Author + information
- Received February 28, 2013
- Revision received May 1, 2013
- Accepted May 6, 2013
- Published online August 6, 2013.
- Jeffrey M. Testani, MD, MTR∗∗ (, )
- Meredith A. Brisco, MD, MSCE†,
- Jennifer Chen, MD‡,
- Brian D. McCauley, MPH§,
- Chirag R. Parikh, MD, PhD∗ and
- W.H. Wilson Tang, MD⋮
- ∗Department of Internal Medicine and Program of Applied Translational Research, Yale University School of Medicine, New Haven, Connecticut
- †Department of Medicine, Cardiovascular Division, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania
- ‡Department of Medicine, Duke University, Durham, North Carolina
- §Cooper Medical School at Rowan University, Camden, New Jersey
- ⋮Section of Heart Failure and Cardiac Transplantation, Cleveland Clinic, Cleveland, Ohio
- ↵∗Reprint requests and correspondence:
Dr. Jeffrey M. Testani, Yale University School of Medicine, 60 Temple Street, Suite 6C, New Haven, Connecticut 06510.
Objectives This study sought to determine if the timing of hemoconcentration influences associated survival.
Background Indicating a reduction in intravascular volume, hemoconcentration during the treatment of decompensated heart failure has been associated with reduced mortality. However, it is unclear if this survival advantage stems from the improved intravascular volume or if healthier patients are simply more responsive to diuretics. Rapid diuresis early in the hospitalization should similarly identify diuretic responsiveness, but hemoconcentration this early would not indicate euvolemia if extravascular fluid has not yet equilibrated.
Methods Consecutive admissions at a single center with a primary discharge diagnosis of heart failure were reviewed (N = 845). Hemoconcentration was defined as an increase in both hemoglobin and hematocrit levels, then further dichotomized into early or late hemoconcentration by using the midway point of the hospitalization.
Results Hemoconcentration occurred in 422 (49.9%) patients (41.5% early and 58.5% late). Patients with late versus early hemoconcentration had similar baseline characteristics, cumulative in-hospital loop diuretic administered, and worsening of renal function. However, patients with late hemoconcentration versus early hemoconcentration had higher average daily loop diuretic doses (p = 0.001), greater weight loss (p < 0.001), later transition to oral diuretics (p = 0.03), and shorter length of stay (p < 0.001). Late hemoconcentration conferred a significant survival advantage (hazard ratio: 0.74 [95% confidence interval: 0.59 to 0.93]; p = 0.009), whereas early hemoconcentration offered no significant mortality benefit (hazard ratio: 1.0 [95% confidence interval: 0.80 to 1.3]; p = 0.93) over no hemoconcentration.
Conclusions Only hemoconcentration occurring late in the hospitalization was associated with improved survival. These results provide further support for the importance of achieving sustained decongestion during treatment of decompensated heart failure.
The primary therapeutic objective in the majority of patients hospitalized with acute decompensated heart failure (ADHF) is optimization of volume status (1,2). Unfortunately, determining the optimal stopping point for decongestive therapies remains a major challenge (1). Notably, the increase in concentration of red blood cells and plasma proteins induced by intravascular volume contraction, known as hemoconcentration, provides a surrogate for changes in intravascular volume status during fluid removal (3–5). In the setting of ADHF treatment, hemoconcentration seems to identify patients who have undergone aggressive decongestion and has been associated with better post-discharge survival (6–8). However, hemoconcentration merely indicates that the plasma refill rate has been exceeded (i.e., fluid has been removed from the intravascular space faster than it could be replaced by extravascular fluid). As a result, hemoconcentration in isolation does not inform total body volume status. Importantly, previous studies of hemoconcentration focused on hemoconcentration at the time of discharge, a time when overall volume status should have been optimized. As such, the majority of patients with hemoconcentration in these studies had objective evidence of improvement in intravascular volume (hemoconcentration) in addition to having the treating physician deem their overall volume status sufficiently optimized to permit discharge.
