Despite widespread testing, the utility of aminoterminal pro-brain natriuretic peptide (NT-pro-BNP) for diagnosis or risk assessment in patients with diabetes mellitus (DM) in the emergency department (ED) remains unclear. NT-pro-BNP was measured in subjects with dyspnea in the ED. A final diagnosis of acute heart failure (HF) was determined by blinded study physicians using all available hospital records. Vital status was assessed at 1 year; independent predictors of death were identified using Cox analysis. Of 599 subjects, 157 (26.2%) had DM, which was an independent predictor of a final diagnosis of acute HF. In patients diagnosed with acute HF, median concentrations of NT-pro-BNP were similar in patients with and without DM (4,784 vs 3,382 pg/ml, respectively, p = 0.93). In dyspneic subjects without acute HF, median concentrations of NT-pro-BNP were significantly higher in patients with DM (242 vs 115 pg/ml, p = 0.01), but this difference was no longer significant after adjusting for relevant covariates. The area under the curve for NT-pro-BNP to diagnose acute HF in subjects with DM was 0.94 (p <0.001). Using age-adjusted cutpoints, NT-pro-BNP was 92% sensitive and 90% specific for the diagnosis of HF in diabetic subjects. In diabetic patients, a NT-pro-BNP level > or =986 pg/ml was independently associated with an increased risk of death at 1 year (hazard ratio 3.42, 95% confidence interval 1.09 to 10.7, p <0.001). In conclusion, NT-pro-BNP testing offers valuable diagnostic and prognostic information in the evaluation of dyspneic patients with DM in the ED, using identical cutpoints as the population as whole.