WORSENING RENAL FUNCTION IN ACUTE DECOMPENSATED HEART FAILURE: A BAD SIGN, OR MAYBE NOT?

Abstract

Introduction: Prognostic role of worsening renal function (WRF) during hospitalization for acute decompensated heart failure remains controversial.

Methods: We reviewed the medical literature on the association of WRF during acute decompensated heart failure with clinical outcomes.

Results: WRF is reported in approximately 25% to 40% of acute decompensated heart failure patients. WRF is usually, but not always associated with worse outcomes in patients with heart failure. Transient WRF accompanied with hemoconcentration, effective decongestion strategies, and initiation of appropriate medical treatment for heart failure with angiotensin-converting enzyme inhibitors or mineralocorticoid receptor antagonists is not associated with worse outcomes.

Conclusions: Multiple mechanisms may contribute to WRF in acute decompensated heart failure, and prognosis will differ according to etiology, patient features, and treatment strategies. During hospitalization, treatment should focus on the patient's clinical status, resolution of symptoms and signs of congestion rather than temporary changes in renal function.

INTRODUCTION

The interaction between chronic congestive heart failure and chronic kidney failure is characterized by hemodynamic and neurohormonal interdependence (1). Conventionally, the presence of chronic kidney disease carries worse prognosis in patients with congestive heart failure. Furthermore, changes in renal function during the hospital admission for acute decompensated heart failure have adverse effects on survival rates (2). However, it is not certain that all changes in renal function during acute decompensated heart failure imply clinical deterioration and are associated with worse outcomes.

METHODS

We reviewed the current literature on the prevalence of worsening renal function, association of worsening of renal function with clinical outcomes in the setting of successful decongestion strategies, hemoconcentration, and initiation of heart failure medications in patients with acute decompensated heart failure.

RESULTS

Prevalence of Worsening Renal Function in Acutely Decompensated Heart Failure Patients

Approximately half of all heart failure patients present with at least moderate chronic kidney disease at baseline (2,3) which serves as a powerful predictor of mortality (4). The reported prevalence of worsening renal function (WRF) in patients admitted with acute decompensated heart failure ranges from 6.8% to 40% (5-26). Studies examining prevalence of WRF during hospitalization for acute decompensated heart failure are listed in Table 1. The large disparity in the prevalence of WRF can be attributed to varying definitions of WRF, such as an increase in creatinine more than 0.03 mg/dL or a decrease in estimated glomerular filtration rate (eGFR) more than 20%. Transient WRF is seen in approximately 30% to 50% of acute decompensated heart failure patients with WRF (23,24). Studies define this pattern of renal function changes by a peak creatinine ≥0.3 mg/dL higher than admission creatinine, which later returns to within 0.3 mg/dL of the baseline values by discharge. Patients with transient WRF have similar rates of mortality at 6 months compared to patients with stable renal function or patients without WRF (12,23). Although these findings suggest that transient WRF is not associated with adverse outcomes in acute decompensated heart failure, providers commonly delay discharge from the hospital until creatinine or other renal parameters normalize. Therefore, transient WRF has been associated with significantly increased costs and lengths of stay in patients hospitalized with in acute decompensated heart failure (24,27).

TABLE 1.

Definition and Prevalence of Worsening Renal Function Reported in Studies of Patients With Acute Decompensated Heart Failure

