Challenges of using natriuretic peptides to screen for the risk of developing heart failure in patients with diabetes: a report from the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) Committee on Clinical Applications of Cardiac Bio-Markers (C-CB)

Introduction

Since the original reports of a peptide in porcine brain sharing structural features and activity with atrial or A-type natriuretic peptide (ANP) [1], considerable scientific progress has been made [2], [3], [4], [5]. It has been established that the concentrations of natriuretic peptides (NPs) in the circulation observed in patients with decompensated congestive heart failure are often markedly elevated, resulting in excellent discrimination between patients presenting with acute shortness of breath with and without a cardiac cause for the symptoms [6], 7]. Recent insights have shown that the NP system is dysregulated in patients with heart failure, generating a large number of B-type natriuretic peptide (BNP) fragments with varying degrees of biological activity [8], [9], [10]. Thus, high concentrations of BNP do not necessarily reflect an adequate counter-regulatory response but may instead reflect the presence of less physiologically active fragments. As a consequence, the NP concentrations measured during development of heart failure will differ dependent on the characteristics of the antibodies used in the assay concerned and the characteristics of the fragments detected [11], 12]. A large variety of assays for BNP, NT (N-terminal)-proBNP and intact proBNP, which have become the clinical markers of choice, have been developed (see https://ifcc.org/ifcc-education-division/emd-committees/committee-on-clinical-applications-of-cardiac-bio-markers-c-cb/biomarkers-reference-tables2025). These assays are not harmonized; i.e., the values from one assay measure the various fragments differently. Thus, they may not reflect the same biology as values from another assay [11], 12].

Multiple studies indicate that BNP and NT-proBNP concentrations, even at levels below decision thresholds for heart failure diagnosis, are strongly associated with the risk of future heart failure and cardiovascular death in patients with chronic coronary syndromes (44) and in the general population [13], [14], [15]. However, the use of BNP and NT-proBNP as markers of risk in the general population or in patients with unrecognized, what has been termed stage B heart failure [16], may provide some additional challenges related to assay standardization, biological interferences and clinical implementation. The present report from the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) Committee on Clinical Applications of Cardiac Bio-Markers (C-CB) aims to educate about the emerging clinical opportunities and analytical issues which should be addressed to support use of NPs at low concentrations to predict the development of disease over time.

Use of natriuretic peptides to predict incident heart failure

Beyond being important diagnostic biomarkers for heart failure, data has demonstrated their ability to identify asymptomatic community dwellers who are at increased cardiovascular risk [17], 18]. A variety of initiatives suggests that NP concentrations well within the putative reference ranges may identify individuals either with subtle heart failure not appreciated clinically or those who might be prone to develop heart failure subsequently [17], 18] due to underlying structural heart disease (so called Stage B heart failure) [16]. As NP concentrations in the circulation begin to increase, even if they stay within the putative reference range, they possess prognostic importance for the development of heart failure. The Dallas Heart Study [14], the Atherosclerosis Risk In the Community (ARIC) study [17] and the Olmsted County study [18] have shown that NP values higher in the reference range manifest prognostic importance in community dwelling individuals. The importance of these lower values was unequivocal regardless of biomarkers being combined with imaging data or not [14], [17], [18], [19].

In the Olmsted County analysis [18], it was shown that age and sex adjusted concentrations above the 80th percentile of the normal range predicted the likelihood of heart failure with a high degree of accuracy long term. The curves depicting the risk of heart failure over time diverged early after a relatively short period of time [18]. This same phenomenon also applies to cardiac troponin (cTn) concentrations, where very low levels are associated with a very low risk of adverse events but even a modest increase within the reference limits is associated with an increased likelihood of the development of cardiovascular disease over time [14], [17], [18], [19], [20], [21].

