Patients hospitalized with acute heart failure, worsening renal function, and persistent congestion are at high risk for adverse outcomes despite current medical therapy

Abstract Introduction Approximately 1/3 of patients with acute decompensated heart failure (ADHF) are discharged with persistent congestion. Worsening renal function (WRF) occurs in approximately 50% of patients hospitalized for ADHF and the combination of WRF and persistent congestion are associated with higher risk of mortality and HF readmissions. Methods We designed a multicenter, prospective registry to describe current treatments and outcomes for patients hospitalized with ADHF complicated by WRF (defined as a creatinine increase ≥0.3 mg/dL) and persistent congestion at 96 h. Study participants were followed during the hospitalization and through 90‐day post‐discharge. Hospitalization costs were analyzed in an economic substudy. Results We enrolled 237 patients hospitalized with ADHF, who also had WRF and persistent congestion. Among these, the average age was 66 ± 13 years and 61% had a left ventricular ejection fraction (LVEF) ≤ 40%. Mean baseline creatinine was 1.7 ± 0.7 mg/dL. Patients with persistent congestion had a high burden of clinical events during the index hospitalization (7.6% intensive care unit transfer, 2.1% intubation, 1.7% left ventricular assist device implantation, and 0.8% dialysis). At 90‐day follow‐up, 33% of patients were readmitted for ADHF or died. Outcomes and costs were similar between patients with reduced and preserved LVEF. Conclusions Many patients admitted with ADHF have WRF and persistent congestion despite diuresis and are at high risk for adverse events during hospitalization and early follow‐up. Novel treatment strategies are urgently needed for this high‐risk population.


| INTRODUCTION
Heart failure (HF) is common, affecting more than six million adults and resulting in over 1.8 million hospitalizations per year in the United States alone, with significant financial impact. 1 HF is a chronic, progressive condition, marked by episodes of acute decompensation, typically requiring hospitalization for stabilization and improvement in symptoms.Almost 90% of hospitalizations are due to symptoms related to volume overload.Unfortunately, outcomes after hospitalization for acute decompensated heart failure (ADHF) are poor.[15] Moderate to severe renal dysfunction or worsening renal failure (WRF) complicates approximately 66% of ADHF admissions. 16The pathophysiology underlying cardiorenal syndrome (CRS) in ADHF is poorly understood and probably multifactorial, reflecting underlying renal dysfunction from HF and other conditions, neurohormonal activation and inflammation, side effects of acute therapies, and renal congestion. 17The degree of acute renal dysfunction has not been shown to correlate with the degree of left ventricular systolic dysfunction in heart failure with reduced ejection fraction (HFrEF), though it may be associated with the severity of diastolic dysfunction in heart failure with preserved ejection fraction (HFpEF). 18Patients with WRF represent a particularly high-risk population with poor outcomes.A previously developed and validated predictive model for mortality from the Acute Decompensated Heart Failure National Registry (ADHERE) database identified laboratory indices of renal function to be closely associated with mortality. 19A subsequent analysis from this registry of >110 000 admissions for ADHF determined that in-hospital mortality increased from 1.9% for patients with normal renal function to 7.6% in patients with severe renal dysfunction. 16A separate analysis of >10 000 patients from the Swedish Heart Failure (SwedeHF) and Stockholm Creatinine Measurement (SCREAM) registries further demonstrated that WRF within one year was strongly associated with subsequent long-term mortality in all left ventricular ejection fraction (LVEF) groups. 20gether, these data suggest WRF represents a separate and distinct risk factor in both HFrEF and HFpEF.
There are no current proven treatments for patients with CRS.
Therapies that reduce congestion without causing WRF have been elusive.The American College of Cardiology Foundation/American Heart Association (ACCF/AHA) Heart Failure Guidelines previously identified the lack of recognition and treatment of CRS as a key gap in heart failure knowledge and evidence. 19There is a need to better understand this unique patient population to develop new treatment strategies that will favorably alter the course of this condition.An increase in creatinine of ≥0.3 mg/dL from baseline was used to define WRF, consistent with prior studies. 20,21The baseline creatinine and WRF were determined by site investigators from data obtained during the hospitalization or in the 90 days before hospital admission.The presence of persistent congestion at 96-h postadmission was defined by signs, symptoms, or objective findings of HF.If hemodynamic measurements were obtained, persistent congestion was defined as PCWP > 22 mmHg or CVP > 10 mmHg.

