Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Biomarkers and Precision Medicine in Heart Failure
Acknowledgments
Dr. Ibrahim is supported in part by the Dennis and Marilyn Barry Fellowship in Cardiology Research. Dr. Gaggin is supported in part by the Ruth and James Clark Fund for Cardiac Research Innovation. Dr. Januzzi is supported in part by the Hutter Family Professorship.
The underlying mechanisms in the development and progression of heart failure (HF) are complex and involve an intricate interplay of cardiac strain and injury, tissue inflammation, neurohumoral activation, oxidative stress, and ventricular remodeling, all compounded by the impact of comorbidities such as cardiorenal interactions and other medical conditions. Traditionally, diagnostic evaluation of patients suspected of having HF has involved history, physical examination, and chest x-ray ( see also Chapter 31 ). However, the complex diagnosis of HF by clinical presentation alone may be challenging because patients often present with signs and symptoms that are vague and nonspecific. In fact, isolated findings from history or physical examination correlate poorly with objective methods of cardiac function, and clinical criteria that combine relevant findings perform poorly in accurately diagnosing HF (sensitivity 50%–73% and specificity 54%–78%). This challenge is associated with delays in definitive diagnosis and treatment, increased health care expenditures, and ultimately with poor prognosis. Although noninvasive and invasive diagnostic studies complement the initial history and physical examination in the evaluation of the HF patient, such methods, including echocardiography and right heart catheterization, have limitations and the routine use of these methods are associated with significant cost and potential risk.
Given the complexity of HF biology—which includes processes that are not easily detectable with physical examination or imaging techniques—increased interest has been given to the use of biomarkers to supplement clinical judgment; many biomarkers are conveniently measured and easily interpretable and may support clinical evaluation, while at the same time reflecting important pathophysiology involved in HF presence, severity, and prognosis.
In addition, beyond the traditional clinical roles played by biomarkers in HF patients (e.g., diagnosis, prognosis estimation, and therapy monitoring), biomarkers are playing an increasing role in the development of therapies for diagnosis. It is likely that testing of biomarker signatures may inform more precise application of therapies.
Biomarkers: Definition and Guidelines for Evaluation
Although biomarkers may be measured from multiple sources, most frequently they are obtained from blood or urine samples; this chapter will focus on biomarkers derived from such sources. In general, a biomarker is typically a protein compound, quantifiable, easily available (at a reasonable cost and turnaround time), and indicative of the biologic process underlying the disease, providing a special insight into the microcosms surrounding the disease.
As articulated by Ibrahim and Januzzi, measurement of an HF biomarker should be easily achieved through the use of assays with acceptable analytical precision and provide accurate results with well-defined biologic variation. The biomarker candidate should primarily reflect important processes involved in HF pathophysiology and should not recapitulate clinical information already available at the bedside. The study of HF biomarkers should be appropriate to the clinical use being evaluated, and design should be rigorous; new assays should be evaluated across a wide range of HF patients, and the statistical methods used to evaluate the biomarker (relative to clinical variables and other biomarkers) should be contemporary and robust.
Major Society Guidelines
Several major societies have set forth clinical practice guidelines for the diagnosis, prognosis, and treatment of HF, including the recent update by the American Heart Association (AHA)/American College of Cardiology (ACC)/Heart Failure Society of America (HFSA) writing group on HF. The use of B-type natriuretic peptide (BNP) and N-terminal B-type natriuretic peptide (NT-proBNP) is now recommended in the routine evaluation of HF for the purposes of diagnosis and in determining prognosis, but other novel applications are now supported.
The most established applications for BNP and NT-proBNP, as noted, are for establishment of diagnosis and estimation of prognosis in HF. The most recent ACC/AHA/HFSA guideline update ( Table 33.1 ) has given natriuretic peptide testing a class I (level of evidence A) for both applications. In patients with acute HF syndromes, the guidelines also put a new emphasis on use of natriuretic peptide testing to estimate risk for adverse events after discharge (class IIa, level of evidence B). As well, the guidelines also added use of BNP or NT-proBNP to identify community-based patients as at risk for new-onset HF (class IIa, level of evidence B–R). As will be discussed later, clinical practice guideline updates have also considered the role of more novel biomarkers; troponin testing for risk assessment was given a class I (level of evidence A), whereas both soluble (s)ST2 and galectin-3 (biomarkers of myocardial fibrosis) were given a class II recommendation.
TABLE 33.1
2017 American College of Cardiology/American Heart Association Heart Failure Guideline Recommendations for the Use of Biomarkers in Heart Failure
| Biomarkers | Class of Recommendation | Level of Evidence | |
|---|---|---|---|
| BNP or NT-proBNP | Diagnosis | I | A |
| Hospital admission prognosis | I | A | |
| Prevention | IIa | B | |
| Hospital discharge prognosis | IIa | B | |
| Guided-therapy (chronic HF) | IIb | B | |
| Troponin T or I (myocardial injury) | Hospital admission prognosis | I | A |
| Soluble ST2, Galectin-3 (myocardial fibrosis) | Prognosis (chronic HF) | IIb | B |
BNP, B-type natriuretic peptide; HF, heart failure; NT-proBNP, N-terminal pro–brain natriuretic peptide.
