Aim: To examine prospectively the relationship between plasma B-type natriuretic peptide (BNP) levels in community-dwelling elderly and their hospitalization.
Methods: A total number of 644 subjects aged 65 years or older were recruited from the annual community health examinations. Those with a history of stroke or neurological findings were not included. After excluding those with old myocardial infarction, left ventricular dysfunction, moderate or severe valvular disorders, atrial fibrillation, renal insufficiency, and history of hospitalization within 1 year, 602 participants (226 men, 376 women; mean age, 80.3 +/- 6.2 years) remained eligible for this study. Antihypertensive medications, activities of daily living (ADL) score and history of hospitalization were assessed by annual interview. Measurement of casual blood pressure, Mini-Mental State Examination, electrocardiography and echocardiography were performed. Plasma BNP, serum creatinine, total cholesterol, albumin and hemoglobin A1c levels were also examined. A follow-up survey was performed for the occurrence and reasons for hospitalization.
Results: During a median follow up of 37 months, 112 subjects were hospitalized. After adjustment for conventional risk factors of hospitalization using the Cox proportional hazard model, each increment of 1 standard deviation in log BNP levels was associated with a 36% increase in the risk of hospitalization (P = 0.02). Plasma BNP levels were significantly higher in the hospitalized subjects due to stroke, heart diseases, dementia, pneumonia and also difficulty to live alone than those of the subjects without hospitalization.
Conclusion: Plasma BNP level is a very useful biochemical marker predictive of future hospitalization in community-dwelling independent elderly people without apparent heart diseases.