P values are from Cox regression. Hazard ratios for the main and the per-protocol analyses are given in eTable 2 in the Supplement.
β-Blocker use analyzed at baseline without consideration of potential crossover during follow-up. The P value is for the interaction between β-blocker and the variable on the y-axis. Squares represent the hazard ratios and lines represent the 95% confidence intervals, for β-blocker yes vs no in the subgroup on the y-axis. Continuous variables were dichotomized at clinically relevant cut-offs. NYHA indicates, New York Heart Association; LVEF, left ventricular ejection fraction; and RAS, renin-angiotensin system.
P values are from Cox regression. Hazard ratios for the main and the per-protocol analyses are given in Table 2.
eFigure 1. Flowchart
eFigure 2. Distribution of propensity scores in the matched and overall HFPEF cohorts
eTable 1A. HFPEF patients, baseline data from the Swedish Heart Failure Registry, the Swedish Patient Registry, and Statistics Sweden, complete from the overall and matched cohorts. Main manuscript lists matched cohort alone
eTable 1B. HFREF patients for “positive control” analysis, baseline data from the Swedish Heart Failure Registry, the Swedish Patient Registry, and Statistics Sweden, complete from the overall and matched cohorts
eTable 2. Association between beta-blocker use and outcomes, “positive control” analyses in HFREF
eTable 3. ICD-10 and procedure codes used for baseline comorbidity and for outcome heart failure hospitalization from the Patient Registry
Customize your JAMA Network experience by selecting one or more topics from the list below.
Lund LH, Benson L, Dahlström U, Edner M, Friberg L. Association Between Use of β-Blockers and Outcomes in Patients With Heart Failure and Preserved Ejection Fraction. JAMA. 2014;312(19):2008–2018. doi:10.1001/jama.2014.15241
Heart failure with preserved ejection fraction (HFPEF) may be as common and may have similar mortality as heart failure with reduced ejection fraction (HFREF). β-Blockers reduce mortality in HFREF but are inadequately studied in HFPEF.
To test the hypothesis that β-blockers are associated with reduced all-cause mortality in HFPEF.
Propensity score–matched cohort study using the Swedish Heart Failure Registry. Propensity scores for β-blocker use were derived from 52 baseline clinical and socioeconomic variables.
Nationwide registry of 67 hospitals with inpatient and outpatient units and 95 outpatient primary care clinics in Sweden with patients entered into the registry between July 1, 2005, and December 30, 2012, and followed up until December 31, 2012.
From a consecutive sample of 41 976 patients, 19 083 patients with HFPEF (mean [SD] age, 76  years; 46% women). Of these, 8244 were matched 2:1 based on age and propensity score for β-blocker use, yielding 5496 treated and 2748 untreated patients with HFPEF. Also we conducted a positive-control consistency analysis involving 22 893 patients with HFREF, of whom 6081 were matched yielding 4054 treated and 2027 untreated patients.
β-Blockers prescribed at discharge from the hospital or during an outpatient visit, analyzed 2 ways: without consideration of crossover and per-protocol analysis with censoring at crossover, if applicable.
Main Outcomes and Measures
The prespecified primary outcome was all-cause mortality and the secondary outcome was combined all-cause mortality or heart failure hospitalization.
Median follow-up in HFPEF was 755 days, overall; 709 days in the matched cohort; no patients were lost to follow-up. In the matched HFPEF cohort, 1-year survival was 80% vs 79% for treated vs untreated patients, and 5-year survival was 45% vs 42%, with 2279 (41%) vs 1244 (45%) total deaths and 177 vs 191 deaths per 1000 patient-years (hazard ratio [HR], 0.93; 95% CI, 0.86-0.996; P = .04). β-Blockers were not associated with reduced combined mortality or heart failure hospitalizations: 3368 (61%) vs 1753 (64%) total for first events, with 371 vs 378 first events per 1000 patient-years (HR, 0.98; 95% CI, 0.92-1.04; P = .46). In the matched HFREF cohort, β-blockers were associated with reduced mortality (HR, 0.89; 95% CI, 0.82-0.97, P=.005) and also with reduced combined mortality or heart failure hospitalization (HR, 0.89; 95% CI, 0.84-0.95; P = .001).
