Factors affecting sodium determinations with direct potentiometry were assessed in 15 patients on dialysis to verify if the method allows measuring the effective gradient of ionic activity of Na between plasma water and dialysate, thus to calculate its transfer by diffusion and its net balance of the session. As expected, changes in the dialysate sodium activity were related to changes in the bicarbonate content and in the ionic strength of the bath. In plasma, the protein/lipid volume displacement effect prevailed over the other factors, although their cumulative effects failed to fully explain the observed change in sodium activity. The bias can be explained by taking into account technical causes inherent in the methodology (mismatch of activity coefficients and liquid junction potentials between calibrants and plasma, inadequate calibration, etc.) and affecting its results, to a greater extent in plasma. However, when properly designed instruments are used, technical interference is negligible on the results of sodium kinetic models applied to HD and HDF, while they may compromise the results of AFB.