https://orcid.org/0000-0002-5655-8201
Proposed natriuretic peptide deficiency syndrome in patients with heart failure and preserved ejection fraction. Several factors (on the left hand side of the figure) have been associated with reduced NP levels. These can lead to a state of intermittent or chronic NP deficiency, which is present in 20–35% of outpatients with HFpEF. Low NP levels can lead to hypertension, increased fluid retention, and increased metabolically unhealthy adipose tissue, which has been shown to result in increased pericardial constraint due to excessive chest wall, pericardial, and epicardial fat. As shown in the study be Verbrugge et al., patients with NP deficiency appear to only have mild intrinsic abnormalities in myocardial structure and function, and yet develop high left atrial pressures during minimal exertion and have worse outcomes compared with dyspnoeic patients without HFpEF. There are several unmet needs in patients with NP-deficient HFpEF (upper right). Treatment with drugs that decrease congestion (e.g. SGLT2 inhibitors), result in profound weight loss (e.g. GLP1 receptor agonists), or increase cyclic guanosine monophosphate downstream of NPs (e.g. PDE9 inhibitors) could all be beneficial in patients with NP-deficient HFpEF. NP, natriuretic peptide; BNP, brain natriuretic peptide; HFpEF, heart failure with preserved ejection fraction; PCWP, pulmonary capillary wedge pressure; SGLT2i, sodium–glucose co-transporter-2 inhibitor; GLP1-RA, glucagon-like peptide receptor agonist; PDE9, phosphodiesterase-9.
This editorial refers to ‘Heart failure with preserved ejection fraction in patients with normal natriuretic peptide levels is associated with increased morbidity and mortality’, by F.H. Verbrugge et al., https://doi.org/10.1093/eurheartj/ehab911.
The discovery of atrial natriuretic peptide (ANP) by Adolfo de Bold and colleagues >40 years ago in 19811 was a seminal event in cardiovascular medicine. By showing that infusion of rat atrial extracts in normal rats resulted in natriuresis and systemic hypotension paved the way for our understanding of the heart as an endocrine organ and the eventual discovery of brain natriuretic peptide (BNP) in 1988.2 The explosion of research on natriuretic peptides (NPs) since their discovery has considerably expanded our understanding of NP biology, the potential for their use as therapeutic agents, and, importantly, the diagnostic and prognostic utility of BNP and N-terminal pro-BNP (NT-proBNP).3 The major advantage of BNP and NT-proBNP as diagnostic markers, which allows for their routine clinical use, comes from their longer half-life in the circulation compared with ANP.3 Several important studies over the past several decades have resulted in consolidation of BNP and NT-proBNP in our diagnostic armamentarium for heart failure (HF).3,4 However, as a tool for clinical reasoning, our steadfast use of NPs as diagnostic aids often causes us to suffer from ‘premature closure’.5
Premature closure, a type of cognitive bias that occurs when the result of a diagnostic test causes a healthcare provider to fail to consider alternative diagnostic possibilities, is one of the most common causes of diagnostic errors in medicine.6 Faulty clinical reasoning can therefore lead to incorrect, missed, or delayed diagnoses, which can directly harm patients. NPs are peculiar as diagnostic aids because they are predominantly beneficial hormones whereas most diagnostic tests in medicine measure abnormalities that have resulted from organ injury (e.g. troponin) or cause disease (e.g. LDL cholesterol). NPs are secreted by cardiomyocytes in response to increased diastolic wall stress, thereby providing the protective effects of natriuresis and vasodilation, which counteract the HF syndrome. Thus, one can hypothesize that there are patients in whom NP deficiency is problematic and is the cause of the HF syndrome because of the inability to counteract increased diastolic wall stress.7 In these patients, a low NP value can cause a diagnostician to exclude the diagnosis of HF, prematurely closing their minds to the possibility that HF, in fact, may exist. Although first recognized in the setting of obesity, it is now known that there are several causes of NP deficiency (Graphical Abstract), including genetic (e.g. polymorphisms in the NPPB gene), African ancestry, increased androgencitiy in women, insulin resistance, hypercortisolism, and certain medications (e.g. spironolactone).7 In addition, diastolic wall stress, one of the main stimuli for BNP secretion by cardiomyocytes, is equal to [left ventricular (LV) diastolic pressure × LV chamber radius] ÷ LV wall thickness. Therefore, for any given elevation in LV end-diastolic pressure, patients with HF and preserved ejection fraction (HFpEF) will have a lower BNP level compared with patients with HF and reduced ejection fraction (HFrEF) because the LV chamber size is smaller, and wall thickness is typically greater, in HFpEF than in HFrEF.