Conclusions and Relevance
In patients with HFPEF, use of β-blockers was associated with lower all-cause mortality but not with combined all-cause mortality or heart failure hospitalization. β-Blockers in HFPEF should be examined in a large randomized clinical trial.
Quiz Ref IDUp to half of patients with heart failure have normal or near-normal ejection fraction,1 termed heart failure with preserved ejection fraction (HFPEF) or diastolic heart failure. The mortality in HFPEF may be as high as in heart failure with reduced ejection fraction (HFREF), also termed systolic heart failure,1 but there is no proven therapy.
Both HFPEF and HFREF share numerous features including catecholamine and other neurohormonal activation,2,3 elevated filling pressures, and classical heart failure signs and symptoms.4 β-Blockers improve outcomes in HFREF5 and may be beneficial in HFPEF by lowering blood pressure and reducing left ventricular hypertrophy6 and diastolic dysfunction7 and by slowing heart rate and reducing myocardial oxygen demand. However, data on β-blockers to treat HFPEF are sparse and inconclusive,8-13 and β-blockers are currently not indicated for treating HFPEF.14,15 Therefore, we tested the hypothesis that β-blockers are associated with reduced mortality in HFPEF.
The Swedish Heart Failure Registry provided the study population and baseline clinical characteristics and medications. This nationwide registry has been previously described.16 Inclusion criteria are clinician-judged heart failure. The protocol, registration form and annual reports are available at http://www.rikssvikt.se. Ejection fraction is categorized as less than 30%, 30% to 39%, 40% to 49%, and 50% or higher. The 2 highest ejection fraction categories define HFPEF, which we used for setting our inclusion criteria (eFigure 1 in the Supplement). An ejection fraction ranging from 40% through 49% is generally not considered normal. It was included herein because randomized clinical trials (RCTs) included only patients with ejection fractions ranging from less than 35% to 40%. However, we also performed prespecified subgroup analyses by ejection fractions from 40% through 49% and 50% or higher.
Patients were included in this study if the index date was between July 1, 2005, and December 30, 2012. The index date was defined as the date of an outpatient visit or hospital discharge; patients who died during the index hospitalization were excluded. Follow-up was until December 31, 2012.
The Swedish Board of Health and Welfare (http://www.socialstyrelsen.se) maintains the Death Registry, the Patient Registry, and the Dispensed Drug Registry. The Death Registry provided date of death. The Patient Registry provided additional baseline comorbidities and the outcome heart failure hospitalization. It contains International Statistical Classification of Diseases, Tenth Revision (ICD-10) codes (eTable 3 in the Supplement) for encounters as inpatients and as outpatients at specialty clinics and is updated and validated annually (last update and thus end of follow-up for this study, December 31, 2012). The positive predictive value for most diagnoses is between 85% and 95%17; a heart failure diagnosis was verified in between 86% and 91% of cases.18 Comorbidities present at baseline (or prior) were defined by corresponding ICD-10 codes in any position between January 1, 1997, when use of ICD-10 codes began, and up to and including the index date (except for malignancy and musculoskeletal and psychiatric disorders, counted only if the corresponding ICD-10 code was present, ie, a health care encounter for this diagnosis had occurred, in the 3 years preceding the index date). The outcomes heart failure hospitalization and death were defined as between the day after the index date and end of follow-up, December 31, 2012, for which a heart failure diagnosis was required as the primary or first secondary diagnosis for hospitalization. In the main analysis, the exposure, β-blocker use, was analyzed at baseline, without consideration of potential crossover during follow-up, but in a per-protocol consistency analysis, the dispensed-drug registry was used to assess β-blocker use throughout follow-up, with censoring at crossover. This registry contains details about every prescription filled in Sweden since July 1, 2005, which was therefore the date for the start of this study. All pharmacies are required to participate by law, ensuring that it is essentially 100% complete because data are transferred electronically when drugs are dispensed. Statistics Sweden maintains socioeconomic data on all Swedish citizens and provided additional baseline data.
