cepia

<p><b>AIMS: </b>To evaluate the interest of brain natriuretic peptide (BNP) for heart failure (HF) diagnosis in very old patients.</p><p><b>METHODS AND RESULTS: </b>A total of 383 patients aged 80 years or older, hospitalized in geriatrics care for dyspnoea, had a BNP measurement at the acute phase. Independent cardiologists blinded to BNP values classified the patients into cardiac vs. respiratory aetiology according to the European Society of Cardiology guidelines. Mean (SD) age was 88.5 (5.4) years, 66% (n = 254) of patients were women, 62% (n = 238) had cardiac dyspnoea and 38% (n = 145) had respiratory dyspnoea. The BNP levels were significantly higher in the cardiac group (median = 385.5 ng/L, interquartile range = 174.0-842.0) than in the respiratory group (median = 172.0 ng/L, interquartile range = 70.8-428.0; P < 0.001). On its own, BNP showed poor discriminative ability [area under the curve (AUC) = 0.68; 95% confidence interval (CI) 0.62-0.73] for the diagnostic. In multivariate analysis, BNP remained independently associated with the cardiac aetiology after full-adjustment (odds ratio 1 log increase = 1.87; 95% CI 1.28-2.74), but did not improve the discrimination between the cardiac and the respiratory aetiologies (ΔAUC = 0.013, P = 0.16). In addition, although adding BNP to the other predictive covariates yielded a significant continuous NRI of 57.8% (95% CI 31.2-83.5%), the mean changes in individual predicted probabilities were too low (<3%) to be clinically relevant.</p><p><b>CONCLUSION: </b>In this population of very old patients with acute dyspnoea, despite being independently associated with the cardiac aetiology, BNP was not useful for better discriminating cardiac vs. respiratory origin.</p>