European Heart Journal Advance Access originally published online on June 2, 2009
European Heart Journal 2009 30(17):2109-2116; doi:10.1093/eurheartj/ehp210
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Relationships between cardiac resynchronization therapy and N-terminal pro-brain natriuretic peptide in patients with heart failure and markers of cardiac dyssynchrony: an analysis from the Cardiac Resynchronization in Heart Failure (CARE-HF) study
1 Department of Cardiology, General Hospital of St Pölten, St Pölten, Austria
2 Department of Cardiology, Medical University of Vienna, Waehringer Guertel 18–20, A-1090 Vienna, Austria
3 School of Health and Population Sciences, University of Birmingham, Birmingham, UK
4 Department of Internal Medicine, Medical University of Graz, Graz, Austria
5 Department of Endocrinology and Metabolism, Medical University of Graz, Graz, Austria
6 Istituto di Ricovero e Cura a Carattere Scientifico, Policlinico San Matteo, Pavia, Italy
7 Department of Cardiology, University of Hull, Kingston upon Hull, UK
Received 13 April 2008; revised 1 April 2009; accepted 4 May 2009; online publish-ahead-of-print 2 June 2009.
* Corresponding author. Tel: +43 140 400 4616, Fax: +43 1408 1148, Email: rudolf.berger{at}meduniwien.ac.at
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
Aims: The Cardiac Resynchronization in Heart Failure (CARE-HF) study showed that cardiac resynchronization therapy (CRT) reduces mortality in HF patients with markers of dyssynchrony. Plasma N-terminal pro-brain natriuretic peptide (NT-proBNP) might predict which patients benefit most from CRT. We evaluated whether the prognostic value of NT-proBNP was influenced by CRT and the effects of CRT stratified according to NT-proBNP.
Methods and results: A total of 813 patients were enrolled in CARE-HF. Baseline log-transformed NT-proBNP independently predicted all-cause mortality, sudden death, and death from pump failure. In a multivariable model including log-transformed NT-proBNP, assignment to CRT remained independently associated with better prognosis without evidence of interaction. Stratifying patients according to the median NT-proBNP and to CRT treatment allocation, all-cause mortality was 12% if <median + CRT, 25% if <median + control group, 35% if 
median + CRT, and 51% if median + control group. There was no evidence of a difference in the relative effect of CRT across different values of NT-proBNP.
Conclusion: NT-proBNP retains its prognostic value in HF patients with CRT. Deploying CRT before the patients have reached end-stage HF may maximize the benefit of treatment.
Key Words: Heart failure • Cardiac resynchronization therapy • Natriuretic peptides • Prognosis
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
In patients with reduced LV systolic function and prolonged QRS, cardiac resynchronization therapy (CRT) improves cardiac function,1–3 heart failure (HF) symptoms2–4 and quality of life,2,3 increases exercise capacity3 and partially reverses maladaptive remodelling.2,3 Two large multicenter trials have demonstrated that CRT reduces morbidity,2,4 one of which, the Cardiac Resynchronization in Heart Failure (CARE-HF) study, also found a substantial reduction of total mortality.2,5 Both studies mainly included patients with moderate (New York Heart Association (NYHA) class III) symptoms at enrolment. Less than 15% of patients had severe symptoms (NYHA class IV). It is unclear if the survival benefits by CRT apply similarly to patients with different severity of disease.
Brain natriuretic peptide (BNP) and its precursor NT-proBNP are secreted from the heart in response to myocardial stretch which is related to atrial and ventricular filling pressures.6 Plasma concentrations are higher in patients with more severe symptoms or worse LV function and are powerful predictors of unfavourable outcomes such as all-cause mortality, sudden death, or readmission for decompensated HF.7,8 As natriuretic peptides often eliminate all other predictors of prognosis in statistical models,7 they are strong candidate markers for staging the severity of HF.