| Study population
In the absence of hemodynamic measurements, physical exam findings (e.g., lower extremity edema, ascites, elevated jugular venous pressure, or pulmonary rales), clinical symptoms (e.g., dyspnea at rest or with minimal exertion, paroxysmal nocturnal dyspnea, or orthopnea), and confirmatory tests (e.g., chest x-ray or echocardiography) were used.
Participants were excluded if they had acute kidney failure, defined as serum creatinine ≥4.0 mg/dL or stage V chronic kidney disease (eGFR ≤ 15 mL/min/1.73m 2 ) on admission or within the prior 90 days.Participants were excluded if they had a single kidney or required dialysis or ultrafiltration in the prior 90 days.Participants were also excluded if they were supported with a mechanical support device or had a prior heart transplant at the time of enrollment.
Enrolled participants were treated according to the standard of care at each enrolling facility and followed for significant clinical events during the hospitalization and through 90 days post-discharge (ADHF + WRF + persistent congestion group only).Eight US sites participated in an economic substudy and participants at those sites signed additional consent approving the collection of billing data.

| Data source
Data were collected from medical records and follow-up phone calls to ascertain vital status and outcomes at 90 days.Records were reviewed for details regarding past medical history, the presence of comorbidities, and any history of HF or prior hospitalizations.Exam findings, blood, and urine laboratory results, imaging studies, invasive testing, and treatments were similarly reviewed.For participants participating in the economic data collection, billing data were collected for the index hospitalization and subsequent rehospitalizations within 90 days of discharge from the index hospitalization.
Participants who were readmitted within 90 days of discharge had all costs of care analyzed, even if the hospitalization extended past 90 days.

| Outcomes
Significant clinical events during the initial hospitalization included escalation of care to an intensive care unit (ICU), mechanical ventilation, extracorporeal membrane oxygenation, dialysis, mechanical circulatory support, discharge on inotropes, addition to the transplant list, or death.
Outcomes at 90 days (±14 days) included survival, HF rehospitalization, and progression of disease, defined as need for evaluation for a left ventricular assist device (LVAD), addition to the heart transplant waitlist, or undergoing a heart transplantation.Costs and length of stay (LOS) were analyzed for subjects participating in the economic substudy.

| Statistical analysis
The data are available on request from the authors.Baseline characteristics, symptoms, and significant clinical events were described by count and percentage of study population.All descriptive data were compared using chi-square test for two-way tables.In cases when frequencies below 5 were met for more than 20% of the cells, the Fisher exact test was applied.All continuous data were compared using a two-independent sample t test.In cases when data failed the Shapiro-Wilk normality test with p value < .15,nonparametric Wilcoxon rank-sum test was applied.
Univariable analysis was performed to assess the relationship between baseline characteristics and the primary study outcome.
Kaplan-Meier curves were generated for event-free survival by analysis of time (days) from 96-h post-admission to the first hospitalization (when persistent congestion despite medical management was confirmed) until the admission to subsequent hospitalization, death from any reasons, LVAD, or heart transplantation, whichever occurred first.This analysis was conducted for all participants within the corresponding subgroup (ADHF + WRF with persistent congestion, HFrEF vs. HFpEF) and censored with the date of follow-up.In cases when follow-up information was not collected or survival status was unknown, the time was censored with the date of discharge from the last hospitalization.In cases when data were available >90-day post-discharge, the time was censored at 90-day post-discharge.Participants with no follow-up were not included (n = 32).
Hospitalization costs were calculated by multiplying all charges by the cost center-specific cost-to-charge ratio obtained from each hospital's Medicare cost report.Physician fees were not analyzed.

| Baseline characteristics
Participants were considered to have persistent congestion based on signs or symptoms at 96 h which included lower extremity edema in 78%, elevated JVP in 45%, and dyspnea at rest or with minimal exertion in 25% (Table 2).

