Background: Mannitol and hypertonic saline are used to ameliorate brain edema and intracranial hypertension during and after craniotomy. We hypothesized that the agreement of measured and calculated serum osmolality during the infusion of hypertonic saline would be better than mannitol. The objective was to determine the accuracy of serum osmolality estimation by different formulas during the administration of hyperosmolar agent.
Methods: A prospective, randomized, double-blinded, controlled trial was conducted in a 30-bed neurosurgical intensive care unit at a university hospital. Thirty-five adult patients requiring the use of hyperosmolar agents for prevention or treatment of brain edema after elective craniotomy were enrolled, and randomly assigned 1:1 to receive 125 mL of either 20 % mannitol (mannitol group) or 3.1 % sodium chloride solution (hypertonic saline group) in 15 min. Serum osmolality, serum sodium and potassium concentration, blood urea nitrogen and blood glucose concentration were measured during the study period. The primary outcome was the agreement of measured and estimated serum osmolality during the infusion of the two experimental agents. We used Bland and Altman's limits of agreement analysis to clarify the accuracy of estimated serum osmolality. Bias and upper and lower limits of agreement of bias were calculated.
Results: For each formula, the bias was statistically lower in hypertonic saline group than mannitol group (p < 0.001). Within group comparison showed that the lowest bias (6.0 [limits of agreement: -18.2 to 30.2] and 0.8 [-12.9 to 14.5] mOsml/kg in mannitol group and hypertonic saline group, respectively) was derived from the formula '2 × ([serum sodium] + [serum potassium]) + [blood urea nitrogen] + [blood glucose]'.
Conclusions: Compared to mannitol, a better agreement between measured and estimated serum osmolality was found during the infusion of hypertonic saline. This result indicates that, if hypertonic saline is chosen to prevent or treat brain edema, calculated serum osmolality can be used as a reliable surrogate for osmolality measurement.
Trial registration: ClinicalTrials.gov identifier: NCT02037815.