Biomarkers in Acute Heart Failure: Diagnosis, Prognosis, and Treatment

INTRODUCTION

Heart failure (HF) poses a substantial global public health problem. It is estimated there are over 64 million individuals living with HF worldwide at a substantial cost, morbidity and mortality.¹⁾ Episodes of acute heart failure (AHF) are especially vulnerable periods of increased morbidity and mortality, and there have been few therapeutic advances that improve outcomes.²⁾,³⁾,⁴⁾ Reasons for this lack of improvement in outcomes and care of AHF patients might be the inability to recognize the different pathophysiologic processes occurring during decompensation, failure to recognize when optimal fluid status is achieved, and difficulty in identifying which patients have a worse prognosis and need of more aggressive interventions. Biomarkers may play an important role in addressing these deficiencies and their routine incorporation into clinical practice may improve outcomes.

Biomarkers have become valuable tools to use in combination with other clinical information for the diagnosis, prognosis, and management of multiple cardiovascular diseases, especially HF. While many biomarkers are well-established for their use in cardiovascular disease and in chronic HF management, there are some specialized and nuanced uses specific to AHF. Furthermore, multiple novel biomarkers have recently shown promising results specifically in AHF that may potentially improve patient management. This review will focus on the use of biomarkers in AHF highlighting the use of established biomarkers and the potential of novel biomarkers to improve diagnosis, prognosis, and management.

NATRIURETIC PEPTIDES

Some of the earliest and most established biomarkers in HF, especially AHF, are the natriuretic peptides (NPs). Their discovery and clinical integration have demonstrated the substantial additive benefit biomarkers have with other clinical information to diagnose, risk-stratify and manage AHF patients. When referring to NPs, individuals are most often referring to B-type natriuretic peptide (BNP) and the N-terminal fragment of proBNP (NT-proBNP), though other NPs exist and can be measured such as atrial natriuretic peptide (ANP) and the mid-regional proANP (MR-proANP). BNP is produced and released primarily from the ventricles of the heart during periods of volume or pressure overload.⁵⁾ BNP is first produced as a 134-amino acid preproBNP peptide.⁵⁾ Subsequently, a 26 amino acid signaling peptide is cleaved to form the 108-amino acid proBNP molecule.⁵⁾ This is then cleaved into the inactive 76-amino acid NT-proBNP and biologically active 32-amino acid C-terminal portion BNP.⁵⁾ Once produced, BNP's half-life is approximately 20 minutes while NT-proBNP's half-life is 120 minutes.⁵⁾ BNP is cleared both by the type-C natriuretic peptide receptor and neutral endopeptidases, like neprilysin. This latter clearance mechanism is why BNP levels are believed to rise when the neprilysin inhibitor sacubitril is used, though there is debate as to how efficient neprilysin is at degrading BNP in humans and if assay differences influence BNP detection with sacubitril.⁶⁾,⁷⁾ Both NT-proBNP and BNP are partially cleared by the kidneys and it was thought that NT-proBNP clearance was more significantly impacted by reduced kidney function, but a well-designed mechanistic study showed BNP and NT-proBNP have equivalent kidney clearance.⁸⁾ There is no set conversion factor for equating BNP and NT-proBNP levels, thus it is preferable to only measure one form in management so serial measurements can be compared.

In 2002, the Breathing Not Properly trial demonstrated that the use of BNP substantially improved the ability to diagnose AHF in patients presenting with acute dyspnea and an unclear diagnosis of AHF.⁹⁾ Soon thereafter in 2005, the Pro-BNP Investigation of Dyspnea in the Emergency Department (PRIDE) study showed that NT-proBNP could also assist with diagnosing AHF in patients with acute dyspnea of an unknown cause.¹⁰⁾ And though the profile and management of AHF patients have changed in the more than 15 years since NPs were first described, the diagnostic utility of NPs has recently been reaffirmed in the modern era of AHF.¹¹⁾ These studies in addition to others have resulted in the 2017 American College of Cardiology (ACC)/American Heart Association (AHA) Heart Failure Guidelines giving a class I recommendation for measuring either BNP or NT-proBNP in patients with possible AHF when a diagnosis is in question based on history and exam findings.¹²⁾ In this situation, the major strength of NPs is their ability to rule out AHF as low levels have a high sensitivity for excluding a diagnosis of AHF.¹³⁾