The CARE-HF study2,5 demonstrated a clear survival benefit from CRT in patients with chronic HF and markers of cardiac dyssynchrony. Despite including mainly patients in NYHA class III, the baseline NT-proBNP levels were distributed over a wide range, indicating different severities of disease and prognosis. N-terminal pro-brain natriuretic peptide might be used to identify high-risk patients who have more to gain, prognostically, from CRT. Also, reduction in NT-proBNP by CRT might lead to a change in the relationship between BNP and prognosis. This analysis investigates these issues.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
Design overview
The CARE-HF trial was a multi-centre, randomized, open-label study evaluating the effect of CRT on morbidity and mortality. Patients were recruited in 82 sites across Europe. Inclusion criteria were LV ejection fraction
35%, a QRS duration 150 ms or QRS ranging from 120 to 149 ms in addition to echocardiographic criteria for dyssynchrony, and NYHA functional class III or IV despite optimized medical therapy. Patients who were not in sinus rhythm were excluded. The protocol was approved by all relevant Ethics Committees, and patients provided written informed consent. The detailed study methods have been described elsewhere.2,5
Randomization and interventions
Using a minimization procedure (Quintiles, Dublin), patients were randomly assigned to receive medical therapy alone or with CRT (Medtronic Bakken Research Center, Maastricht, The Netherlands). Blood samples for measurement of NT-proBNP levels were drawn at baseline and at 3 months. The extracted plasma was analysed at the neurohumoral core-laboratory (Medical University Graz) using a Roche Diagnostics proBNP radioimmunoassay on an Elecsys 2010 analyser.9 As the analysis of NT-proBNP was performed after completion of the main study, the investigators were not aware of NT-pro-BNP values throughout the study.
Follow-up and outcomes
Patients were evaluated prior to discharge, at 1 month, 3 months, every 3 months for the first year, and every 6 months thereafter. Detailed assessments were performed at baseline and at 3 months. Investigators were asked to report all adverse events, which were adjudicated by an endpoints committee in a blinded fashion. The endpoints of this secondary analysis were all-cause mortality, sudden death, and death from pump failure using the extended follow-up period.5 The follow-up of the extension period ended on 30 September 2004.
|
Statistical analysis |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
All analyses were conducted according to the intention-to-treat principle and patients were analysed according to the group to which they were randomized, regardless of whether they received the treatment they were allocated to. All efforts were made to minimise missing data. However, some NT-proBNP values were missing because either the sample was mislaid during storage at investigator sites or inadequately labelled when received in the core laboratory. Data on certain echocardiographic variables were missing due to poor quality examination, damaged tapes, and the core laboratory not receiving certain tapes. The principal analyses were conducted on available data, and additional supportive data were conducted using multiple imputation techniques to assess the extent to which missing data may have affected the observed results. Multiple imputation was undertaken using the MI and MIANALYZE procedures in SAS V9.1 (SAS Institute, Cary, NC, USA). Tests were two-sided at P < 0.05. Cox proportional Hazards models were used to determine independent predictors of all-cause mortality, sudden death, and death from pump failure. Patients, who died from other causes than analysed, were censored at time of death. The models included patient level covariates that were pre-specified in the statistical analysis plan, namely age, sex, baseline clinical (aetiology, NYHA functional class, heart rate, supine systolic blood pressure, glomerular filtration rate), electrocardiographic (QRS duration), and echocardiographic characteristics (ejection fraction, mitral regurgitation area, end-systolic volume index, inter-ventricular mechanical delay), baseline medical therapy (use of an angiotensin converting enzyme-inhibitor or an angiotensin receptor blocker, use of a beta-blocker), NT-pro-BNP, and CRT. Variables were transformed using the logarithm and cubic spline to test for non-linear relationships with the outcome. The Akaike Information Criteria was used to determine the most appropriate transformation. The multivariable analysis was carried out using Cox proportional hazards model, with a stepwise approach. A P-value of 0.1 was set as the criterion for entry of each variable into the model, and a P-value of 0.05 was set as the criterion to stay in the model. The proportional hazards assumption was tested using Kolmogorov-type supremum tests computed on 1000 simulated patterns and was satisfied for all variables. Simple models including CRT treatment, BNP and an interaction term (log BNP*CRT treatment) were used to test whether BNP retains its prognostic power for these outcomes, irrespective of CRT treatment. Time-dependent covariate models were used to account for short-term changes in NT-proBNP. In these models, baseline values of NT-proBNP were updated with 3 months values for those individuals who had still not reached the endpoint under consideration, i.e. were still at risk of an event after this time. In the next step, patients were stratified according to the median NT-proBNP and treatment allocation, to illustrate the effects of CRT on patients with different stages of disease. Analyses were performed using SAS V9.1 (SAS Institute). The multivariable models were validated using the design library in the statistical package R (R Foundation for Statistical Computing, 2008) using the approach suggested by Harrell et al.10 Two hundred bootstrap samples were drawn from the original dataset. For each sample, a model was fitted using backward stepwise selection. This model was then frozen and applied to the original dataset. The differences between Somers's D for the model calculated on the bootstrap sample and the original dataset were averaged and this was subtracted from the value of D for the original model. This provides an estimate of the model optimism.
