• diabetic ketoacidosis
• diabetes
• fluid management
• cerebral oedema

Inward and Chambers¹ have called for a rethink of the management of diabetic ketoacidosis. In their article they quote a study by Grove L M and colleagues² suggesting that pediatricians overestimated the quantum of dehydration in DKA. Over correction of dehydration is implicated in precipitating cerebral oedema. On the face of it, it seems implausible that pediatricians who are so adapt at estimating dehydration in the context of gastroenteritis, diarrhea, and vomiting should err in estimating the dehydration in DKA, unless the dehydration of DKA has special features. Hypertonicity may be that special feature. We hypothesize that hypertonic dehydration can result in the tongue appearing dry and parched and when this is combined with acidotic respiration of DKA, the treating pediatrician may classify the child as more severely dehydrated than he or she actually is.

In a study of DKA we found that the mean osmolality at admission was 318 (SD 12.9; range 291–337).³ Further, we also found that the calculated osmolality (calculated osmolality=1.86(Na + + K +) + Urea + Glucose)⁴ was only 289 (range 282–304). This suggests hypertonicity is common in DKA and calculated osmolality underestimates the true osmolality. The mean osmolar gap was 29 (range 14–48). The osmolar gap between true and the calculated osmolality, is made up of unmeasured substances like ketoacids. The osmolality of ketoacids have been ignored in the past, as they are considered to be osmotically inactive and not contributing to osmotonicity.⁵ A study done by us (submitted for publication) has demonstrated that ketoacids (acetoacetates) are osmotically active. (Acetoacetate can influence fluid shifts across a semipermeable membranes. This is in contrast to urea, which is not osmotically active.)Osmolality, osmolar gap, and ketone bodies are not measured routinely during the management of DKA. A rapid fall in ketone body level can result in a fall in osmolality and osmoticity of the serum and lead to cerebral edema. In a recent paper looking at the risk factors for development of cerebral edema in DKA⁶ the author noted that since none of the “relevant variables” (serum glucose concentration at presentation, change in serum glucose concentration during therapy, rate of fluid and sodium administration) were associated with the risk of cerebral edema, their data did not support the theory that a rapid decrease in extra cellular osmolality during treatment results in osmotic mediated swelling. Osmolality and osmolar gap were not measured, nor was ketone body levels studied (personal communication, Glaser N). Our studies demonstrates that the ketone level is probably a “relevant variable” that needs to be estimated before we can be certain that rapid decrease in extracellular osmolality did not occur. In summary we suggest that changes in ketone body levels be considered, as a factor that can be partially responsible for the cerebral edema often seen during treatment of DKA. We will be glad to share our data at any summit of experts convened to study the enigma of cerebral edema in DKA.