However, there are many important caveats to remember when interpreting NP levels. Elevated NP levels are less specific for confirming a diagnosis of AHF unless they are substantially elevated, and both NPs have a ‘gray zone’ in which sensitivity and specificity are lower and the diagnosis of HF is less certain (Table 1).¹⁴⁾,¹⁵⁾ It is in this ‘gray zone’ that many other conditions can mimic AHF and cause a strain on the heart provoking an elevation in NPs (Table 2).¹⁶⁾ There are also certain AHF states and HF-like conditions where NPs may not be elevated (Table 3).¹⁶⁾ Thus, this gray zone is an area where additional biomarkers are needed to assist with diagnosis.

There are additional variables to consider when interpreting NP levels. NPs are higher in patients with heart failure with reduced ejection fraction (HFrEF) than patients with heart failure with preserved ejection fraction (HFpEF). This can sometimes make the diagnosis of HFpEF difficult or falsely rule out a diagnosis of HF because NP levels fall into the gray zone or could even be in the normal range with decompensated HFpEF.¹⁷⁾,¹⁸⁾ Other factors can alter NP levels causing levels to be higher including increasing age, declining kidney function and atrial fibrillation, or cause levels to be lower including obesity and black race.¹⁹⁾ Lastly, in a patient with established HF and chronically elevated NP levels, it can be difficult to determine if new or worsening shortness of breath, fatigue or edema is from HF decompensation or another condition such as pneumonia. In general, NPs levels that are substantially higher than previous levels, such as >50% higher than levels measured when the HF patient was compensated, are consistent with worsening HF status and volume overload.¹³⁾ While it is important to remember these caveats when interpreting NPs, their remarkable utility for assisting with the diagnosis of AHF in a patient where the diagnosis is still equivocal based on history and exam is unquestionable. It has now become standard of care to measure NPs in any patient presenting with acute dyspnea or other symptoms suggestive of AHF, in which the diagnosis is not certain based on history and exam, much as the electrocardiogram is essential to evaluating any patient with chest pain or equivalent symptoms to help diagnose an acute coronary syndrome (ACS).

Further research has shown that NP's usefulness in AHF extends beyond diagnosis and their assessment is instrumental in prognostication and potentially in managing AHF. In two large analyses, one of 48,629 patients from Acute Decompensated Heart Failure National Registry (ADHERE) and the other of 99,930 patients from the Get With The Guidelines-Heart Failure Registry (GWTH-HFR), have demonstrated unequivocally that higher admission values of NPs are associated with an increased risk of in-hospital mortality.²⁰⁾,²¹⁾ This point is pertinent to remember when interpreting an elevated NP not just as a strong indicator that a diagnosis of AHF is likely, but also that the more elevated the level, the higher the risk of death is for that patient. While the admission value is important for initial risk stratification, discharge values are even more valuable for long-term risk stratification. Multiple studies, including one of over 7,000 patients from Organized Program to Initiate Lifesaving Treatment in Hospitalized Patients with Heart Failure (OPTIMIZE-HF) registry, have demonstrated discharge NP levels are more prognostic for mortality and HF readmission after discharge than the admission NP level.²²⁾,²³⁾,²⁴⁾ These studies focused on BNP and evaluated different values as cut-points for risk, finding that the risk for mortality and HF readmission was lowest when BNP was less than 250–350 pg/mL at discharge. Fewer studies have evaluated NT-proBNP, but the results have been similar with discharge values better at identifying lower risk cohorts than admission values.²⁵⁾,²⁶⁾ Notably, these studies primarily looked at the percent change in NT-proBNP level from admission to discharge and found that a reduction of ≥30% in NT-proBNP from the admission level was associated with a lower risk of death and readmission. Similarly, a percent change (≥30% reduction) in BNP from admission to discharge can also identify a lower risk cohort; although, the absolute discharge BNP value appears to be more prognostic when compared to percent reduction.²²⁾,²³⁾,²⁷⁾

