BACKGROUND: Patients complaining of chest pain are frequently referred to secondary care, although the majority have pain of non-cardiac origin. AIM: To investigate whether B-type natriuretic peptide (BNP) levels are predictive of a diagnosis of non-cardiac pain. DESIGN: Cross-sectional study. METHODS: Consecutive patients (n = 296) presenting to a rapid-access chest pain clinic (RACPC) received the usual clinical assessment plus near-patient BNP testing, with the assessor blinded to the result. After clinical assessment (including exercise stress testing if clinically indicated), pain was diagnosed likely/definitely cardiac or non-cardiac. RESULTS: Median BNP was higher in those diagnosed with likely/definite cardiac chest pain (26.5 vs. 8 pg/ml) (p < 0.0001, Wilcoxon rank sum test). The odds ratio for cardiac pain in those with BNP <20 pg/ml was 0.25 (95%CI 0.14-0.47) (p < 0.0005); adjusting for age and sex reduced this to 0.41 (95%CI 0.20-0.83) (p = 0.01). The area under the curve (AUC) for the model including BNP, age and sex was 0.70. With BNP as a continuous variable, the AUC for the same model was 0.72. DISCUSSION: In typical patients presenting to a RACPC, those with a BNP < or =20 pg/ml were significantly less likely to be diagnosed with cardiac pain. Near-patient BNP testing may have a role as a 'rule out test' for angina in patients presenting to a RACPC.