Page 113 - Biomarkers for risk stratification and guidance in heart failure

Chapter 5
Composite NT-proBNP Score
By combining the independent prognostic NT-proBNP parameters (ie, inpatient decrease in NT-proBNP concentration below vs above the median, early outpatient increase vs decrease in NT-proBNP, and NT-proBNP concentration at 1 month after discharge above vs below the median), a composite NT-proBNP score was formed by giving 1 point for each parameter that was elevated. This resulted in patients being divided into 4 groups: from 0 parameters elevated (group 1) to 3 parameters elevated (group 4). Baseline characteristics of all groups are depicted in Table 1. With increasing NT-proBNP parameters elevated, patients were older, more often had a previous episode of heart failure, and more often had ischemic cause of HF. Increasing composite NT-proBNP score was also associated with lower blood pressure and more impaired renal function. Interestingly left ventricular ejection fraction did not differ among the 4 groups.
The composite NT-proBNP score strongly predicted events: both short- and long-term prognosis differed significantly among the 4 groups regarding mortality and the combined end point HF-related readmission or mortality (Table 2; Fig. 1). All patients without NT-proBNP parameters elevated (group 1; n= 62) survived 1 year follow-up, whereas 41% of patients with all 3 NT-proBNP parameters elevated (group 4; n= 70) died within 1 year of follow-up.
Prognostic Impact of Small Outpatient Changes in NT-proBNP Concentration
Small changes in NT-proBNP concentration are associated with outcome. In multivariate analysis including the reference model, inpatient change in NT- proBNP, and NT-proBNP concentration at 1 month after discharge, early outpatient increase in NT-proBNP concentration <30% was associated with worse outcome compared with early outpatient decrease in NT-proBNP <30% (HR for mortality 2.05, 95% CI 1.02-4.13, Wald 4.1 \[P=.04\], HR for the combined end point 2.59, 95% CI 1.45-4.64, Wald 10.2 \[P= .001\]). Interestingly, there was no significant difference in mortality or the combined end point between patients with an early outpatient decrease of <30% vs >30% (HR for the combined end point 1.04, 95% CI 0.50-2.18, Wald 0.01; P= .914). Likewise, an increase in NT-proBNP concentration of <30% yielded a clinically similar hazard for events compared with an increase >30% (HR for the combined end point 0.96, 95% CI 0.62-1.47, Wald 0.04; P= .837).
112