Page 70 - Biomarkers for risk stratification and guidance in heart failure

Management of chronic heart failure guided by individual NT-proBNP targets.
true nadir for most patients. Findings from observational studies and the control
arm of earlier biomarker-guided treatment trials indicate that in the absence of
clinical deterioration, BNP or NT-proBNP levels continue to fall for months after discharge.5,7,23 The PRIMA investigators report that 80% of subjects achieved
levels lower than their individual target NT-proBNP levels by 1 year and that NT-
proBNP levels were only above target on 23% of occasions during follow-up.
Therefore, on the remaining 77% visits where NT-proBNP levels were at or below
target, up-titration of therapy would not have been triggered, despite the fact that
NT-proBNP levels may have been at levels generally associated with increased 3 risk of adverse events. By comparison, previous studies showing benefits from biomarker-guided therapy used comparatively stringent absolute levels of NT-
proBNP or BNP, and although target NT-proBNP levels were achieved in only 40% to 70% of subjects, up-titration of therapy was triggered on more occasions than in the current study.5-7
Second, the definition of an “off-target” NT-proBNP may have further reduced the occasions when up-titration of therapy for patients in the NT-proBNP group was triggered. The PRIMA study design mandated that to trigger an increase in treatment, NT-proBNP levels at follow-up had to be higher than target level by at least 10% and by a minimum of 850 pg/ml. For one-half of the cohort with NT-proBNP levels at or below the median of 2,491 pg/ml, a rise of 850 pg/ml, required to trigger increasing treatment, actually represents an increase of 30% or more. A patient in PRIMA with a median NT-proBNP level at baseline would have therefore required an NT-proBNP level of 3,350 pg/ml at follow-up to trigger treatment titration, a level that is nearly 3X higher than required in the biomarker- guided arm of BATTLESCARRED (NT-proBNP-Assisted Treatment To Lessen Serial Cardiac Readmissions and Death) trial and between 4X and 8X higher than agestratified targets in TIME-CHF.5,7 As a result, for many patients, this feature of the PRIMA study design may have further limited the opportunity to apply the treatment strategy, possibly diluting its effect. Whether target levels should take account of variability in peptide measurements on serial testing is debatable. In our view, this is unnecessary, especially for NT-proBNP, where analytical variability is very low. Although innate within-patient variability in peptide levels has been recognized in apparently stable patients, the causes are poorly understood. Many contributing factors such as myocardial ischemia or subclinical arrhythmia are clinically relevant and could contribute to adverse outcomes. Regardless,
69