Sodium removal is one of the main factors affecting intradialytic cardiovascular stability and interdialytic hypertension, and its removal should therefore be individualized. The aims of this study were: (1) to test the ability of a single-pool variable volume (SPVV) sodium kinetic model (NaKM) to optimize sodium removal in paired filtration dialysis (PFD), and (2) to test a SPVV conductivity kinetic model (CKM) in order to verify whether CKM can be used as an alternative for NaKM in estimating sodium balance. The mean difference between the NaKM-predicted and measured end-PFD plasma water ionized sodium concentrations was 0.00 +/- 0.55 mEq/l, which means that the model has an imprecision of < or = 1.1 mEq/ l. The mean difference between predicted and measured sodium removal was 0.21 +/- 16.86 mEq/session, which means a model overestimate of 0.21 mEq/session. The mean difference between the CKM-predicted and measured end-PFD ultrafiltrate conductivity was 0.01 +/- 0.05 mS/cm, which means an inaccuracy of the model of 0.01 mS/cm and an imprecision of < or = 0.1 mS/cm. The regression in the ionized sodium concentration measured in plasma or blood on the conductive values of the ultrafiltrate shows an error of < or = 2 mEq/l in the prediction of the ionized sodium concentration in blood by means of ultrafiltrate conductivity measurements. These results demonstrate that both models make it possible to obtain a level of dialytic sodium removal that is almost equivalent to interdialytic sodium loading. Moreover, given that it does not require blood sampling and the possibility of making repeated and inexpensive ultrafiltrate conductivity measurements, the CKM allows online monitoring of programmed sodium removal.