|
Results |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
Patients
Of 813 patients enrolled, 409 were randomly assigned to CRT and medical therapy and 404 to medical therapy alone. The patient characteristics have been published.2,5 In brief, patients were characterized by a median LV ejection fraction of 25% (IQR 22–29), a median NT-proBNP level of 1814 pg/mL (IQR 744–4199), and median QRS duration of 160 ms (IQR 152–180). Ischaemic heart disease was present in 42% of patients, 94% of patients were in NYHA class III and 6% in class IV. Almost all patients (95%) were treated with an ACE-inhibitor or ARB, 72% of patients had a beta-blocker, and more than half of the patients were treated with spironolactone. (Table 1).
|
There was no missing data for treatment allocation, age, sex, use of ACE-inhibitors or ARB, and beta-blocker use. There was <5% missing data for heart rate, QRS, systolic blood pressure, and ischaemic cause for HF. There was <10% missing data for end-systolic volume index, NT-proBNP, intraventricular mechanical delay, glomerular filtration rate, and ejection fraction. Mitral regurgitation index was not available in 25.6% of subjects.
Outcome
Details of the main result and extension phase have been published showing that CRT reduces both morbidity and mortality in this patient population.2,5 The present analysis uses the extended follow-up period. The median follow-up was 37.6 months (IQR 31.5–42.5 months). One hundred and fifty four of 404 patients died in the control group compared with 101 of 409 patients in the CRT group (Kaplan–Meier analysis—P < 0.0001). Cardiac resynchronization therapy reduced both the risk of sudden death (55 vs. 32 patients; Kaplan–Meier analysis—P = 0.0040) and death due to pump failure (64 vs. 38 patients; Kaplan–Meier analysis—P = 0.0028).
Univariate and multivariable predictors
All variables except sex, heart rate, and QRS duration were univariate predictors of all-cause mortality as well as death from pump failure. Univariate predictors of sudden death were CRT, ischaemic heart disease, log-transformed NT-proBNP, log-transformed end-systolic volume index, and inter-ventricular mechanical delay. All independent predictors of the multivariable analyses are given in Table 2. In multivariable analyses, CRT and log-transformed NT-proBNP emerged as independent predictors of all-cause mortality, death from pump failure, and sudden death. Model optimism for all-cause mortality was 1.0%, for sudden death was 2.4%, and for HF death was 2.2%.
|
The application of multiple imputation techniques to assess the potential influence of missing data on the statistical models performed led to no material differences in models for sudden death and death from HF. The analysis of all-cause mortality was also substantially unchanged; however, although the prognostic effects of mitral regurgitation index remained statistically significant, the strength of the association was reduced. In addition, NYHA class was included in the final model with a modest effect, which was just statistically significant.