Subsequent studies have evaluated whether NP measurements can be used to identify high-risk individuals who may benefit from referral to cardiologists and therapeutic intervention. In STOP-HF [22], BNP was measured in individuals with clinical risk factors for developing heart failure and a value of 50 ng/L was used to identify those considered to be at increased risk who were randomized to extensive evaluation vs. usual care. In those with BNP concentrations above 50 ng/L, additional evaluations, usually via a referral to a cardiologist, improved outcomes. In particular, less heart failure was seen during the seven years of follow-up. In PONTIAC [23], 300 individuals with type 2 diabetes and an NT-proBNP concentration greater than 125 ng/L were randomized to a maximum tolerated dose of ACE-inhibitor and beta blocker compared to usual care. This is a relatively low NT-proBNP concentration that overlaps with some population-based reference intervals and is far lower than any of the heart failure decision limits deployed in the emergency department to evaluate patients with acute shortness of breath (e.g., NT-proBNP>300 pg/mL; ng/L). By two years after enrollment in the Pontiac study there was a benefit in the group receiving the ACE-inhibitors and beta blockers, with lower rates of hospitalization and death due to cardiac disease.

An important issue is establishing why increases in NPs and cTn within the reference range are associated with the development of disease over time. It could be suggested that values that have prognostic significance should not be considered “normal”. However, this approach of using decision limits that overlap with reference intervals to titer treatment is used for other biomarkers like total cholesterol and LDL cholesterol where secondary prevention with cholesterol lowering is often initiated at values far below the reference intervals for the test. This is exactly the situation with the NPs thresholds that have been suggested to increase awareness of heart failure in symptomatic outpatients by guideline groups [24]; i.e., overlaps substantially with commonly reported reference intervals for the assays, especially for subgroups like women.

Although one could argue that values within the reference interval that are prognostic should be called abnormal, that may not be the only possibility. Another more parsimonious explanation might be that all systems have ranges of flexibility and that over time those with a lesser extent of reserve are at risk. In addition, some biomarkers, population-based reference intervals might be too broad to effectively identify individuals at risk. Biological variation studies which help to define the upper bound of the 95 % confidence interval of a random change for NP indicate that this interval is large. In those situations, it might be better for biomarkers with high between subject biological variation compared to within subject biological variation to determine setpoints for individual patients and then to monitor changes from this baseline instead of using reference intervals as indication of disease or increased risk [25], 26]. A recent paper explored the clinical utility of risk prediction based on setpoints for hematological parameters, demonstrating that subclinical disease was predicted in individuals with low or high setpoints who moved significantly within the normal range [25], 26]. Applying individual setpoints would preserve the concept of normality but would suggest that when individuals move from their setpoint to higher values within the reference intervals they lack whatever reserve might be present compared to other individuals and therefore could be at increased risk for cardiovascular events. This would need to be done with care since conjoint analytical and biological variation is high at lower levels [27]. This should be the topic of future studies. Also, one should investigate if individuals with physiological high set points for NPs have increased risk or might be protected from future heart disease [28], 29] due to the cardiac protective abilities of these hormones.

Clinical implications

The above findings suggest that concentrations of NPs, even within the reference range, identify patients at increased risk for adverse cardiovascular events, in particular heart failure and cardiovascular death. Patients with chronically elevated levels or with levels in the high-normal range should be considered for further investigations, including cardiac imaging. If functional or structural cardiac disease is detected, the suggestion is to start or intensify therapy. This approach is supported by results from trials of SGLT2 inhibitors showing that groups with values above certain concentrations as a group benefit from therapy [30], 31]. Their primary analysis used a value of 125 pg/mL but the benefit was not reserved for that group, which is more in keeping with the continuous relationship between NTproBNP and prognosis similar to the results of previous analyses [19]. However, the optimal threshold values for identifying subclinical disease or increased risk are unclear and will require additional information in this area.

