ABSTRACT
BACKGROUND AND AIMS: The diagnosis of heart failure (HF) in elderly patients is often difficult, due to overlap of typical signs and symptoms with those of comorbidities. B-type Natriuretic Peptide (BNP) predicts diagnosis and prognosis of HF, but little is known on its predictive role of short-term prognosis when admission diagnosis is other than HF. METHODS AND RESULTS: We prospectively recruited 404 consecutive patients (aged≥65 years) hospitalized in the Unit of Internal Medicine, University of Catania, Catania, Italy, with an admission diagnosis other than HF. Clinical examination, laboratory data and BNP were evaluated at the admission. The predictive value of BNP and other variables for in-hospital mortality, thirty-day mortality and three month re-hospitalization was assessed. During hospitalization 48 (12%) patients died; by logistic regression analysis, in-hospital mortality was not predicted by BNP>600 pg/ml (OR = 1.36; CI 95% = 0.60-2.80; p = 0.4), while it was by chronic kidney disease (CKD, p < 0.001), WBC count (p < 0.001), immobilization syndrome (p < 0.008) and age (p = 0.012). After discharge, 54 patients (15%) died within 30 days; in these patients thirty-day mortality was significantly predicted by BNP>600 pg/ml (OR = 2.70; CI 95% = 1.40-5.00; p = 0.001), CKD (p < 0.001), malnutrition (p = 0.029) and age (p = 0.033). Re-hospitalized patients were 97 (32%); three month re-hospitalization was predicted by BNP>600 pg/ml (OR = 12.28; CI 95% = 6.00-24.90; p < 0.001) and anamnestic HF (p = 0.002). CONCLUSIONS: Our study shows that BNP>600 pg/ml, CKD, malnutrition and age predict thirty-day mortality after discharge in elderly patients with an admission diagnosis other than HF, while CKD, WBC count, immobilization syndrome and age predict in-hospital mortality. Three-month re-hospitalization was predicted by BNP>600 pg/ml and anamnestic HF.
