In their study recently published in Diabetes Care, Epshteyn et al. (1) found that plasma B-type natriuretic peptide (BNP) was able to discriminate between diabetic patients with and without left ventricular (LV) dysfunction, even among the subset without any clinical suspicion of heart failure. It is this last observation that supports the use of BNP as a screen for LV dysfunction among people with diabetes. Detection of LV dysfunction, an early feature of diabetic heart disease, presents an important opportunity for prevention of downstream morbidity and mortality (2). The question that begs to be answered is when should screening for LV dysfunction take place?
Epshteyn et al.’s (1) sample of 91 patients without clinical suspicion of heart failure did, however, contain a significant number with vascular disease. A history of hypertension, coronary artery disease, and myocardial infarction was present in 81, 23, and 18% of the sample, respectively. The study reinforced the value of screening with BNP in a high-risk diabetic population. What remains unknown is whether the utility of BNP for the detection of LV dysfunction extends to asymptomatic individuals without overt vascular disease.
We addressed this by undertaking a substudy within the large Australian Diabetes, Obesity and Lifestyle (AusDiab) study (3). A random sample of 100 adults with type 2 diabetes but free of overt cardiovascular disease and hypertension were matched 1:1:1 by age and sex to subjects with normal glucose tolerance (NGT) and impaired glucose tolerance (IGT), who were also without overt cardiovascular disease and hypertension. In contrast to the study by Epshteyn et al. no differences were found in mean (±SE) levels of plasma N-terminal BNP across the three groups: type 2 diabetes, 155 ± 33; IGT, 172 ± 40; and NGT, 162 ± 51 pg/ml (P = 0.96). However, there were significant differences in urinary protein: type 2 diabetes, 102 ± 15; IGT, 64 ± 7; and NGT, 50 ± 4 mg/day (P < 0.001).
There are two possible explanations for the observed findings with N-terminal BNP: 1) among diabetic patients without overt cardiovascular disease and hypertension, LV dysfunction is not more common compared with that of those with IGT and NGT, or 2) plasma N-terminal BNP is not sensitive to its presence. Echocardiographic studies would be needed to confirm which explanation holds, but either way plasma N-terminal BNP appears to have little utility for early screening of LV dysfunction in patients with diabetes (in the absence of cardiovascular disease and hypertension). This contrasts with early screening for renal dysfunction by urinalysis.
