Diabetes mellitus (DM) is a strong risk factor for cardiovascular (CV) disease. Plasma B-type natriuretic peptide (BNP) levels are elevated in various types of cardiac diseases. Increased plasma BNP levels have been reported to be associated with CV risk in apparently healthy individuals. However, no studies have yet examined the specific value of plasma BNP for predicting CV incidence in unselected DM subjects in a community-based population.In a community-based DM cohort (n = 1,059, mean = 66 years), baseline BNP levels were determined, and CV events were followed and captured prospectively. The cohort was divided by plasma BNP quintiles. The Cox proportional hazard model was used to determine the relative hazard ratios (HR) among the quintiles. In addition, the effects of adding the plasma BNP or urine albumin-to-creatinine ratio (UACR) to an established CV risk scoring model was examined by calculating the area under the receiver operating characteristic (ROC) curve (AUC).During the 5.7 year follow-up period, CV events were identified in 65 of the DM cohort. There was a significant association between plasma BNP levels and CV event rate (P < 0.001). HR was significantly increased in the highest quintile compared to the lowest (HR = 4.38; 95%CI 1.69 -11.84). The AUC generated from ROC analysis of the Framingham risk score for predicting general CV events was improved by adding BNP testing (from 0.66 to 0.74; P = 0.05), but not by adding UACR (from 0.66 to 0.67; P = 0.49).In a community sample of people with DM, plasma BNP levels above the 80 percentile are directly associated with CV risk, and measurement of plasma BNP alone or in conjunction with an established CV risk score is of value in predicting CV events in these subjects.