Factors Associated With and Prognostic Implications of Cardiac Troponin Elevation in Decompensated Heart Failure With Preserved Ejection Fraction: Findings From the American Heart Association Get With The Guidelines–Heart Failure Program

Key Points

Question  What are the prognostic implications of elevated troponin levels among patients with decompensated heart failure and preserved ejection fraction (HFpEF)?

Findings  In this observational cohort study, abnormally elevated troponin levels in patients with decompensated HFpEF were associated with higher risk of in-hospital and postdischarge adverse outcomes, independent of other risk predictors.

Meaning  Assessment of troponin levels among patients with decompensated HFpEF may have important prognostic value independent of other established clinical predictors.

Importance  Elevated levels of cardiac troponins are associated with adverse clinical outcomes among patients with heart failure (HF) and reduced ejection fraction. However, the clinical significance of troponin elevation in the setting of decompensated HF with preserved ejection fraction (HFpEF) is not well established.

Objective  To determine the clinical predictors of troponin elevation and its association with in-hospital and long-term outcomes among patients with decompensated HFpEF.

Design, Setting, and Participants  Observational analysis of Get With The Guidelines–HF registry participants who were admitted for decompensated HFpEF (ejection fraction ≥50%) from January 2009 through December 2014 and who had quantitative or categorical (elevated vs normal based on institution’s reference laboratory) measures of troponin level (troponin T or troponin I, as available).

Main Outcomes and Measures  In-hospital outcomes (mortality, length of stay, and discharge destination) and postdischarge outcomes (30-day mortality, 30-day readmission rate, 1-year mortality).

Results  We included 34 233 patients with HFpEF from 224 sites with measured troponin levels (33.4% men; median age, 79 years): 78.6% (n = 26 896) with troponin I and 21.4% (n = 7319) with troponin T measurements. Among these, 22.6% (n = 7732) had elevation in troponin levels. In adjusted analysis, higher serum creatinine level, black race, older age, and ischemic heart disease were associated with troponin elevation. Elevated troponin was associated with higher odds of in-hospital mortality (odds ratio [OR], 2.19; 95% CI, 1.88-2.56), greater length of stay (length of stay >4 days OR, 1.38; 95% CI, 1.29-1.47), and lower likelihood of discharge to home (OR, 0.65; 95% CI, 0.61-0.71) independent of other clinical predictors and measured confounders. Presence of elevated troponin I levels was also significantly associated with increased risk of 30-day mortality (hazard ratio [HR], 1.59; 95% CI, 1.42-1.80), 30-day all-cause readmission (HR, 1.12; 95% CI, 1.01-1.25), and 1-year mortality HR, 1.35; 95% CI, 1.26-1.45).

Conclusions and Relevance  Troponin elevation among patients with acutely decompensated HFpEF is associated with worse in-hospital and postdischarge outcomes, independent of other predictive variables. Future studies are needed to determine if measurement of troponin levels among patients with decompensated HFpEF may be useful for risk stratification.

Up to 50% of hospitalized patients with heart failure (HF) also have preserved ejection fraction (HFpEF).¹ This heterogeneous syndrome encompasses multiple different pathophysiological components in varying combinations and carries variable risks of mortality and complications such as readmission. Few tools are currently available for risk stratification of this heterogeneous group of patients. There has been an increasing interest in the prognostic utility of cardiac troponins T and I, biomarkers of myocardial injury, for risk stratification of hospitalized patients with HF. Quiz Ref IDSeveral studies have demonstrated a consistent association between elevated troponin levels and risk of adverse clinical outcomes among patients with HF with reduced ejection fraction (HFrEF), even in absence of chest pain or myocardial infarction.^2-6 Furthermore, the current HF guidelines recommend early assessment of troponin levels among patients with acute HF for risk stratification.⁷ However, these recommendations are based on large cohort studies that predominantly included patients with HFrEF, and the clinical significance of abnormal troponin elevation among hospitalized patients with HFpEF in the absence of a precipitating acute coronary syndrome (ACS) is not well established. This is particularly important, considering the significant pathophysiological differences in the cardiac substrate of HFpEF and HFrEF.^8-10 Against this background, we aimed to identify the clinical characteristics associated with troponin elevation and to determine the association between troponin level and clinical outcomes in a large cohort of patients hospitalized with decompensated HFpEF using the Get With The Guidelines (GWTG)–HF registry.