Author, Publication Year	Worsening Renal Function Definition	Prevalence of Worsening Renal Function (%)
Gottlieb, 20025	≥0.3 mg/dL increase in Cr	39.0
	≥20% decrease in eGFR	30.0
Butler, 20046	≥0.3 mg/dL increase in Cr	33.0
Logeart, 20087	increase in ≥ 25 umol/L in Cr	37.0
Nohira, 20088	≥0.3 mg/dL increase in Cr	29.5
Mullens, 20099	≥0.3 mg/dL increase in Cr	40.0
Testani, 201010
Testani, 201011	≥20% change in eGFR	21.2
Aronson, 201012	≥0.5 mg/dL in Cr or ≥25% change in eGFR	15 (persistent WRF)
	≥0.5 mg/dL Cr or ≥25% change in eGFR that returned to within 0.5 mg/dL in Cr or 25% change in eGFR	8.3 (transient WRF)
Testani, 201114	≥20% decrease in eGFR	20.5
	≥0.3 mg/dL increase in Cr	18.1
Voors, 201115	≥0.3 mg/dL increase in Cr	30.0
Blair, 201116; Konstam, 200717	≥25% decrease in GFR	10.0
	≥0.3 mg/dL increase in Cr	15.0
Greene, 201319	≥0.3 mg/dL increase in Cr	13.8
	≥25% decrease in eGFR	11.6
Brisco, 201320	≥25% decrease in GFR post-discharge	39.2
Damman, 201421	≥0.3 mg/dL increase in Cr	6.8
	≥25% decrease in eGFR	11.5
Krishnamoorthy, 201423	≥0.3 mg/dL increase in Cr, returned to within 0.3 mg/dL Cr at discharge	11.7 (transient WRF)
	≥0.3 mg/dL increase in Cr	20.3 (persistent WRF)
Palmer, 201524	≥0.3 mg/dL and a ≥25% increase in Cr that did not persist at discharge	13.0 (transient WRF)
	≥0.3 mg/dL and >25% increase in Cr	10.0 (persistent WRF)
Kula, 201626	≥20% decrease in eGFR	17.0

WRF, worsening renal function; eGFR, estimated glomerular filtration rate; Cr, creatinine.

Studies evaluating patient factors that contribute to the development of WRF show that a history of diabetes and a baseline serum creatinine greater than 2.5 mg/dL are independent risk factors for developing WRF in patients hospitalized for acute decompensated heart failure (6,9,14). Reductions in systolic blood pressure have also been associated with WRF (15), suggesting that decreased cardiac output and reduced renal perfusion could be key contributors of the development WRF. Paradoxically, there are reports showing that WRF accompanied with a decrease in systolic blood pressure is not associated with mortality (14). Of all the hemodynamic variables, elevated central venous pressure has been reported to be strongly associated with WRF, and appears to be a stronger predictor of mortality than elevated pulmonary capillary wedge pressure or cardiac output (9,28). Thus, emphasis is shifting away from a reduced cardiac output model to a mechanism rooted in venous congestion resulting in WRF and being associated with adverse outcomes.

Importance of Decongestion

Current literature suggests that WRF accompanied with successful decongestion strategies is not associated with adverse outcomes. Patients with WRF and signs of effective decongestion with resolution of jugular venous pressure or edema at discharge have comparable survival rates to patients without WRF (29). On the other hand, WRF with incomplete decongestion, with one or more signs of congestion such as rales, jugular venous distension, or edema at discharge, is significantly associated with increased mortality rates (29). Treatment-induced hemoconcentration, commonly defined as an increase in plasma hemoglobin or albumin levels in acute decompensated heart failure, has been studied as a surrogate marker of effective decongestion strategies in acute decompensated heart failure. There is a growing body of evidence showing that hemoconcentration is associated with improved survival in patients hospitalized for heart failure (10,14,19,30). Furthermore, patients with WRF and markers of hemoconcentration have comparable clinical outcomes compared to patients without WRF (19,30). Patients with hemoconcentration and WRF frequently experience renal function recovery after discharge (19). To better understand the relationship between timing of fluid equilibration and hemoconcentration, the timing of hemoconcentration during admission has been characterized in acute decompensated heart failure patients (31). Late hemoconcentration with a peak of hemoconcentration at 4 days was reported to be associated with a better survival compared to early hemoconcentration with a peak hemoconcentration at 2 days or earlier (31). Those with early hemoconcentration underwent less aggressive diuretic regimens and diuresis was de-escalated too early compared to patients with late hemoconcentration, resulting in ineffective relief of fluid overload (31). These studies underscore the importance of continued diuresis and effective decongestion strategies, and complete resolution of congestion in hospitalized patients with decompensated heart failure, despite a transient increase in creatinine values. Future studies are needed on the role of hemoconcentration as a surrogate marker for targeting effective decongestion strategies in acute decompensated heart failure.

