Author + information
- Received March 18, 2016
- Revision received May 2, 2016
- Accepted May 3, 2016
- Published online September 1, 2016.
- Susan Stienen, MDa,
- Khibar Salah, MDa,
- Luc W. Eurlings, MDb,
- Paulo Bettencourt, MD, PhDc,
- Joana M. Pimenta, MD, PhDc,
- Marco Metra, MD, PhDd,
- Antoni Bayes-Genis, MD, PhDe,
- Valerio Verdiani, MD, PhDf,
- Luca Bettari, MDg,
- Valentina Lazzarini, MDh,
- Jan P. Tijssen, PhDa,
- Yigal M. Pinto, MD, PhDa and
- Wouter E. Kok, MD, PhDa,∗ ()
- aDepartment of Cardiology, Academic Medical Center, Amsterdam, the Netherlands
- bDepartment of Cardiology, VieCuri Medical Center, Venlo, the Netherlands
- cDepartment of Internal Medicine, Hospital S. João, University of Porto Medical School, Porto, Portugal
- dDepartment of Medical and Surgical Specialties, Radiological Sciences and Public Health, University of Brescia, Brescia, Italy
- eDepartment of Cardiology, Hospital Universitari Germans Trias i Pujol, Barcelona, Spain
- fDepartment of Internal Medicine and Emergency, Careggi University Hospital, Florence, Italy
- gDepartment of Cardiology, Azienda Istituti Ospitalieri di Cremona, Cremona, Italy
- hDepartment of Cardiology, Ospedale San Pellegrino, Castiglione delle Stiviere, MN, Italy
- ↵∗Reprint requests and correspondence
: Dr. Wouter E. Kok, Department of Cardiology, Academic Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands.
Objectives The aim of this study was to analyze the prognostic value and attainability of N-terminal pro-brain natriuretic peptide (NT-proBNP) levels in young and elderly acute decompensated heart failure (ADHF) patients.
Background Less-effective NT-proBNP-guided therapy in chronic heart failure (HF) has been reported in elderly patients. Whether this can be attributed to differences in prognostic value of NT-proBNP or to differences in attaining a prognostic value is unclear. The authors studied this question in ADHF patients.
Methods Our study population comprised 7 ADHF cohorts. We defined absolute (<1,500 ng/l, <3,000 ng/l, <5,000 ng/l, and <15,000 ng/l) and relative NT-proBNP discharge cut-off levels (>30%, >50%, and >70%). Six-month all-cause mortality after discharge was studied for each level in Cox regression analyses, and compared between elderly (age >75 years) and young patients (age ≤75 years). Thereafter, we compared percentages of elderly and young patients attaining NT-proBNP levels (= attainability).
Results A total of 1,235 patients (59% male, 45% >75 years of age) was studied. Admission levels of NT-proBNP were significantly higher in elderly versus younger patients. The prognostic value of absolute and relative NT-proBNP levels was similar in elderly and young patients. Attainability was significantly lower in elderly patients for all absolute levels and a >50% relative reduction, but not for >30% and >70%. For absolute levels, attainability differences between age groups were decreased to a large extent after correction for admission NT-proBNP and anemia at discharge. For relative levels, attainability differences disappeared after correction for HF etiology and anemia at discharge.
Conclusions In young and elderly ADHF patients, it is not the prognostic value of absolute and relative NT-proBNP levels that is different, but the attainability of these levels that is lower in the elderly. This can largely be attributed to factors other than age.
After hospital admissions for acute decompensated heart failure (ADHF) there are high percentages of mortality and readmissions (1). Both a lack in reduction in N-terminal pro-brain natriuretic peptide (NT-proBNP) levels during hospitalization and higher absolute discharge NT-proBNP levels predict a worse prognosis after discharge for ADHF (2,3). A role for NT-proBNP-guided ADHF treatment is therefore suggested (4).
Pivotal studies investigating the effect of NT-proBNP-guided treatment in chronic heart failure (HF) patients stratified by age raised the suggestion that guided therapy is less effective in elderly (>75 years of age) than in younger patients (≤75 years of age) (5,6). In contrast, another study observed no interaction between age and treatment allocation for outcome (7). Nonetheless, 2 meta-analyses demonstrated that the beneficial effect of NT-proBNP guidance on reduction of all-cause mortality was only seen in patients ≤75 years of age (8,9). Subsequently, the use of age-adjusted natriuretic peptide targets in future HF trials has been proposed (10). At present, it is not clear whether this lesser effectivity can be attributed to age itself, with different comorbidities (11,12), to a difference in prognostic value of an NT-proBNP target (13), or to a difference in attaining a prognostic value (14).