All Swedish citizens have unique personal identification numbers that enable linking of disease-specific health registries and governmental health and statistical registries. Establishment of the heart failure registry and this analysis with linking to the above registries were approved by a multisite ethics committee. Individual patient consent was not required, but patients were informed of entry into national registries and allowed to opt out.
A propensity score for treatment with β-blockers was estimated for each patient with logistic regression19 using 52 relevant baseline variables (Table 1, Table 2, and Table 3). The propensity score is the propensity from 0 to 1 of receiving a treatment given a set of known variables and is used to attempt to adjust for potential selection bias, confounding, and differences between treatment groups in observational studies.19,20
Continuous variables were modeled using restricted cubic splines with 3 dfs. Missing data were handled by estimating separate logistic regressions for each missing variable pattern on available observations (yielding a total of 1248 unique missing data patterns and thus logistic regression models). Each individual received the propensity score from the model that incorporated all nonmissing variables for that individual. Using matching without replacement,21 a cohort was constructed matching each untreated patient to the 2 closest treated patients in which age differed by 5 or fewer years and the propensity score differed by 0.01 or less. The ability of matching to balance the groups was assessed visually (eFigure 2 in the Supplement) and the ability of the propensity score to balance baseline characteristics was assessed by absolute standardized differences (the difference in percentage between the means for the 2 groups divided by the mutual standard deviation; Table 1). Standard differences of less than 10% are considered inconsequential.22
The prespecified primary outcome was all-cause mortality; the secondary outcome was all-cause mortality or first heart failure hospitalization. Crude outcomes in the overall cohort were assessed with Kaplan-Meier analysis and Cox regression. Adjustment for confounders was made by assessing outcomes in the matched cohort also with Kaplan-Meier analysis and Cox regression, for which the matched pairs were modeled using a frailty term. The scaled Schoenfeld residuals and df βs from the models were investigated to detect violations to the proportional hazards assumption and possible influential outliers, respectively; none were detected. Interactions between β-blocker use and prespecified clinically important variables were estimated by Cox regression and displayed in a forest plot.
Statistical analyses were performed using R version 3.0.3 (R Foundation for Statistical Computing). The level of significance was 5% and all P values and confidence intervals were 2-sided.
In HFREF, the magnitude of mortality reduction is greater for β-blockers5 compared with renin-angiotensin system antagonists.23 In a previous study from our registry, renin-angiotensin system antagonists were associated with a hazard ratio (HR) of 0.91 for mortality, and 1-year survival among propensity score–matched treated patients was 77% and untreated patients was 72%.16 Assuming a similar or higher benefit associated with β-blockers, an accrual period of 7 years (from our population) and a power of 80% and 2-sided significance level of .05, at least 2204 patients would be needed in the smallest group.
We performed numerous combinations of prespecified consistency analyses for the primary and secondary outcomes.
In per-protocol analysis, continued β-blocker use was followed prospectively in the Dispensed Drug Registry and patients were censored at crossover. In untreated patients, crossover was defined as de novo β-blocker (Anatomical Therapeutic Chemical code C07) dispension. Prescriptions are normally for 3-month durations but pills may last longer. Therefore, for a treated patient, crossover was defined as failure to refill, ie, absent β-blocker dispension, within 6 months from baseline or from last dispension, where crossover and thus censoring was set 3 months from index or last refill. Matching minimizes confounding but may impair generalizability. Therefore we also performed a Cox regression in the overall HFPEF cohort with adjustment for the propensity scores as a covariate.
