1Universitätsklinikum Regensburg Klinik und Poliklinik für Innere Med. II, Kardiologie Regensburg, Deutschland; 2Universitätsklinikum Regensburg Lehrstuhl für Genetische Epidemiologie Regensburg, Deutschland; 3Universitätsklinikum Regensburg Institut für Klinische Chemie und Laboratoriumsmedizin Regensburg, Deutschland; 4Krankenhaus Barmherzige Brüder Regensburg Klinik für Kardiologie Regensburg, Deutschland
Background: NT-proBNP is an established marker of heart failure and a predictor of mortality in the general population. However, data on reference values and risk prediction in the elderly remains scarce, despite of their increasing risk of heart failure diagnosis and death.
Aim: We set out to derive reference values of NT-proBNP, to define the specificity for heart failure diagnosis, and to assess NT-proBNP as risk predictor for mortality in the mobile elderly population.
Methods: In a population-based cohort study (AugUR), inhabitants at least 70 years of age were recruited. The local residents’ registers provided a random sample of inhabitants, who were invited by mail. NT-proBNP was determined in serum. Cardiac morphology and function were evaluated by echocardiography (left ventricular mass, ratio of the E and e’ waves, ejection fraction). Survival status was queried from the data of the residents' registers (censoring at date 2022/01/19). Causes of death were collected from death certificates.
Results: A total of 1.014 subjects aged 70 to 95 years were included at baseline (55% men). NT-proBNP values were higher in female subjects and increased with age (NT-proBNP median for women 70-74/75-79/80-84/85-89/>90years: 171.6/201.3/319.9/421.9/541.4pg/ml; for men 70-74/75-79/80-84/85-89/>90years: 148.8/175.4/244.0/456.3/407.2pg/ml). European guidelines endorse a rule-out value of <125pg/ml to exclude heart failure in a non-acute setting. In our study, 69.2% of all subjects showed values above this cut-off. In 320 apparently healthy participants free of cardiovascular risk factors (e.g., diabetes, obesity, hypertension), renal failure, coronary artery disease and atrial fibrillation, 55% (70 to 79 years) and 80% (> 80 years) still showed NT-proBNP values above 125pg/ml. In a subgroup of 73 subjects, in whom additionally left ventricular mass, E/e’-ratio and ejection fraction were measured and normal, 51% of subjects (70 to 79 years) or 71% (>80 years) were determined with false-positive results for heart failure based on NT-proBNP. Together, the specificity of the established cut-off value for heart failure was poor in the elderly population.
After a median follow-up time until death or censoring of 6.6 years (up to 8.5years), 171 of the 1.1014 subjects had died. Using Cox regression analyses, mortality risk significantly increased with increased baseline NT-proBNP values (hazard ratio HR=1.60 per unit increment in log(NT-proBNP) adjusted for age and sex; 95% confidence interval CI=1.42-1.81; p<0.001). The increased risk persisted after further adjustment for cardiovascular risk factors, comorbidities (e.g., renal failure, cardiovascular disease), left ventricular diastolic and systolic dysfunction, left ventricular hypertrophy, left atrial enlargement and atrial fibrillation (HR=1.50; 95%CI=1.22-1.84; p<0.001). Using the same model of adjustments, the risk for cardiovascular death was slightly higher (HR=1.79; 95%CI=1.26-2.54; p<0.001).
Conclusions: In the elderly population, NT-proBNP levels continue to increase with age, also among “healthy” elderly. Accordingly, specificity of the established, age-independent exclusion cut-off value of < 125pg/ml for chronic heart failure is poor in old-aged subjects. However, NT-proBNP remains a strong predictor for all-cause and cardiovascular mortality in the elderly population, independent of age, cardiovascular risk factors, comorbidities, atrial fibrillation, diastolic and systolic dysfunction.