RESUMO
INTRODUCTION: This prospective study was designed to evaluate whether preoperative plasma brain natriuretic peptide (BNP) could identify significant preoperative cardiovascular disease in elderly hip-fractured patients. PATIENTS AND METHODS: Preoperative plasma BNP measurement and rest transthoracic echocardiography (TTE) were performed within 24 h after admission in consecutive hip-fractured patients aged ≥65 years. The major echocardiographic abnormality (MEA) group included patients with at least one TTE abnormality, defined as systolic pulmonary artery pressure (PAP(s)) ≥50 mmHg, left ventricular (LV) systolic dysfunction, increased LV filling pressure (LVFP) or severe valvular disease. The control group included the remaining patients. RESULTS: Seventy-five patients (mean±SD (range) age=85±5 (69-97) years) were included during a 6-month period. Twenty-four (32%) patients constituted the MEA group (17 elevated PAP(s), three LV systolic dysfunctions, 10 increased LVFP, one severe aortic stenosis and one severe mitral regurgitation). Median (interquartile) preoperative BNP value was significantly greater in MEA than in the control group (527 (361) vs. 119 (154) pg ml(-1); p<0.0001). A preoperative plasma BNP cut-off value at 285 pg ml(-1) predicted well MEA with an area under the receiver operating characteristic (ROC) curve equal to 0.895 (p<0.0001) and with a hazard ratio (HR) (confidence interval, CI) of 23.8 (3.7-142.9) (p=0.0008) on multivariate analysis. The presence of MEA or BNP≥285 pg ml(-1) was associated with high mortality. DISCUSSION: The incidence of echocardiographic signs of elevated PAP(s) or elevated LVFP in elderly hip-fractured patients was high. A preoperative BNP value ≥285 pg ml(-1) can discriminate between elderly hip-fractured patients with or without MEA.