Comparison of dialysate and plasma NTproBNP in prediction of clinical outcomes of diabetic and nondiabetic peritoneal dialysis patients .

Background: Plasma level of N-terminal pro-brain natriuretic peptide (P-NTproBNP) is a useful marker in prediction of mortality in peritoneal dialysis (PD) patients. However, the predictive value of spent dialysate counterpart (D-NTproBNP) of plasma NTproBNP on mortality and dropout is not known.

Materials And Methods: Simultaneous P-NTproBNP and D-NTproBNP assays were performed after an overnight dwell in 44 scheduled ambulatory PD patients. Patients were followed for ~ 47 months. Deceased patients or patients who were transferred to hemodialysis were regarded as dropouts.

Results: 14 patients (31.8%) dropped out at ~ 4 years (9 deaths and 5 transfers to hemodialysis). Diabetics, males, and patients with higher membrane permeability had higher dropout rates. Patients with P- and D-NTproBNP higher than median values had higher mortality and dropout rates (Kaplan-Meier test, log-rank Test p < 0.05). Odds ratios of D-NTproBNP for death and dropouts were (3.807 (0.907 - 15.971), p = 0.068) and (2.87 (1.009 - 8.138) p = 0.048), respectively; odds ratios of P-NTproBNP for death and dropouts were (4.652 (0.914 - 23.693), p = 0.064) and (2.67 (0.924 - 7.716), p = 0.07), respectively; in ROC analysis for death, AUC for P- and D-NTproBNP were 0.762 (0.578 - 0.946, p = 0.016) and 0.765 (0.590 - 0.940, p = 0.015), respectively. Exclusion of diabetic patients from the analyses resulted in significant changes in the predictive value P- and D-NTproBNP. Although death and dropout rates were still higher in nondiabetic patients with higher NTproBNP levels, the differences between groups lost statistical significance.

Conclusions: Both P-NTproBNP and D-NTproBNP are significant predictors of outcomes of interest. Predictive value of NTproBNP might be different in diabetics and non-diabetic CAPD patients.
.