We linked data from the ADHERE registry for 2003 through 2006 with Medicare claims to identify patients ?5 years who were hospitalized with HF, did not have AF, and did not receive warfarin at discharge. We estimated rates of all-cause mortality, thromboembolic events, myocardial infarction (MI), and stroke using Kaplan-Meier methods and the cumulative incidence function. We used Cox models to assess associations between log BNP level and each outcome after adjustment for potential confounders.
The study population included 11,679 patients from 146 sites. Patients in the highest quartile of BNP level were older and more often male and African American. They had higher rates of coronary artery disease, renal insufficiency, and peripheral vascular disease and lower rates of diabetes mellitus and chronic obstructive pulmonary disease. After multivariable adjustment, each 30 % increase in BNP level was associated with increased risks of death (hazard ratio 1.07, 95 % CI 1.05-1.08) and MI (1.07, 1.04-1.10) but not thromboembolism or stroke.
Higher BNP level upon admission with HF among older patients without AF was associated with increased risks of MI and mortality; however, higher BNP level was not associated with subsequent thromboembolism or stroke.