Influence of cardiac resynchronization therapy on the prognostic performance of N-terminal pro-brain natriuretic peptide
Simple multivariable models tested the effects of CRT treatment, log NT-proBNP and an interaction term (log NT-proBNP*CRT treatment) on outcomes including all-cause mortality, sudden death, and death from pump failure. The addition of this interaction term did not improve the fit of the models: NT-proBNP retained its prognostic power and the interaction term was not significant. These results indicate that a higher baseline plasma concentration of NT-proBNP predicts a higher risk of all-cause mortality, sudden death, and death from pump failure, irrespective of treatment group. We stratified patients according to the median NT-proBNP and to CRT treatment-allocation, to illustrate the effects of CRT on two patient groups with different stages of disease. Patient characteristics of these four subgroups are given in Table 1. In patients with <median NT-proBNP, there were 48 deaths (25%/annual mortality rate 7.4%) in 189 patients assigned to the control group and 22 deaths (12%/3.4%) in 177 patients assigned to CRT. In patients with median NT-proBNP, there were 93 deaths (51%/17.7%) in 181 patients assigned to the control group and 65 deaths (35/14.1%) in 185 patients assigned to CRT. Patients with below median NT-proBNP had similar absolute (13% vs. 16%) and somewhat greater relative reduction (52% vs. 31%) in mortality compared with those with levels above the median. The Kaplan–Meier estimates of the time to all-cause mortality in these four subgroups are given in Figure 1. Data on the mode of death according to subgroups are presented in Table 3 and Figure 2.
|
|
|
Time-dependent covariate models
In a time-dependent covariate model, log-transformed NT-proBNP, updated from baseline to 3 months values for 669 subjects still at risk of an event after this time, was the strongest independent predictor of all-cause mortality (P < 0.0001). Inter-ventricular mechanical delay (P = 0.0024), log-transformed mitral regurgitation area (P = 0.0034), CRT (P = 0.0040), and presence of ischaemic heart disease (P = 0.0118) provided additional prognostic information. Note that patients with greater inter-ventricular mechanical delay had a better outcome.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
This analysis of the CARE-HF study demonstrates that NT-proBNP measured before device implantation retains its prognostic value in patients with chronic HF treated with CRT. However, measurement of NT-proBNP, 3 months after implantation, reflecting the effect of CRT9 and other treatments and changes in the underlying disease, is a stronger predictor than baseline NT-proBNP of subsequent outcome. Potentially of greatest interest is the observation that patients with less markedly elevated NT-proBNP benefited similarly in relative terms to those patients with higher levels. Within CARE-HF, it is not possible to determine a single cut-off value for BNP to select patients for CRT. Finding the appropriate spectrum of disease severity for intervention is part of the art, science, and complexity of medicine.
Cardiac resynchronization therapy in patients with different stages of disease
The CARE-HF study enrolled patients with HF in NYHA class III (94%) or IV (6%) at the start of the run-in period. However, NYHA class has a large subjective element and is not fixed. In CARE-HF, the pre-randomization assessment indicated that 22% of patients had improved with medical treatment alone and had no or only mild symptoms (NYHA class I or II).11 The severity of symptoms appeared a poor guide to therapeutic effect.12 N-terminal pro-brain natriuretic peptide is a more objective measure of cardiac dysfunction than symptoms or echocardiographic ejection fraction and a much stronger predictor of prognosis. Outcome according to symptom severity and plasma concentration of NT-proBNP should be viewed in the context of other cornerstone HF trials. The annual mortality rates of patients assigned to the control group of CARE-HF with below and above median NT-proBNP were 7.4% and 17.7%, respectively. In the MERIT-HF trial,13 which proved efficacy of the beta-blocker metoprolol-succinate in patients with mild to moderate HF [NYHA II (41%)/III (55%)], the annual mortality rate of patients assigned to beta-blockers was 7.2%—similar to that of the control patients of CARE-HF with below median NT-proBNP levels. In the RALES trial, which demonstrated a survival benefit by treatment with spironolactone compared with placebo in advanced HF patients (NYHA III/IV), the 1 year mortality of patients assigned to spironolactone was 17.5%—similar to that of the control patients of CARE-HF with above median NT-proBNP.14