Heart Failure Biomarkers
A vast number of biomarkers in HF have been examined to date (examples are shown in Table 33.2 ). No single, standardized categorization of HF biomarkers exists, although several have been proposed. Braunwald suggested that markers may be considered in the following categories: (1) myocardial stretch, (2) myocyte injury, (3) extracellular matrix remodeling, (4) inflammation, (5) renal dysfunction, (6) neurohumoral activation, and (7) oxidative stress ( Fig. 33.1 ). Clearly, significant overlaps among these various processes exist, which may allow further simplification of biomarker categories to (1) myocardial insult, (2) neurohormonal activation, (3) myocardial remodeling, and (4) markers of comorbidity.
TABLE 33.2
Biomarkers of Heart Failure
| Myocardial Insult |
|
| Neurohormonal Activation |
|
| Remodeling |
|
| Markers of Comorbidity |
|
BNP, B-type natriuretic peptide; BUN, blood urea nitrogen; CKMB, creatinine kinase, myocardial bound; eGFR, estimated glomerular filtration rate; GDF, growth differentiation factor; KIM, kidney injury molecule; MMP, matrix metalloproteinases; MR-proANP, midregional propeptide assay for atrial natriuretic peptide; MR-proADM, midregional proadrenomedullin; NAG , N -acetyl β-( d )-glucosaminidase; NGAL, neutrophil-gelatinase associated lipocalin; NT-proBNP, N-terminal pro–brain natriuretic peptide; PINP, procollagen type I amino-terminal; PIIINP, procollagen type III amino-terminal; RDW, red blood cell distribution width; TNF, tumor necrosis factor; VEGFR, vascular endothelial growth factor receptor.
Myocardial Insult
Various myocardial insults are among the initiating event in the cascade of changes that occur in HF. Within the context of myocardial insult are subcategories of biomarkers that reflect myocardial stretch, myocyte necrosis, and oxidative stress.
Myocardial Stretch
Dense granules were noted in tissue derived from the atria of the heart by electron microscopy in the 1950s. Furthermore, stretching of the canine left atrium was shown to increase urine output, and injection of atrial homogenates into rats caused diuresis and natriuresis. Atrial natriuretic peptide (ANP) was subsequently purified, sequenced, and reproduced. In the 1980s, a homologous peptide with similar biologic activity was discovered in porcine brain and named BNP. Soon other natriuretic peptides, sharing similar structural features, were discovered: urodilatin (a renally active isoform of ANP), C-type natriuretic peptide, and Dendroaspis natriuretic peptide, the latter found in snake venom. Although each of the natriuretic peptide family of markers is of potential use in the evaluation and management of the patient with HF, BNP and NT-proBNP are most studied, and ANP has been increasingly examined.
BNP and NT-proBNP
Left ventricular (LV) wall stretch from increased pressure or volume is the most potent inducer of BNP gene transcription. One of the early products of the BNP gene is a 108–amino acid peptide, proBNP 1-108 , which is subsequently cleaved into the biologically active 32–amino acid peptide, BNP, and a biologically inert 76 amino acid peptide, NT-proBNP. Both BNP and NT-proBNP are released into the blood stream within minutes of their synthesis. In addition, varying amounts of uncleaved proBNP 1-108 are released. Although the cause of this observation is not understood, it is now known that such concentrations of proBNP 1-108 are produced in increasing degrees in more advanced HF. Furthermore, the conventional assays for BNP or NT-proBNP measurement cross-react with circulating proBNP 1-108 , which means that the overall measured natriuretic peptide value in patients evaluated clinically contains a mixture of cleaved and uncleaved peptide.
BNP binds to membrane-bound natriuretic peptide receptors (NPRs) type A and B, activating intracellular cyclic guanosine monophosphate (cGMP), beginning a cascade of events leading to natriuresis, diuresis, vasodilation, inhibition of renin and aldosterone, and inhibition of fibrosis. Besides being removed by receptor-mediated mechanisms (including by the NPR C), BNP is also degraded by various enzymatic processes, including neprilysin, meprin-A, and dipeptidyl peptidase-IV. In addition, BNP is cleared passively by numerous organs, such as the kidneys, with high blood flow. Due to the multitude of means by which BNP is removed from circulation, the half-life of BNP is approximately 20 minutes. In contrast, NT-proBNP is only passively cleared by multiple organs, including the kidneys.
A major misconception about clearance of BNP and NT-proBNP regards to the degree of dependence on renal function for their removal from the circulation. Mechanistic studies actually suggest the degree of renal clearance to be identical for both BNP and NT-proBNP, with approximately 25% of both being cleared by renal mechanisms, down to an estimated glomerular filtration rate (eGFR) of less than 15 mL/min/1.73 m 2 .
Extensive studies have established BNP and NT-proBNP as the “gold standard” biomarkers for the diagnosis and prognostication of HF, and emerging data suggest value for their use in the management of patients with HF.
Diagnosis
In healthy adults, circulating concentrations of BNP are quite low and women tend to have slightly higher values than men (14 vs. 8 pg/mL), with similar findings observed with NT-proBNP (<300 pg/mL). Other conditions that affect concentrations of natriuretic peptides beyond cardiac structure and function include factors that may lead to higher values (advancing age, renal dysfunction), as well as those that may lead to lower than expected values (obesity). These factors are discussed later ( Table 33.3 and 33.4 ). In the context of pressure or volume overload states such as acutely decompensated HF (ADHF), BNP or NT-proBNP concentrations typically dramatically increase and sex-dependent differences becomes less relevant.
