All 4 NPs were measured in patients admitted to the emergency unit with SOB (in 2 centers) or acute heart failure (AHF) (1 FINN-AKVA cohort) and in a control population of stable chronic HF. Follow-up was 1 (2 centers) and 5 years (1 FINN-AKVA cohort). Area under the curve (AUC) was used to assess diagnostic properties. AUC, multivariate Cox regression, net reclassification improvement (NRI), and Kaplan-Meier analyses were used to assess mortality.
We included 710 patients (¡°Biomarcoeurs¡± cohort n = 336; FINN-AKVA study, n = 306; stable chronic HF, n = 68). Pro-BNP was almost as powerful as BNP to diagnose AHF (AUC 0.953 vs 0.973 respectively, p = 0.003), NT-proBNP also performed well (0.922, p < 0.001 vs BNP). MR-proANP performed less well (0.901). AUC over time showed greater MR-proANP values over the first year. At 5 years, MR-proANP had the best prognostic value (AUC 0.668 vs 0.604 for BNP, p = 0.042). Kaplan Meier analysis confirmed better survival with MR-proANP ¡Ü 416.8 pmol/L at 5 years. NRI at 5 years was greater for MR-proANP (0.23, p < 0.05) than for proBNP, BNP or NTproBNP (p = NS).
Our study provides firm evidence that all NPs perform equally well for diagnostic purposes, and that MR-proANP has long term prognostic value in patients with acute heart failure.