The study was performed using data from the GWTG-HF registry, the database of an ongoing, national, voluntary hospital-based quality improvement program initiated by the American Heart Association in 2005.The details of the GWTG-HF registry have been previously published.¹¹ Briefly, the registry includes patients hospitalized at participating centers with a primary diagnosis of new or worsening HF or patients who developed significant HF symptoms during the hospitalization such that HF was the primary discharge diagnosis. The registry is representative of hospitals from all regions and includes community and large tertiary-care hospitals. Participating centers submit information either on consecutive patients with HF or by random sample (if the annual HF burden >75 cases). Trained personnel collect patient-level data using an internet-based patient management tool (Quintiles Real-World and Late Phase Research) in compliance with the Joint Commission on the Accreditation of Healthcare Organizations and Centers for Medicare and Medicaid Services standards. Data collected include patient demographics, socioeconomic status, medical history, medications, laboratory data, in-hospital treatment, in-hospital outcomes, discharge medications, discharge status, and postdischarge follow-up (available from 2009 onwards). Information on hospital characteristics was obtained from the American Hospital Association annual survey files.

All participating centers are required to follow local regulatory and privacy guidelines and to obtain institutional review board approval for the GWTG-HF protocol. As data collection is primarily meant for quality improvement purposes, all participating centers are granted a waiver of informed consent under the Common Rule. Quintiles serves as the data collection coordination center for American Heart Association GWTG programs. The Duke Clinical Research Institute serves as the data analysis center and has approval to analyze the aggregate deidentified data for research purposes.

Among GWTG-HF participants older than 65 years with fee-for-service Medicare coverage, data on postdischarge outcomes were obtained by linking patient-level data in the registry with Medicare denominator and Part A inpatient claims files using dates of admission and discharge, hospital name, date of birth, and sex.¹²

In the present study, we included patients with HF who were hospitalized between January 2009 and December 2014 at GWTG-HF participating centers and had troponin measurements available. It is noteworthy that all patients included in the GWTG-HF registry are primarily admitted for decompensated HF and not ACS by enrollment criteria. The details of the study cohort derivation are shown in eFigure 1 in the Supplement. Consistent with previous studies, patients with significant renal dysfunction (serum creatinine >2 mg/dL or hemodialysis requirement) were excluded because the reduced renal clearance of troponin in these patients may confound the association between troponin and clinical outcomes.⁵ The final study population included 34 233 patients with HFpEF (ejection fraction ≥50%) from 224 sites.

Postdischarge clinical outcomes were assessed in the subgroup of study participants with linked Medicare follow-up data. This included 14 819 patients enrolled from 209 centers between January 1, 2009, and November 30, 2014 (21% with troponin T and 79% with troponin I measurement during index hospitalization).

The primary exposure variable of interest for this study was the peak troponin level measured during the HF hospitalization. Both quantitative (in nanograms per milliliter or micrograms per liter) and categorical (elevated vs normal based on each institution’s upper reference limit) measures of peak troponin levels (troponin T or troponin I, as available) were recorded. For categorical analysis (elevated vs normal), we pooled cardiac troponin T and troponin I measurements because they are used interchangeably in the clinical setting for qualitative assessment. For quantitative analysis, we used troponin I levels since it was the most commonly reported (in 78.6% of the study participants).

The primary outcome of interest for our study was all-cause in-hospital mortality. Secondary outcomes included length of hospital stay and discharge destination (home vs hospice-home, hospice-health care facility, other health care facility, or died), and postdischarge clinical outcomes (30-day all-cause readmission, 30-day mortality, and 1-year mortality). Data on postdischarge mortality was obtained from Medicare enrollment files, and readmission data were obtained from Part A inpatient claims files. We also evaluated the adherence to process of care measures that are applicable to patients with HFpEF. These include discharge instructions, smoking cessation counseling, use of anticoagulation for atrial fibrillation, postdischarge follow-up referral, and adequate blood pressure control at discharge.

