Heart Failure is one of the fastest growing cardiovascular diseases of the 21st century. Echocardiogram is considered the gold standard for diagnosis, but is costly, time consuming and not readily accessible to all patients. Our aim was to assess the diagnostic utility of BNP to risk stratify patients for ECHO. Seventy-four GP referred, non-pregnant patients of > or = 18 years with a working diagnosis of HF were recruited. Patients were given two appointments to attend the Cardiology Department and at each, were examined by the same cardiologist, had their medications recorded and blood drawn for BNP analysis. ECHO was performed at the second visit. The diagnosis of HF was confirmed in 49 of 74 patients (66%). The clinical utility of BNP to rule-in HF was evaluated using ROC curve analysis. The AUC was satisfactory at 0.691 (C.I. 0.573-0.793). The positive likelihood ratio (+LR) was 5.87, negative likelihood ratio (-LR) was 0.58, the positive predictive value was 92% and a negative predictive value was 47%. One-third of patients (n = 25) had a BNP >178 pg/mL, 23 of whom had HF confirmed. At this decision threshold BNP correctly classified 23 of 25 patients who were confirmed not to have HF (Specificity for HF of 92%). A BNP of > or = 178 pg/mL can be used to prioritise GP patients for ECHO.