Background: The role of serial NT-proBNP measurements in patients suffering from chronic systolic heart failure (CHF) who already receive individually optimized pharmacotherapy is still unresolved.
Methods: NT-proBNP was assessed at baseline and at 6 months follow-up in 504 stable CHF patients treated with individually optimized pharmacotherapy. After assessment of clinical stability at 6 months, patients were followed up for at least 1 year. The combined primary endpoint was defined as death, hospitalization due to cardiac reasons or heart transplantation in 1-year follow-up. We stratified our patients according to two principles: first, a percent change of value (CV) between the first and second measurement of NT-proBNP and secondly, the transformed logarithm of NT-proBNP measured at 6 months.
Results: During the follow-up period of 1 year, 50 patients (9.9%) reached the combined primary endpoint. Stratification according to percentage CV was less accurate in predicting endpoint-free survival compared to a classification in categories of lnNT-proBNP measured at 6 months (ROC AUC = 0.615; 95% CI 0.525-0.70 vs. ROC AUC = 0.790; 95% CI 0.721-0.856, respectively). When entered into proportional hazard regression analysis, lnNT-proBNP measured at 6 months remained an independent predictor of the combined primary endpoint with an associated HR of 2.53 (95% CI 1.385-4.280).
Conclusion: To date, this is the largest analysis of serial NT-proBNP measurements in patients with CHF receiving individually optimized medical therapy. These data suggest that a single NT-proBNP measurement after 6 months in stable clinical conditions may have higher predictive value than stratification of change in serial measurements.