In ADHF patients, it is currently not known whether NT-proBNP−guided treatment is effective, and whether differences in effect of guiding should be expected for young and elderly patients. In ADHF patients, age >75 years predicts short-term prognosis after discharge (3,15), and—although there is a clear association between age and NT-proBNP suggestive of an interaction (16)—any interaction between age and NT-proBNP levels has not been reported thus far.
To define possible age-dependent differences in NT-proBNP−related outcomes in an ADHF population, we addressed the following questions. First, does attaining a certain NT-proBNP level or percentage NT-proBNP reduction at discharge provide the same prognostic information for young and elderly ADHF patients? Second, does attainability of discharge NT-proBNP cut-off levels differ between young and elderly patients, even without the effect of guiding therapy?
We studied these questions in 2 predefined age groups (≤75 years and >75 years) in a large database of 7 prospective European registries of ADHF patients.
The study cohort was assembled from individual patient data from 7 ADHF cohorts (N = 1,301). Details of the source gathering and development of the final database have been reported previously (3). For the present study, unique patients were included with NT-proBNP measurements available at admission and discharge (N = 1,235). Data on all-cause mortality and cardiovascular readmissions within 6 months after discharge was present for all patients. All studies were approved by the ethical commission in their respective centers. Demographic characteristics are presented as frequencies and percentages for categorical data, and compared using Fisher exact test. Normally distributed, continuous variables are reported as mean ± SD, and were compared using the Student t test. Not normally distributed continuous variables, expressed as median with interquartile ratio (IQR), were compared using the Mann-Whitney U test.
All patients were categorized as being discharged with absolute discharge NT-proBNP levels of <1,500 ng/l, <3,000 ng/l, <5,000 ng/l, and <15,000 ng/l, and with relative NT-proBNP reductions of >30%, >50%, and >70% from admission to discharge (17).
The endpoint was all-cause mortality within 6 months after discharge. The distribution of the endpoint among the above-mentioned categories of absolute discharge and relative reduction levels was studied for young (age ≤75 years) and elderly (age >75 years) patients using Fisher exact test. A cut-off level of 75 years for age was used in this study as this was the cut-off in previous studies indicating a difference in effect of NT-proBNP−guided therapy in chronic HF patients (5,6,8,9), as well as having significance to prognosis in studies on ADHF-related outcome (3,15).
To compare the risk of 6-month all-cause mortality in young and elderly ADHF patients associated with several NT-proBNP cut-off levels, we calculated adjusted hazard ratios (HRs), and 95% confidence intervals using Cox proportional hazards regression models for each absolute and relative NT-proBNP level. Our aim was to construct models to study the potential effect of guidance on prognosis using a single NT-proBNP cut-off level, and not to determine independent predictors for prognosis. Therefore, only 1 NT-proBNP level was added per model. This in contrast to the European collaboration on acute decompensated heart failure (ELAN-HF) score for risk stratification of ADHF patients in which both a >30% NT-proBNP reduction and absolute discharge NT-proBNP categories were incorporated (3). In all multivariate models, adjustments were made for several variables. First, predictors of 6-month all-cause mortality were selected, based on previous research: peripheral edema, systolic blood pressure (SBP) ≤115 mm Hg, and serum sodium <135 mmol/l at admission, and serum urea ≥15 mmol/l, and New York Heart Association functional class (NYHA) III/IV at discharge (3). Second, although not a significant prognostic variable in ADHF studies with available discharge NT-proBNP (2,3), we added admission NT-proBNP level as a variable to our models because we demonstrated in our previous study that admission NT-proBNP levels influence the proportion of mortality that, in theory, could be prevented by attaining an NT-proBNP target (17).
Interaction terms of age >75 years and each of the absolute and relative discharge NT-proBNP levels on 6-month all-cause mortality were studied. Multivariate regression models were constructed incorporating all aforementioned variables, age >75 years, and an interaction term for age >75·NT-proBNP cut-off level (dichotomous: discharge NT-proBNP level attained/not attained). Thereafter, a secondary analysis was performed by adding an interaction term for age >75·admission NT-proBNP level (dichotomous: NT-proBNP level below/above median (6,446 ng/l)). To check for the proportional hazards assumption, we used the log-minus-log function for all regression models. No deviations from parallelism were observed.