TABLE 33.3
Factors Influencing the Clinical Interpretation of BNP or NT-proBNP Values
| Factors That Decrease BNP or NT-proBNP |
|---|
|
| Factors That Increase BNP or NT-proBNP |
Left ventricular dysfunction
|
| Previous heart failure |
Arrhythmia
|
| Acute coronary syndromes |
Cardiotoxic drugs
|
Significant pulmonary disease
|
| Advanced age |
| Renal dysfunction |
| Anemia |
Critical illness
|
High output states
|
BNP, B-type natriuretic peptide; LV , left ventricular; NT-proBNP, N-terminal pro–brain natriuretic peptide.
When considering the interpretation of BNP or NT-proBNP, numerous data exist currently to suggest that their concentrations reflect multiple aspects of physiology. Thus, although wall stress is a prime factor responsible for their release, a broad range of cardiac structural and functional abnormalities may trigger elevation in BNP or NT-proBNP (see Table 33.4 ). Appropriate interpretation of these natriuretic values in the context of each patient is crucial. Table 33.5 details the recommended cutoff points for natriuretic peptide testing in HF.
TABLE 33.4
Cardiac Abnormalities Associated With Increased Natriuretic Peptide Concentrations
| Myocardial dysfunction |
|
| Valvular abnormalities |
|
| Cardiac chamber size |
|
| Filling pressures |
|
| Ischemic heart disease |
|
| Heart rhythm abnormalities |
|
| Pericardial diseases |
|
| Congenital abnormalities |
|
TABLE 33.5
Suggested Natriuretic Peptide Cut-Points in Heart Failure
| Cutoff Value | Sensitivity | Specificity | Positive Predictive Value | Negative Predictive Value | |
|---|---|---|---|---|---|
| To Exclude Acutely Decompensated HF: | |||||
| BNP NT-proBNP MR-proANP |
<30–50 pg/mL <300 pg/mL <57 pmol/L |
97% 99% 98% |
– – – |
– – – |
96% 99% 97% |
| To Identify Acutely Decompensated HF: | |||||
| Single cutoff point strategy | |||||
| BNP NT-proBNP MR-proANP |
<100 pg/mL <900 pg/mL <127 pmol/L |
90% 90% 87% |
76% 85% 79% |
79% 76% 67% |
89% 94% 93% |
| Multiple cut-point strategy | |||||
| BNP, “gray zone” approach NT-proBNP, “age-stratified” approach MR-proANP, “age-stratified” approach |
<100 pg/mL to exclude 100–400 pg/mL, “gray zone” >400 pg/mL, to rule in <450 pg/mL for age <50 years <900 pg/mL for age 50–75 years <1800 pg/mL for age >75 years <104 pmol/L for age <65 years 214 pmol/L for age ≥65 years |
90% – 63% 90% 82% |
73% – 91% 84% 86% |
75% – 86% 88% 75% |
90% – 74% 66% 91% |
| Outpatient Application | |||||
| BNP NT-proBNP, “age stratified” approaches MR-proANP |
20 pg/mL (asymptomatic) or 40 pg/mL (symptomatic) <125 pg/mL for age <75 years <450 pg/mL for age ≥75 years or <50 pg/mL for age <50 years <75 pg/mL for age 50–75 years <250 pg/mL for age >75 years Unknown |
– – – – – – – Unknown |
– – – – – – – Unknown |
– – – – – – – Unknown |
96% 98% 91% 98% 98% 93% Unknown |
BNP, B-type natriuretic peptide; HF, heart failure; MR-proANP, midregional propeptide assay for atrial natriuretic peptide; NT-proBNP, N-terminal pro–brain natriuretic peptide.
BNP and NT-proBNP: ADHF diagnosis
In 1586 patients presenting to the emergency department with acute dyspnea, the Breathing Not Properly Study Multinational Study showed that patients diagnosed with ADHF had higher BNP levels compared with those without HF (mean, 675 ± 450 vs. 110 ± 225 pg/mL; P < .001). Furthermore, increasing concentration of BNP was associated with increasing severity of HF as evidenced by New York Heart Association (NYHA) functional class ( P < .001). In a multivariable logistic regression analysis, a BNP greater than 100 pg/mL was the single most accurate predictor of the diagnosis of ADHF than any other single findings from history, physical examination, chest x-ray, or laboratory tests. A BNP cutoff value of 100 pg/mL had overall sensitivity of 90%, specificity of 76%, and accuracy of 85%, performing better than either the NHANES (National Health and Nutrition Examination Survey) criteria (accuracy, 67%) or the Framingham criteria for the diagnosis of HF (accuracy, 73%). BNP added independent and additive information in the diagnosis of ADHF when added to the traditional evaluation of patients with HF. The performance of BNP in the diagnosis of HF can be summarized by its receiver operating characteristic (ROC) curve area under the curve (AUC) of 0.91 (95% confidence interval [CI] 0.90–0.93; P < .001).