Concern exists that patients with advanced disease may not benefit from CRT, as the procedure of device implantation may have an adverse effect on cardiac or renal function15 and, thereby, worsen short-term outcome. However, this analysis suggests that high-risk patients with above median NT-proBNP have a similar gain in mortality expressed in relative terms. The Comparison of Medical, Pacing, and Defibrillation Therapies in Heart Failure (COMPANION) trial evaluated the effects of CRT alone or combined with a defibrillator. Similar to the CARE-HF trial, the COMPANION trial enrolled mainly patients in NYHA class III.4 Subset analysis of this trial demonstrated an extension of the time to death or rehospitalization and a trend for improved mortality by CRT in the small proportion of patients in NYHA class IV (14%).16
The effect of CRT on survival has not been established in patients with milder stages of disease. However, in the CONTAK-CD trial, which assessed the efficacy of CRT in symptomatic HF patients with malignant ventricular tachyarrhythmias, CRT induced LV reverse remodelling but no improvement in symptoms or exercise capacity in a subgroup of 263 patients in NYHA class I and II.3 Similarly, in the MIRACLE-ICD II study, which evaluated the effects of CRT on disease progression in NYHA II patients with an indication for an ICD, CRT did not alter exercise capacity but improved cardiac structure and function.1 The REsynchronization reVErses Remodeling in Systolic left vEntricular dysfunction (REVERSE) study assessed the effects of CRT on HF progression in NYHA I and II patients with previous HF symptoms: thereby, its long-term findings demonstrated a progressive reverse remodelling, which in turn reduced the risk for HF hospitalization (Linde C. ESC Congress 2008; Session number 4484). Analysis of the CARE-HF trial shows that patients who had improved to NYHA class I/II prior randomization had lower baseline NT-proBNP.11 Accordingly, it is likely that NT-proBNP can be used as an objective marker indicating a greater probability to be in NYHA class I or II, less severe cardiac dysfunction and better prognosis. The observation that patients with values of NT-proBNP below median gain similar benefits from CRT as those with higher levels is consistent with a previous report from CARE-HF indicating that patients with milder symptoms obtain similar prognostic benefit to those with more severe symptoms.11
The absolute and relative risk of sudden death in relation to the severity of the disease
Considerable confusion exists about proportional and absolute risk of sudden death in patients with HF. It is widely recognized that sudden death is the commonest mode of death in patients with mild symptoms. Less well appreciated is that the rate of sudden death is higher in patients with more severe symptoms of HF, even though the proportion of patients who die suddenly is lower.13 In other words, the rate of death due to both, pump failure and arrhythmias, increase with worsening HF, but the rate of pump failure death increases faster, thus reducing the proportion of patients who die suddenly.
Defibrillators can prevent sudden death in patients with HF, but may not improve overall survival when pump failure death is likely.17 This hypothesis is supported by an analysis of the Sudden Cardiac Death in Heart Failure trial (SCD-HeFT) revealing a survival benefit from a defibrillator in NYHA class II patients but not in NYHA class III patients.18 However, our data demonstrate that CRT reduces the risk of dying from pump failure in high (as well as low) risk patients, and that a substantial proportion of high-risk patients die suddenly. Accordingly, for many patients who are not at high risk of dying from non-cardiac disease, CRT-D(efibrillator) may be preferred to CRT alone even if symptoms and LV dysfunction are severe. Nevertheless, another approach could be to leave the D out, wait for a positive response to occur, and then discuss with the patient to potentially upgrade to a CRT-D device. Some patients might prefer to die suddenly than slowly through a protected pump failure death.
N-terminal pro-brain natriuretic peptide predicts outcome irrespective of cardiac resynchronization therapy
Our data indicate that NT-proBNP predicted all-cause mortality, sudden death, and death from pump failure, whether they were assigned to CRT or the control group. Similar results have been demonstrated previously with neurohormonal antagonists that are known to improve survival in patients with HF.8,19 As mortality rates remain substantially higher in patients with high baseline NT-proBNP levels even if they receive CRT, patients with high NT-proBNP levels should be monitored more carefully to ensure optimal device function and pharmacological treatment.