Baseline patient characteristics and in-hospital management strategies were compared among participant groups first according to the reported qualitative peak troponin level, elevated vs normal based on institution’s reference laboratory, and then by data-derived quartiles of peak troponin I, the most commonly available troponin measurement among the study participants (78.6%). Categorical variables were reported for the study groups as percentages and compared using the Pearson χ² test. Continuous variables were reported as medians with interquartile ranges and compared using Kruskal-Wallis tests. Clinically meaningful differences in participant characteristics among participants with elevated vs normal troponin levels were evaluated using absolute standardized difference. By convention, an absolute standardized difference greater than 10 is considered significant. Factors associated with elevation in troponin levels were identified by constructing a logistic regression model. Stepwise selection was used for the model with α = .15 to enter and α = .05 to remove, using all the clinical and demographic characteristics as detailed in eMethods in the Supplement. Multiple imputation was used for missing covariates, with 25 imputed data sets generated using fully conditional specification methods to generate the final estimates. The overall rate of missing data for covariates was low (<5%) except admission B-type natriuretic peptide (BNP) levels (28% missing) and estimated glomerular filtration rate (10% missing). All continuous variables were evaluated for nonlinearity with the outcome, and linear splines were used for those that did not meet the linear relationship criteria. Generalized estimating equations were used in final multivariable logistic model to account for in-hospital clustering. The associations between elevated peak troponin level (vs normal) and in-hospital outcomes such as in-hospital mortality, length of stay (>4 vs ≤4 days), and discharge to home (vs others) were assessed using multivariable adjusted logistic regression models with generalized estimating equations to account for in-hospital clustering. The potential confounders that were adjusted for in these models included year of enrollment, demographic characteristics (age, sex, race, insurance status, cigarette smoking in the past year), medical history (systolic blood pressure on admission, anemia, ischemic heart disease, diabetes, hyperlipidemia, hypertension, atrial fibrillation), estimated glomerular filtration rate and BNP level at admission, and hospital characteristics (region, hospital type, number of beds, rural vs urban). Similar analysis was also performed to assess the association of different troponin I quartiles (quartiles 2, 3, and 4 vs quartile 1) with these outcomes. Furthermore, the adjusted association between continuous measures of troponin I levels and in-hospital mortality was assessed using a multivariable logistic regression model with generalized estimating equations to account for in-hospital clustering. Restricted cubic splines with 3 knots (5th, 50th, and 95th percentiles) were used to account for the nonlinearity in the association.

In the subset of participants with available Medicare-linked follow-up data, associations between troponin I levels (elevated vs normal) and postdischarge clinical outcomes (30-day mortality, 30-day readmission, 1-year mortality) were assessed using multivariable adjusted Cox proportional hazards models. These models were adjusted for year of enrollment, demographic characteristics (age, sex, race, insurance status, cigarette smoking in the past year), medical history (systolic blood pressure on admission, anemia, ischemic heart disease, diabetes, hyperlipidemia, hypertension, atrial fibrillation), estimated glomerular filtration rate and BNP level at admission, and hospital characteristics (region, hospital type, number of beds, rural vs urban).

Several sensitivity and subgroup analyses were performed to test the robustness of our study findings. First, the association between troponin levels and continuous measure of in-hospital length of stay was evaluated using multivariable adjusted linear regression models. Second, among participants who had only troponin T measurements available, sensitivity analysis was performed to determine the association between presence of elevated troponin T levels with clinical outcomes. Third, sensitivity analysis was also performed to determine the association between elevation in troponin levels and in-hospital outcomes among participants without a history of myocardial infarction or revascularization (percutaneous or surgical). Finally, the prevalence of troponin elevation and its association with in-hospital outcomes was also evaluated in HFrEF patients from the GWTG-HF centers included in our study. All statistical tests were 2 sided, and P < .05 was considered statistically significant. All analyses were performed using SAS software, version 9.4 (SAS institute Inc).

Overall, troponin levels were measured in 35% of all patients with HFpEF hospitalized during the study period. We included 34 233 patients with HFpEF from 224 sites with measured troponin levels during the index hospitalization. There were few meaningful differences in the baseline characteristics and in-hospital management of patients with HFpEF with vs without troponin assessment (eTable 1 in the Supplement). Prevalence of hypertension, mean BNP levels at admission, and use of vasodilators were higher in patients with troponin assessment. In-hospital clinical outcomes were not significantly different between patients with vs without troponin assessment at baseline (eTable 1 in the Supplement).

