NT-proBNP: Can We Better Utilize Biomarkers in Patients With Diabetes?

B-type natriuretic peptide (BNP) is secreted from the atria of the heart as a response to myocardial stretch that occurs with increases in volume, afterload, and other hemodynamic changes that result in increased cardiovascular stress. The atria of the heart secrete BNP in its proform and it is subsequently cleaved into an inactive form (N-terminal [NT]-proBNP) and an active form (BNP) (1). Although there are some differences with regard to interpretation of the assays for these different metabolites, both NT-proBNP and BNP (referred to collectively as natriuretic peptides [NP]) can be measured and quantified and play an important role in the diagnosis and management of both acute and chronic heart failure.

While the role of NP in the management of heart failure is well established, the role of NP outside of heart failure is less well defined. Prior studies consistently show a clear relationship between elevated levels of NP and subsequent risk of cardiovascular events, including atrial fibrillation (2), diastolic dysfunction (3), coronary heart disease, and stroke (4), and the development of diabetes (5). Given these associations, it is possible to use NP to assist in the risk stratification of patients into those who are at high and low risk of cardiovascular events in populations at risk for cardiovascular disease, with stable coronary artery disease, or following a myocardial infarction (6). Furthermore, it is possible that these biomarkers could also help to tailor therapy such that patients at high risk have intensification of therapy, while those patients at lower risk have therapies that are de-intensified (7).

In this issue of Diabetes Care, Birukov et al. present additional data from the European Prospective Investigation Into Cancer and Nutrition (EPIC)-Potsdam cohort, which included almost 30,000 subjects, on the association of NT-proBNP with risk of developing diabetes and the risk of vascular complications in those patients with diabetes (8). They performed a nested case-control study in which they identified 1,338 patients who developed diabetes or cardiovascular disease during the follow-up of this cohort and matched them with a random cohort of patients from within the overall Potsdam cohort. Out of this cohort, approximately 20% had NT-proBNP levels that were below the level of detection for the assay used, while the others had detectable levels of circulating NT-proBNP. The authors then examined the relationship between NT-proBNP and the risk of developing diabetes, cardiovascular events (cardiovascular death, myocardial infarction, or stroke), and the risk of diabetes-related complications in the cohort who went on to develop diabetes.

Models were developed that adjusted for a variety of demographic, social, and clinical factors including biomarkers other than NT-proBNP. The study resulted in several interesting findings. First, consistent with prior studies (9), the authors found that NT-proBNP was inversely correlated with BMI, such that as BMI increased, levels of NT-proBNP were lower. In a likely related manner, higher levels of NT-proBNP were associated with a small statistically lower risk of developing type 2 diabetes. This relationship was complicated by a further observation of a significant interaction between sex, NT-proBNP, and the risk of developing diabetes, such that the relationship between increased NT-proBNP and decreased future risk of diabetes was only seen in women.

When the relationship between NT-proBNP and future cardiovascular risk was examined, NT-proBNP had no relationship with future cardiovascular events but was associated with an increased risk of complications from diabetes in the cohort that went on to develop diabetes. These findings may be due to the relatively small number of cardiovascular events included in this analysis.

Understanding who will develop of diabetes (and then identifying prevention strategies) is a topic of considerable public health importance. While there have been some suggestions that genetics may play a role, recent studies have supported the idea that obesity is the most important and potentially modifiable risk factor (10). The relationship between BMI and subsequent risk of diabetes may also be an important relationship to consider when interpreting the results of this particular study.

NT-proBNP levels are dynamic and influenced by multiple different factors (11). For example, the presence of atrial fibrillation increases NP due to increased atrial stress (12). Decreased glomerular filtration rate is associated with increased NP (and NT-proBNP in particular) likely both from decreased clearance of natriuretic peptides and elevated circulating plasma volume (13). Most relevant in this particular study is the interdependent relationship between BNP and obesity. The inverse relationship between BMI and NT-proBNP is well described and replicated in this study. As BMI increases, NT-proBNP tends to decrease (14). Thus, BMI, NT-proBNP, diabetes, and cardiovascular events are interconnected in a manner that is difficult to untangle.

In this particular study, the authors utilized robust statistical methods to attempt to control for the potential confounding that BMI may cause in the relationship between NT-proBNP, the development of diabetes, and the ensuing complications. The overall models developed in this study had good discriminatory ability to identify patients who would subsequently develop diabetes. However, these complex and interdependent relationships may be one reason NT-proBNP in isolation was no better than the play of chance in identifying those at risk for subsequently developing diabetes.

One is left with multiple questions, including how can we use this information to better understand the underlying biology of diabetes, and how can we better utilize biomarkers to predict risk? Unfortunately, this article cannot definitively answer the two fundamental questions about this relationship. First, is there residual confounding by BMI in the relationship between NP and diabetes that cannot be accounted for with modeling? Second, even if NP are associated with diabetes independent of BMI, are NP simply markers of better metabolic health or are they causal actors in the complicated hormonal metabolic physiology that leads to insulin resistance and diabetes?

In the current era, we have the ability to collect more phenotypic data than ever before. The use of wearable technologies gives one the ability to characterize physical activity, heart rates both at rest and with exertion, and sleep patterns. Established biomarkers such as creatinine, NT-proBNP, and high-sensitivity troponin can be combined with measures of metabolic health including adiponectin, insulin resistance, and hemoglobin A_1c to better characterize current levels of health (15). However, it is not yet understood how these individual factors are related to different genetic environments or how biomarkers can be combined with genetic information to better predict long-term outcomes. The data from this important study will help provide some insight into these relationships and help as we figure out how we can use available information to better predict future events.