Interpretation of these observations is challenging because it is unclear to what degree this survival advantage stems from a cause-and-effect association with aggressive decongestion or simply from the fact that hemoconcentration may occur more commonly in patients who are diuretic responsive and are therefore healthier. However, if a patient underwent diuresis early in an ADHF hospitalization with sufficient rapidity to exceed the plasma refill rate, intravascular volume contraction and hemoconcentration would occur. This outcome would be true even if extravascular volume overload persisted at the time of hemoconcentration. However, as diuresis was slowed (perhaps prematurely in response to the improved signs of intravascular volume status), extravascular fluid would equilibrate, leading to recurrence of intravascular volume overload and loss or reduction in the degree of hemoconcentration. Importantly, both early hemoconcentration (early HC) and late hemoconcentration (late HC) should identify treatment-responsive patients. However, the degree of volume optimization achieved over the hospitalization would likely be superior in patients with the peak degree of hemoconcentration in proximity to discharge, when extravascular volume status should be optimized to a greater degree. As a result, comparing the survival of patients with peak hemoconcentration early versus late in the hospitalization could provide some insight into the relative importance of aggressive decongestion versus treatment responsiveness in the improved survival associated with hemoconcentration. As such, the primary objective of the current study was to evaluate the association between the timing of peak hemoconcentration and subsequent survival.
Consecutive admissions from 2004 to 2009 to noninterventional cardiology and internal medicine services at the Hospital of the University of Pennsylvania with a primary discharge diagnosis of congestive heart failure were reviewed. Inclusion into this single-center study required an admission B-type natriuretic peptide (BNP) level of >100 pg/ml within 24 h of admission along with serial hematocrit (Hct) and hemoglobin (Hgb) levels. To focus primarily on the physiology and timing of decongestion, patients with a length of stay <3 days (who likely underwent limited decongestion) and patients with length of stay >14 days (who likely had either atypical degrees of congestion or nondiuresis-related problems driving the length of stay) were excluded. Patients undergoing renal replacement therapy were also excluded. In the event of multiple hospitalizations for a single patient, only the first admission meeting the aforementioned inclusion criteria was retained. Online Figure 1 provides additional details on patient selection.
Hemoconcentration was defined as an increase in both Hct and Hgb levels above admission values at any time during the hospitalization. Two positive markers of hemoconcentration were required to improve the signal-to-noise ratio and to maintain some consistency with our previous hemoconcentration definition (7). For the primary analyses, the timing of hemoconcentration was determined by averaging the percentage of the hospital stay that had elapsed at the time of the peak Hct and Hgb levels, then dichotomizing this value into early HC and late HC by using the 50% point. The relative timing of hemoconcentration was chosen to focus on a patient’s overall improvement in intravascular volume status, rather than the degree of euvolemia achieved after an arbitrary period of time. This distinction is of practical importance for 2 reasons: 1) the amount of baseline extravascular volume overload present in each patient is highly variable; and 2) extravascular and intravascular volumes are in equilibrium. As a result, intravascular euvolemia will on average take longer to achieve if severe extravascular volume overload is present at baseline and therefore confound the absolute time to peak hemoconcentration by the degree of baseline volume overload. Secondary analysis focused on the absolute time to peak hemoconcentration with adjustment for length of stay.
