Time to rethink fluids in DKA?

A comment on “Clinical Trial of Fluid Infusion Rates for Pediatric Diabetic Ketoacidosis” by Kuppermann et al. (2018)

Jessica Wall, MD, MPH

In medical school and residency we are taught to be vigilant for mental status changes and cerebral edema in our pediatric patients with diabetic ketoacidosis (DKA).  It is medical dogma that we must rehydrate them gently and slowly to prevent such a devastating complication, but the little evidence that we have for this teaching is observational and/or decades old.  Recently, “Clinical Trial of Fluid Infusion Rates for Pediatric Diabetic Ketoacidosis” was published in The New England Journal of Medicine, with which Nate Kuppermann and the PECARN DKA FLUID Study Group have taken a major first step in changing how we think about fluid administration in DKA.

The structure of their study is elegant in that they completed a large, randomized study in which there was a 2-by-2 factorial design, allowing for the assessment of both fluid type (0.9% sodium chloride or 0.45% sodium chloride) and rate of administration (10% of body weight replaced in a 36 hour regimen or 5% body weight replaced in a 48 hour regimen).  This design allowed for an assessment of both sodium chloride concentration and fast replacement with a higher volume versus slow replacement with a lower volume.

And drum roll please…  there was no significant difference in proportion of patients with a drop in Glasgow Coma Score or neurocognitive assessment after recovery between all four treatment groups.  Interestingly, there was a trend to a higher incidence of clinically apparent brain injury in the slow hydration groups, but this was not statistically significant.   Now there were a few differences between non-neurologic adverse events, mainly higher rates of hyperchloremic acidosis in the 0.9% sodium chloride group and/or rapid administration group, and higher rates of hypocalcemia and hypophosphatemia in the 0.9% sodium chloride groups.  However, the take home point from this study is that isotonic versus hypotonic intravenous hydration and slow versus fast hydration does not affect the incidence of neurologic compromise in diabetic ketoacidosis.  We can hydrate these patients when they are in extremis without having that nagging thought in the back of our minds that we are increasing the risk of cerebral edema.

It is time to rethink our protocols and open discussion with our endocrinology and critical care colleagues about fluids in DKA.



Kuppermann N, Ghetti S, Schunk JE, et al. Clinical Trial of Fluid Infusion Rates for Pediatric Diabetic Ketoacidosis. N Engl J Med. 2018;378(24):2275-2287. doi:10.1056/NEJMoa1716816.

Leave a Reply

Your email address will not be published.