SEVERE CHRONIC HYPONATREMIA: A Pathophysiological Rumination (Part 1)

6551 5

Severe chronic hyponatremia (<120 mEq/L) remains the #1
reason nephrologists lose sleep on call nights and rightly so.  The fear of overcorrection and the risk of
central pontine myelinolysis (CPM) or osmotic demyelination syndrome (ODS),
however uncommon they actually might be, has been drilled into our brains since
the beginning of medical school.
In this post and the few that follow, I will attempt to
address some aspects of chronic severe hyponatremia which have traditionally
not been included when hyponatremia is taught or written about or have only
recently been backed by some evidence and have not yet made their way into the
textbooks. The recently released guidelines
also did not address some of these issues. While this is not, by any means, an
exhaustive discussion of the topic, I hope that these posts will not only help
the readers enhance their understanding of the pathophysiology of severe
hyponatremia but also help them manage it more effectively with a lot less
stress and mental anguish.
The Schrier-Berl algorithm for diagnosis of hyponatremia has
been used successfully for decades for teaching, and for reasoning through the
differential diagnosis at the bedside. 
It is an integral part of every medicine textbook and pocket handbook. It
takes us to our diagnosis through 3 decision points: Plasma osmolarity, volume
status and urine sodium sequentially. 
While the algorithm holds true for the garden variety mild to moderate
hyponatremia, it almost invariably breaks down in case with very severe

During our extensive review of cases of severe hyponatremia
treated with 3% saline, we seldom came across a case in which there was only
one isolated cause for hyponatremia. 
Rather, they were almost always two or more possible etiologies.  In addition, many patients that were
initially considered to be euvolemic by experienced nephrologists, responded to
3% saline as if they were volume depleted. 
Lastly, the clinical course of these patients during the hospitalization
seemed to suggest many of these causes of inappropriate release of ADH were
transient and reversible (SSRIs, acute nausea, postop state etc.). The
patients’ physiology seemed to change from time to time, with overcorrection of
sodium invariably accompanied by large-volume water diuresis as these transient
sources of ADH were “switched off.”
It was, however, the paper by Sood et al that finally looked
at the possible different etiologies for cases of severe hyponatremia as shown
in the table (see image), which I consider the single most important table in
all of recent hyponatremia literature. 
They showed for the first time for multiple etiologies and processes are
at play, some fixed and some transient in generating severe hyponatremia.

It is of paramount importance, that in the workup of
hyponatremia, especially severe cases, we do not limit our reasoning through
the differential diagnosis to comply with the algorithm that we have so
familiar with but constantly look for multiple etiologies and transient causes
of SIADH, especially subclinical volume depletion (low urinary sodium can be
helpful here), as often it is the resolution of these causes that leads to the
large-volume water diuresis and overcorrection. 
Posted by Hashim Mohmand


  1. I agree. Response to a bolus of NS is often very helpful.

  2. Perhaps, but I don't think it would be practical. It is much easier to do a trial of IV NS and see if hyponatremia improves.

  3. Do you think the use of bio-impedance spectroscopy can help stratify patient's volume status and guide treatment in hyponatremia?

  4. Do you think the use of bio-impedance spectroscopy can help stratify patient's fluid status and guide treatment for hyponatremia?

  5. Nice post. The classic algorithm (hypotonic/isotonic/hypertonic/hypovolemic/euvolemic/hypervolemic) is also actually pretty ineffective in diagnosing hyponatremia( and it is very difficult to determine volume status ( I think hyponatremia and all other electrolyte abnormalities should be taught based on pathophysiology and not based on cookbook medicine algorithms. I particularly like the pathophysiological approach using the Edelman equation as the starting point.

Leave a Reply