Estimated glomerular filtration rate (eGFR) was calculated by using the 4-variable Modification of Diet in Renal Disease equation (9). Worsening renal function (WRF) was defined as a ≥20% decrease in eGFR at any time during the hospitalization (7,10–14). All-cause mortality was determined via the Social Security Death Index (15). Loop diuretic doses were converted to furosemide equivalents with 1 mg of bumetanide = 20 mg of torsemide = 80 mg of furosemide for oral diuretics, and 1 mg of bumetanide = 20 mg of torsemide = 40 mg of furosemide for intravenous diuretics. Average daily loop diuretic doses were calculated by dividing the total dose of loop diuretics over the hospitalization (in furosemide equivalents) by the hospital length of stay. Given that hospital admissions over a 5-year period were analyzed, a sensitivity analysis was undertaken to ensure that the primary findings were consistent over this observation period and thus the cohort could be analyzed as a whole. There was no significant difference in the findings between the first quartile enrolled and the fourth quartile enrolled (p interaction = 0.52) or continuously across the enrollment interval (p interaction = 0.23). The study was approved by the institutional review board of the Hospital of the University of Pennsylvania.
The primary goal of the current analysis was to evaluate the association between timing of hemoconcentration and post-discharge survival. As such, the primary analysis focused on risk for all-cause mortality in patients with early HC compared with late HC. Values reported are mean ± SD, median (quartile 1 to quartile 4), and percentile. Independent Student t test or the Wilcoxon rank sum test was used to compare continuous variables. The chi-square test was used to evaluate associations between categorical variables. Proportional hazards modeling was use to evaluate time-to-event associations with all-cause mortality. Candidate covariates entered in the model were baseline characteristics with <10% missing values and a univariate association with all-cause mortality at p ≤ 0.2. Covariates that had a p value >0.2 but a theoretical basis for potential confounding were forced into the model. Models were built by using backward elimination (likelihood ratio test) in which all covariates with a p value <0.2 were retained (16). Adjusted survival curves were plotted for patients with early HC, late HC, and no hemoconcentration; the x axis was terminated when the number at risk was <10%.
Statistical analysis was performed by using SPSS version 20.0 (IBM SPSS Statistics, IBM Corporation, Armonk, New York).
Baseline characteristics of the 845 patients meeting inclusion criteria are presented in Table 1. Overall, 422 (49.9%) patients had an increase in both Hgb and Hct levels from admission to their highest in-hospital level; 323 (38.2%) had no increase; and 100 (11.8%) had an increase in only 1 of the 2 parameters. There was a greater incidence of an increase in Hgb (58.5%) than Hct (53.3%) level. Among patients whose Hct and Hgb levels increased above their admission value, the mean increases in Hct and Hgb were 10.2 ± 9.0% and 9.3 ± 9.5%, respectively. Patients with increases in levels of both Hct and Hgb during hospitalization had parameters consistent with aggressive decongestion, including greater volume/weight loss, exposure to loop diuretics, and deterioration in renal function compared with those with no evidence of hemoconcentration (Online Table 1).
The mean time to peak Hct levels was 3.8 ± 2.8 days and to peak Hgb was 3.8 ± 2.4 days. This equated to the mean peak Hct and Hgb levels occurring 62.7 ± 29.5% through the hospitalization, with 41.5% of patients having peak hemoconcentration in the first 50% of the hospitalization (early HC) and 58.5% with peak hemoconcentration after the 50% mark (late HC). In patients with early HC, the median time to peak in both Hgb and Hct was 2 days (interquartile range [IQR]: 1 to 3 days) in the early HC group and 4 days (IQR: 3 to 6 days) in the late HC group. Overall, baseline characteristics were similar between patients with early and late HC (Table 1). The cumulative amount of loop diuretic administered throughout the hospitalization was not significantly different between groups; however, the average daily quantity of both intravenous and total loop diuretic was significantly greater in patients with late HC compared with early HC (Table 2). In patients with early HC, the maximum dose of intravenous diuretics was administered earlier in