Troponin I levels were measured in 78.6% of participants, and troponin T in 21.4%. An elevated peak troponin level during the index hospitalization was detected in 22.6% of participants. Characteristics of the patients according to presence of elevated vs normal troponin levels are summarized in Table 1. Meaningful differences in certain patient characteristics across the 2 groups were noteworthy. Patients with elevated troponin levels had lower BMI, higher BNP levels, higher heart rate, and lower estimated glomerular filtration rate at presentation than patients with normal troponin levels. Similar trends in patient characteristics were also observed on comparison across troponin I quartiles (eTable 2 in the Supplement).

Factors associated with troponin elevation are shown in Figure 1. In multivariable adjusted analyses, higher creatinine, black race, smoking, ischemic heart disease, elevated heart rate and blood pressure (among patients with systolic blood pressure >150 mm Hg) at presentation, increasing QRS duration (among patients with QRS <120 milliseconds), older age, and higher BNP levels were associated with greater likelihood of troponin elevation. In contrast, higher BMI and female sex were associated with lower odds of elevation in troponin levels.

Quiz Ref IDAdherence to process of care measures such as discharge instructions, smoking cessation counseling at discharge, and adequate blood pressure control at discharge was not different among patients with elevated vs normal troponin levels (eFigure 2 in the Supplement). Referral for postdischarge follow-up was higher among patients with elevated troponin levels. In contrast, anticoagulation use for AF was significantly lower in patients with HFpEF with elevated vs normal troponin levels (eFigure 2 in the Supplement).

Use of revascularization strategies such as coronary artery bypass graft and percutaneous coronary intervention was very low in our study population (0.26% and 0.18%, respectively), with slightly higher use among those with elevated troponin levels. Among other in-hospital management strategies, use of right heart catheterization, intravenous vasodilators, and mechanical ventilation was also higher among patients with HFpEF with elevated vs normal troponin levels (Table 1). However, the between-group differences for use of these interventions were modest and not clinically meaningful (absolute standardized difference <10 for each; Table 1).

Patients with elevated troponin levels had a higher rate of in-hospital mortality than those with troponin in the normal range (3.95% vs 1.84%; Table 1). In multivariable adjusted logistic regression analysis, those with elevated troponin levels had a significantly higher likelihood of in-hospital mortality than those with normal levels (adjusted odds ratio [OR], 2.19; 95% CI, 1.88-2.56; Table 2). Similarly, among patients with available troponin I levels, the risk of in-hospital mortality was found to increase significantly with increasing troponin I levels by both categorical (quartiles 2, 3, and 4 vs quartile 1; Table 2) and continuous analysis (Figure 2), independent of other predictive risk factors, including BNP levels at admission.

Among other in-hospital clinical outcomes, patients with elevated troponin levels had a longer length of stay and lower proportional discharge to home than those with troponin in the normal range (Table 1). In adjusted analysis, elevated troponin level (vs normal level) was significantly associated with longer length of stay and lower likelihood of discharge to home (Table 2; eTable 3 in the Supplement). Similarly, increasing levels of troponin I (quartiles 2, 3, and 4 vs quartile 1) were associated with significantly higher likelihood of these adverse in-hospital outcomes among participants with available troponin I measurements (Table 2). Among post-discharge outcomes, elevation in troponin I levels was strongly associated with higher risk of 30-day mortality, 30-day all-cause readmission, and 1-year mortality in multivariable adjusted Cox proportional hazards analysis (Table 3).

Several sensitivity analyses were performed to test the robustness of our study findings. In the subgroup of patients with HFpEF without history of myocardial infarction or revascularization (22 620 patients from 217 sites), 21.8% had elevated peak troponin levels during index hospitalization, with a significant independent association between elevated troponin levels and greater likelihood of adverse in-hospital clinical outcomes, similar to the overall study population (eTable 4 in the Supplement).

Among participants who had only troponin T measurement available, a significant association was observed between higher levels of troponin T and likelihood of adverse in-hospital outcomes (Table 2). Furthermore, elevation in troponin T was also significantly associated with greater risk of 30-day and 1-year mortality but not 30-day all-cause readmission (Table 3).