Quiz Ref IDIn HFREF, β-blockers reduce mortality.5 Therefore, we performed a positive-control analysis in HFREF to confirm that results in our registry are representative of those in RCTs. The HFREF patients were selected and outcomes analyzed with the same methods as for HFPEF.
Between July 1, 2005, and December 30, 2012, there were 76 133 registrations from 67 of approximately 75 hospitals and 94 of approximately 1000 primary care clinics in Sweden. We included 19 083 individual patients with HFPEF, yielding 15 786 treated and 3297 not treated with β-blockers in the overall HFPEF cohort. After propensity score matching 2 treated to 1 untreated patient, there were 8244 patients (5496 treated and 2748 un-treated; eFigure 1 in the Supplement).
Baseline characteristics are shown for the matched HFPEF cohort in Table 1, Table 2 and Table 3 and for the overall and matched HFPEF cohorts in eTable 1A in the Supplement. In the overall cohort, patients were a mean (SD) age of 76 (12) years and 46% were women. Treated patients were younger, received more specialized care and planned follow-up, and given more renin-angiotensin system antagonists, but also had lower ejection fraction. The distributions of propensity scores were therefore different (eFigure 2 in the Supplement). In the matched cohort, patients were a mean (SD) age of 78 (11) years and 46% were women. The standardized differences between the groups were considerably smaller, with only 1 of the 52 variables having a standardized difference of more than 10% (location of care: outpatient physician, which was more common in untreated patients) and a majority in the range of 0% to 3% (Table 1). The distributions of propensity scores were therefore essentially identical (eFigure 2 in the Supplement).
Table 4 lists and Figure 1 shows outcomes in the main analysis. In the overall HFPEF cohort, the median follow-up time was 755 days (range, 1-2770 days), for a total of 46 976 patient-years of follow-up. Among those who received β-blockers, the total number of deaths was 5639 (36%) vs 1518 (46%) among those who did not receive them. The β-blocker group had 143 (95% CI, 140-147) deaths per 1000 patient-years vs 198 (95% CI, 189-209) among those who did not receive the agents (P < .001). Survival at 1 year was 84% (95% CI, 83%-85%) among those who received β-blockers vs 78% (95% CI, 76%-79%) among those who did not. At 5 years, survival was 51% (95% CI, 50%-52%) among those who received β-blockers vs 41% (95% CI, 39%-44%) among those who did not. The unadjusted HR throughout follow-up was 0.73 (95% CI, 0.69-0.77; P < .001).
In the matched cohort, the median follow-up time was 709 days (range, 1-2770 days), for a total of 19 391 patient-years of follow-up. The total number of deaths for those who received β-blockers was 2279 (41%) vs 1244 (45%) among those who did not, and the number of deaths per 1000 patient-years was 177 (95% CI, 170-184) among those who received β-blockers vs 191 (95% CI, 181-202) among those who did not (P = .03). At 1 year, survival was 80% (95% CI, 79%-81%) among those who received β-blockers vs 79% (95% CI, 78%-81%) among those who did not. At 5 years, it was 45% (95% CI, 43%-47%) among those who received β-blockers vs 42% (95% CI, 40%-45%) among those who did not. The HR throughout follow-up was 0.93 (95% CI, 0.86-0.996, P = .04; Table 4 and Figure 1), for a numbers-needed-to-treat to save 1 life of 100 at 1 year and 33 at 5 years.
In the overall cohort, crude survival free from heart failure hospitalization at 1 year was 62% (95% CI, 61%-62%) among those who received β-blockers vs 58% (95% CI, 57%-60%) among those who did not, with an unadjusted HR throughout follow-up of 0.86 (95% CI, 0.82-0.91; P < .001). In the matched cohort, it was 58% (95% CI, 57%-59%) vs 59% (95% CI, 57%-61%), with an HR throughout follow-up of 0.98 (95% CI, 0.92-1.04; P = .46; Table 4 and Figure 1).