the hospital course (25% [IQR: 12.5% to 37.5%]) vs. 33% [IQR: 20% to 50%] of the way through the hospitalization; p < 0.001), and the transition to oral diuretics also occurred earlier in the hospital course (61% [IQR: 39% to 80%]) vs. 67% [IQR: 50% to 83%] of the way through the hospitalization; p = 0.034). The odds of the transition to oral diuretics falling in the first half of the hospitalization was significantly higher for patients with early HC (odds ratio [OR]: 2.2 [p = 0.005]). The relative frequency of hypokalemia (serum potassium <3.5 mEq/l; p = 0.99) and the daily amount of supplemental potassium administered (p = 0.89) was similar between early HC and late HC. The change in eGFR from admission to the worst in-hospital value (Table 2) and the overall incidence of WRF was similar between late HC and early HC (OR: 1.1; p = 0.65). However, the greatest deterioration in renal function occurred further through the hospital course (33% [IQR: 0% to 67%] vs. 25% [IQR: 0% to 67%] of the way through the hospitalization; p = 0.003) and the odds of WRF occurring in the latter 50% of the hospitalization was greater in the late HC group (OR: 2.5; p = 0.006). Notably, the total and average daily fluid/weight losses were significantly greater in patients with late HC (Figs. 1A and 1B), but the length of stay was shorter (Table 2). Although the increase from baseline to peak Hct and Hgb level was only slightly greater among patients with late HC versus early HC, the change in baseline to discharge Hct and Hgb level was substantially greater in the late HC group (Fig. 2). Notably, in patients with early HC, hemoconcentration was largely transient such that only 34.5% with early HC continued to meet criteria for hemoconcentration at the time of discharge. In patients with early HC, the mean peak to discharge hemodilution (change in Hct –9.6 ± 8.0%; change in Hgb –9.0 ± 7.8%) did not differ from the initial degree of hemoconcentration in these patients (p > 0.61 for both comparisons).
Over a median follow-up of 3.4 years (IQR: 1.3 to 5.3 years), 439 (52%) patients died of any cause. The time from hospital admission to ascertainment of all-cause mortality was not significantly different between patients with early HC, late HC, or no hemoconcentration (p = 0.2). Hemoconcentration at any time during the hospitalization (i.e., either early HC or late HC) was not associated with a significantly better survival both before (hazard ratio [HR]: 0.85 [95% confidence interval (CI): 0.70 to 1.0]; p = 0.078) and after (HR: 0.89 [95% CI: 0.72 to 1.1]; p = 0.25) adjustment for baseline characteristics (age, race, sex, hypertension, systolic blood pressure, heart rate, edema, BNP, serum sodium, eGFR, Hgb, Hct, loop diuretic dose, beta-blocker use, and coronary artery disease). Interestingly, late HC was associated with significantly better survival compared with both early HC (HR: 0.73 [95% CI: 0.56 to 0.96]; p = 0.026) and no hemoconcentration (HR: 0.74 [95% CI: 0.59 to 0.93]; p = 0.009). Mortality did not differ between patients with early HC and no hemoconcentration (HR: 1.0 [95% CI: 0.80 to 1.3]; p = 0.93). After adjusting for baseline characteristics, late HC remained significantly associated with better survival compared with both early HC and no HC (Fig. 3). The survival advantage with late HC compared with early HC remained significant after controlling for the absolute magnitude of admission to peak change in Hct and Hgb levels and the length of stay (HR: 0.69 [95% CI: 0.52 to 0.92]; p = 0.012). In addition, extensive adjustment for baseline characteristics, in-hospital treatment-related parameters, and discharge medications (the aforementioned baseline characteristics plus length of stay, admission to peak change in Hct and Hgb, net fluid output, total loop diuretic administered, and inotrope use, as well as discharge beta-blocker, angiotensin-converting enzyme/angiotensin receptor blockers, spironolactone, digoxin, and loop diuretic dose) minimally affected these results (HR: 0.65 [95% CI: 0.45 to 0.92]; p = 0.016). Exploratory analysis comparing patients with late HC who maintained hemoconcentration through discharge (79.4% of late HC patients, n = 196) versus early HC patients who no longer met criteria for hemoconcentration at discharge (62.3% of early HC patients, n = 109) revealed a particularly pronounced survival advantage with sustained late HC (adjusted for the aforementioned admission, in-hospital, and discharge parameters; HR: 0.43 [95% CI: 0.25 to 0.72]; p = 0.002).