NT-proBNP has also been shown to be useful to aid in the diagnosis of ADHF. In the ProBNP Investigation of Dyspnea in the Emergency Department (PRIDE) study, patients with ADHF had much higher NT-proBNP values compared with patients without HF (median, 4054 [interquartile range 1675–10028] vs. 131 [interquartile range 46–433] pg/mL; P < .001), increasing NT-proBNP correlated well with increasing severity of HF ( P = .001) and was the strongest predictor of ADHF diagnosis, compared with any other single traditional findings. Despite the excellent diagnostic performance of NT-proBNP (AUC, 0.94) compared with clinical judgment alone (AUC, 0.90), the best approach in the evaluation of patients suspected of HF is a combination of both the NT-proBNP and clinical judgment (AUC, 0.96), as shown in Fig. 33.2 . Although an NT-proBNP cutoff value of 900 pg/mL provided identical performance to that reported for a BNP of 100 pg/mL in the Breathing Not Properly Multinational Study, the International Collaborative on NT-proBNP (ICON) investigators reported that age stratification of NT-proBNP reference limits (≥450, ≥900, and ≥1800 pg/mL for ages <50, 50–75, and >75 years) improved performance even further. Importantly, an NT-proBNP threshold value of less than 300 pg/mL was found to exclude ADHF with high negative predictive value. The ICON Re-evaluation of Acute Diagnostic Cut-Offs in the Emergency Department (ICON-RELOADED) Study has affirmed the value of NT-proBNP for diagnostic evaluation of ADHF in a contemporary population.
Important insights from these studies include the fact that BNP and NT-proBNP were superior to radiographic standards for HF diagnosis, and both may also identify unsuspected HF in patients with underlying lung disease. Although renal function may impair the diagnostic performance of both BNP and NT-proBNP (as discussed later), with careful adjustment of reference limits, careful analyses suggest both retain utility for evaluation of patients with acute dyspnea.
As 50% of modern HF is composed of patients with preserved left ventricular ejection fraction (LVEF), it is helpful to understand the performance of BNP and NT-proBNP in those with heart failure and preserved ejection fraction (HFpEF) versus those with heart failure and reduced ejection fraction (HFrEF). Given the generally smaller LV chamber size in HFpEF, there tends to be less wall stress in such patients; accordingly, natriuretic peptide concentrations are typically lower in those with HFpEF; however, lower BNP or NT-proBNP concentrations are not pathognomonic for HFpEF by any means. The same cutoff values for BNP and NT-proBNP are recommended for the diagnosis of HF in patients with HFpEF as well as HFrEF, with the recognition that the sensitivity may be reduced in those with preserved LV function. In this setting, values of BNP or NT-proBNP are rarely normal, but if “low,” they are more likely in a range between the “rule-out” and “rule-in” thresholds (the so-called gray zone; see later). It is important to know the normal or baseline values of BNP or NT-proBNP in comparison for a given patient.
Relative to the added value of natriuretic peptide testing when put together with standard clinical evaluation, the B-type natriuretic peptide for Acute Shortness of breath EvaLuation (BASEL), the Improved Management of Patients With Congestive Heart Failure (IMPROVE-CHF), and the NT-proBNP for EValuation of dyspneic patients in the Emergency Room and hospital (BNP4EVER) studies all indicated an advantage to use of either BNP or NT-proBNP when added to clinical judgment. For example, in both BASEL and IMPROVE-CHF, the use of BNP and NT-proBNP, respectively, led to considerable cost savings, findings echoed by Siebert and colleagues in a report from the PRIDE study. In BASEL, the use of BNP was associated with less use of intensive care unit admission, without excess hazard with such care. In the IMPROVE-CHF study, not only was NT-proBNP–supplemented evaluation superior diagnostically, but patients in the NT-proBNP arm also had better short-term outcomes.
When considering the appropriate application of BNP or NT-proBNP for diagnostic evaluation of ADHF, it is worth reviewing the circumstances where testing is most valuable. As shown by Green and colleagues, indecision when evaluating patients with acute dyspnea occurs in approximately 30% of cases seen in the emergency department and is associated with considerably higher short-term risk. Results by Steinhart and colleagues from the IMPROVE-CHF study lend important clarity regarding the importance of natriuretic peptide testing in this setting; although NT-proBNP correctly (and significantly) reclassified diagnoses in patients judged with confidence in this study, the value of the biomarker was considerably greater in those with uncertain diagnosis. These findings inform the current class I guideline recommendations for use of BNP or NT-proBNP for ADHF diagnosis.
BNP and NT-proBNP: chronic HF diagnosis
Not surprisingly, both BNP and NT-proBNP have been shown to be useful in the diagnosis of HF in the outpatient setting. However, in contrast to the diagnostic application of BNP or NT-proBNP in the acute environment, both peptides have been mainly examined relative to their negative predictive value to exclude the diagnosis, rather than to confirm it. In this regard, the optimal reference limits for use in this setting are considerably lower than in patients with acute dyspnea (see Table 33.5 ). For NT-proBNP, the ICON-Primary Care group showed that age stratification again improves diagnostic accuracy in this setting. If a patient is found to be greater than the BNP or NT-proBNP cutoffs, further diagnostic testing such as echocardiography is likely needed. Causes of falsely low BNP or NT-proBNP in the outpatient setting are comparable to those with acute dyspnea.
Another potential use of BNP or NT-proBNP in the nonacute setting is for screening at-risk patients for the presence of underlying structural heart disease. Although influenced by numerous cardiac correlates (see Table 33.4 ), a single measurement of BNP or NT-proBNP may be able to identify reduced LV function in asymptomatic individuals; in addition, in recognition of their dependence on diastolic indices, as well for their concentrations, both peptides may be useful also in screening for diastolic ventricular dysfunction.