Figure 2 shows the HRs for mortality associated with β-blocker use for clinically relevant prespecified subgroups. There were no interactions between β-blockers and any subgroup, including with ejection fraction.
The HFPEF consistency analyses are shown in Table 2. In HFPEF per-protocol analysis, censoring occurred due to crossover from not using a β-blocker to using one in 1292 of 3297 patients (39%) at a median of 115 days (interquartile range [IQR], 46-390) and from using β-blockers to not using them in 6176 of 15 786 (39%) at 398 days (IQR, 189-778). Hazard ratios were more favorable for β-blocker use, but the patterns in the main analysis (without consideration of crossover) were repeated. When adjusting for rather than matching by propensity score, the HR was 0.97 (95% CI, 0.91-1.03; P = .32).
The HFREF positive-control consistency analysis is shown in eTable 2 in the Supplement and Figure 3 and yielded 22 893 patients with HFREF overall and 6081 matched (4054 treated and 2027 untreated) patients. In the HFREF overall cohort, crude survival at 1 year was 84% (95% CI, 84%-85%) in treated vs 73% (95% CI, 71%-75%) in untreated patients, with an unadjusted HR throughout follow-up of 0.63 (95% CI, 0.59-0.67; P < .001).
In the HFREF matched cohort, the total number of deaths for those who received β-blockers was 1798 (44%) vs 1000 (49%) among those who did not, and the total number of deaths per 1000 patient-years was 190 (95% CI, 181-199) for those who received β-blockers vs 209 (95% CI, 196-222) among those who did not (P = .02). At 1 year, survival was 78% (95% CI, 76%-79%) vs 75% (95% CI, 73%-77%), with an HR throughout follow-up of 0.89 (95% CI, 0.82-0.97; P = .005). In HFREF, β-blockers were associated also with reduced combined heart failure hospitalization or mortality, with an HR in the matched cohort of 0.89 (95% CI, 0.84-0.95; P = .001). In the per-protocol analysis, HRs were again more favorable for β-blocker use but the patterns in the analysis without consideration of crossover were repeated.
Quiz Ref IDIn this large prospective propensity score–matched registry analysis of unselected patients with HFPEF, β-blockers were associated with reduced all-cause mortality. Findings in HFPEF were also similar to findings in a positive-control consistency analysis in patients with HFREF, in whom β-blockers have been proven to reduce mortality. However, β-blockers used to treat HFPEF were not associated with improved combined all-cause mortality or heart failure hospitalization.
Quiz Ref IDThere are few previous studies of β-blockers used to treat HFPEF. The Study of Effects of Nebivolol Intervention on Outcomes and Rehospitalization in Seniors With Heart Failure (SENIORS) demonstrated reduced combined all-cause mortality or cardiovascular hospitalization but not mortality alone.8 This was irrespective of ejection fraction (≤35% or >35%), but only a small minority had preserved ejection fraction. Among those with an ejection fraction higher than 35%, 77% had coronary disease,24 which may explain the benefit. In small RCTs, β-blockers improved the ratio of early to late ventricular filling velocity7 but not clinical outcomes.9 Among observational studies, 2 reports from the Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) registry did not find an association between β-blocker use and outcomes,10,11 but 2 small studies12,13 and a meta-analysis25 suggested improved outcomes.
