Aim: To assess the accuracy of B-type natriuretic peptide (BNP) plasma levels for the diagnosis of left ventricular hypertrophy (LVH) in hypertensive patients.
Participants and methods: We studied a sample of 409 adults aged 45 years or older, recruited from residents of Porto by random digit dialing. Data were collected by clinical interview and physical examination, ECG, echocardiogram and venous blood sampling for the measurement of plasma concentrations of BNP. Hypertension (HT) was defined as blood pressure > or = 140/90 mmHg on the day of interview and/or self-reported HT if treated with any antihypertensive medication; LVH was defined as left ventricular mass index (LVMI) > or = 125 g/m2 in men and 110 g/m2 in women. The participants were further classified in four strata according to left ventricular morphology--normal, concentric remodeling, eccentric LVH or concentric LVH.
Results: Two hundred and thirty-two (56.7%) individuals were hypertensive, and among these 73 (31.5%) had LVH. BNP levels were significantly higher in these individuals (median [P25-P75] = 55.8 pg/ml [22.6-88.4]) than in hypertensive patients without LVH (29.9 pg/ml [10.0-62.8]), p = 0.003. BNP levels also differed significantly across strata of left ventricular geometry, the main difference depending on the presence or absence of LVH. There was a positive correlation between plasma BNP levels and LVMI (Spearman's P 0.185, p = 0.005). The area under the ROC curve--a parameter for diagnostic accuracy quantification--was 0.62 (95% confidence interval 0.54-0.70), indicating low discriminatory power between normal and abnormal LVMI.
Conclusion: In the assessed population, BNP levels were higher in hypertensive patients with LVH than in the absence of LVH. However, BNP did not perform well in discriminating between the presence or absence of LVH.