| Clinical events during index hospitalization and 90-day follow-up
During the index hospitalization, 7.6% of participants were transferred to the ICU, 3.4% were discharged on inotropes, and 3.4% of participants died (Table 3).HFrEF and HFpEF participants had similar rates of events during the index hospitalization, except only HFrEF participants were discharged on inotropes (5.6% of HFrEF participants, 0% of HFpEF participants; Supporting Information: Table ).
Overall, serum creatinine increased during admission and decreased back to baseline before discharge.The mean creatinine at baseline was 1.7 ± 0.7 mg/dL, 2.2 ± 0.9 mg/dL at peak, and 1.8 ± 0.8 mg/dL closest to discharge, though this was different for participants with HFrEF compared to HFpEF as described below.
Overall, 53% of participants were discharged with clinical signs of persistent congestion.During the 90-day follow-up period, 53 participants (23%) had at least one rehospitalization and 29 participants (12%) died.
At 90-day follow-up, 33% of participants with ADHF + WRF + persistent congestion were readmitted for ADHF or died, and 41% were readmitted for ADHF, had an LVAD placed, underwent heart transplant, or died.Event-free survival until death, LVAD, heart transplantation, or rehospitalization was similar between participants with HFrEF compared to HFpEF (Figure 1).In this study of patients with ADHF, a large proportion, 39%, developed WRF and of that population, a majority, 71%, also had persistent congestion after 96 h of standard therapy.These participants with ADHF, WRF, and persistent congestion had a high clinical event rate that includes both in-hospital and short-term mortality at 90 days, readmission to the hospital for recurrent ADHF, and disease progression leading to LVAD or heart transplantation.When analyzed in this prospective cohort, the cumulative incidence of these events occurred in 42% of our cohort.This study adds to prior findings that patients with HF and renal dysfunction are at a high risk for poor outcomes. 21Further, this study is unique in the inclusion and comparison of patients with HFrEF and HFpEF.Despite different  Most patients admitted with ADHF have volume overload on admission. 22The presence of congestion is associated with adverse outcomes [23][24][25][26][27][28][29][30] ; and thus a goal of therapy for patients with ADHF is safe and effective decongestion. 23,31Moreover, filling pressures have been shown to be more prognostically important than even cardiac index in the management of patients with ADHF, and congestion is associated with worse outcomes independent of cardiac index. 32rrent recommendations for the treatment of ADHF emphasize the need to assess the trajectory of treatment throughout the hospital course to ensure the patient is progressing toward the goal for discharge. 33Despite these efforts, a substantial portion of patients hospitalized with volume overload are discharged with persistent congestion, and persistent congestion is associated with worse outcomes. 14,15,32,33In prior studies, the percentage of patients discharged with congestion varies from <10% to almost 50% in part due to the variable ways of measuring and defining congestion.
In the current study, 53% of patients with WRF and persistent congestion at 96 h were discharged with remaining clinical signs of congestion, emphasizing that persistent congestion is a high-risk feature and highlighting the need for effective treatment options for decongestion.
Previous studies have shown that WRF occurs in approximately 1/3 of patients admitted to the hospital for ADHF; the present study observed a similar percentage with 39% of patients having WRF.
Those patients with HF and renal dysfunction are at the highest risk for poor outcomes including in-hospital death and those with renal dysfunction are at high risk of not achieving adequate diuresis. 20,215][36] Although earlier studies suggested that the risk of readmission for HF and mortality in ADHF was largely mediated by the presence of WRF, follow-up analyses suggested that the effect of WRF was mediated by the presence of persistent congestion. 20,21e combination of WRF and persistent congestion is associated with a greater increase in the risk of mortality and hospital readmission. 20In one study, a 30% reduction in eGFR was associated with a 20% increase risk of death; however, the risk of death was modulated by decongestion status, with the risk of death decreasing with decreasing severity of congestion. 376][7] There is an unmet need for in-hospital therapies for the treatment of ADHF to augment diuresis and achieve adequate decongestion, including in patients without cardiogenic shock.Therapeutic options may be new medical treatments, such as new diuretic strategies, 39 or novel devices, such as the Aortix (Procyrion) or preCARDIA (Abbott) devices, intended to improve decongestion. 40,41spitalization costs of ADHF patients, as demonstrated by the findings of the economic substudy, are substantial, with half of all patients requiring 10 or more days of inpatient care.While these findings are directionally in line with prior studies that reported costs per HF hospitalization in the range of $15 000-$20 000, 42,43 our data suggest care for ADHF patients with WRF might be somewhat more resource-intensive and costly than hospitalizations of less complex HF populations.Among other factors, this is evidenced by the study observed LOS of 10.0 days, which is more than twice as high as the reported stay of mean 5.1 days in Medicare's diagnosisrelated group 291, which captures HF patients with major comorbidities and complications.This also results in payments that will likely not cover the costs of care for the patients investigated in the current study.

