Abstract 331: Elevated Pre-discharge B-type Natriuretic Peptide Predicts 30-day Heart Failure-specific Rehospitalization

Abstract

Background: Studies support the use of B-type natriuretic peptide (BNP) to guide both the diagnosis and treatment of heart failure. However the variation in clinical use of BNP, and its impact on outcomes outside the clinical trial setting, has not been well characterized. Our aim was to evaluate the clinical use of BNP across the Veterans Affairs (VA) Health Care System, and to identify factors associated with 30-day rehospitalization. Methods: We determined both the frequency and timing of BNP and amino-terminal pro-B-type natriuretic peptide (NT-proBNP). We examined all hospital discharges with the diagnosis of heart failure in the VA Health Care System from 2006 - 2009 (n=109,875). The relationship between pre-discharge BNP and 30 day rehospitalization was evaluated using generalized estimating equations that adjusted for covariates and clustering of patients within facilities. Results: There were 59,720 out of 109,875 patients (54%) with a BNP (50%) or NT-ProBNP (4%) obtained during admission. Use of a BNP or NT-ProBNP test increased from 41% in 2006 to 60% in 2009. Patients evaluated with a BNP test had an older mean age (71 vs, 68 years, p<0.001). Black patients were less likely to have a BNP (50%) compared to non-black patients (58%, p<0.001). There were 125 VA hospitals with at least 80 heart failure discharges in the time period studied. Among these hospitals, the use of BNP (or NT-proBNP) during admission ranged from 0% to 83% of heart failure hospitalizations (median 57%, interquartile range 42% to 72%). BNP within 2 days of discharge (pre-discharge) was measured in 25,988 (24%) heart failure admissions. Higher pre-discharge BNP was associated with a greater risk for rehospitalization with heart failure as the principal diagnosis (Table; p<0.001 for trend). Risk for heart failure readmission tripled from a pre-discharge BNP of 200 to 1000 pg/ml. Pre-discharge BNP was not a significant predictor of other causes of readmission. There were too few NT-proBNP values for comparison. Similar findings were seen after adjustment for patient characteristics. Conclusions: There is significant variability in BNP usage across the VA Health Care System, although use of the test is growing. Pre-discharge BNP is a strong predictor of heart failure readmission but not of other causes of rehospitalization. These data may help clinicians identify high-risk patients prior to hospital discharge, and perhaps indicate a need for more aggressive heart failure-specific therapy and outpatient monitoring.