Early HC versus late HC, defined by using an absolute rather than relative time to peak hemoconcentration, was not associated with significant differences in survival (p ≥ 0.16 for all cutoff points using days 1 through 14). However, these findings were influenced by a significantly longer length of stay in the late HC groups. For example, defining late HC according to the median time to hemoconcentration (3 days), the length of stay was significantly greater with late HC versus early HC (7 days [5 to 10] vs. 4 days [3 to 6]; p < 0.0001). After adjustment for length of stay, late HC (defined by using the 3-day cutoff point [patient characteristics are presented in Online Table 2]) was associated with significantly better survival than early HC (HR: 0.69 [95% CI: 0.51 to 0.95]; p = 0.022), an association that persisted after adjustment for admission, in-hospital, and discharge parameters (HR: 0.59 [95% CI: 0.39 to 0.89]; p = 0.011). Notably, late HC, which was defined as a peak in hemoconcentration occurring at very late time points, such as after 7 days (adjusted HR: 0.48 [95% CI: 0.25 to 0.93]; p = 0.028), 8 days (adjusted HR: 0.30 [95% CI: 0.13 to 0.71]; p = 0.007), 9 days (adjusted HR: 0.24 [95% CI: 0.09 to 0.66]; p = 0.006), and 10 days (adjusted HR: 0.26 [95% CI: 0.07 to 0.91]; p = 0.035), was actually associated with a progressively greater survival advantage compared with early HC after controlling for length of stay.
The principal finding of the current study was that the timing of hemoconcentration during the treatment of decompensated heart failure strongly influences the associated mortality, with significantly better survival for patients with late HC compared with those with early HC or no hemoconcentration. Interestingly, baseline disease severity indicators, total in-hospital loop diuretics administered, and the overall incidence of WRF were similar between patients with early HC and late HC. However, the pattern of diuretic approach seemed to differ with higher average daily doses of loop diuretics, later transition to oral diuretics, later time to peak deterioration in renal function, and greater fluid and weight loss with late HC versus early HC. Notably, the majority of early HC patients experienced substantial hemodilution subsequent to their early HC, such that their discharge Hct and Hgb levels were similar to the admission values. In conjunction with the lesser degree of overall weight and volume loss, these observations suggest that diuresis may have been de-escalated before complete elimination of extravascular volume overload in early HC patients. Overall, these findings indicate that the association between hemoconcentration and better survival is contingent not only on whether hemoconcentration can be induced but also on the timing/durability of the hemoconcentration.
As a basic physiological principal, removal of intravascular fluid at a rate faster than it can be refilled from extravascular stores will lead to hemoconcentration. This physiology has been used by nephrologists for decades to monitor intravascular volume status during hemodialysis (3,4). Analogously, some degree of hemodilution after cessation of fluid removal is also to be expected as extravascular fluid comes to equilibrium with intravascular volume. However, in the setting of ADHF with severe extravascular volume overload, recurrence of intravascular volume overload can be nearly complete after abrupt cessation of early aggressive fluid removal. In a study by Guazzi et al. (17), severely volume-overloaded patients were exposed to a brief period of high rate of ultrafiltration (500 cc/h), resulting in a 22% decrease in plasma volume. Despite a stable weight maintained over the next 48 h with oral diuretics, plasma volume completely returned to pre-treatment values. Notably, both right- and left-sided filling pressures were reduced with the intervention, but unlike plasma volume, filling pressures did not rebound and actually remained at the immediate post-treatment levels. Similarly, in a study by Marenzi et al. (18), 24 h after removal of 7.5 liters of fluid (ultrafiltrate + urine), plasma volume was unchanged from pre-treatment levels, but filling pressures were maintained at the improved immediate post–fluid removal levels. These observations highlight the fact that via the substantial and dynamic compliance of the capacitance vessels of the venous system, filling pressures and blood volume may have little correlation (19). Importantly, in the study by Marenzi et al. (18), the mean right atrial pressure 24 h after ultrafiltration was reduced to ∼8 mm Hg. This finding indicates that a significant percentage of these patients likely had a normalization of jugular venous pressure despite an unchanged blood volume. As such, it is understandable how transient or even a complete absence of hemoconcentration can occur in the setting of substantial diuresis and how hemoconcentration could be transient even in the setting of continued slower diuresis. These findings reinforce the fact that the vast majority of our surrogates for decongestion, including right heart catheterization parameters and even directly measured blood volume, only provide a snapshot of the intravascular environment rather than a true measure of global decongestion. The implication of this physiology is that application of a one-size-fits-all strategy of aggressive diuresis to a prespecified/nonpatient specific time to hemoconcentration is unlikely to be successful. Support for this concept is provided by the finding that late HC defined in absolute terms (i.e., number of days to hemoconcentration) was only associated with better survival after adjustment for length of stay.