Heart Failure Medications and Renal Function

In addition to diuretic therapy, neurohormonal antagonism with angiotensin converting enzyme inhibitors (ACEIs), angiotensin receptor antagonists or mineralocorticoid receptor antagonists (MRAs) are part of the standard heart failure therapies. However, changes in creatinine frequently serve as an impetus for discontinuing these medications. These medications are beneficial regardless of the modest and usually transient changes in renal function after initiation. For example, it has been shown that despite a transient 10% to 15% increase in serum creatinine after initiation, ACEI therapy has been associated with a significant survival benefit in patients with heart failure (13). Whereas this modest creatinine increase commonly occurs within the first 2 weeks of ACEI, creatinine levels usually stabilize thereafter (32). Similarly, MRAs spironolactone or eplerenone retain mortality benefits in heart failure patients with reduced ejection fraction despite modest increases in creatinine levels after initiation of these medications (18,22). Of course, it should be kept in mind that MRAs are contraindicated in heart failure patients with severe renal dysfunction, with serum creatinine greater than 2.5 mg/dL in men or greater than 2.0 mg/dL in women (or eGFR < 30 mL/min/1.73 m²), and/or potassium greater than 5.0 mEq/L. Caution should be exercised in patients with persistent and continued increase in creatinine levels after initiation of ACEI or MRAs, as this may imply intrinsic kidney disease and/or advanced comorbidities. These patients require close monitoring, dose adjustment, and consideration for discontinuation of these medications if a pattern of persistent or accelerated worsening of renal function is noted.

DISCUSSION

Approximately 25% to 40 % of patients who are hospitalized for decompensated heart failure have WRF, commonly detected by a transient increase in creatinine or a decrease in eGFRs during hospitalization. Physicians commonly react to these changes as signs of acute kidney injury and delay discharge and tailor treatment accordingly. Multiple mechanisms may contribute to WRF in acute decompensated heart failure, and prognosis will differ according to etiology, patient features, and treatment strategies. WRF accompanied with hemoconcentration, effective decongestion strategies, or initiation of appropriate medical treatment with ACEIs or MRAs is usually not associated with worse prognosis in patients with heart failure. Future studies are needed to define the role of hemoconcentration as a surrogate target for effective decongestion strategies in acute decompensated heart failure.

SUMMARY

During hospitalization for heart failure, treatment should focus on the patient's clinical status, resolution of symptoms, and signs of congestion rather than temporary changes in kidney function. Successful resolution of congestion and appropriate treatment strategies for heart failure is associated with favorable outcomes regardless of the temporary changes in kidney function during hospitalization for acute decompensated heart failure.

DISCUSSION

Konstam, Boston: That was a terrific overview of the ridiculously complicated set of data. You've touched on the last slide — which is right-heart failure — right ventricular function, and it's more and more recognized that the renal vein pressure may be the dominant feature driving this. So, two questions: How do you account for the fact that decongesting the right heart seems to be good for renal function? But on the other hand, hemoconcentration is also good and associated with worse renal function very often? Secondly, just generally, I mean, should we be paying more attention to the right ventricle in kind of stratifying the significance of renal impairment?

Bozkurt, Houston: I'll start with the second one. The answer is absolutely yes. We should pay more attention. The problem is we don't know how to treat the RV and you alluded to the complex paradox. It's a U-shaped curve. Too much preload reduction will result in a detriment. We do know that of all the variables that are associated with worsening renal function — CVP or the right atrial pressure — are the most important hemodynamic variables associated with development of worsening renal function and also with adverse clinical outcomes according to studies published by Kevin Damman, Wilfried Mullen, and Wilson Tang. It is also important to recognize that there are certain patients who are responders to RV unloading and there are certain patients who are not. The ones who are not are likely the ones who develop worsening renal function. I propose a new paradigm where we should use the first response to therapy as a surrogate for our target. So, if they are responding, advance it and enhance it to the level of maximum optimization. That is a marker. Responsiveness to therapy can be a biomarker of the RV. And if they don't respond, don't push it; you're on the other curve part of the U-shaped arm of the RV. So, I try to address the paradox that we see with RV; it is the responsiveness of the RV that is the Achilles heel. If it works, wonderful. The declining creatinine with the decongestion looks like temporarily it's better. Regarding the first question, yes the elevation of central venous pressure can be transmitted back to the renal veins, resulting in direct impairment of renal function. Increased CVP also causes an increase in renal interstitial pressure, which might lead to a hypoxic state of the renal parenchyma; and furthermore, increased intra-abdominal pressure caused by visceral edema or ascites may also play a role. But in essence, we don't know how to measure the abdominal or renal congestion. Maybe we'll do a Swan of the kidneys and maybe some of the nephrologists will teach us how to do the renal blood flow to assess renal congestion. Our knowledge of decongestion of certain vascular beds such as the splanchnic area, or elimination of renal venous congestion or gut edema is very limited.

