Dialysis adjustment for hyponatremia

This has only ever come up once, in Question 24.3 from the first paper of 2019. The  question put forward a scenario where the patient is severely hyponatremic, and needing dialysis. How do you plan to protect his myelin, they asked. Name three ways.

Avoid dialysis altogether. It is possible to skirt around dialysis in some situations. Perhaps this is one of those? Even those rare situations where urgent dialysis is strongly indicated (severe hyperkalemia, acute pulmonary oedema) can sometimes be addressed with something other than dialysis. If you aren't able to put it off forever, at least you can delay it, so that you can - 

Give sodium, then dialyse. If you are able to delay dialysis for some period of time, then this period of time should be used to correct the sodium. Every 24 hours is a potential rise by 12mmol/L. If one can spare 48-72 hours, it will allow the sodium to come up to a level where demyelination will not be such a massive risk (eg. 125mmol/L or so). After that, it will be possible to carry on with dialysis as usual.

Use low dose dialysis. The dialysate flow rate can be turned down to the point where electrolyte exchange across the membrane is slow and gentle. The downside of this would be slower solute clearance, potentially so slow that it has no clinical benefit. As sodium ions are more likely to exchange across the membrane than the relatively larger urea molecules, one may find oneself in a position were no useful metabolic waste clearance has taken place, but the patient's sodium has corrected dangerously quickly anyway. Even if this dangerous outcome does not occur, one could still advance the logical argument that having pointless dialysis has zero advantage over having no dialysis.

Give systemic 5% dextrose. The patient will gain sodium via the circuit, so you give them water at the same time. This works just fine, so long as you keep giving enough 5% dextrose to keep up with the sodium exchange rate. With the dialysate flow low enough and the dextrose rate high enough, you should be able to achieve any desired rate of rise of sodium. Of course, this is not without its disadvantages. For one, that's 66 mmol/L of dextrose you're giving, which will have an undesirable effect on their BSL. Secondly, that's potentially several litres of extra fluid you are giving, which can't possibly be benign for this patient's circulatory system. Obviously at some later stage you will have to take that water out via the circuit, as the renally incompetent patient is probably not up to excreting it themselves. And when you take it out with the circuit, the sodium will go up again.  

Give sterile water or 5% dextrose into the return line of the circuit. That's functionally the same step as above, but one less access line.

Alter the sodium content of the dialysate. This takes some people out of their comfort zone, as one needs to create one's own bags of sterile dialysate. To quote the case report by Bender et al (1998)"Each 2 liters of dialysate was prepared by mixing 1 liter of 5% dextrose and 0.45% NaCl (half normal saline) with 1 liter of 0.675% NaCl (three-quarters normal saline) to which 50 mEq NaHCO3 was added; the final sodium concentration was thus 121 mEq/L"

This, the simplest of javascript forms, will help you calculate the fluid which you need to remove from each bag and then add as sterile water. Just change the "desired sodium concentration" value and the form will automatically update. Theoretically you could also mess with the dialysate bag volume and sodium content, but practically almost all commercially available dialysate comes in 5L bags and contains 140 mmol/L sodium.

Volume of each dialysate bag (L)
Sodium content of each dialysate bag (mmol/L)
Desired sodium concentration (mmol/L)
Volume to be removed from each bag,
and then added as sterile water (ml)


From a practical perspective, this might sound easy to prescribe, but is difficult for nurses to execute. After some minutes of painfully aspirating and discarding fluid using a 50ml syringe (and potentially losing count of the number of 50ml volumes discarded), they will usually come to the conclusion that it would be easier to empty and then fill the dialysate bag using a standard peristaltic drug pump, which can be easily programmed with a volume, and which can move a litre of fluid every forty minutes.

Another factor to remember is that the dilution of dialysate bags will also unhelpfully decrease the concentration of the other electrolytes. It is easily forgotten about in a rich Western ICU where plenty of ABGs are available to track the electrolytes, but  it could easily sneak up on somebody who does not have access to quick point of care testing. 

Use SLED: SLED prescriptions can usually be amended to include whatever electrolyte concentration you want (within reason). If the patient can tolerate SLED from a circulatory perspective, this could be the most convenient method, and one which will not make the ICU pharmacist anxious in the same way as violating bags sterile dialysate would. If at all possible, this should be the option of first choice.


Bender, Filitsa H. "Successful treatment of severe hyponatremia in a patient with renal failure using continuous venovenous hemodialysis." American journal of kidney diseases 32.5 (1998): 829-831.

Wendland, Erik M., and Andre A. Kaplan. "A Proposed Approach to the Dialysis Prescription in Severely Hyponatremic Patients with End‐Stage Renal Disease." Seminars in dialysis. Vol. 25. No. 1. Oxford, UK: Blackwell Publishing Ltd, 2012.

Rosner, Mitchell H., and Michael J. Connor. "Management of severe hyponatremia with continuous renal replacement therapies." Clinical Journal of the American Society of Nephrology 13.5 (2018): 787-789.