Recommendations for natriuretic peptide screening for heart failure in the outpatient setting

In the outpatient setting NP measurement is used in patients with symptoms of heart failure, to identify those who need further follow-up and confirmation of the diagnosis using echocardiography. In 2022 the American Diabetes Association (ADA) recognized that many patients with diabetes are at risk of stage B heart failure. They defined being at risk despite the absence of symptoms as having at least one of the following: 1) evidence of structural heart disease, 2) abnormal cardiac function, or 3) elevated NP levels and/or elevated cTn levels. To facilitate earlier diagnosis to enable the implementation of therapies to prevent adverse outcomes, they recommended the measurement of NPs and hs-cTn on at least a yearly basis to identify the presence of stage B heart failure and to determine the risk for progression to symptomatic heart failure. For BNP, a concentration of 50 pg/mL was recommended, and for NT-proBNP a concentration of 125 pg/mL was suggested [32]. These cut offs have also been suggested for use to confirm the diagnosis of outpatient heart failure in patients with diabetes [33].

Subsequently, in 2023, a clinical consensus statement from the Heart Failure Association of the European Society of Cardiology proposed rule-in and rule-out concentration cutoffs for outpatient heart failure. They also defined a threshold for what they called “heart stress” for NT-proBNP. This left a grey zone in between these two thresholds [24]. A threshold value of <125 pg/mL, ng/L was recommended to rule-out the presence of heart failure in the outpatient setting (Figure 1). This threshold was suggested because it had been recommended by guidelines groups. Additional age-specific rule-in thresholds were also recommended. It was suggested that the patients with values above the “heart stress” threshold should proceed rapidly to echocardiography (within six weeks) and thereafter a decision in regard to treatment would be considered. Regarding heart stress, this group suggested rule-out and rule-in thresholds to identify those with increased risk of developing future cardiovascular events and in particular heart failure (Figure 1). The rule-out threshold was 50 pg/mL and it was suggested that patients above this level should have repeated NT-proBNP testing within six months (Figure 2) whilst patients with concentrations above the age adjusted cutoffs for rule-in (ranging from 75 to 300 pg/mL, ng/L) should undergo echocardiography and clinical evaluation to consider the need for risk reduction therapy. The rule out value proposed is similar to that recently proposed in an extensive analysis of patients with heart failure with preserved ejection fraction [34]. However, the optimal threshold values for identifying subclinical disease or increased risk are much more ambiguous.

Figure 1:

Consensus recommendations from the Heart Failure Association of the European Society of Cardiology concerning when to consider the possibility of heart failure in the outpatient setting. Note that age adjusted thresholds are suggested with suggestions for subsequent management. Reproduced with permission from reference [30].

Figure 2:

Consensus recommendations from the Heart Failure Association of the European Society of Cardiology concerning the evaluation of asymptomatic diabetic patients. Note that age adjusted thresholds are suggested with suggestions for subsequent evaluation. Reproduced with permission from reference [30].

Likewise, in 2023, the Consensus Report from the Diabetes Technology Society recommended routine biomarker screening with BNP or NT-proBNP at age 30 years in patients with type 1 diabetes and at any age of diagnosis in patients with type 2 diabetes, with annual measurements recommended [32]. The timing for such evaluations is shown in Table 1. Although patients with diabetes are easy to define as a high-risk group for cardiovascular disease, these sorts of guidelines might be expanded to include a variety of other subsets in the future who are either at risk or are in need of cardiovascular screening.

Analytical and clinical considerations

The current threshold recommendations rely predominantly on clinical trials that identify groups of patients who are at risk and demonstrate the benefits of treatment in these groups. We are in general supportive of these approaches to facilitate and treat early disease. However, given the multiple variables that impact NP levels, extrapolation from groups to individual patients may be challenging. For example, one must be careful to acknowledge the multiplicity of issues that can increase or lower NP levels [11], 12] (Table 2). The higher the natriuretic peptide cutoffs endorsed, the more specific they are likely to be. However, if too high, there is a risk one could diminish sensitivity and thus could fail to identify all of those with unrecognized heart failure. Careful consideration is needed as the risk of excessive testing is high, though the risk of missing those in need of testing may be substantial as well. The following areas are ones that require additional data and consideration:

Educational guidance

1. Data on the use of BNP assays for population screening are sparse and more are needed. Assay specific thresholds will likely be necessary due to lack of standardization.