Bishopric, Washington DC: I wanted to ask you how your study broke out in terms of the numbers of patients who had preserved versus reduced ejection fraction, and for how many people did you have actual baseline creatinine levels? I think the study is more complicated than it would look like at face value, because you probably got a bunch of different subpopulations that are responding differently to their diuretic therapy.

Bozkurt, Houston: All our patients were heart failure with reduced ejection fraction (EF), with EF of less than 35% per our inclusion criteria. Though similar trends — actually worse trends — have been reported in patients with heart failure with preserved ejection fraction (HFPEF)in other studies…We excluded CKD-stage 5. Three hundred fifty-eight of our patients had paired admission and discharged creatinine levels. All had admission creatinine levels within 24 hours.

Bishopric, Washington DC: But I mean before they were admitted…

Bozkurt, Houston: Before they are admitted, what percent had the creatinine levels at baseline? Since it's a closed system at the VA, I would say probably patients other than the new-onset acute heart failure admissions, which was one-third of our population, had former creatinine levels measured before admission. So, the remaining two-thirds of the population would have had former creatinine levels.

Golden, Baltimore: So, I was interested in one of your subgroups. I am an endocrinologist and certainly patients with diabetes and heart failure do much more poorly and have higher rates of readmission and are harder to diurese. So, I am just curious: like, looking at the BNP and your other parameters, are there any differences by diabetes status, and how should we be focusing on trying to decongest those patients better?

Bozkurt, Houston: We didn't look at the clustering or the subgroup analysis of diabetes. But in other studies, diabetes is always associated with worse outcomes, worsening of the kidney function, diuretic resistance, and a very rapid rise type (i.e., type 4 renal tubular acidosis, patients with worsening renal function and hyperkalemia. Diabetes is truly one of our biggest challenges. The other important thing is most of our very promising treatment strategies in terms of prevention of heart failure, is coming from glucose lowering agents. We're now finding out that those agents are not only treating diabetes, they're actually preventing heart failure. Promising agents include SGLT-2 inhibitors. Additionally, I think we need to focus on prevention and elimination of diabetes rather than just treating the blood sugars, which may help with the heart. I think there's definitely a double target concept — prevention of injury to the end organs while treating or preventing diabetes. Definitely the heart and kidney commonly gets impaired in diabetes, and prevention or treatment of diabetes may help prevent development of heart failure.

Luke, Cincinnati: Did you look at the proteinuria or microalbuminuria? Right heart failure does change protein excretion. The second question, did you look at the long-term renal function say a year later? Because we're very worried about AKI in terms of that.

Bozkurt, Houston: We didn't, but there are data. There's big Medicare data looking at 1 year outcomes in terms of what happens to the kidneys especially when they have demonstrated worsening renal function during the hospitalization. The transient worsening of renal function is very similar in the Medicare outpatient population. One-third normalizes at 30-day follow-up, one-third goes persistently up, and one-third is stable. Regarding proteinuria, it's a known subgroup of the individuals with worsening renal function that probably has the most adverse outcomes in heart failure. Yes, it goes hand-in-hand with diabetes and somewhat with hypertension. We didn't look at proteinuria in our patients. It is probably one of the scariest patient subgroups that we worry about. We don't know exactly what would prevent it. It is seen more with RV failure too, as you would see with most of the patients with hepatic-congestion and the gut congestion. Though there are data with prevention and treatment of proteinuria with ACE inhibitors or angiotensin receptor blockers in patients with chronic kidney disease and diabetes, we don't know whether treatment of heart failure by itself without direct effects on the kidney would result in improvement of proteinuria. I'm not aware of proteinuria benefit with decongestion alone. I think the kidney needs to get better as well.