While most studies have explored either absolute cut-offs or percent reductions in NP levels at discharge, one study explored both absolute cut-offs and percent changes to try and determine which method was better for prognostication by analyzing NT-proBNP levels from 1,266 AHF from 7 different cohorts.²⁸⁾ While both absolute value cut-offs and percent changes from admission to discharge were prognostic, the authors noted that the prognostic impact of absolute cut-offs varied based on the admission value while the prognostic significance of percent changes were independent of the initial admission value, thus the authors concluded that percent change should be used.²⁸⁾ Of note, the authors only evaluated NT-proBNP in this study, and whether the same findings would be seen with BNP still needs to be determined. As critical as NPs are for assisting with diagnosis and initial risk stratification at admission, discharge values are even more important for determining long-term prognosis and potentially developing a future care plan based on an AHF patient's risk.

An ideal biomarker is one that can be used for diagnosis, prognosis, and management with biomarker levels changing in response to therapeutic interventions.¹⁶⁾ In AHF, NPs clearly assist with diagnosis and prognosis, and may be able to fulfill the final role of an ideal biomarker by guiding management. Biomarker guided therapy with NPs has primarily been evaluated in an outpatient setting with mixed results. Two prior meta-analyses have generally shown improved outcomes with reduced mortality and HF hospitalization using an NP guided strategy.²⁹⁾,³⁰⁾ To more definitively answer this question, The Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure (GUIDE-IT) trial was initiated in 2013 and reported in 2017.³¹⁾ This trial was to enroll 1,100 patients with chronic HF and randomize them to either usual care or an NT-proBNP guided arm with the goal of achieving an NT-proBNP ≤1,000 pg/mL. The trial was stopped early though after enrolling just under 900 patients as there was no difference between groups for the primary outcome of time to first HF hospitalization or cardiovascular mortality.³¹⁾ These results were thought to potentially be an end to NP-guided therapy; however, there have been concerns with the GUIDE-IT trial design that may make its findings not as definitive. First, many HF patients enrolled seemed to have advanced HF that was refractory to medical therapy. Second, patients were treated at advanced HF centers where practitioners already focused on aggressive titration of medical therapy, thus patients in the ‘usual care’ arm likely did not receive the usual care actually practiced outside of advanced HF centers. Lastly, most trial patients were unable to achieve target dosing of medical therapy, thus failing to achieve a main goal of the trial. A post-hoc analysis of GUIDE-IT though did show the potential benefits of NP guided therapy. This analysis showed that the patients who successfully reached an NT-proBNP ≤1,000 pg/mL, regardless of randomization arm, had a substantially improved survival and reduced time to first HF hospitalization compared to patients unable to reach an NT-proBNP ≤1,000 pg/mL.³²⁾ Furthermore, a step-wise increase in risk for the primary outcome was seen with higher NT-proBNP levels.³²⁾ Thus, these latter findings emphasize the importance of reducing NP levels with medical therapy in HF patients. While most studies for NP-guided therapy have been in an outpatient setting, a few studies have attempted to evaluate the role of NP-guided therapy in hospitalized patients with AHF.