Natriuretic peptides: caveats
Despite the promising information conveyed by BNP or NT-proBNP in the evaluation of HF patients, there are several important limitations that need to be considered in the interpretation of their results; several factors have been shown to alternately lead to higher than expected BNP or NT-proBNP values, as well as lower than expected values (see Table 33.3 ). Beyond age, as discussed previously, a number of diagnoses have been associated with increased natriuretic peptide levels, as detailed in Table 33.3 . Clinical judgment when interpreting natriuretic peptide concentrations is crucial: as with any diagnostic test, a differential diagnosis should be kept in mind when interpreting an elevated BNP or NT-proBNP value.
Although discussed previously, the importance of renal function in the interpretation of BNP or NT-proBNP deserves more detail. Patients with chronic kidney disease typically have higher BNP and NT-proBNP values. Mechanistically, this is due both to peptide accumulation, as well as increased release of BNP or NT-proBNP due to shared comorbidities such as hypertension, LV hypertrophy, and chronic volume overload. In patients with renal insufficiency (GFR <60 mL/min/1.73 m 2 ), a BNP cutoff of 200 pg/mL or NT-proBNP of 1200 pg/mL provides good diagnostic performance; alternatively, the age-stratified NT-proBNP cutoffs may be used with good accuracy and without adjustment. Furthermore, both BNP and NT-proBNP are considerably prognostic in patients with chronic kidney disease, even in the absence of overt cardiovascular disease; NT-proBNP may even predict progression of renal dysfunction in population studies.
Notably, BNP and NT-proBNP levels are lower in overweight and obese patients. This phenomenon is hypothesized to be due to suppression of synthesis or release of natriuretic peptides in obese subjects. However, regardless of body mass index (BMI), BNP, or NT-proBNP concentrations are typically higher in patients with HF compared with patients without, and age-adjusted cutoff points retain usefulness for the diagnosis of acute HF. In the PRIDE study of acute dyspnea patients, ROC analyses of NT-proBNP for diagnosis of acute HF had AUC of 0.94 for lean, 0.95 for overweight, and 0.94 for obese patients. An NT-proBNP value less than 300 pg/mL still had excellent diagnostic performance in ruling out acute HF across all BMI categories.
When a patient has a BNP or NT-proBNP between the optimized cutoff to exclude HF and the optimized cutoff to diagnose it, this is referred to as a “gray zone” result. van Kimmenade and colleagues reported that patients with a gray zone NT-proBNP in ICON were more likely to have ADHF when physical findings, such as elevated jugular venous pressure or pulmonary rales, were present. In addition, although complicating biomarker-based evaluation of the patient with suspected HF, gray zone values have prognostic meaning, above that of patients with lower concentrations of BNP or NT-proBNP. This informs another strength of BNP and NT-proBNP: the estimation of prognosis across the spectrum of HF.
An important development has been approval of a novel class of HF therapy, the angiotensin receptor/neprilysin inhibitor (ARNI; e.g., sacubitril/valsartan); this drug was recently embedded as a class I, level of evidence (LOE) B-R therapy into HF treatment guidelines ( see also Chapter 37 ). The mechanism of action of this class of drug includes blockade of the angiotensin II receptor together with neprilysin. The latter is a ubiquitous zinc-dependent metalloproteinase involved in degradation of numerous vasoactive peptides involved in cardiovascular regulation, including biologically active natriuretic peptides. Thus concentrations of ANP, BNP, and C-type natriuretic peptide (CNP) are all thought to be affected by inhibition of neprilysin; in the context of therapy with sacubitril/valsartan, concentrations of BNP tend to rise. In contrast, NT-proBNP is not a substrate for neprilysin, so its concentrations tend to fall in the setting of neprilysin inhibition.
In the phase 2 Prospective Comparison of ARNI with ARB on Management of HFpEF (PARAMOUNT) trial, 301 patients with chronic HFpEF, NYHA class II–III symptoms, and elevated natriuretic peptide concentrations were treated with sacubitril/valsartan versus valsartan alone. At 12 weeks, ARNI therapy reduced NT-proBNP and increased BNP concentrations compared with valsartan alone. More recently, in the Angiotensin-Neprilysin Inhibition versus Enalapril in Heart Failure (PARADIGM-HF) trial, measurement of BNP and NT-proBNP concentrations in patients with HFrEF were made at baseline, 4 weeks, and 8 months. A significant increase in measured BNP concentrations was seen at 4 weeks after treatment with sacubitril/valsartan, and at 8 months patients treated with neprilysin inhibition still had higher concentrations of BNP when compared with those treated with enalapril. Conversely, those treated with sacubitril/valsartan had early and sustained reduction in NT-proBNP across the duration of the study.
A post hoc analysis of 2080 patients in the PARADIGM-HF demonstrated that higher posttreatment NT-proBNP concentrations predicted outcomes such as cardiovascular death or HF hospitalization. Those HFrEF patients with significant posttreatment NT-proBNP reductions had lower subsequent rates of such adverse outcomes, independent of whether the patients were treated with angiotensin-converting enzyme inhibition (i.e., enalapril) or with neprilysin inhibition (i.e., sacubitril/valsartan).
Uncertainties abound with respect to how neprilysin inhibits BNP, including vagaries about how much “rise” is expected from treatment alone, whether the peptide remains prognostic in those taking sacubitril/valsartan, and how durable the effect of the “rise” in BNP may be over time. In addition, to the extent the numerous available BNP immunoassays are based on different antibody pairs, it is not even certain all assays are affected similarly by neprilysin inhibition; given that neprilysin degrades the BNP in multiple sites ( Fig. 33.3 ), it is reasonable to suspect differences will be modest, but this is presently unknown.
