Background: In the early identification of cardiovascular risk, it is essential to establish a biological marker for cardiac complications that is comparable to albuminuria for nephropathy. We tested the hypothesis that N-terminal pro-brain natriuretic peptide (NT-proBNP) might be a marker for silent myocardial ischaemia in diabetes.
Methods: In forty consecutively recruited subjects without evident coronary artery disease, serum NT-proBNP was measured together with multi-slice computed tomography. With patients suspected of having significant coronary artery stenosis by multi-slice computed tomography, coronary angiography was performed. Silent myocardial ischaemia was defined as the presence of significant coronary artery stenosis with more than 50% luminal narrowing by angiography.
Results: Thirteen patients (32.5%) had silent myocardial ischaemia. NT-proBNP levels were significantly higher in these patients (181.1 ± 43.8 versus 55.2 ± 9.7 pg/mL, p < 0.005) but HbA(1c), lipid profiles, and creatinine were similar in the two groups. Moreover, log NT-proBNP was identified as an independent predictor of silent myocardial ischaemia (R(2) = 0.502, p < 0.05) after adjustment for HbA(1c), creatinine, albuminuria, hypertension, hyperlipidaemia, or smoking. After stratifying patients by NT-proBNP, the upper tertile compared to the lowest tertile was significantly associated with silent myocardial ischaemia (odds ratio: 26.7, p < 0.05). Receiver operation characteristics analysis with a cut-off value of 52 pg/mL showed 92% sensitivity and 75% specificity for predicting silent myocardial ischaemia (positive predictive value 64.7%, negative predictive value 94.3%).
Conclusions: The outstandingly high negative predictive value of NT-proBNP enables us to focus on diabetic patients with occult coronary disease, independently of microalbuminuria.