I'm not 100% sure about this, but I am of the understanding that RO membranes, at least filmtec anyway, are desalination filters, and as such I have a hard time believing that sodium was getting through the filters.
http://www.dow.com/liquidseps/prod/app_home.htm
According to Dow, sodium is reduced/rejected by their membranes. My gut feeling is if there really was sodium leaking through his membrane in enough quantities to cause problems, it would be a membrane problem.
To answer the question though, this is taken from
http://www.remco.com/ro_quest.htm, and refers specifically to industrial RO:
Q. Is any pretreatment required?
A. There are various pretreatment configurations that will work on the front of an reverse osmosis water system. Part of the selection is based on the capabilities and experience of you maintenance staff. The better preventative maintenance you have, the easier it will be to maintain a chemical addition system. Chemical metering systems require more daily maintenance and calibration to insure consistent operation. Fixed bed systems such as softeners and carbon beds require little daily maintenance.
Water must have a very low silt (solids) content to keep the membranes from plugging up. This can be accomplished by removing the solids or keeping them in suspension while passing through the system. Chemicals can be added to the incoming water to keep the solids in suspension or efficient filtration can be used. We prefer to remove all solids before the system, which results in the lowest rate of membrane plugging.
As the water passes through the reverse osmosis system, the ionic content of the reject stream increases as water permeates the membranes. This increase in TDS can results in calcium and magnesium (the hardness ions) precipitating out in the system and plugging the membranes. Again, either the Calcium and Magnesium can be removed or a chemical can be added to keep them in solution. We prefer using a water softener to remove the hardness ions and replace them with sodium.
Chlorine must be removed for thin film membranes and should be minimal for CTA membranes. Either it can be removed by carbon treatment or reduced with a chemical addition of sodium metabisulfite. The carbon is preferred because the chemical addition can enhance bacterial growth in the system which can plug the membranes.
HTH,
Ben