Purpose: Profound total body potassium deficits occur in children admitted with diabetic ketoacidosis (DKA). However, amount required to maintain normal serum concentration during treatment of DKA is largely unknown. At our institution 40 mEq/L and 60 mEq/L is administered during treatment of DKA, based on initial serum potassium. We aimed to determine the effect of these infusions on potassium levels in children during treatment of DKA. Methods: A retrospective chart review was conducted, after IRB approval, in children admitted to the pediatric ICU from December 2010 to December 2013 with diagnosis of DKA. Duration of insulin infusion was considered as treatment period. Clinical characteristics and laboratory data were analyzed. Statistical analysis was done using means, standard deviations, t-test, and Chi-square. Results: Out of 124 eligible patients, complete data was available (and analyzed) in 94 (67% female). Average age was 12±4 years, mean weight 47±22 kg. Lab values at admission included mean serum bicarbonate 8.3±3.5 mEq/L, potassium 4.6±1 mEq/L, pH 7.13±12, glucose 352±172 mg/dL, anion gap 22±6. Average insulin infusion duration (treatment period) was 14.3±11.4 hours. Lab values at the end of this period included serum bicarbonate 17±3.4 mEq/L, potassium 3.8±0.6 mEq/L, pH 7.3±0.06, glucose 205±93 mg/dL, chloride 111±6.4 mEq/L, anion gap 11±3. Total fluid administered minus initial bolus was mean 2.8±2.3 L/kg, total initial bolus mean 23±13 ml/kg. Total potassium administered until resolution of DKA was mean 2.3±1.6 mEq/kg. Thirteen patients required additional potassium supplementation of mean 0.8±0.5 mEq/kg. Eighty-two percent of patients received fluids with 40mEq/L of potassium and 18% received 60 mEq/L. There were 35 low potassium values (<3.5 mEq/L) and 9 high values (>5.5 mEq/L). Patients that received 60 mEq/L potassium infusion had significantly higher number of low serum values (82.4%) at any given time point during the treatment period than those receiving 40 mEq/L (22.7%) (p=0.0001). Initial mean potassium values in children receiving 60 mEq/L were significantly lower 3.7±0.6 mEq/L as compared to 4.8±0.9 in 40 mEq/L group, (p<0.0001). However, total potassium received, treatment period, and additional potassium supplementation were all similar between the 2 groups. Conclusions: This study suggests high potassium infusions are needed when treating DKA. Close monitoring of serum potassium is needed even when administering fluids containing 40 mEq/L and 60 mEq/L potassium.