In the second part of the analysis, we compared the percentage of patients attaining discharge NT-proBNP cut-off levels (= attainability) between young and elderly patients, using the Fisher exact test. To further investigate the attainability differences that we observed between age groups, we compared the prevalence of clinical variables that are known to predict not attaining the lowest absolute (<1,500 ng/l) and the greatest relative (>70%) discharge NT-proBNP category between age groups (17). Subsequently, attainability rates for absolute and relative discharge NT-proBNP levels were compared between age groups after correction for those clinical variables with significantly different prevalence (and hence potential reasons for differences in attainability). For the purpose of this study, 1 of these predictors—discharge hemoglobin (Hb) level—was studied as categorical (anemia yes/no) instead of as continuous variable. Anemia was defined as Hb <129 g/l (8 mmol/l) in men, and Hb <121 g/l (7.5 mmol/l) in women (18).
A total of 1,235 patients were included in the study, of whom 558 (45%) patients were age >75 years. Baseline characteristics of young and elderly ADHF patients are listed in Table 1. Elderly patients more often were women, and more often had a nonischemic HF etiology, atrial fibrillation at admission, a preserved left ventricular ejection fraction (LVEF), and higher admission SBP. In addition, elderly patients had higher admission and discharge NT-proBNP levels, and a lower admission estimated glomerular filtration rate (eGFR) when compared to younger patients. At discharge, elderly patients received angiotensin-converting enzyme inhibitor (ACEI) and angiotensin II receptor blockers (ARBs) less often than their younger counterparts (respectively, 87% vs. 91%, p = 0.03). In selected patients with LVEF ≤45%, there were no significant differences in prescription rates of ACEI and ARBs at discharge between elderly and young (88% vs. 91%, respectively, p = 0.20). Furthermore, the prescription rates of beta blockers and aldosterone antagonists at discharge for patients with LVEF ≤45% were not significantly different between young and elderly (66% vs. 60%, p = 0.10, and 56% vs. 61%, respectively, p = 0.40). There was a trend towards a smaller median percentage NT-proBNP reduction during admission in elderly versus young patients (respectively, 40% (IQR 14% to 66%) versus 47% (IQR 15% to 69%), p = 0.09).
Association between NT-proBNP cut-off levels at discharge and outcome in young and elderly ADHF patients
The 6-month all-cause mortality rate was significantly higher in elderly compared to younger patients (18% vs. 12%, respectively, log rank test, p = 0.004). Figure 1 depicts mortality rates for young and elderly patients within categories of absolute and relative discharge NT-proBNP levels. In both age groups, mortality rates decreased when patients were discharged with lower absolute levels (p < 0.001) and greater percentage reductions (p < 0.001). Mortality rates per category of absolute and relative NT-proBNP levels were slightly higher in elderly compared to younger patients, although these differences (except for a >70% NT-proBNP reduction) did not reach significance.
In young and elderly patients, adjusted HRs for mortality for each absolute and relative NT-proBNP cut-off level are depicted in Figure 2. There was no significant interaction between age (≤75 years of age vs. >75 years of age) and both absolute (Figure 2A) and relative NT-proBNP levels (Figure 2B). There was also no interaction between age (≤75 years of age vs. >75 years of age) and admission NT-proBNP (below vs. above median) for mortality (data not shown).
Attainability of NT-proBNP cut-off levels in young versus elderly patients
In both young and elderly patients, attainability of NT-proBNP cut-off levels decreased with lower absolute discharge NT-proBNP levels or greater percentages NT-proBNP reduction (Figure 3). For each absolute NT-proBNP level, attainability was significantly lower for elderly compared to younger patients (21% vs. 33%, p < 0.001 for <1,500 ng/l; 40% vs. 54%, p < 0.001 for <3,000 ng/l; 57% vs. 70%, p < 0.001 for <5,000 ng/l; 88% vs. 92%, p = 0.01 for <15, 000 ng/l) (Figure 3A). For relative NT-proBNP levels (Figure 3B), attainability was significantly lower for elderly compared to younger patients for >50% (39% vs. 47% respectively, p = 0.01). No significant differences in attainability between elderly and younger patients were observed for >30% and >70% (61% vs. 63%, p = 0.50, and 21% vs. 24%, respectively, p = 0.30).
