Abstract
Background:
To minimize the risk of cerebral edema in the management of hyperglycemic emergencies, guidelines recommend gradually decreasing serum glucose levels, but there is limited literature validating these reduction goals or evaluating outcomes associated with different correction rates.
Objective:
This study evaluated cerebral edema rates associated with 2 different serum glucose correction rates in adults with hyperglycemic emergencies.
Methods:
This retrospective study’s primary endpoint was the incidence of cerebral edema. Secondary endpoints included average hourly changes in sodium, potassium, chloride, bicarbonate, anion gap, glucose, blood urea nitrogen, creatinine, and calculated osmolality, over the treatment course. Patients were divided into 2 groups based on their average hourly serum glucose correction rate (> 75 mg/dL and ≤ 75 mg/dL), and their outcomes were compared. Patients were included if they were ≥18 years old, received the institutions’ diabetic ketoacidosis or hyperglycemic hyperosmolar state insulin infusion order set within 12 hours of presenting to the hospital, and had a pre-insulin infusion glucose > 600 mg/dL.
Results:
One hundred thirty-four patients had a slow correction rate, and 51 had a rapid correction rate. The median time from starting the insulin infusion to achieving a glucose < 300 mg/dL in the slow and rapid correction rate groups was 9.9 and 6.1 hours, respectively (p < 0.001). In the total population, 2 (1.5%) cerebral edema events occurred. Both patients were in the slow correction rate group and had radiographic evidence of cerebral edema on post-return of spontaneous circulation imaging. There were significantly larger decreases in glucose and osmolality levels in the rapid correction rate group at several time points during treatment.
Conclusion and relevance:
Rapid serum glucose correction rates were not associated with increased cerebral edema events. Our findings suggest glucose and osmolality levels can be corrected faster than what is currently recommended.
Get full access to this article
View all access options for this article.