N-terminal pro-brain natriuretic peptide at 3 months
Few data exist about the prognostic significance of changes in NT-proBNP. An analysis of the Val-HeFT study demonstrated that changes in BNP at 4 months of follow-up were associated with corresponding changes in morbidity and mortality.19 Maeda et al.20 evaluated the prognostic value of BNP before and 3 months after optimization of medical therapy in an observational trial of 102 patients and reported that follow-up values predicted mortality better than baseline values. Our data are entirely consistent with these findings and support a role for monitoring patients with HF using NT-proBNP to evaluate the likely outcome after treatment.
Limitations
Our analyses used all-cause mortality and different modes of death as endpoints, although there are many other possible endpoints, which are positively influenced by CRT. Although the level of missing data was modest for most variables included in the statistical modelling, there was substantial missing data for the mitral regurgitation index. However, multiple imputation techniques that enable the inclusion of data from all subjects in the statistical modelling were unchanged for sudden death and death from HF, and consistent with the non-imputed model for the outcome of all-cause mortality.
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
N-terminal pro-brain natriuretic peptide is a powerful marker of prognosis in patients who are being considered for CRT. However, within the general range of values observed in CARE-HF, NT-proBNP is not a useful method to select patients because the relative benefit of CRT appears similar in patients regardless of their level of NT-proBNP. Our data argue strongly that the survival benefit of CRT applies similarly to patients with different stages of disease. Measurement of NT-proBNP, 3 months after implantation, was a better predictor of prognosis than baseline values, suggesting a role for NT-proBNP in monitoring the effect of CRT. The effects of CRT on survival in patients with asymptomatic LV dysfunction or mild HF remain to be determined and are being addressed in the Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy (MADIT-CRT trial).
|
Funding |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
The Medtronic Corporation.
Conflicts of interest: J.G.F.C., L.T., N.F., F.F., and R.P. reported having received research grants and speakers' bureau/honoraria for Medtronic. J.G.F.C., L.T., and N.F reported having been involved as consultants/advisory boards for Medtronic. The other authors reported no conflicts.
|
References |
|---|
|
TopAbstractIntroductionMethodsStatistical analysisResultsDiscussionConclusionFundingReferences |
|---|
- Abraham WT, Young JB, Leon AR, Adler S, Bank AJ, Hall SA, Lieberman R, Liem LB, O'Connell JB, Schroeder JS, Wheelan KR. Effects of cardiac resynchronization on disease progression in patients with LV dysfunction, an indication for an implantable cardioverter-defibrillator, and mildly symptomatic heart failure. Circulation (2004) 110:2864–2868.
[Abstract/Free Full Text] - Cleland JGF, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappenberger L, Tavazzi L. The effect of cardiac resynchronization on morbidity and mortality in heart failure. N Engl J Med (2005) 352:1539–1549.
[Abstract/Free Full Text] - Higgins S, Hummel J, Niazi I, Giudici MC, Worley SJ, Saxon LA, Boehmer JP, Higginbotham MB, De Marco T, Foster E, Yong PG. Cardiac resynchronization therapy for the treatment of heart failure in patients with inter-ventricular conduction delay and malignant ventricular tachyarrhythmias. J Am Coll Cardiol (2003) 42:1454–1459.
[Abstract/Free Full Text] - Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De Marco T, Carson P, DiCarlo L, DeMets D, White BG, DeVries DW, Feldman AM. Cardiac resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N Engl J Med (2004) 350:2140–2150.
[Abstract/Free Full Text] - Cleland JGF, Daubert JC, Erdmann E, Freemantle N, Gras D, Kappenberger L, Tavazzi L. Longer-term effects of cardiac resynchronization therapy on mortality in heart failure. Eur Heart J (2006) 27:1928–1932.
[Abstract/Free Full Text] - Hall C. NT-ProBNP: the mechanism behind the marker. J Card Fail (2005) 11:S81–S83.[CrossRef][Web of Science][Medline]
- Doust JA, Pietrzak E, Dobson A, Glasziou P. How well does B-type natriuretic peptide predict death and cardiac events in patients with heart failure: systematic review. Br Med J (2005) 330:625.