Distribution of variables influencing attainability of absolute and relative NT-proBNP levels is compared between age groups in Table 2. The strongest known predictor for not attaining absolute NT-proBNP targets in ADHF patients — admission NT-proBNP within the highest tertile — was significantly more frequently observed in elderly compared to younger patients. In addition, another predictor for not attaining absolute NT-proBNP levels, discharge Hb level, was significantly lower in elderly compared to younger patients. Figure 4 demonstrates that differences in attainability of absolute NT-proBNP levels between age groups decreased to a large extent after correction for admission NT-proBNP (tertiles) and anemia at discharge. The contribution of anemia at discharge to attainability of absolute NT-proBNP levels was smaller than that of admission NT-proBNP level. For example, attainability rates for <5,000 ng/l of patients within the third tertile of admission NT-proBNP levels were higher in nonanemic versus anemic patients (38% vs. 21%, respectively, p = 0.006). However, a larger difference in attainability for <5,000 ng/l was observed between patients presenting within the first and third tertile of admission NT-proBNP levels (95% vs. 28%, p < 0.001).
For relative NT-proBNP levels, the clinical variables predictive of nonattainability that were found to be more frequently present in elderly patients were nonischemic HF and lower discharge Hb level. The difference in attainability of >50% NT-proBNP reduction between age groups disappeared after correction for HF etiology and anemia at discharge (Figure 5). The contribution of etiology of HF to attainability of relative discharge NT-proBNP levels was smaller than that of anemia at discharge. For example, attainability rates for >30% NT-proBNP reduction were lower in patients with nonischemic versus ischemic HF (47% vs. 53%, respectively, p = 0.001). A larger difference was however observed in the percentage of anemic versus not anemic patients attaining a >30% NT-proBNP reduction (56% vs. 70%, p < 0.001). The strongest predictor for not attaining relative NT-proBNP cut-off levels, admission NT-proBNP levels within the first tertile, was more often present in younger than in elderly patients, and could therefore not have contributed to the lower attainability rates in elderly patients. The influence of admission NT-proBNP on attainability rates in the entire cohort is reflected by the finding of significantly less patients within the first tertile of admission NT-proBNP levels that attained a >30% NT-proBNP reduction compared to patients within the third tertile (56% vs. 66%, respectively, p = 0.002).
In this study, we evaluated the prognostic value and attainability of several absolute and relative discharge NT-proBNP cut-off levels in young (≤75 years of age) and elderly (>75 years of age) ADHF patients to assess the need for age-adjusted approaches for NT-proBNP−guided treatment. First, we demonstrated that the prognostic value of a single absolute or relative reduction NT-proBNP cut-off level on 6-month all-cause mortality is not significantly different between young and elderly patients. Second, for all relevant absolute NT-proBNP levels, attainability was significantly lower in elderly patients compared to younger patients. For relative NT-proBNP levels this difference in attainability was of lesser significance, and only significant for >50% reduction, which was less often attained in elderly patients. These differences in attainability can largely be attributed to factors other than age.
NT-proBNP as independent predictor of mortality irrespective of age
Similar to previous studies (2,3,17), we found that mortality after admission for ADHF is highly related to discharge NT-proBNP levels. Our study shows that the concept of “the lower NT-proBNP, the better the prognosis” applies to both young and elderly ADHF patients.
This is the first study to demonstrate that the prognostic value of attaining a single absolute or relative NT-proBNP cut-off level in an ADHF population does not depend on age >75 years. For either absolute or relative discharge NT-proBNP levels, HRs overlapped between young and elderly patients, and no interaction with age >75 years was seen. Our results are in line with current studies showing that HF therapies should have no differential age-dependent effect on outcome, at least not for patients with reduced ejection fraction (19).
Attainability of absolute NT-proBNP discharge targets
Knowledge of potential obstacles for attaining a prognostically favorable NT-proBNP discharge level is relevant for future guiding studies, for both young and elderly patients. We demonstrated that the attainability of favorable absolute NT-proBNP levels was significantly lower in elderly compared to younger patients. However, we demonstrated that higher age may not be the best or only explanation for this difference in attainability.
Our finding of significantly higher admission NT-proBNP levels and lower discharge Hb levels in elderly patients led us to hypothesize that differences in attainability between age groups for absolute NT-proBNP levels are due to differences in these variables, and not age. For this purpose, patients were categorized within tertiles of admission NT-proBNP levels and the presence of anemia at discharge. Subsequently, we observed that differences in attainability rates between age groups almost (although not completely) disappeared. Admission NT-proBNP level was a stronger contributor to differences in attainability of absolute NT-proBNP levels than anemia at discharge. Other variables that are known to influence attainability were not different between age groups (admission urea ≥15 mmol/l) or were present in a manner that attainability should have been increased instead of decreased in elderly patients given the higher frequency of elderly patients with SBP ≥115 mm Hg and higher LVEF. A significantly worse renal function at admission and somewhat lower prescription rates of ACEI or ARBs at discharge were more frequently present in elderly compared to younger patients in our study. However, in our previous analysis we could not demonstrate that these variables independently predict attainability of absolute or relative NT-proBNP targets (17).