The association between hemoconcentration and improved survival provides reassurance that the net clinical benefit from aggressive decongestion may outweigh the adverse effects of decongestive therapies (6,7). However, interpretation of this observational data is challenging because it is unclear to what degree the improved survival results directly from the aggressive decongestion or if patients who respond to treatment with hemoconcentration are healthier patients. Given that diuretic resistance is a powerful prognostic indicator, it is intuitive that patients responding to diuretics with an effective diuresis may have lower disease severity (20–22). There are several findings in the current analysis that argue that this is not the primary factor driving these observations. First, both early and late HC patients were likely diuretic responsive because hemoconcentration occurred in both. Notably, the group with early HC actually experienced hemoconcentration with less aggressive treatment. Second, early HC patients had similar maximum intravenous loop diuretic doses but earlier administration of that maximum dose and an earlier transition to oral diuretics. This suggests that the inability to maintain hemoconcentration was not purely a function of a greater degree of diuretic refractoriness. Third, baseline characteristics were remarkably similar, but essentially all in-hospital treatment-related parameters trended toward a more sustained/aggressive diuresis in the late HC group. Fourth, mortality differences between groups persisted after extensive adjustment for baseline, in-hospital, and discharge parameters. These findings lend incremental support to the concept that an association between hemoconcentration and improved survival is not solely driven by treatment refractoriness in the comparison group.
Although it seems probable that there was a difference in the timing of de-escalation of therapy between the early HC and late HC groups, which was probably not primarily driven by diuretic refractoriness, the motivation for this de-escalation is unknown. Although admission, discharge, and available nondiuretic in-hospital characteristics were similar between groups, differences in the ability to tolerate therapy may have emerged during treatment. As a result, the assumption that more aggressive/sustained diuresis in the early HC group would have led to improved outcomes may be incorrect. For example, factors such as hypotension, WRF, adequacy of perfusion, or even a general ill-appearance of the patient can (appropriately) factor into the physician’s decision-making regarding subsequent diuretic therapy. As such, early de-escalation of therapy could have resulted from a physician’s erroneous conclusion that the patient no longer required diuresis or the astute recognition that the patient could no longer tolerate further diuresis. As a result, it is possible that the differences in survival were derived from differences in the ability of the patients to tolerate sustained aggressive diuresis rather than the sustained diuresis itself.
