March 13, 2013 by EmerJencyWEBB
I’ve recently had several head-scratching moments when discussing admission for patients who just so happen to have diabetes, and are hyperglycemic on admission. Lab testing reveals no concern for DKA. The conversation with the IM resident usually goes like this…
IM resident: “so…did you see that their glucose is 400”
Me: “yes, they are getting some fluid and sub-cutaneous insulin per their correction factor”
IM resident: “did you know they have an elevated anion gap?”
Me: “Their anion gap is 8…” (?)
IM resident: “well, if you took the time to correct their sodium, you would see they actually have an anion gap of 18.”
For those of you who have run into the same scenario, and have become both confused as well as a little angry at the assertion, I have a great chart and article for you to back up your claim this is only simple hyperglycemia AND that your anion gap is actually normal.
The table below shows that even though the sodium level may change based on hydration status, it is roughly equal in relation to the change of measured anions.
This nice summary from the Cleveland Clinic Journal of Medicine summarizes:
Hyponatremia or hypernatremia (due to water gain or loss) affects the anion gap, but only slightly. Table 1 demonstrates the theoretic effect on the normal anion gap in the case of a 10% gain or 10% loss of water. Since all of the components of the anion gap calculation are concentrated or diluted to the same extent, the anion gap is minimally altered—too little to interfere with proper interpretation of the acid-base status.
So, do not correct the sodium level due to hyperglycemia when calculating anion gap. It will always skew your results if you do. The article ONLY recommends using corrected sodium when evaluating hydration status.
Enjoy the read, and the next time you find yourself in this situation, you can correct the other person’s knowledge gap.
Beck, LH. Should the actual or the corrected serum sodium be used to calculate the anion gap in diabetic ketoacidosis? CLEVELAND CLINIC JOURNAL OF MEDICINE 2001; 68 (8) 673-674.