In a non-randomized retrospective study of 186 AHF patients, therapy was titrated with bioimpedance and BNP levels to try and achieve a BNP <250 pg/mL at discharge.³³⁾ The investigators found that AHF patients unable to reach this goal, despite further adjustments in diuretics and neurohormonal therapy, had a greater than three-fold increased risk of death or HF readmission in the 6 months following discharge.³³⁾ In a randomized trial, 271 AHF patients were randomized at the time of clinical stability prior to discharge to usual care or NT-proBNP guided therapy with the goal of intensifying therapies in patients who still had an NT-proBNP ≥3,000 pg/mL.³⁴⁾ With this study design though, only 63 of the 137 AHF patients in the NT-proBNP guided therapy arm and 53 of the 134 subjects in the usual care arm had an NT-proBNP ≥3,000 pg/mL.³⁴⁾ No differences in outcomes were found between the NT-proBNP guided arm and usual care arm at 6 months; however, overall those subjects that did have NT-proBNP decrease from a pre-discharge value ≥3,000 pg/mL to <3,000 pg/mL at discharge had improved outcomes compared to those whose NT-proBNP did not decrease.³⁴⁾ In a subsequent randomized study, 405 AHF patients were similarly randomized to either usual care or an NT-proBNP guided strategy with the goal of achieving a ≥30% reduction in the discharge NT-proBNP level from the admission value.³⁵⁾ Patients were enrolled at admission, but not randomized until later in the hospitalization when clinical stability had been established. The primary analysis found no difference in all-cause mortality or HF hospitalization at 6-months. However, similar to the other randomized trial, because patients were enrolled at admission and only randomized when clinical stability was established, the majority of patients had already achieved a ≥30% reduction in NT-proBNP at the time of randomization resulting in a smaller ‘actionable’ patient population. Only 71 of the original 202 patients randomized to the biomarker guided therapy arm had a <30% reduction in NT-proBNP and thus could actually undergo further intensive modification of therapy to see if biomarker guidance improved outcomes. Because of this reduced power, the authors performed a post-hoc analysis between the patients in the usual care arm and the 71 patients that actually underwent NT-proBNP guided therapy. They found those patients who had already achieved a ≥30% reduction in NT-proBNP at time of randomization did the best, while those patients with a <30% reduction in NT-proBNP that subsequently achieved ≥30% reduction with NT-proBNP guided therapy had an intermediate outcome, and patients who were unable to obtain a 30% reduction had the worse outcome.³⁵⁾ The authors noted the difficulties with study design and highlighted that future studies should focus on high-risk patients with a <30% reduction in NT-proBNP at time of ‘stability’ to determine if NP-guided therapy can improve outcomes. These few studies do demonstrate some benefit in aiming to achieve a low BNP or NT-proBNP at discharge, but also partially temper the enthusiasm for using NPs to guide therapy in AHF because of the negative study results. Larger randomized studies focusing exclusively on high-risk patient populations are needed to determine if NPs can guide therapy in AHF. However, before these large studies can be appropriately performed, further research is needed to best define a high-risk patient population to study and determine what the optimal target endpoint is for NP-guide therapy: a percent change in NP levels or an absolute NP value at discharge.

A summary of how the diverse uses of NPs in AHF can be assimilated into a management pathway is displayed in Figure 1. NPs can significantly assist in confirming or excluding a diagnosis of AHF in a patient presenting with symptoms or signs suggestive of AHF but still an equivocal diagnosis after considering history and exam. Additionally, the degree of NP elevation can assist with prognosis. Recognizing that patients with higher levels of NPs are at an increased risk for in-hospital mortality might lead to triaging these patients to a higher level of care, such as the intensive care unit, consideration of invasive hemodynamic monitoring, and potentially more aggressive cardiac support such as mechanical circulatory support. During hospitalization, intermittent assessment of NPs might assist with determining whether an AHF patient's clinical status is improving or not. For example, when serum creatinine increases during an AHF hospitalization, the change in BNP from admission can assist in determining the prognostic importance of this creatinine change.³⁶⁾,³⁷⁾ With resolution of symptomatic AHF and as patients approach discharge, reassessment of NPs can assist with prognostication and guiding therapy. An inability to achieve a ≥30% reduction in NPs or an absolute BNP <250–350 pg/mL or NT-proBNP <1,000 pg/mL identifies a higher-risk patient cohort who might need to remain hospitalized for further optimization and have early follow-up after discharge for close monitoring. If these NP goals cannot be achieved despite further intensification of diuretic therapy, adjustment of neurohormonal therapy, and optimization of co-morbidities, then clinicians might need to consider advanced therapies or even possibly a goals of care discussion. It is from this foundation of NP use in AHF that the other biomarkers will be discussed for their benefits to further improve diagnosis, prognostication and management.

