Trends in serum sodium concentration and effective osmolality during repair of paediatric diabetic ketoacidosis

Abstract

Objective: To assess the effect of resuscitation fluid on trends in serum sodium and effective osmolality during repair of paediatric diabetic ketoacidosis. Design: Retrospective. Setting: A paediatric intensive care unit (PICU) in a tertiary academic medical centre. Subjects: Sixteen patients (age range: 7–20 years; mean: 14 +/− 3 years) admitted to the PICU over 19 months for 35 episodes of diabetic ketoacidosis. Interventions: We reviewed trends in serum sodium and effective osmolality as well as total fluid and sodium intake and mean resuscitation fluid sodium concentration (total sodium intake/total fluid intake) during the first day of therapy. Measurements and main results: Within the range of resuscitation fluid composition observed (mean resuscitation fluid [sodium]: 63–149 mmol/l), we noted a weak, positive correlation between trends in uncorrected serum [sodium] and mean resuscitation fluid [sodium]. No association was found between trends in corrected serum [sodium] or effective osmolality and fluid intake, sodium intake, or mean resuscitation fluid [sodium]. Near-isotonic resuscitation fluid (mean resuscitation fluid [sodium] >125 mmol/l) was associated with both positive and negative trends in uncorrected and corrected serum [sodium], as well as negative trends in effective osmolality. Trends in uncorrected serum [sodium] and effective osmolality were unrelated, while trends in corrected serum [sodium] and effective osmolality, though related, were not concordant in all episodes with a negative trend in effective osmolality. Conclusions: A strategy to limit precipitous declines in effective osmolality during repair of diabetic ketoacidosis should not rely solely on the trends in uncorrected or corrected serum [sodium] as markers of effective osmolality, or on changes in resuscitation fluid volume or content to bring about a change in effective osmolality.