Given these uncertainties, although data are limited, it is fair to suggest NT-proBNP as the preferred biomarker when measuring natriuretic peptides in patients taking sacubitril/valsartan.
BNP and NT-proBNP: prognosis
Concentrations of BNP or NT-proBNP are strong predictors of future clinical outcomes in a variety of populations spanning all stages of HF articulated by the ACC/AHA guidelines: from patients without any cardiac dysfunction but at high risk of developing HF in the future (stage A), through patients with asymptomatic LV dysfunction (stage B), to symptomatic HF (stage C) and advanced HF (stage D).
The largest body of data supporting the use of BNP and NT-proBNP is in patients with ADHF. The Acute Decompensated Heart Failure National Registry (ADHERE) showed that higher admission BNP values were associated with increased in-hospital mortality among 48,629 patients admitted to the hospital with ADHF. Moreover, there was a linear relationship between increasing quartile of BNP and in-hospital mortality, even after adjusting for potential confounders such as age, gender, systolic blood pressure, pulse, renal function, sodium, and dyspnea in both HFpEF and HFrEF patients. In a similar manner, admission NT-proBNP concentrations were found to be strongly predictive of both short- and long-term clinical outcomes. NT-proBNP values greater than 986 pg/mL predicted death at 1-year ( P < .001, 79% sensitivity and 68% specificity), even after adjusting for relevant traditional factors. Some studies have examined the role of follow-up BNP or NT-proBNP measurement compared with admission. Indeed, discharge natriuretic peptides obtained after inpatient treatment for ADHF appeared to be even more predictive of future mortality and/or rehospitalization when compared with admission values. This application for hospital discharge risk assessment is discussed further, later.
In chronic HF, serial assessment with BNP and NT-proBNP for prognosis has been shown to be of value. In a large population of ambulatory patients with chronic stable HF enrolled in the Valsartan Heart Failure Trial (Val-HeFT) study, NT-proBNP was measured at baseline and at 4 months; although baseline values were prognostically important, changes in NT-proBNP concentrations over 4 months and relationship to all-cause mortality were more important in providing prognostic information. Thus, in analogy to those patients with acute symptoms, serial assessment of BNP or NT-proBNP appears to provide incremental data regarding likelihood for adverse outcome. Mechanistically, elevated values of NT-proBNP in chronic HF not only predict adverse outcome but also identify patients at highest risk for deleterious LV remodeling.
An emerging application for BNP or NT-proBNP is to predict cardiovascular risk in apparently unaffected patients in the population. In 3346 asymptomatic subjects without HF from the Framingham Offspring Study, baseline elevated concentrations of BNP or NT-proBNP strongly predicted future clinical outcomes, including all-cause mortality, first major cardiovascular event, HF, atrial fibrillation, and stroke or transient ischemia attack; most notably, BNP levels greater than 80th percentile (20.0 for men and 23.3 pg/mL for women) were associated with a 62% increase in the risk of death, 76% increase in first cardiovascular event, and 307% increase in the risk of HF (all P < .05). Similar results were found with NT-proBNP, where values greater than 80th percentile (497 for men and 541 pg/mL for women) were associated with a 76% increase in the risk of death, 52% increase in first cardiovascular event, and 502% increase in the risk of HF (all P < .05). Much as in acute and chronic coronary syndromes, serial measurement of natriuretic peptides in the community may inform risk for future HF better than a single assessment. For example, among ambulatory elderly patients without prevalent HF at baseline, DeFilippi and colleagues demonstrated that a baseline NT-proBNP greater than 190 pg/mL predicted incident HF during follow-up; however, a second measurement added considerable prognostic information; those with rising concentrations of NT-proBNP had substantial risk for incident HF (adjusted hazard ratio [HR] 2.13; 95% CI 1.68–2.71) and cardiovascular death (HR 1.91; 95% CI 1.43–2.53) compared with those with sustained low concentrations.
BNP and NT-proBNP: management of HF
Coupled with their dynamic and prognostically meaningful behavior in the context of HF treatment is the interesting observation that natriuretic peptide concentrations appear to fall in the context of treatment with therapies shown to improve long-term mortality in HF, including beta-blockers, angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, and mineralocorticoid receptor antagonists, as well as cardiac resynchronization therapy. Significant reduction in BNP or NT-proBNP concentrations typically occur within 2 to 4 weeks of successful therapy titration, allowing for a defined window for resampling and assessment for benefit of therapy. The obvious exception to this rule is if patients are treated with an ARNI, wherein a rise in BNP is expected, with reduction in NT-proBNP.
The concept of using BNP or NT-proBNP as a “guide” to HF care, with a goal of achieving excellent guideline-derived medical therapy plus natriuretic peptide suppression, has been the subject of several clinical trials.
Following the seminal pilot study by Troughton and colleagues from the Christchurch Cardioendocrine Group, numerous studies have explored the topic of BNP- or NT-proBNP–guided HF care, with mixed outcomes. The largest and most recent of these studies was the Guiding Evidence Based Therapy Using Biomarker Intensified Treatment in Heart Failure (GUIDE-IT) trial of 894 high-risk patients with chronic HFrEF. This study did not find NT-proBNP–guided therapy more effective than usual care in improving outcomes including HF hospitalization and cardiovascular mortality.