In our overall HFPEF cohort, 83% were receiving β-blockers, which was nearly as high as the 90% in our overall HFREF cohort. This is higher use than that in previous observational HFREF and HFPEF reports.1,10,11,26,27 Only 30% had prior myocardial infarction; 50%, atrial fibrillation. β-Blocker use was only slightly higher when treating ejection fraction levels ranging from 40% through 49% compared with ejection fraction levels of 50% or higher. The high use in HFPEF may reflect Swedish tradition or a perceived benefit in HFPEF, despite lack of evidence and recommendations for its use.14,15
With broad coverage in cardiology, internal medicine, geriatrics, and primary care throughout Sweden, our study was unselective and the findings likely representative of and generalizable to a broad general HFPEF population. Patients were old with median age of 76 years, overall, and 78 years in the matched cohort compared with a median age of 76 years of participants in the SENIORS trial, for which the inclusion criterion was that patients were at least 70 years. Furthermore, 46% were women compared with 37% in the SENIORS trial.8 Patients were quite ill, with a vast majority with a New York Heart Association class of II or III; a median N-terminal pro-brain natriuretic peptide of approximately 2000 ng/L; and 80% requiring chronic diuretic use.
In unadjusted analysis, there was a separation of survival curves, with a 23% reduction in mortality. This may partly be explained by differences in baseline characteristics. Thus, the reduction in mortality was attenuated after matching by age and propensity score, for which patient characteristics were virtually identical, and the separation in survival curves was more subtle. In the main analysis, with consideration of β-blocker use at baseline only, the mortality reduction was 7%. However, unlike in RCTs, there were no specific efforts to maintain treatment, and considerable crossover occurred in both directions; in the per-protocol analysis, the mortality-reduction was greater at 14%.
The mortality-reduction in HFPEF was significant but may seem numerically low. However, in the HFREF positive-control consistency analysis, mortality-reductions were only slightly greater: 11% in the main analysis (vs 7% in HFPEF) and 21% in the per-protocol analysis (vs 14% in HFPEF). This was lower than the 35% mortality reduction in a meta-analysis of RCTs that assessed treatment of patients with HFREF,5 suggesting that the numerically low mortality reduction in HFPEF was due to study design and population rather than a lesser potential effect of β-blockers used to treat HFPEF than to treat HFREF.
This difference may have several explanations. Most RCTs recruited stable outpatients with chronic heart failure. In our study, more than 60% were inpatients, for whom β-blocker initiation and continuation is considered safe,28 but may be associated with complications.29 Outcomes after discharge are distinctly worse for inpatients than for outpatients with chronic heart failure,30 and may be related to the recent hospitalization itself rather than heart failure,31 and thus less likely to be prevented by β-blockers. The rigorous follow-up and monitoring in randomized trials may also minimize adverse events such as hypotension, bradycardia, presyncope, or syncope. Our patients were also unselected and thus older with more comorbidity than the selected patients in randomized trials,5 which may have made our patients more susceptible to complications and death from noncardiovascular causes, again unlikely to be prevented by β-blockers.
β-Blockers were associated with reduced mortality across subgroups. With only 30% prior myocardial infarction and no interaction with β-blockers, a potential benefit is unlikely mediated via ischemic heart disease. There was no interaction with ejection fraction, suggesting potential benefits are similar in ejection fraction levels of 50% or higher compared with ejection fraction values ranging 40% through 49%.
In HFREF RCTs, the reduction in heart failure hospitalization was similar to the reduction in mortality.5 Thus, it was surprising that β-blockers were not associated with reduced combined mortality or heart failure hospitalization in HFPEF. There are several potential explanations.
Hospitalization for heart failure was not adjudicated, and we included ICD-10 codes for the main but also for alternative diagnoses in the first position, for which the presence of chronic heart failure may be reflected even if it was unrelated to a noncardiovascular hospitalization, which may be common both in elderly patients and in patients with heart failure.31 Indeed, in HFPEF, noncardiovascular comorbidity is more common and a stronger predictor of hospitalization, even heart failure hospitalization.32,33Quiz Ref IDThe potential complications prevented by monitoring in RCTs may explain the relatively lower risk reduction for mortality in our study, but may be especially relevant for hypotension, bradycardia and presyncope, which may lead to hospitalization and β-blocker discontinuation to a relatively greater extent than it leads to death. Indeed, there was a nonsignificant signal toward lesser risk reduction in mortality for the elderly, which may be even more pronounced for hospitalization.