Attainability of relative NT-proBNP discharge targets
Given the finding of a lower attainability of absolute NT-proBNP levels in the elderly compared to younger patients, physicians might also expect that attainability of relative NT-proBNP levels is also lower in these patients. To our surprise, attainability differences between young and elderly patients of relative NT-proBNP levels were of lesser significance than those of absolute levels, and only significantly lower in elderly for an >50% NT-proBNP reduction. The following explanation may be given: because elderly patients more often had admission NT-proBNP levels within the highest tertile, a better attainability of relative NT-proBNP levels in elderly compared to younger patients would have been expected (17). Therefore, we argue that attainability of relative NT-proBNP levels is lower (than expected) in elderly compared to younger patients and was influenced by other factors than admission NT-proBNP. Other predictors for not attaining relative NT-proBNP levels in elderly as well as young patients are nonischemic HF etiology, and lower discharge Hb level (17), which more frequently occurred in our elderly population. We found that, after correction for HF etiology and anemia at discharge, attainability differences that were present between young and elderly patients disappeared. The strongest contributor to attainability problems for relative NT-proBNP levels in elderly patients was the presence of anemia at discharge. In another study, lower admission and discharge Hb level was found to be prognostically important, and it was suggested that anemic patients were less likely to receive HF medication at discharge (20). However, we did not observe a difference in prescription rates of HF medications at discharge between anemic versus nonanemic patients with LVEF ≤45% in our cohort (data not shown).
The prevalence of peripheral edema at admission was similar between age groups, and therefore could not explain differences in attainability. In addition, eGFR and SBP <115 mm Hg do not predict attainability of relative NT-proBNP cut-off levels (17), and their uneven distribution among young and elderly patients cannot be used to explain the difference in attainability. However, we still may not fully understand all obstacles for guiding HF patients towards favorable relative NT-proBNP levels. Ribeira et al. (21), for example, studied predictors for a ≤30% BNP reduction during ADHF admission and found that only cholesterol level was an independent predictor, whereas admission BNP and age were not (21).
Our results indicate that the effect of NT-proBNP−guided treatment on prognosis should be similar between young and elderly ADHF patients, provided that there is an even distribution in percentage of patients attaining an NT-proBNP target. However, the distribution in percentage of patients attaining NT-proBNP targets in our cohort was different between young and elderly patients: elderly less often attained favorable absolute discharge NT-proBNP levels and a relative NT-proBNP reduction of >50% at discharge compared to younger patients. The main explanations for not attaining absolute and relative discharge NT-proBNP levels in the elderly were a higher admission NT-proBNP level and the presence of anemia at discharge, respectively. The clinical significance of improving attainability by guiding HF treatment in the young and elderly is illustrated by the finding that only ∼20% to 30% of young and elderly patients attained the lowest absolute and relative NT-proBNP levels.
But how should we guide ADHF therapy? It seems mandatory to consider admission NT-proBNP levels but also other variables that could potentially influence attainability of NT-proBNP targets in the population at stake before choosing or accepting a NT-proBNP level to guide treatment. Attainability of NT-proBNP targets should therefore be the focus in future investigations in all HF patients, irrespective of age.
Although our study population consists of acute and not chronic HF patients, our results suggest that the less beneficial effect of NT-proBNP−guided treatment in elderly compared to younger HF patients observed in previous chronic HF studies (5,6,8,9) may also have been caused by differences in attainability rather than by differences in prognostic value of NT-proBNP. Suggestive of difficulties in attainability of NT-proBNP targets in elderly HF patients is the observation by Troughton et al. (9) of a significantly greater fall in NT-proBNP levels in chronic HF patients aged <75 years versus ≥75 years, whether guided (43% and 26%, respectively) or unguided (41% and 20%, respectively). Instead of targeting different NT-proBNP levels based on age group, attainability of NT-proBNP targets should also be investigated in chronic HF patients.
