This study was performed to determine whether the N-terminal pro-brain natriuretic peptide (NT-pro-BNP) is a useful biomarker in predicting cardiovascular congestion, mortality, and cardiovascular death and event in chronic peritoneal dialysis (PD) patients. A prospective cohort study was conducted in 230 chronic PD patients in a dialysis unit of a university teaching hospital. Serum NT-pro-BNP was measured at baseline together with echocardiography and dialysis indices. Each patient was followed for 3 yr from the day of enrollment or until death. Time to develop first episode of cardiovascular congestion and other cardiovascular event and time to mortality and cardiovascular death were studied in relation to NT-pro-BNP. NT-pro-BNP showed the strongest correlation with residual GFR, followed by left ventricular ejection fraction and left ventricular mass index. In the univariate Cox regression model, NT-pro-BNP was a significant predictor of cardiovascular congestion, mortality, and cardiovascular death and event. In the fully adjusted multivariable Cox regression analysis that included residual GFR, left ventricular ejection fraction, and left ventricular mass index, the hazard ratios for cardiovascular congestion, mortality, composite end point of mortality and cardiovascular congestion, and cardiovascular death and event for patients of the fourth quartile were 4.25 (95% confidence interval [CI] 1.56 to 11.62; P = 0.005), 4.97 (95% CI 1.35 to 18.28; P = 0.016), 5.03 (95% CI 2.07 to 12.26; P < 0.001), 7.50 (95% CI 1.36 to 41.39; P = 0.021), and 9.10 (95% CI 2.46 to 33.67; P = 0.001), respectively, compared with the first quartile. These data showed that NT-pro-BNP is an important risk predictor of cardiovascular congestion, mortality, and adverse cardiovascular outcomes in chronic PD patients and adds important prognostic information beyond that contributed by left ventricular hypertrophy, systolic dysfunction, and other conventional risk factors.