Finally, the prevalence of troponin elevation was higher among patients with decompensated HFrEF vs HFpEF hospitalized at the GWTG-HF centers included in our study (34.0% vs 22.6%). The rates of in-hospital mortality, discharge home, and median length of stay across different strata of troponin levels were comparable among patients with decompensated HFrEF and HFpEF (eTable 5 in the Supplement).

In this cohort of patients with decompensated HFpEF in whom troponin T or I was measured as part of clinical care, more than one-fifth had elevated peak troponin levels during the acute hospitalization. Presence of elevated troponin levels was associated with greater risk of adverse clinical outcomes. Compared with patients with normal troponin levels, those with elevated peak troponin levels during index hospitalization had significantly longer length of stay and higher risk of in-hospital, 30-day, and 1-year mortality. Taken together, these finding suggest that troponins are potentially useful prognostic markers in decompensated HFpEF.

Troponin levels were measured in up to one-third of hospitalized patients with HFpEF during the study period. We observed troponin elevation in 22.6% of patients with acutely decompensated HFpEF, with a similar proportion among those with vs without history of ischemic heart disease. The use of revascularization procedures in our study population of patients with decompensated HFpEF was very low, suggesting that the observed troponin elevation was not likely related to a concomitant ACS event. Quiz Ref IDPrevious small cohort studies have reported troponin elevation among 26% to 50% of patients with acutely decompensated HFpEF.^13-16 This variability in the probability of troponin elevation could be related to differences in assays and cutoffs used to define troponin elevation, the patient populations studied, and whether troponin was measured routinely or selectively. Furthermore, the marked heterogeneity in the pathophysiologic characteristics of HFpEF may also contribute to the variability in the probability of troponin elevation across studies.^17,18

Previous studies have established the prognostic value of elevated troponin levels among patients with acute coronary syndrome, stable coronary artery disease, and decompensated HFrEF.^6,19,20 In a study from the ADHERE registry, Peacock et al⁵ reported that positive cardiac troponin test findings among patients with acute decompensated HF was independently associated with higher in-hospital mortality, but patients with HFrEF and HFpEF were not separately evaluated. In a recent secondary analysis from the BARI-2D study, Everett et al²⁰ demonstrated that cardiac troponin T concentration was an independent predictor of major adverse cardiovascular outcomes among patients with type 2 diabetes and stable ischemic heart disease. The present study adds important findings to the existing literature by demonstrating the prognostic role of cardiac troponin levels in a large cohort of patients with HFpEF.

A few small studies have previously evaluated the prognostic significance of elevated troponin levels among hospitalized patients with HFpEF. Sanders-van Wijk et al¹⁵ reported a significant association between higher troponin levels and risk of HF hospitalization or death among symptomatic patients with HFpEF included in the TIME-CHF trial. Similar findings were also reported from another European cohort of patients with HFpEF with 84% higher risk of death or HF readmission among patients with elevated cardiac troponin levels.¹⁴ However, these studies were limited by small sample size and small numbers of outcome events, which limited the ability to adjust for several potential confounders. Furthermore, these studies included only European patients, and their findings may not be generalizable to a more ethnically diverse HFpEF population. This is particularly relevant considering the recently observed geographical heterogeneity in HFpEF characteristics and associated clinical outcomes.^21-23 The present study, including 34 233 patients with HFpEF (20% blacks or Hispanics) from 224 US sites, represents the largest and most comprehensive evaluation to our knowledge of the prognostic significance of cardiac troponin levels in this patient population.

Quiz Ref IDWe identified significant clinical predictors of troponin elevation in patients with HFpEF. In particular, male sex, black race, presence of renal function abnormalities, ischemic heart disease, and other clinical markers of neurohormonal stress were independently associated with troponin elevation in patients with HFpEF in the absence of a precipitating ACS. It is plausible that the presence of ischemic heart disease or subclinical nonobstructive coronary artery disease may exaggerate the demand-supply mismatch in the setting of decompensated HF and volume overload and lead to latent myocardial injury. Renal dysfunction may be associated with elevated troponin levels among patients with HFpEF owing to decreased clearance of troponin and greater burden of myocardial abnormalities, including higher left ventricular mass and severe diastolic dysfunction in patients with chronic kidney disease.²⁴Quiz Ref ID Furthermore, presence of infiltrative cardiomyopathy from conditions such as cardiac amyloid in patients with HFpEF may also contribute to the elevated troponin levels.^25,26 The demographic profile of male sex and African American ethnicity that was independently associated with elevation in troponin levels among patients with HFpEF is also common to patients with transthyretin cardiac amyloid.²⁷ Future studies with race-stratified and genotype analysis are needed to further test this hypothesis. Finally, direct cellular toxic effects from elevated neurohormones in the decompensated state can cause cardiomyocyte apoptosis and autophagy and lead to elevated troponin levels.²⁸