One of the first indications of the presence of BNP deficiency in some patients came from an early publication on our mechanistic understanding of BNP in the setting of HF. Iwanaga and colleagues demonstrated that LV end-diastolic wall stress correlated well with BNP levels in outpatients with HF.8 In this study, patients with HFpEF had lower BNP levels than those with HFrEF, and some patients, despite elevated LV end-diastolic pressure, had low levels of BNP. Subsequently, we showed that this phenomenon is common in HFpEF, occurring in approximately one-third of consecutive patients (total n = 159) with documented elevations in pulmonary capillary wedge pressure (PCWP) at rest and a prior history of HF hospitalization.9 In our study, 29% of patients had a BNP < 100 pg/mL (the threshold for its use as a diagnostic test for HF). As expected, BNP < 100 pg/mL was more common in obese individuals.9
Intriguingly, prior HFpEF clinical trials have shown that patients with lower values of NPs seem to respond more favourably to medications such as mineralocorticoid receptor antagonists (TOPCAT trial) and angiotensin receptor blockers (I-PRESERVE trial). Furthermore, sacubitril/valsartan reliably reduces NT-proBNP levels, but does not result in improved exercise capacity or health status (PARALLAX trial), and only has modest effects on reduction in HF hospitalizations (PARAGON trial). Conversely, sodium–glucose co-transporter-2 (SGLT2) inhibitors have little effect on NP values and yet appear to reduce HF hospitalizations and improve exercise capacity and health status (EMPEROR-Preserved and PRESERVED-HF trials, respectively). These findings call into question our dependence on elevated NP values for diagnosis and assessment of treatment response in the setting of HFpEF.
Given these pitfalls in the use of NP values as a diagnostic measure in HFpEF, guideline statements and consensus documents advocate the use of very low BNP and NT-proBNP thresholds to exclude the possibility of a HF diagnosis.10,11 Furthermore, NT-proBNP did not make the cut of variables that were used to create the H2FPEF score for HFpEF diagnosis because of its weak performance in outpatients with unexplained dyspnoea (using the gold standard of exercise invasive haemodynamics for diagnosis).12 Nonetheless, many clinicians still believe NP levels to be the best test for HF diagnosis, including in HFpEF, and large-scale HFpEF clinical trials often require elevated NT-proBNP values as an inclusion criterion to ensure the veracity of the HFpEF diagnosis. Additionally, the magnitude of NP elevation is consistently associated with risk for adverse outcomes in the general population and in patients with HF across the ejection fraction spectrum.3,4
For these reasons, even if we ultimately can more accurately diagnose these NP-deficient HFpEF patients, should we conclude that they are low-risk individuals? In this issue of the European Heart Journal, Verbrugge and colleagues answered this question by performing a secondary analysis of a large prospective study of 581 patients referred to the Mayo Clinic for invasive exercise haemodynamics for unexplained dyspnoea.13 The authors compared patients with HFpEF (diagnosed if resting PCWP was ≥15 mmHg or exercise PCWP was ≥25 mmHg) with low NP values (NT-proBNP <125 pg/mL; n = 157) vs. elevated NP values (NT-proBNP ≥125 pg/mL; n = 263); they also compared these two groups with patients without HFpEF (n = 161). The authors found that patients with low levels of NP carried an intermediate pathophysiological and risk profile between dyspnoeic non-HFpEF controls and HFpEF patients with high NP values. High NP HFpEF patients had limited cardiac output reserve during exercise whereas low NP HFpEF patients did not, despite the presence of similar and strikingly high values of exercise PCWP in both groups of HFpEF patients. Importantly, although event rates, as expected, correlated with NP values in HFpEF patients, the patients with low NP HFpEF had an almost a three-fold higher event rate (mortality or HF hospitalizations) compared with patients without HFpEF after adjusting for age, sex, and body mass index. These results underscore the importance of accurately diagnosing HFpEF, not only because there is now a proven therapeutic for HFpEF (SGLT2 inhibition) but also because, from a risk profile perspective, the diagnosis of HFpEF, even in the setting of a low NP level, is associated with adverse outcomes.