Another explanation may be the paradoxical nature of rehospitalization. Mortality and heart failure hospitalization have decreased over time,34 but rehospitalization remains common.30,35 For more than 60% of the inpatients in our study, β-blockers may have been associated with complications29 and resulted in early rehospitalization. Indeed, in heart failure, readmission may be paradoxically inversely related to mortality.36 Rehospitalization of patients with heart failure overwhelmingly early30,37 and early rehospitalization may be caused by a posthospital syndrome, because of which re-hospitalization results from causes other than worsening heart failure.31
Our observational study is subject to selection bias and confounding. Our registry contains extensive baseline variables, and with access to nationwide health and statistical registries, we had complete data on comorbidities such as malignancy and markers of frailty, and importantly socioeconomic variables such as level of education and income and family status. By matching patients by propensity scores generated from 52 variables, we accounted for most conceivable confounders. Nevertheless, we cannot rule out inaccuracies in registry data or residual unmeasured confounding. The reason for β-blocker use or nonuse was not available. Untreated patients may have been at higher risk due to past intolerance or perceived risk of future intolerance or have been at lower risk due to absent indication. With decreasing numbers of patients at risk and increasing crossover over time, findings at longer durations of follow-up should be interpreted with caution but should, if anything, underestimate potential benefits associated with β-blocker use. The low magnitude of risk reduction (also in HFREF) and absence of reduction in combined mortality and heart failure hospitalization and mortality in covariate-adjusted analysis all have potential explanations but are reasons to interpret our findings with caution. We report all-cause mortality and heart failure hospitalization but did not have access to cause-specific mortality. The positive-control HFREF analysis strengthens our findings but still cannot replace a randomized study design.
In patients with HFPEF, use of β-blockers was associated with lower all-cause mortality but not with lower combined all-cause mortality or heart failure hospitalization. β-Blockers in HFPEF should be studied in a sufficiently powered RCT.
Corresponding Author: Lars H. Lund, MD, PhD, Department of Cardiology, Section for Heart Failure, Karolinska University Hospital, N305, 171 76 Stockholm, Sweden (Lars.Lund@alumni.duke.edu).
Author Contributions: Drs Lund and Benson had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr Dahlström is the founder of and registrar for the Swedish Heart Failure Registry. Dr Edner is the cofounder of the Swedish Heart Failure Registry.
Study concept and design: Lund, Dahlström, Edner, Friberg.
Acquisition, analysis, or interpretation of data: Lund, Benson, Dahlström, Edner, Friberg.
Drafting of the manuscript: Lund.
Critical revision of the manuscript for important intellectual content: Benson, Dahlström, Edner, Friberg.
Statistical analysis: Benson, Friberg.
Obtained funding: Lund, Dahlström.
Administrative, technical, or material support: Lund, Dahlström, Edner.
Study supervision: Lund.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Lund reported that he has received research grants to his institution, served on the speakers bureau or received consulting fees from AstraZeneca, Novartis, Vifor Pharma, and Boston Scientific. Dr Dahlström reports that he has served on the speakers bureau or received consulting fees from Novartis and Vifor Pharma. No other disclosures reported.
Funding/Support: The Swedish Heart Failure Registry is funded by the Swedish National Board of Health and Welfare, the Swedish Association of Local Authorities and Regions, the Swedish Society of Cardiology, and the Swedish Heart-Lung Foundation. This study was supported in part by grants 2013-23897-104604-23 from the Swedish Research Council, 20080409 and 20100419 from the Swedish Heart Lung Foundation, and 20090556 and 20110120 from the Sockholm County Council to Dr Lund’s institution.
Role of the Funder/Sponsor: The funders 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.
Additional Contributions: We thank all local center study personnel for data collection and entry and Eva Wallgren, BS, Department of Medicine, Karolinska Institutet, for assistance with figure formatting, who was not compensated for her contributions.
Create a personal account or sign in to: