Background: In the general population, the plasma concentrations of atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) are useful to predict left ventricular hypertrophy (LVH) and LV systolic dysfunction. Whether these cardiac hormones have a similar diagnostic potential in dialysis patients is unknown.
Methods: We studied the diagnostic value of ANP and BNP for alterations in LV mass and function in a cohort of 246 dialysis patients without clinical evidence of heart failure.
Results: Both ANP and BNP were independently related to left ventricular mass (P < 0.0001) as well as to ejection fraction (P < 0.0001). In an analysis based on a prospectively defined threshold (95th percentile of the normal range), BNP had a significantly higher (P < 0.01) sensitivity (88%) than ANP (51%) for the diagnosis of LVH, but the positive predictive value of the two peptides was very similar (92 and 87%, respectively, P = NS). However, the negative predictive value of BNP for excluding LVH was 22% higher than that of ANP (53 vs. 31%, P = 0.05). Both natriuretic peptides had a high sensitivity for the detection of LV dysfunction (87 and 94%), but their positive predictive value was low (25 and 15%). Importantly, both ANP and BNP proved to be very useful for excluding this alteration (negative predictive value 97 and 96%, respectively). An analysis based on the "best cut-offs" of each peptide as identified on the basis of the ROC curves augmented the positive and negative prediction values of BNP for the diagnosis of LVH to 95 and 61%, respectively. This approach also raised the BNP-positive prediction value for the identification of LV dysfunction to 31% but did not modify the diagnostic potential of ANP (either for LVH or for LV dysfunction).
Conclusions: Measuring the plasma concentration of cardiac natriuretic hormones, particularly BNP, may be useful for the identification of dialysis patients with LVH or for excluding systolic dysfunction.