One interesting aspect of the study by Verbrugge and colleagues is the finding that some patients without HFpEF had elevated NT-proBNP values (above the 125 pg/mL threshold). Thus, in stable outpatients with unexplained dyspnoea, it is possible that there are patients with elevated NP values despite the absence of haemodynamic evidence of HFpEF. In the Atherosclerosis Risk in Communities population-based study, the presence of an NPPB promoter polymorphism (which causes higher NP levels) was associated with lower blood pressure, less hypertension, reduced cardiovascular mortality, and increased life span,14 suggesting a protective effect of high lifelong NP levels in some individuals (i.e., the opposite of NP deficiency syndrome).
A few limitations of the study by Verbrugge et al. merit consideration. First, the study was single centre, and potentially subject to bias due to inclusion of individuals referred for invasive haemodynamic evaluation at a quaternary referral centre. However, the observations of the high proportion of stable HFpEF outpatients with NP values below HF diagnosis thresholds have been made by other investigators in alternative settings, throughout the world, and in patients from different race/ethnic backgrounds. Additionally, the finding that low NP HFpEF patients had worse outcomes than patients without HFpEF is not surprising given the fact that these patients have more comorbidities than their non-HFpEF counterparts, and the multivariable models in the study did not adjust for these comorbidities (which would have probably attenuated the association of low NP HFpEF and adverse outcomes, when compared with non-HFpEF patients; of note, the authors did perform age- and comorbidity-matched analyses and found similar differences in echocardiographic and haemodynamic characteristics between groups).
In the future, several steps should help mitigate the imperfection of NPs as diagnostic and prognostic markers in HFpEF. First and foremost, HF clinicians and cardiologists should raise awareness of the problem of NP deficiency, its pathophysiology, and its unmet needs in HFpEF ((Graphical Abstract). NPs perform poorly as tests to rule out HFpEF, especially in the outpatient setting, but also at times in the inpatient setting. Algorithms now exist to diagnose HFpEF without NPs (H2FPEF score12) or with alternative criteria besides NPs (HFA-PEFF score10). Some HFpEF clinical trials are already employing a sliding scale of lower diagnostic thresholds for NPs based on increasing body mass index,15 and several trials have allowed for additional pathways of enrolment besides NPs (including prior HF hospitalization or documented evidence of elevated PCWP at rest or during exercise). Finally, in the future, we may have access to individual biomarkers or proteomic panels that are better at diagnosing HFpEF without the pitfalls of NPs.
In conclusion, to prevent premature closure and subsequent diagnostic errors, clinicians would be well served by remembering the following mnemonic when ordering the BNP test: Biomarker Not Perfect. The importance of the initial discovery and subsequent wealth of publications on NPs over the past 40+ years in the realms of diagnosis, prognosis, therapeutics, and HF biology cannot be understated; nevertheless, NPs remain only one aspect of our diagnosis and prognosis toolbox. Like any clinical test, NPs should be carefully evaluated within the context of all other clinical information, and invasive exercise haemodynamics should be performed in patients with unexplained dyspnoea if the underlying diagnosis is in question.
Conflict of interest: S.J.S. has received research grants from the 10.13039/100000002National Institutes of Health (U54 HL160273, R01 HL107577, R01 HL127028, R01 HL140731, and R01 HL149423), Actelion, AstraZeneca, Corvia, Novartis, and Pfizer; and has received consulting fees from Abbott, Actelion, AstraZeneca, Amgen, Aria CV, Axon Therapies, Bayer, Boehringer-Ingelheim, Boston Scientific, Bristol-Myers Squibb, Cardiora, Coridea, CVRx, Cyclerion, Cytokinetics, Edwards Lifesciences, Eidos, Eisai, Imara, Impulse Dynamics, Intellia, Ionis, Ironwood, Lilly, Merck, MyoKardia, Novartis, Novo Nordisk, Pfizer, Prothena, Regeneron, Rivus, Sanofi, Shifamed, Tenax, Tenaya, and United Therapeutics.
References
Author notes
The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.