2. The present NT-proBNP data are mostly derived from studies using one assay (Roche Diagnostics); thus, care is necessary in applying these criteria to other NTproBNP assays.

3. Making sure the same assay is used for the same patient and making more consistent the timing of obtaining these values would to some extent reduce the variability that can plague the interpretation of these measurements.

4. A number of demographic and clinical factors may impact NP concentrations and additional scrutiny of the threshold values that have been recommended is necessary, especially for women and potentially for Black patients as well.

5. Obtaining a baseline value in individuals at high risk such as those with diabetes should be considered to establish an individual baseline for subsequent follow-up over time.

6. Given the variability of the measurements of NPs, clinical implications can only be based on large changes or serial measurements showing a clear trend.

7. Better education concerning the factors influencing NP values will be necessary for clinicians. Considering the confidence intervals of the ranges being deployed which are available for normal values might help. Even better still would be an approach taking biological and analytical variability into account. In addition, using the alert functions of electronic medical records and developing AI algorithms in this area could improve decision making.

8. Current clinical data apply mostly to patients with diabetes but there are many individuals who are at high risk for cardiovascular disease and eventually extrapolation to those populations is likely. Data should be made available investigating whether the same values as those recommended for patients with diabetes will apply to other groups.

Members of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) Committee on Clinical Applications of Cardiac Biomarkers (C-CB)

Allan S. Jaffe, Department of Cardiovascular Medicine, Mayo Clinic, Rochester, MN, United States; Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, MN, United States; Yader Sandoval, Minneapolis Heart Institute, Abbott Northwestern Hospital, and Center for Coronary Artery Disease, Minneapolis Heart Institute Foundation, Minneapolis, MN, USA; Nicholas L. Mills, BHF Centre for Cardiovascular Science, University of Edinburgh, Edinburgh, United Kingdom; Usher Institute, University of Edinburgh, Edinburgh, United Kingdom; Louise Cullen, Emergency and Trauma Centre, Royal Brisbane and Women’s Hospital Health Service, Herston, Queensland, Australia; Lori B. Daniels, Division of Cardiovascular Medicine, Department of Medicine, University of California, San Diego, La Jolla, CA, United States; Ola Hammarsten, Department of Clinical Chemistry Sahlgrenska University Hospital, Gothenburg, Sweden; Jens P. Goetze, Department of Clinical Biochemistry, Rigshospitalet University Hospital, Copenhagen, Denmark; Blanca Fabre-Estremera, Department of Laboratory Medicine, La Paz University Hospital, Madrid, Spain; IdiPaz-Hospital La Paz Institute for Health Research, Madrid, Spain; Paul Collinson, City St George’s University of London, London, United Kingdom; Fred S. Apple, Department of Laboratory Medicine and Pathology, University of Minnesota, Minneapolis, MN, United States; Clinical and Forensic Toxicology Laboratory, Hennepin Healthcare/Hennepin County Medical Center, Minneapolis, MN, United States; Torbjørn Omland, Department of Cardiology, Division of Medicine, Akershus University Hospital, Lørenskog, Norway; K.G. Jebsen Center for Cardiac Biomarkers, Institute of Clinical Medicine, University of Oslo, Oslo, Norway; Kristin M. Aakre, Department of Medical Biochemistry and Pharmacology and Department of Heart Disease, Haukeland University Hospital, Bergen, Norway; Department of Clinical Science, University of Bergen, Bergen, Norway for the Committee on Clinical Applications of Cardiac Bio-Markers (C-CB) of the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC).