Page 69 - Biomarkers for risk stratification and guidance in heart failure

Chapter 3
proBNP of >10% and at least 850 pg/ml during hospitalization (not a requirement of earlier studies). For the 174 subjects randomized to the NT-proBNP-guided group, an individualized target NT-proBNP level was identified based on the lowest NT-proBNP level obtained at discharge or within 2 weeks after discharge. For these subjects, up-titration of treatment was triggered if their NT-proBNP level at scheduled 3-monthly visits was >10% and at least 850 pg/ml above their individual baseline level. For the 171 subjects in the comparator clinically-guided group, treatment was up-titrated on the basis of standard clinical assessment. After a median follow-up of 702 days, no difference was found in the primary end point of days alive and out of hospital, despite the fact that there was greater up- titration of treatment in the NT-proBNP-guided group— especially in the use of inhibitors of the renin-angiotensin system and in diuretic dosing—and despite the fact that 80% of subjects achieved NT-proBNP levels below their individualized target level at 1-year follow-up. The investigators did observe fewer deaths in the NT-proBNP-guided group, particularly in patients younger than age 75 years and in subjects with a left ventricular ejection fraction below 45%, although none of these comparisons achieved statistical significance. Neither was there any statistically significant difference between groups in hospitalization rates, quality of life scores, or estimated glomerular filtration rates.
The PRIMA study is the largest to study an individualized NT-proBNP target, and the investigators should be congratulated for addressing this question and undertaking such a comprehensive study of this strategy. Why did the PRIMA study not show a significant benefit from treatment guided by an individualized NT-proBNP target? The first possibility is that there is indeed no significant benefit to be gained by the approach taken in this study. Perhaps more likely is the possibility that aspects of the PRIMA study design may have served to obscure benefits of biomarker guidance.
First, the study highlights 1 potential limitation of treatment based on an individualized NT-proBNP target. If the target is derived from a BNP or NT-proBNP level that is set too high and therefore is not a reasonable estimate of the nadir, this will reduce the occasions when up-titration of therapy is prompted. In the PRIMA study, the NT-proBNP target level was based on the lowest plasma level for an individual either at baseline or 2 weeks after hospital discharge. The timing of this assessment may partly explain the relatively high median target level of 2,491 pg/ml. It is likely that levels measured at this time would not represent the
68