Background: Amino-terminal pro-B-type natriuretic peptide (NT-proBNP) level is useful to diagnose or exclude acutely decompensated heart failure (ADHF) in dyspnoeic patients presenting to the emergency department (ED). Aim: To evaluate the impact of ED NT-proBNP testing on admission, length of stay (LOS), discharge diagnosis and longterm outcome. Methods: Dyspnoeic patients were randomized in the ED to NT-proBNP testing. Admission and discharge diagnoses, and outcomes were examined. Results: During 17 months, 470 patients were enrolled and followed for 2.0±1.3 years. ADHF likelihood, determined at study conclusion by validated criteria, established ADHF diagnosis as unlikely in 86 (17%), possible in 120 (24%), and likely in 293 (59%) patients. The respective admission rates in these subgroups were 80, 91, and 96%, regardless of blinding, and 61.9% of blinded vs. 74.5% of unblinded ADHF-likely patients were correctly diagnosed at discharge (p=0.029), with similar LOS. 2-year mortality within subgroups was unaffected by test, but was lower in ADHF-likely patients with NTproBNP levels below median (5000 pg/ml) compared with those above median (p=0.002). Incidence of recurrent cardiac events tracked NT-proBNP levels. Conclusion: ED NT-proBNP testing did not affect admission, LOS, 2-year survival, or recurrent cardiac events among study patients but improved diagnosis at discharge, and allowed risk stratification even within the ADHF-likely group. (ClinicalTrials.gov#NCT00271128)