| Limitations
This study has several limitations.There were only 12 sites in the United States and Australia and thus the results may not be generalizable to other regions.In addition, all the sites had dedicated advanced HF programs and their results may be better than in the general clinical community.Though the intention of the study was to enroll consecutively, this was not possible due to local site-specific informed consent requirements.Data analyzed were limited to the data collected as part of the study, and we relied on accuracy of data entry by study coordinators.Furthermore, associations with the outcome may be influenced by residual measured and unmeasured confounders.Our economic data considered only costs for hospital care, but not for associated physician fees.These might amount to an increase in ~10%-20% additional cost per case.

| CONCLUSIONS
In conclusion, ADHF with WRF and persistent congestion is common among patients hospitalized with ADHF and represents a high-risk and high-cost population for which there are inadequate treatment options.These HFrEF and HFpEF patients have a high in-hospital and 90-day follow-up clinical event rate.The findings from this study add to our understanding of the clinical progression of HF and demonstrate the unmet clinical need for novel treatment strategies in CRS.
This multicenter, prospective registry study enrolled participants from 12 clinical sites in the United States and Australia.All adult patients aged ≥21 years admitted with a primary diagnosis of ADHF, regardless of LVEF, were screened.Informed consent was obtained from participants who met enrollment criteria and agreed to participate.Participants were categorized into one of three groups, ADHF alone, ADHF + WRF, and ADHF + WRF + persistent congestion at 96-h post-admission.Data collection varied by group, with the ADHF + WRF + persistent congestion group having the most extensive data collection.Due to the observational nature of this study, informed consent requirements were based on the local requirements of each participating institution and varied in level.For example, some sites require that all participants provide written informed consent before any screening took place, while others only required participants to provide written informed consent if they were assigned to the full data collection group (ADHF + WRF + persistent congestion).The institutional review boards at each of the 12 study sites approved the protocol and consent forms.Participants were enrolled from April 2019 to March 2020.

A
total of 1210 participants were enrolled in this study of which there were 41 screen failures and 210 participants who could not be classified.Out of the remaining 959 participants, 378 (39%) had WRF and of that population 269 (71%) had persistent congestion.Of the 269 classified to ADHF + WRF + persistent congestion, 237 provided consent to participate in the full data collection (Supporting Information: Figure).
baseline demographics, vitals, and laboratory values, patients with HFrEF and HFpEF had similar outcomes.This study also adds to prior work by including an economic analysis showing that the cost for hospitalization for these high-risk patients is substantial, and that the cost of care remains high even after the first hospitalization.

F
I G U R E 1 Event-free survival after hospitalization until rehospitalization, left ventricular assist device implantation, heart transplantation, or death stratified by participants with heart failure with reduced ejection fraction (blue) and heart failure with preserved ejection fraction (red).HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction.F I G U R E 2 (A) The relationship between cost and length of stay (LOS) for the index hospitalization for acute decompensated heart failure.(B) Boxplot of index hospitalization cost.(C) Boxplot of LOS.
Baseline patient characteristics.
T A B L E 2 Findings of persistent congestion assessment at 96 h.
a Patient could have more than one finding of congestion.bSeventy-fivepatients (31.6%) had only one finding of persistent congestion, and 162 patients (68.4%) had two or more.T A B L E 3 Significant clinical events during initial ADHF hospitalization.extracorporeal membrane oxygenation; ICU, intensive care unit; LVAD, left ventricular assist device; MCS, mechanical circulatory support; PC, persistent congestion; WRF, worsening renal function.a Patient could have more than one significant clinical event.If the patient had more than one significant clinical event in one category, the patient is counted only once per category.b Consisted of, but not limited to, anemia, arrhythmia, catheterization/ cardiac surgery, and hypotension.