Many claims have been made in recent years regarding the utility of plasma B-type natriuretic peptide (BNP) concentration measurements in the diagnosis, risk stratification and monitoring of patients with heart failure. This paper summarizes the current evidence and provides guidance for practising clinicians. Overall, plasma BNP testing appears to be of most value in the diagnostic arena, where it is likely to improve the performance of non-specialist physicians in diagnosing heart failure. In clinical practice, BNP testing is best used as a 'rule out' test for suspected cases of new heart failure in breathless patients presenting to either the outpatient or emergency care settings; it is not a replacement for echocardiography and full cardiological assessment, which will be required for patients with an elevated BNP concentration. Although work is ongoing in establishing the 'normal' values of BNP, heart failure appears to be highly unlikely below a plasma concentration of 100 pg/ml. However, as BNP levels rise with age and are affected by gender, comorbidity and drug therapy, the plasma BNP measurement should not be used in isolation from the clinical context.