First, given the observational nature of the current study, causality is impossible to demonstrate, and residual confounding cannot be excluded. Furthermore, the data were generated from a single center, and given that local practice patterns can vary across institutions, it is possible that these findings will not generalize to other centers. Due to the modest size of our cohort, the ability to rule out temporal variations in the association between early HC and late HC with mortality is limited. Larger studies will be needed to determine if changing practice patterns over time influence these observations. In addition, exclusion of patients who did not have a baseline BNP level available further limits generalizability. Serial measures of total protein and albumin were not available in this population, markers which seem to have better operating characteristics than Hct and Hgb levels (7). Notably, in previous reports, Hct level alone was not significantly associated with WRF or mortality, and Hgb level alone did not remain significantly associated with mortality after multivariate adjustment in a different study (6). Although this limitation was likely overcome somewhat by increasing the number of patients studied (n = <340 in the previous studies) and using both Hct and Hgb levels to define hemoconcentration, it is unclear if the results would differ significantly if total protein and albumin data had been available. Furthermore, the primary assumption is that increases or decreases in the concentration of Hct, Hgb, and plasma proteins are predominantly driven by changes in volume status, but this hypothesis has never been tested. Factors such as bleeding, phlebotomy, and the rate of production of new red blood cells have an uncertain influence, and we cannot exclude the possibility that these factors do not similarly affect early HC and late HC; thus, the lack of availability of these parameters is a significant limitation to the study. Although the analyses herein may provide proof of concept when analyzed in large cohorts, on the individual patient level, biological and assay variability will likely preclude serial Hct and Hgb measures as a reliable gauge of intravascular volume shifts. Although total and peak doses of intravenous diuretics did not support the hypothesis that the early HC group was highly diuretic resistant, these are crude metrics to assess diuretic responsiveness, and it is possible that there were differences between groups that may have influenced outcomes. In addition, the lack of data on rehospitalizations is a significant limitation. Perhaps most importantly, the observational nature of this study makes it impossible to exclude the possibility that patients with early HC were not just “sicker,” and the observed differences in survival actually had little to do with the aggressiveness of diuresis. As a result of these noted limitations, our findings should be considered hypothesis-generating and serve primarily to motivate further consideration of randomized trials of decongestive strategies that can address causality.
The association between hemoconcentration and survival is strongly influenced by the timing at which peak hemoconcentration occurs, with better survival in patients with late HC compared with early HC or no hemoconcentration. It cannot be determined from these data to what degree an aggressive sustained diuretic strategy versus a durable response to that strategy drove these observations. However, these data do suggest that even when initial diuresis is rapid enough to produce hemoconcentration, outcomes are inferior if it is not maintained at a rate adequate to progressively contract plasma volume throughout the hospitalization. Given that volume management is the primary therapeutic objective in most decompensated heart failure admissions, and its completeness may have substantial influence on subsequent outcomes, additional research is necessary to develop more objective and reproducible therapeutic endpoints.
For a supplemental figure and tables, please see the online version of this article.
This research was funded by grants from the National Institutes of Health (5T32HL007843-15, 1K23HL114868-01, and K24DK090203). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute decompensated heart failure
- B-type natriuretic peptide
- confidence interval
- estimated glomerular filtration rate
- hazard ratio
- interquartile range
- odds ratio
- worsening renal function
- Received February 28, 2013.
- Revision received May 1, 2013.
- Accepted May 6, 2013.
- American College of Cardiology Foundation
- Adams K.F. Jr..,
- Fonarow G.C.,
- Emerman C.L.,
- et al.
- Leypoldt J.K.,
- Cheung A.K.,
- Steuer R.R.,
- Harris D.H.,
- Conis J.M.
- Testani J.M.,
- Chen J.,
- McCauley B.D.,
- Kimmel S.E.,
- Shannon R.P.
- van der Meer P.,
- Postmus D.,
- Ponikowski P.,
- et al.
- Testani J.M.,
- Kimmel S.E.,
- Dries D.L.,
- Coca S.G.
- Mickey R.M.,
- Greenland S.
- Guazzi M.D.,
- Agostoni P.,
- Perego B.,
- et al.
- Marenzi G.,
- Lauri G.,
- Grazi M.,
- Assanelli E.,
- Campodonico J.,
- Agostoni P.
- Testani J.M.,
- Cappola T.P.,
- Brensinger C.M.,
- Shannon R.P.,
- Kimmel S.E.