Change of BNP between admission and discharge after ST-elevation myocardial infarction (Killip I) improves risk prediction of heart failure, death, and recurrent myocardial infarction compared to single isolated measurement in addition to the GRACE score.

Abstract

In ST-elevation myocardial infarction, 7-15% of patients admitted as Killip I will develop symptomatic heart failure or decreased ejection fraction. However, available clinical scores do not predict the risk of severe outcomes well, such as heart failure, recurrent myocardial infarction, and sudden death in these Killip I individuals. Therefore, we evaluated whether one vs two measurements of BNP would improve prediction of adverse outcomes in addition to the GRACE score in ST-elevation myocardial infarction/Killip I individuals. Consecutive patients with ST-elevation myocardial infarction/Killip I (n=167) were admitted and followed for 12 months. The GRACE score was calculated and plasma BNP levels were obtained in the first 12 h after symptom onset (D1) and at the fifth day (D5). Fifteen percent of patients admitted as Killip I developed symptomatic heart failure and/or decreased ejection fraction in 12 months. The risk of developing symptomatic heart failure or ejection fraction <40% at 30 days was increased by 8.7-fold (95% confidence interval: 1.10-662, p=0.046) per each 100 pg/dl increase in BNP-change. Both in unadjusted and adjusted Cox-regressions, BNP-change as a continuous variable was associated with incident sudden death/myocardial infarction at 30 days (odds ratio 1.032 per each increase of 10 pg/dl, 95% confidence interval: 1.013-1.052, p<0.001), but BNP-D1 was not. The GRACE score alone showed a moderate C-statistic=0.709 (p=0.029), but adding BNP-change improved risk discrimination (C-statistic=0.831, p=0.001). Net reclassification confirmed a significant improvement in individual risk prediction by 33.4% (95% confidence interval: 8-61%, p=0.034). However, GRACE +BNP-D1 did not improve risk reclassification at 30 days compared to GRACE (p=0.8). At 12 months, BNP-change was strongly associated with incident sudden death/myocardial infarction, but not BNP-D1. Only BNP-change following myocardial infarction was associated with poorer short- and long-term outcomes. BNP-change also improves risk reclassification in addition to the GRACE score.