Caveats exist about the result of GUIDE-IT. Aggregate experience would dictate that for natriuretic peptide–guided HF care to be successful, a low target value must be sought (BNP <100 pg/mL; NT-proBNP <1000 pg/mL), therapies must be titrated to lower the natriuretic peptide concentrations, and significant lowering of the biomarkers must occur. In those trials that had these three characteristics, substantial improvement in outcomes was observed. In GUIDE-IT, no differences were seen with respect to guideline-directed medical therapy (GDMT) and no differences in achieved NT-proBNP concentration: 46% of participants in the biomarker-guided arm and 40% of the usual care group achieved an NT-proBNP less than 1000 pg/L at 12 months ( P = .21). To achieve this substantial NT-proBNP lowering in the “usual care” arm, patients were seen nearly monthly, on average. In addition, most of the GUIDE-IT study investigators practiced at academic tertiary care referral centers. Thus it remains uncertain if the “usual care” in GUIDE-IT was a fair representation of usual HFrEF care in nonacademic centers.
Meta-analyses combining findings from existing studies have shown a 20% to 30% mortality reduction associated with biomarker-guided HF management over standard HF care, nonetheless suggesting merit of this approach. The GUIDE-IT study suggests aggressively managed, frequently seen patients may not profit from this approach; future studies focusing the intervention on patients treated in a more typical fashion are needed.
Biomarkers have also been used to identify at-risk patients who may benefit from aggressive titration of guideline-directed medical therapies. Huelsmann and colleagues examined 300 patients with type 2 diabetes and an elevated NT-proBNP (<125 pg/mL) but free of cardiac disease and determined that those randomized to the intensified group who had uptitration of renin-angiotensin system antagonists and beta-blockers in a cardiologist’s office had significant reduction in the primary endpoint of hospitalization/death due to cardiac disease at 2 years compared with the control group (HR 0.351; 95% CI 0.127–0.975; P = .044). The same was true for other end points: all-cause hospitalization and unplanned cardiovascular hospitalizations/death ( P < .05 for all).
Emerging natriuretic peptide tests: MR-proANP
In parallel to the B-type peptides, circulating levels of ANP rapidly increase with cardiac stretch; unlike BNP, whose production is induced after myocyte stretch, ANP is premade and stored in the myocardium, predominantly in the atrium. However, reliable detection of circulating ANP is challenging because its half-life is only 2 to 5 minutes due to effects of neprilysin. However, its immediate precursor protein, proANP, is stable and has a longer half-life, which makes serum measurement possible. The development of a midregional propeptide assay for atrial natriuretic peptide (MR-proANP) assay has led to the examination of its use for HF applications. Table 33.5 details suggested cutoff points for MR-proANP for clinical use.
The role of MR-proANP in the diagnosis of ADHF was first examined in 1641 patients with acute dyspnea in the Biomarkers in Acute Heart Failure (BACH) trial. MR-proANP performed well in diagnosing ADHF and was noninferior to BNP or NT-proBNP; a MR-proANP cutoff point of greater than or equal to 120 pmol/L had a sensitivity of 97% and specificity of 60% with accuracy of 74%, whereas BNP with a cutoff point of 100 pg/mL had a sensitivity of 96%, specificity of 62%, and accuracy of 73%. In the PRIDE study analysis of MR-proANP, NT-proBNP performed slightly better than MR-proANP in the diagnosis of ADHF (AUC of 0.94 for NT-proBNP vs. 0.90 for MR-proBNP; P = .001 for difference); however, MR-proANP was found to be an independent predictor of HF diagnosis even with NT-proBNP in a multivariable model (odds ratio, 4.34; 95% CI 2.11–8.92; P < .001) and when added to NT-proBNP measurement, correctly reclassified patients who had false-negative and false-positive results by NT-proBNP testing alone.
In the PRIDE study, MR-proANP strongly and independently predicted 1- and 4-year mortality (adjusted HR 2.99, P < .001, and 3.12, P < .001, respectively) and addition of NT-proBNP to these models did not attenuate the predictive power of MR-proANP. Adding MR-proANP to base models containing NT-proBNP significantly improved the C-statistic at 1 and 4 years and reclassified mortality risk as a part of a multimarker strategy in determining prognosis.
In chronic HF, the Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienza Cardiaca–Heart Failure (GISSI-HF) study examined the predictive power of MR-proANP in stable chronic HF patients. Investigators found that an MR-proANP greater than or equal to 278 pmol/L had the best prognostic accuracy for 4-year mortality among several novel and established biomarkers including NT-proBNP, midregional proadrenomedullin (MR-proADM), C-terminal pro-vasopressin (copeptin), and C-terminal pro-endothelin-1 (AUC of 0.74, 95% CI 0.70–0.76). In addition, MR-proANP added independent prognostic information beyond NT-proBNP and relevant clinical characteristics in a reclassification analysis. Using the same biomarkers, only the change in MR-proANP over 3 months was found to be significant in predicting mortality.
Although initial data from the BACH study suggested that MR-proANP was less likely to be affected by covariates that reduce diagnostic accuracy of BNP or NT-proBNP (such as age, renal function, or obesity), subsequent data from other sources suggest that factors influencing BNP or NT-proBNP are quite likely to exert a similar effect on MR-proANP. As an example, Richards and colleagues recently reported that atrial fibrillation reduced the diagnostic accuracy of MR-proANP for ADHF diagnosis just as much as it did BNP or NT-proBNP.
