Thirty day post discharge BNP levels predict outcomes better than discharge levels - Italian RED study

Introduction: Heart failure (HF) patients often relapse into acute decompensation and consequent rehospitalization resulting in social and economic burdens. BNP has been previously shown to be prognostic at hospital discharge, but it has not been well characterized at time points after discharge from the emergency department (ED). Hypothesis: To determine the prognostic utility of BNP levels and changes in the emergency department (ED) and follow-up time-points. Methods: 276 HF patients were enrolled from 8 centers in Italy. The primary endpoints were cardiac associated rehospitalizations and deaths. Results: The Mean/Median BNP concentration at ED presentation was (mean±SE) 707.94±23.44pg/ml. Changes in BNP levels from presentation to 24h: –549.54±26.91pg/ml (p<0.001); to 72h: –426.57±26.91pg/ml (p<0.001); and at discharge: –323.59±29.51pg/ml (p<0.001). The mean/median BNP levels at follow-up time-points were 407.38±24.54 at 30 days, 363.07±29.51 at 60 days and 281.83±26.91pg/ml at 180 days. The BNP level at the 30 day follow-up was found to be highly predictive for events up to 180 days with an Area Under Curve (AUC) of 0.716 (p< 0.008) with a cut-point of 355pg/ml BNP and a hazard ratio of 8.56 (p = .014). The other considered BNP measurements did not significantly contribute to the prediction of the outcome. The discharge BNP level failed to significantly predict short term (within 30 days) outcomes (AUC: 0.647; p=0.054), and shows similar prognostic ability for long term (180 days) (AUC: 0.605; p<0.039).Conclusions: Tracking BNP levels post-discharge from the hospital may be equally important as the discharge BNP. Our results suggest that an increase in BNP at 30 days from the discharge BNP can be highly prognostic and may be a useful tool to discriminate between patients at higher risk for future events. This time period may be a crucial window to monitor BNP stability in the outpatient setting.