Our study findings have important clinical implications. We observed a significant association between presence of elevated peak troponin levels and risk of adverse clinical outcomes among patients with decompensated HFpEF independent of other well-established risk predictors, including BNP. This association was significant even among patients without a history of coronary artery disease and was consistent for both in-hospital and long-term outcomes. Taken together, our study findings suggest that troponin elevation may identify sicker patients with decompensated HFpEF and highlight its utility as an early risk-stratification tool in this patient population.

Current guidelines recommend early risk stratification of patients with decompensated HF using BNP, troponin assessment, and multivariable clinical risk score to facilitate informed decision making.⁷ However, no separate recommendations are provided for uses of biomarkers to risk stratify patients with decompensated HFpEF. In the GWTG-HF registry, assessment of troponin levels was performed in only one-third of patients with decompensated HFpEF during the study period. Addition of troponin measurement to risk-stratification models for patients with acutely decompensated HFpEF may identify those with increased in-hospital and long-term mortality risk who could benefit from early implementation of aggressive therapies as well as closer postdischarge follow-up. Furthermore, it may also identify more resource-intensive cases that would require longer length of stay, and thus it could be useful for efficient allocation of hospital resources and help in providing cost-effective care. Finally, our study findings may also have important implications for future HFpEF research. Similar to elevated BNP levels, troponin elevation may be useful as an objective criterion to select high-risk patients and as a treatment target for future clinical trials evaluating novel drug therapies for HFpEF.²⁹

Several limitations to our study are noteworthy. First, owing to the retrospective nature of the study there is a potential for residual or unmeasured confounding, and a cause-effect relationship between elevations in cardiac troponin and risk of adverse clinical outcomes cannot be established. However, the associations observed in our study were strong, independent of other predictive risk factors, consistent across categorical as well as continuous measures of troponin, and in agreement with the findings from previous small studies.^13,14

Second, cardiac troponin levels were measured using a variety of assays across participating centers with different cutoff points for elevated levels. However, we observed a consistent pattern of results across different measures of troponin level, including laboratory-defined elevated vs normal, data-derived categories of troponin I, and continuous measures of troponin I levels. This consistency in our observation highlights the potential robustness of our study findings. Furthermore, use of laboratory-specific cutoffs to identify elevated troponin levels increases the generalizability of our study findings in general practice.

Third, cardiac troponin was not measured in all patients with HFpEF, and it is likely that clinical factors influenced the decision to measure troponins in many patients. However, the possibility of significant selection bias in our study findings is low because there were no major clinically relevant differences in most of the baseline demographic and clinical characteristics and in-hospital outcomes of patients with HFpEF with vs without troponin assessment.

Fourth, only a single measure of peak troponin was available in each case; thus, a true rise and fall in the biomarker levels due to underlying acute coronary syndrome cannot be excluded. However, GWTG-HF participants all have primary HF discharge diagnoses, and the possibility of significant underlying ACS is low, as reflected by the very low use of revascularization procedures during the index hospitalization in our study population. Finally, participation in the GWTG-HF program is voluntary, and our study findings may not be generalizable to other hospitals with different patient case mix, resource availability, and care patterns from GWTG-HF participating centers.

In conclusion, troponin elevation is common among hospitalized patients with decompensated HFpEF and is associated with poor in-hospital and long-term clinical outcomes. Taken together, these findings suggest a role of early troponin assessment as an important risk-stratification tool during the initial evaluation of these patients. Future studies are needed to determine if troponin levels could be used to guide targeted therapies among high-risk patients with HFpEF.