Cardiac Necrosis
Cardiac troponins
In the context of cardiomyocyte necrosis, disruption of normal cardiomyocyte membrane results in the inner contents of the damaged cells to be released into the extracellular space; a variety of cellular and structural proteins such as troponin, creatine kinase, myoglobin, and cardiac fatty acid binding protein is released into circulation and become detectable in peripheral blood. Of these, cardiac troponins have rapidly become the standard of care biomarker in diagnosing myocardial infarction (MI). Cardiac troponins levels have also been found to be elevated in nonacute MI settings; one such setting is in HF.
The exact mechanism behind the release of troponins in HF is unclear but may be related to either increased myocyte membrane permeability or necrosis. A variety of HF mechanisms are involved in the release of troponin: inflammation, neurohormonal activation, ventricular stretch, increased wall tension, supply-demand mismatch, cytotoxicity, cellular necrosis, apoptosis, or autophagy. Regardless, increased circulating levels of cardiac troponins in HF patients have been shown to be closely linked to future clinical outcomes. With the development and application of highly sensitive cardiac troponin (hsTn) assays that can accurately detect even minute concentrations, most patients with HF are found to have detectable levels of cardiac troponins with a substantial percentage demonstrating concentrations of the biomarker above the upper reference limit for a normal patient population, even in the absence of ischemic heart disease.
In the ADHERE, 4240 patients (6.2%) out of 69,259 patients with ADHF had an elevated troponin, and in this context, higher in-hospital mortality (8.0% vs. 2.7%, P < .001) was observed, with an adjusted odds ratio for death of 2.55 (95% CI 2.24–2.89, P < .001) compared with patients with negative troponin measurement. Using an hsTnT assay, 30.6% of patients with ADHF were found to have elevated troponin values. In a multivariable model that included NT-proBNP and the interleukin (IL) receptor family member, sST2, hsTnT remained a significant and independent predictor of all-cause mortality with an HR of 1.16 (95% CI 1.09–1.24, P < .001). In a multimarker strategy, patients with all three biomarkers less than their optimal cutoff point had the best survival (0% death) at a median follow-up of 739 days, whereas 53% of those with elevation of all three biomarkers died. In integrated discrimination analyses, the use of all three markers in a multimarker approach was the best model for mortality prediction. As might be expected, the use of a highly sensitive assay was particularly helpful in determining the prognosis in patients with undetectable conventional TnT concentrations.
Although the prognostic ramification of an elevated troponin in patients with ADHF is clear, the therapeutic steps to follow with such an elevated value remain less well defined. Nonetheless, given the great importance of acute MI in the precipitation of ADHF, current position statements recommend the universal measurement of troponin in patients with acute symptoms, to primarily diagnose or exclude an ischemic cause for the presentation.
In chronic HF, cardiac troponin levels are also frequently increased, and in analogy to ADHF, such elevations are prognostic. When using a conventional assay, only approximately 10% of patients had detectable TnT out of the 4035 stable chronic HF patients in the Val-HeFT study. Not surprisingly, detectable troponin concentration was associated with an increased risk of death (HR 2.08, 95% CI 1.72–2.52) and first hospitalization for HF (HR 1.55, 95% CI 1.25–1.93) at 2 years in a model that adjusted for traditional risk factors. When hsTnT was measured in this same cohort, troponin was detectable in 92% of the cohort and predicted future adverse outcomes (HR 1.05, 95% CI 1.04–1.07, P < .001) as well as LV remodeling. Adding hsTnT to a baseline model including BNP and relevant clinical predictors significantly improved prognostic discrimination. Combining patients from both the Val-HeFT study and the GISSI-HF study, investigators looked at the role of serial hsTnT measurement in 5284 patients with chronic HF. Increases in hsTnT over 3 to 4 months of follow-up strongly predicted all-cause mortality (adjusted HR 1.59, 95% CI 1.39–1.82 and 1.88, 95% CI 1.50–2.35 after adjustment for traditional risk factors, baseline hsTnT, and baseline NT-proBNP), but improvement in test performance was only modest over a single baseline measurement of hsTnT.
Concentrations of troponin—particularly when measured with a highly sensitive method—may be of use to predict the onset of HF in apparently healthy individuals. For example, hsTnT measurement was found to be useful in predicting future development of HF in 4221 older, community-dwelling adults; hsTnT >12.94 pg/mL was associated with HF incidence rate of 6.4 per 100 person-years (95% CI 5.8–7.2) and an adjusted HR of 2.48 (95% CI 2.04–3.00). An elevated hsTnT was also predictive of future cardiovascular death (incidence rate of 4.8, 95% CI 4.3–5.4 and adjusted HR 2.91, 95% CI 0.9–1.2 compared with those with undetectable hsTnT). A repeat hsTnT measurement in 2 to 3 years showed that among patients with detectable hsTnT at baseline, greater than 50% change in hsTnT was associated with an even higher risk for HF (adjusted HR 1.61, 95% CI 1.32–1.97) and cardiovascular death (adjusted HR 1.65, 95% CI 1.35–2.03), whereas a decrease in hsTnT was associated with a lower risk for developing HF and having cardiovascular death. In a lower risk cohort from the Framingham Heart Study, Wang and colleagues similarly demonstrated the prognostic importance of hsTnI for predicting death and HF onset, even when extensively adjusted for relevant covariates, including other biomarkers.
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here
