Aims: Cardiac disease is the major cause of death in patients undergoing chronic haemodialysis. Recent studies have found that B-type natriuretic peptide (BNP) levels accurately reflect the cardiovascular burden of dialysis patients. However, the prognostic potential of BNP measurements in dialysis patients remains unknown.
Methods and results: The study included 113 chronic dialysis patients who were prospectively followed up. Levels of BNP were measured at baseline and every 6 months thereafter. The potential of baseline BNP and annual BNP changes to predict all-cause and cardiac mortality were assessed as endpoints. Median follow-up was 735 (354-1459) days; 35 (31%) patients died, 17 (15%) of them from cardiac causes. Baseline BNP levels were similar among survivors and non-survivors, and failed to predict all-cause and cardiac death. Cardiac death was preceded by a marked increase in BNP levels. In survivors BNP levels remained stable [median change: +175% (+20-+384%) vs. -14% (-35-+35%) over the 18 months preceding either death or the end of follow-up, P< 0.001]. Hence, annual BNP changes adequately predicted all-cause and cardiac death in the subsequent year {AUC(all-cause) = 0.70 [SD 0.05, 95% CI (0.60-0.81)]; AUC(cardiac) = 0.82 [SD 0.04, 95%CI (0.73-0.90)]}. A BNP increase of 40% provided the best cut-off level. Cox regression analysis confirmed that annual increases over 40% were associated with a seven-fold increased risk for all-cause and cardiac death.
Conclusions: Annual BNP increases above 40% predicted all-cause and cardiac death in the subsequent year. Hence, serially measuring BNP levels may present a novel tool for risk stratification and treatment guidance of end-stage renal disease patients on chronic dialysis.