Objectives: Cerebral edema in diabetic ketoacidosis is a devastating complication with significant morbidity and mortality. This entity has traditionally been treated with mannitol, but use of 3% hypertonic saline has become an accepted alternative. We sought to assess if changes in the use of hyperosmolar therapies for treatment of cerebral edema in diabetic ketoacidosis may have influenced mortality over the last decade. Design: Retrospective cohort study. Setting: Patients discharged between 1999 and 2009 from 41 children's hospitals that provided data to the Pediatric Health Information System database. Patients: A total of 43,107 children (age < 19) with diagnosis codes related to diabetic ketoacidosis were identified and further classified as having cerebral edema if treated with mannitol and/or 3% hypertonic saline. Interventions: None. Measurements and Main Results: Billing for 3% hypertonic saline and mannitol was quantified, and mortality associated with both diabetic ketoacidosis and cerebral edema in diabetic ketoacidosis was examined. Overall mortality in diabetic ketoacidosis was 0.25% and significantly decreased (p < 0.001) over the study period, whereas the frequency of treatment with hyperosmolar agents (3.8%) was unchanged. Use of mannitol as a sole agent decreased from 98% to 49%, 3% hypertonic saline as a sole agent increased from 2% to 39%, and combined therapy increased from 0% to 10%. Use of 3% hypertonic saline alone was associated with a higher mortality than mannitol alone (adjusted odds ratio, 2.71 [95% CI, 1.01-7.26]) in patients treated for cerebral edema. Similar results were obtained after adjustment for the propensity to receive hypertonic saline (adjusted odds ratio, 2.33 [95% CI, 1.07-5.07]) and in the subset of subjects receiving mechanical ventilation (adjusted odds ratio, 3.27 [95% CI, 1.12-9.60]). Conclusion: Hypertonic saline has replaced mannitol as the most commonly used agent at many institutions for treatment of cerebral edema in diabetic ketoacidosis. In our analysis, however, use of hypertonic saline as a sole agent was associated with an increased risk of mortality. Recognizing the limitations of administrative data, we conclude that equipoise regarding choice of therapy for treatment of cerebral edema in diabetic ketoacidosis should be maintained until a more definitive study is performed to guide therapy of this potentially lethal complication.
