A randomized controlled trial of isotonic versus hypotonic maintenance intravenous fluids in hospitalized children

This was a masked controlled trial. Randomization was stratified by admission type: medical patients and post-operative surgical patients, aged 3 months to 18 years, who required IV fluids for at least 8 hours. Patients were randomized to receive either 0.45% or 0.9% saline in 5.0% dextrose. Treating physicians used the study fluid for maintenance; infusion rate and the use of additional fluids were left to their discretion.Sixteen children were randomized to 0.9% saline and 21 to 0.45% saline. Baseline characteristics, duration (average of 12 hours) and rate of study fluid infusion, and the volume of additional isotonic fluids given were similar for the two groups. [Na] increased significantly in the 0.9% group (+0.20 mmol/L/h [IQR +0.03, +0.4]; P = 0.02) and increased, but not significantly, in the 0.45% group (+0.08 mmol/L/h [IQR -0.15, +0.16]; P = 0.07). The rate of change and absolute change in serum [Na] did not differ significantly between groups.When administered at the appropriate maintenance rate and accompanied by adequate volume expansion with isotonic fluids, 0.45% saline did not result in a drop in serum sodium during the first 12 hours of fluid therapy in children without severe baseline hyponatremia. Confirmation in a larger study is strongly recommended.NCT00457873 (http://www.clinicaltrials.gov/ webcite)For almost half a century, pediatricians have ordered "maintenance" intravenous (IV) fluids for children according to the guidelines set out by Holliday and Segar: 100 cc/kg/day for the first 10 kg, plus 50 cc/kg/day for the next 10 kg, plus 20 cc/kg/day for each remaining kilogram [1]. Based on these recommendations for water intake, and on the estimated daily sodium and potassium needs of 3 milliequivalents and 2 milliequivalents per 100 kcal per day respectively, a hypotonic solution (0.2% saline) was recommended.The wisdom of this approach to IV maintenance fluid therapy has been questioned recently [2-9]. Several authors have argued that admini