The purpose of this study was to investigate the feasibility of bicarbonate as a substitute for acetate or lactate in hemofiltration solutions using a new bag for the bicarbonate substitution fluid. We analysed 24 hemofiltration sessions with different HCO3- concentrations (30, 35 and 40 mEq/L) in the substitution fluid. The increase in the HCO3- concentration in the substitution fluid resulted in a more positive HCO3- balance. The Net Base Gains (NBG) were, respectively, 73.7 +/- 92 with 30, 138.2 +/- 97 with 35 (p < 0.05 vs 30) and finally 201 +/- 65.9 with 40 mEq/L (p < 0.001 vs 30). The physical separation between the base losses and gains could facilitate the modelling approach in hemofiltration. By means of a stepwise regression analysis, we studied a series of variables that could influence end-treatment HCO3-, which was significantly and directly dependent (F = 6.003, r = 0.747, p = 0.0027) on the HCO3- concentration in the substitution fluid and the apparent HCO3- space. HCO3- values predicted by the statistical model correlated well with those actually measured (r = 0.757; p < 0.001). This mathematical modelling approach allowed us to predict the quantities and concentrations of HCO3- to be infused in order to obtain an ideal acidosis correction, tailored to individual patient needs.