome Death
(236 events )
CV death
(166 events )
MI
(290 events )
HF
(148 events )
Stroke
(70 events )

1st 2nd

3rd

Variables Variables
+LVEF

log2NT-proBNP (χ2:107, HR 1.46)

COPD (χ2:9, HR 1.81)

log2NT-proBNP (χ2:107, HR 1.46)

COPD (χ2:9, HR 1.81)

Male (χ2:9, HR 1.58)
Male (χ2:9, HR 1.58)

Variables Variables
+LVEF

log2NT-proBNP (χ2:115, HR 1.57)

HbA1c (χ2:14, HR 1.27)

log2NT-proBNP (χ2:68, HR 1.49)

HbA1c (χ2:13, HR 1.26)

Male (χ2:9, HR 1.71)
AF (χ2:10, HR 1.92)

Variables Variables
+LVEF

prior MI (χ2:44, HR 2.24)

log2NT-proBNP (χ2:30, HR 1.18)

prior MI (χ2:44, HR 2.24)

log2NT-proBNP (χ2:30, HR 1.18)

NSTEMI (χ2:25, HR 1.18)
NSTEMI (χ2:25, HR 1.18)

Variables Variables
+LVEF

log2NT-proBNP (χ2:86, HR 1.56) log2NT-proBNP (χ2:86, HR 1.56)

BMI per 5 (χ2:25, HR 1.33)
BMI per 5 (χ2:25, HR 1.33)

CABG (χ2:19, HR 2.48)
CABG (χ2:19, HR 2.48)

Variables Variables
+LVEF

log2NT-proBNP (χ2:14, HR 1.26) log2NT-proBNP (χ2:14, HR 1.26)

TIA (χ2:10, HR 3.36)
TIA (χ2:10, HR 3.36)

LDL per 10 (χ2:6, HR 1.07)
LDL per 10 (χ2:6, HR 1.07)

Table S13. Discriminatory changes in best risk models with and without BNP and NT-proBNP with LVEF and coronary intervention procedure added to base model

C statistics in each model (n=3390)

Death (236 events)
CV death (166 events)
MI (290 events)
HF (148 events)
Stroke (70 events)

Base model

BNP NT-pr oB NP

0.75 (0.72-0.78)

BNP NT-pr oB NP

0.77 (0.73-0.81)

BNP NT-pr oB NP

0.70 (0.67-0.73)

BNP NT-pr oB NP

0.86 (0.83-0.89)

BNP NT-pr oB NP

0.72 (0.66-0.78)

BNP/ NT-proBNP in
model
0.79 (0.76-0.82)* 0.79 (0.75-0.82)*
0.83 (0.80-0.86)* 0.82 (0.79-0.85)*
0.71 (0.67-0.74) 0.71 (0.68-0.74)
0.88 (0.85-0.90)* 0.88 (0.85-0.90)*
0.76 (0.70-0.82) 0.75 (0.70-0.81)

NRI IDI

31.8% (25.1-37.4)* 25.5% (19.4-31.9)*

5.3% (3.5-7.4)* 3.5% (2.3-5.1)*

34.8% (28.0-41.7)* 29.9% (22.4-36.8)*

5.7% (3.7-8.4)* 3.9% (2.3-6.2)*

14.3% (9.3-19.5)* 10.6% (5.7-16.6)*

1.2% (0.6-2.1)* 0.8% (0.3-1.6)*

35.4% (24.7-40.6)* 29.9% (21.8-36.6)*

5.0% (3.0-7.6)* 3.8% (2.2-5.8)*

19.3% (8.8-29.9)* 17.2% (6.3-28.1)*

0.4% (0-1.2) 0.2% (0-0.8)

*p<0.05, co mparison between base model and /NT-p roBNP model. NRI – Net Reclassification Index. IDI – Integrated Discrimination Index. NRI and IDI summarized as mean percent improvement ±95%CI.

Myocardial infarction summary criteria for positive adjudication:
Spontaneous MI: Elevated cardiac markers (CM) and either new electrocardiographic (ECG) changes or a clinical presentation consistent with an acute MI. ○PCI-related MI: Elevated CM (or other criteria in the absence of elevated CM). ○Coronary artery bypass graft (CABG)–related MI: Elevated CM and new ECG changes (or other criteria). Detailed criteria for positive adjudication: a. Spontaneous MI: Cardiac markers: ○Troponinp>upper limit of normal (ULN) or ○CK-MB>ULN
and at least 1 of the following: ○Ischemic symptoms: rest or accelerated symptoms (pain, dyspnea, and pressure) consistent with myocardial ischemia. ○ECG changes consistent with infarction: •New significant Q waves (or R waves in V1-V2)in 2 contiguous leads in absence of previous left ventricular hypertrophy or conduction abnormalities. OR •Evolving ST-segment to T-wave changes in≥2 contiguous leads. •Development of new left bundle-branch block. •ST-segment elevation requiring thrombolytics or PCI. b. PCI-related MI: Cardiac markersq: 1. Assuming baseline value>ULN 2. Within 48 hours of procedure a. Troponinp>3× ULN OR b. CK-MB>3× ULN c. CABG-related MI: Cardiac markers: 1. Assuming baseline value>ULN 2. Within 72 hours of procedure a. Troponinp>5× ULN OR b. CK-MB>5× ULN AND c. New pathologic Q waves or left bundle-branch block, new native or graft vessel occlusion, or imaging evidence of loss of viable myocardium. 3. Hospitalization for UA a. Unplanned hospitalization for worsening angina defined as rest or accelerated symptoms (pain, dyspnea, and pressure) consistent with myocardialischemia AND b. Cardiac markers (CK-MB or troponin) suggestive of myocardial injury but not meeting MI criteria. Note: if abnormal troponin, value must be in the suggestive (middle) range and below the threshold for MI.

Baseline variables used in risk models:
Log2BNP, Log2NT-proBNP
Race, ethnicity, region, randomization to lixisenatide, PCI at index ACS, age, gender, BMI, systolic blood pressure, diastolic blood pressure (above/below 70 mmHg), heart rate, smoking (current/never/former), history of MI, history of HF, history of AF, history of PAD, history of TIA, history of stroke, history of ventr. tachycardia, history of CABG, pacemaker implanted, carotid disease, history of hypertension, index ACS (STEMI, NSTEMI, UAP), insulin use (yes/no), duration of T2D, retinopathy, neuropathy, asthma, COPD, albuminuria (no/micro/macro), logCRP, HbA1c, HDL, LDL, eGFR, Hgb, Na (above/below 140 mmol/L), albumin.

Figure S1. The association of BNP concentrations and risk of all-cause death according to history of heart failure

BNP concentration and risk of death

Patients with no prior HF

Patients with prior HF

10 15 20

Hazard Ratio (Reference = 35) 10 20 3040 0

10 15 20

Hazard Ratio (Reference = 35) 10 20 3040 0

Percentage

Percentage

5

5

0

0

0

500

1000

1500

BNP (pg/ml)

0

500

1000

1500

BNP (pg/ml)

The hazard of death is depicted with 95%CI. The Co x spline model was fully adjusted for all significant variables. The reference of HR=1.0 corresponds to a BNP concentration of 35 pg/ml.