[Abstract/Free Full Text] - Hartmann F, Packer M, Coats AJ, Fowler MB, Krum H, Mohacsi P, Rouleau JL, Tendera M, Castaigne A, Anker SD, Amann-Zalan I, Hoersch S, Katus HA. Prognostic impact of plasma N-terminal pro-brain natriuretic peptide in severe chronic congestive heart failure. Circulation (2004) 110:1780–1786.
[Abstract/Free Full Text] - Fruhwald FM, Fahrleitner-Pammer A, Berger R, Leyva F, Freemantle N, Erdmann E, Gras D, Kappenberger L, Tavazzi L, Daubert JC, Cleland JGF. Early and sustained effects of cardiac resynchronization therapy on N-terminal pro-B-type natriuretic peptide in patients with moderate to severe heart failure and cardiac dyssynchrony. Eur Heart J (2007) 28:1592–1597.
[Abstract/Free Full Text] - Harrell FE Jr, Lee KL, Mark DB. Tutorial in biostatistics multivariable prognostic models. Stat Med (1996) 15:361–387.[CrossRef][Web of Science][Medline]
- Cleland JFG, Freemantle N, Daubert JC, Toff WD, Leisch F, Tavazzi L. Long-term effect of cardiac resynchronisation in patients reporting mild symptoms of heart failure. Heart (2008) 94:278–283.
[Abstract/Free Full Text] - Ingle L, Rigby AS, Carroll S, Butterly R, King RF, Cooke CB, Cleland JFG, Clark AL. Prognostic value of the 6 min walk test and self-preceived symptom severity in older patients with chonic heart failure. Eur Heart J (2007) 28:560–568.
[Abstract/Free Full Text] - The MERIT-HF Study Group. Effect of metoprolol CR/XL in chronic heart failure. Lancet (1999) 357:2001–2007.
- Pitt B, Zannad F, Remme WJ, Cody R, Castaigne A, Perez A, Palensky J, Wittes J. The effect of spironolactone on morbidity and mortality in patients with severe heart failure. N Engl J Med (1999) 341:709–717.
[Abstract/Free Full Text] - Cowburn PJ, Patel H, Pipes RR, Parker JD. Contrast nephropathy post cardiac resynchronization therapy. Eur J Heart Fail (2005) 7:899–903.
[Abstract/Free Full Text] - Lindenfeld J, Feldman AM, Saxon L, Boehmer J, Carson P, Ghali JK, Anand I, Singh S, Steinberg JS, Jaski B, DeMarco T, Mann D, Yong P, Galle E, Ecklund F, Bristow M. Effects of cardiac resynchronization therapy with or without a defibrillator on survival and hospitalizations in patients with New York Heart Association Class IV Heart Failure. Circulation (2007) 115:204–212.
[Abstract/Free Full Text] - Stevenson LW, Desai AS. Selecting patients for discussion of the ICD as primary prevention for sudden death in heart failure. J Card Fail (2006) 12:407–412.[CrossRef][Web of Science][Medline]
- Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R, Domanski M, Troutman C, Anderson J, Johnson G, McNulty SE, Clapp-Channing N, Davidson-Ray LD, Fraulo ES, Fishbein DP, Luceri RM, Ip JH. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N Engl J Med (2005) 352:225–237.
[Abstract/Free Full Text] - Anand IS, Fisher LD, Chiang JT, Latini R, Masson S, Maggioni AP, Glazer RD, Tognoni G, Cohn JN. Changes in brain natriuretic peptide and norepinephrine over time and mortality and morbidity in the Valsartan Heart Failure Trial (Val-HeFT). Circulation (2003) 107:1278–1283.
[Abstract/Free Full Text] - Maeda K, Tsutamoto T, Wada A, Mabuchi N, Hayashi M, Tsutsui T, Ohnishi M, Sawaki M, Fujii M, Matsumoto T, Kinoshita M. High levels of plasma brain natriuretic peptide and interleukin-6 after optimized treatment for heart failure are independent risk factors for morbidity and mortality in patients with congestive heart failure. J Am Coll Cardiol (2000) 36:1587–1593.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||