DIAGNOSIS OF ACUTE HEART FAILURE: THE ‘GRAY ZONE’

While NPs are powerful adjunctive tools for assisting with a diagnosis of AHF, there are still approximately 20% of patients that fall into the ‘gray zone’ where the diagnosis of AHF may remain uncertain. This is an area where additional biomarkers to assist with diagnosing or excluding AHF are needed. An incorrect diagnosis of AHF could lead to an adverse outcome, such as giving diuretics to a patient with an acute pulmonary embolism who is dependent on a higher right ventricular preload. Additionally, this is a zone that patients with HFpEF often fall into since they traditionally have lower levels of NPs from reduced wall stress and comorbidities, such as obesity. Only a few biomarkers though have been shown to assist with improving diagnosis of AHF in this gray zone with recent studies identifying a promising candidate.

PROGNOSIS IN ACUTE HEART FAILURE: IMPROVING RECOGNITION OF HIGH-RISK HEART FAILURE

Research in HF, whether it is novel biomarkers, medications, devices, or treatment pathways, often focuses on improving the outcomes of mortality and HF readmission. While it is important to reduce mortality and readmission whenever possible, it is equally important to identify those subjects that may have progressed in their disease course and have no further options to modify their disease process—that is, the end-stage HF or Stage D HF patient as defined by the ACC/AHA Heart Failure Guidelines.⁵¹⁾ An AHF hospitalization may be the sentinel event that identifies these patients and their disease trajectory might actually be changed through either evaluation and treatment with advanced therapies, such as heart transplant or left ventricular assist devices, or end-of-life discussions and initiation of hospice. Identifying high-risk AHF patients or Stage D HF patients while also predicting risk of death have classically been difficult. Biomarkers can provide a valuable tool for helping further risk stratify patients and identify those at the highest risk for future adverse outcomes.

BIOMARKERS FOR MANAGING ACUTE HEART FAILURE: IDENTIFYING CONGESTION

The majority of patients presenting with AHF present with symptoms and signs of volume overload or congestion leading to more than 90% of patients receiving diuretic therapy.⁸⁵⁾ Multiple studies have demonstrated the importance of achieving adequate decongestion, as residual congestion is associated with increased morbidity and mortality.³⁶⁾,⁸⁶⁾,⁸⁷⁾,⁸⁸⁾,⁸⁹⁾ When congestion is evaluated though, it should be evaluated in two different compartments: intravascular and within tissues.⁹⁰⁾,⁹¹⁾ High NP levels are associated with high intravascular pressures, and NP levels decrease in conjunction with reductions in intravascular congestion. Thus they serve as a useful surrogate to detect congestion in the intravascular compartment.⁹²⁾,⁹³⁾,⁹⁴⁾,⁹⁵⁾ As noted, at least a 30% reduction in NP levels from the admission value is recommended to reduce morbidity and mortality and this should be viewed as a biomarker goal for monitoring intravascular congestion.²²⁾,²⁵⁾,²⁸⁾,⁸⁹⁾,⁹⁰⁾ To evaluate tissue congestion though, two other biomarkers are showing substantial promise as surrogates for tissue congestion and might be used in conjunction with NPs to monitor response to therapy in AHF in the future.