There are some limitations to this study that should be acknowledged. First, heterogeneity in the 7 ADHF cohorts should be considered, such as percentages of patients age >75 years, and median admission and discharge NT-proBNP levels (3). However, this heterogeneity also provides us with a wide range of NT-proBNP levels, differences in patient characteristics and therapeutic approaches, reflective of patients seen in daily clinical practice. Second, although prescription of HF medication at admission and discharge was reported, detailed information on dosages of HF medication was not available. Prescription of singular HF medication at admission and discharge did not independently predict attainability of NT-proBNP targets (17), but we find the possibility likely that the dosage of HF medication is an important determinant of attainability of NT-proBNP targets.
We are well aware of the assumption we make that the effect of lowering NT-proBNP levels on all-cause mortality is primarily due to a reduction in cardiovascular (CV) mortality. Although no formal comparisons in causes of death between young and elderly ADHF patients were made thus far, it is clear that CV mortality explains 60% to 70% of mortality after ADHF (22). Considering the findings in our present study of significantly higher admission and discharge NT-proBNP levels in elderly versus younger patients and a similar percentage NT-proBNP reduction in both age groups, it may be hypothesized that CV mortality rates are higher in elderly than in younger patients. Evidently, more research investigating modes of death in young and elderly patients is warranted in order not to overestimate any effect of lowering NT-proBNP.
In this context, no studies have yet been performed investigating the effect of NT-proBNP−guided treatment in ADHF, and patients in our study originated from prospective ADHF studies in which treatment was not guided by NT-proBNP. Both the prognostic value of attaining an NT-proBNP targets and the attainability in young versus elderly ADHF patients in a setting where therapy is guided by a predefined NT-proBNP target warrant further study.
In ADHF patients, the prognostic value of attaining NT-proBNP levels (both absolute and relative discharge levels) is similar between young and elderly patients. However, the attainability of these NT-proBNP levels is lower in the elderly, which can largely be attributed to factors other than age. Attainability of NT-proBNP targets should be the focus of further investigation in young and elderly patients alike.
COMPETENCY IN MEDICAL KNOWLEDGE: In ADHF patients, it is currently not known whether differences in prognostic effect of symptom- or biomarker-guided treatment should be expected for young and elderly patients. In chronic HF patients, a less beneficial effect of natriuretic peptide−guided treatment has been suggested in patients age >75 years compared to ≤75 years. This study examines possible differences in the prognostic value and attainability of several potential absolute and relative discharge NT-proBNP targets between young and elderly ADHF patients. Our results suggest that the prognostic value of attaining an NT-proBNP level is similar between young and elderly patients. The attainability of these NT-proBNP levels is lower in the elderly, which can largely be attributed to factors other than age. The focus of research should lie on attainability of treatment targets, in young and elderly patients alike.
TRANSLATIONAL OUTLOOK: Our study highlights the importance to investigate determinants and measures to improve attainability of predefined NT-proBNP targets when guiding treatment in all ADHF patients, irrespective of age.
Dr. Pinto has received consultancy fees from Roche Diagnostics. Dr. Pimenta has received payment for lectures from Bayer and Tecnimede. Dr. Metra has received consulting honoraria from Amgen, Bayer, Novartis, and Servier. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme inhibitor
- angiotensin II receptor blocker
- acute decompensated heart failure
- estimated glomerular filtration rate
- heart failure
- hazard ratio
- interquartile range
- left ventricular ejection fraction
- New York Heart Association functional class
- systolic blood pressure
- Received March 18, 2016.
- Revision received May 2, 2016.
- Accepted May 3, 2016.
- American College of Cardiology Foundation
- Bettencourt P.,
- Azevedo A.,
- Pimenta J.,
- Frioes F.,
- Ferreira S.,
- Ferreira A.
- Salah K.,
- Kok W.E.,
- Eurlings L.W.,
- et al.
- Stienen S.,
- Salah K.,
- Moons A.H.,
- et al.
- Lainchbury J.G.,
- Troughton R.W.,
- Strangman K.M.,
- et al.
- Berger R.,
- Moertl D.,
- Peter S.,
- et al.
- Troughton R.W.,
- Frampton C.M.,
- Brunner-La Rocca H.P.,
- et al.
- Richards A.M.,
- Troughton R.W.
- Troughton R.,
- Michael F.G.,
- Januzzi J.L. Jr..
- Raymond I.,
- Groenning B.A.,
- Hildebrandt P.R.,
- et al.
- Stienen S.,
- Salah K.,
- Eurlings L.W.,
- et al.
- McMurray J.J.,
- Adamopoulos S.,
- Anker S.D.,
- et al.
- Metra M.,
- Dei Cas L.,
- Massie B.M.