Accepted for Publication: October 16, 2016.

Corresponding Author: Gregg C. Fonarow, MD, Ahmanson-UCLA Cardiomyopathy Center, Ronald Reagan UCLA Medical Center, 10833 LeConte Ave, Room 47-123 CHS, Los Angeles, CA 90095-1679 (gfonarow@mednet.ucla.edu).

Published Online: December 28, 2016. doi:10.1001/jamacardio.2016.4726

Author Contributions: Dr Fonarow had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Pandey, Golwala, DeVore, Yancy, Fonarow.

Acquisition, analysis, or interpretation of data: Pandey, Sheng, Hernandez, Bhatt, Heidenreich, de Lemos, Fonarow.

Drafting of the manuscript: Pandey, Golwala, Fonarow.

Critical revision of the manuscript for important intellectual content: Pandey, Sheng, DeVore, Hernandez, Bhatt, Heidenreich, Yancy, de Lemos, Fonarow.

Statistical analysis: Pandey, Sheng, DeVore.

Obtained funding: Hernandez.

Administrative, technical, or material support: Sheng, Hernandez, Fonarow.

Database leadership: Yancy.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. The following conflicts were reported: Dr Bhatt, Advisory Board: Cardax, Elsevier Practice Update Cardiology, Medscape Cardiology, Regado Biosciences; Board of Directors: Boston VA Research Institute, Society of Cardiovascular Patient Care; Chair: American Heart Association Quality Oversight Committee; Data Monitoring Committees: Duke Clinical Research Institute, Harvard Clinical Research Institute, Mayo Clinic, Population Health Research Institute; Honoraria: American College of Cardiology (Senior Associate Editor, Clinical Trials and News, ACC.org), Belvoir Publications (Editor in Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical trial steering committees), Harvard Clinical Research Institute (clinical trial steering committee), HMP Communications (Editor in Chief, Journal of Invasive Cardiology), Journal of the American College of Cardiology (Guest Editor; Associate Editor), Population Health Research Institute (clinical trial steering committee), Slack Publications (Chief Medical Editor, Cardiology Today’s Intervention), Society of Cardiovascular Patient Care (Secretary/Treasurer), WebMD (CME steering committees); Other: Clinical Cardiology (Deputy Editor), NCDR-ACTION Registry Steering Committee (Vice-Chair), VA CART Research and Publications Committee (Chair); Research Funding: Amarin, Amgen, AstraZeneca, Bristol-Myers Squibb, Eisai, Ethicon, Forest Laboratories, Ischemix, Medtronic, Pfizer, Roche, Sanofi Aventis, The Medicines Company; Royalties: Elsevier (Editor, Cardiovascular Intervention: A Companion to Braunwald’s Heart Disease); Site Co-Investigator: Biotronik, Boston Scientific, St. Jude Medical; Trustee: American College of Cardiology; Unfunded Research: FlowCo, PLx Pharma, Takeda. Dr Adrian F. Hernandez – Research: Janssen, Novartis, Portola, Bristol-Myers Squibb; Consulting: Bristol-Myers Squibb, Gilead, Boston Scientific, Janssen, Novartis. Dr James de Lemos – Research: Abbott Diagnostics and Roche Diagnostics; Consulting: Roche Diagnostics, Abbott Diagnostics, Siemen’s Health Care, and Radiometer, Inc. Dr Gregg C. Fonarow – Research: Agency for Healthcare Research and Quality, National Institutes of Health; Consulting: Amgen, Baxter, Bayer, Janssen, Novartis, and Medtronic. Dr Adam DeVore – Research: American Heart Association, Amgen, and Novartis. All other authors have no conflicts of interest to disclose.

Funding/Support: The American Heart Association provides the Get With The Guidelines–Heart Failure program (GWTG-HF), which has been previously funded through support from Medtronic, GlaxoSmithKline, Ortho-McNeil, and the American Heart Association Pharmaceutical Roundtable.

Role of the Funder/Sponsor: The funders/sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Disclaimer: While Dr Hernandez is Associate Editor, Dr Yancy Deputy Editor, and Dr Fonarow Associate Editor for Health Care Quality and Guidelines for JAMA Cardiology, they were not involved in any of the decisions regarding review of the manuscript or its acceptance.