NEW FRONTIERS IN CARDIOGENIC SHOCK: DIPEPTIDYL PEPTIDASE 3

The most severe presentation of AHF is in the form of cardiogenic shock, a state associated with a high mortality that has been difficult to manage despite developments in mechanical circulatory support and other therapies.¹⁰⁹⁾ There are multiple reasons for a failure to improve outcomes in patients with cardiogenic shock, but one potential reason is an inability to distinguish the severity and trajectory of shock. Dipeptidyl peptidase 3 (DPP3) is an intracellular protease that has been shown to have myocardial depressant properties when found in circulation (cDPP3).¹¹⁰⁾,¹¹¹⁾ Two studies have recently explored its prognostic utility in cardiogenic shock and notably may have identified it also as a target for pharmacotherapy.

In the Study Comparing the Efficacy and Tolerability of Epinephrine and Norepinephrine in Cardiogenic Shock (OptimaCC), cDPP3 was serially measured in 57 patients with cardiogenic shock after an acute myocardial infarction.¹¹⁰⁾ cDPP3 levels were found to be significantly higher at admission, 24 and 48 hours in patients with refractory cardiogenic shock (defined as major organ dysfunction and sustained hypotension despite treatment with high dose inotropes or intra-aortic balloon pump support) compared to those in non-refractory cardiogenic shock.¹¹⁰⁾ Additionally, higher admission cDPP3 levels (≥59.1 ng/mL) were associated with an increased risk of death at 90-days.¹¹⁰⁾ In another study, cDPP3 levels were evaluated in 174 cardiogenic shock patients in the CardShock Study.¹¹¹⁾ Similar to findings in OptimaCC, higher levels of cDPP3 on admission were associated with increased mortality at 90-days.¹¹¹⁾ When cDPP3 levels were serially assessed, patients with higher mortality had higher cDPP3 levels on the serial assessments.¹¹¹⁾ Notably, those patients with high cDPP3 on admission but with a subsequent reduction in cDPP3 levels at 24 hours had a lower mortality than patients whose cDPP3 levels remained elevated.¹¹¹⁾ Those with low cDPP3 levels at admission that increased at 24 hours had an increased mortality compared to those patients where cDPP3 levels remained low.¹¹¹⁾

The investigators further examined the physiologic effects of DPP3 by injecting DPP3 and administering procizumab, an antibody that binds and blocks physiologic effects of cDPP3, in a mouse model.¹¹¹⁾ Injection of DPP3 resulted in a decline in left ventricular function and an increased renal resistive index.¹¹¹⁾ In an AHF mouse model induced by isoproterenol infusion, cDPP3 levels increased with the development of AHF and injection of procizumab resulted an improvement in left ventricular systolic function and cardiac output.¹¹¹⁾ Thus, cDPP3 is not just a biomarker for severity in cardiogenic shock, but an actual mediator of shock that may be a target for therapy.

While the sample sizes are small, these two studies provide compelling evidence that cDPP3 might be a very useful biomarker in AHF and cardiogenic shock. Admission levels in severe AHF or cardiogenic shock might help guide therapeutic interventions in the future with higher levels prompting initiation of more aggressive hemodynamic support. Serial assessment might be used to again determine when hemodynamic support should be initiated or escalated and improve risk assessment for risk of death. Furthermore, measuring cDPP3 might lead to a therapeutic intervention by administering a cDPP3 neutralizing antibody such as procizumab. Future studies are certainly needed to confirm these findings and potentially identify a biomarker that is both prognostic and directly actionable upon.

CONCLUSIONS

Biomarkers have become invaluable adjunctive tests for the diagnosis, prognosis, and management of AHF. NPs alone have substantially improved the evaluation and management of AHF patients; however, given the complex pathophysiology of AHF, no one biomarker can adequately address diagnosis, prognosis, and management. Thus, the use of multiple biomarkers can improve each of these dimensions by applying a multifaceted pathophysiologic approach (Figure 5). Both well-established and novel biomarkers are finding increasing utility in these dimensions in AHF. This review has covered some of established and more promising biomarkers in AHF (Table 4); however, each year novel biomarkers continue to emerge. Improving clinicians' understanding of the abilities and uses of these different biomarkers will improve outcomes in AHF in the future.