Increased intracranial pressure (ICP) resulting in death or neurologic morbidity continues to complicate traditional management of diabetic ketoacidemia (DKA) in pediatric patients. When ICP or cerebrospinal fluid pressures have been measured, correction of hyperglycemia in animals and treatment of DKA in humans have consistently resulted in pathologic increases in ICP. We hypothesized that elevations in ICP can be minimized if changes in effective osmolality (Eosm) are controlled during treatment of DKA. During a six-hour study period, three groups of rabbits were studied: a normal control group of nondiabetic animals (Cnor, n = 10), a control group of animals with DKA (CDKA, n = 8), and an experimental group of animals with DKA (EDKA, n = 8). There was no significant difference between the two groups with DKA regarding pretreatment degree of dehydration, blood pressure, hyperglycemia, acidemia or ICP. During the treatment period, Cnor received maintenance fluids only. CDKA received insulin and an assumed volume of deficit (150 ml/kg) along with maintenance fluids and urinary output replacement with 0.45% NaCl. EDKA received insulin and one-half the volume of deficit calculated by the weight lost with 0.9% NaCl plus maintenance fluids. There was no significant difference between CDKA and EDKA regarding the rate at which DKA was corrected. While CDKA demonstrated a progressive and statistically significant increase in ICP, EDKA exhibited no such increase in ICP compared to normal, nondiabetic controls (Cnor) during treatment. Changes in Eosm during treatment in CDKA compared to Cnor and in CDKA compared to EDKA were significantly greater (p < .01), however, changes in EOSM in EDKA compared to Cnor were not significant. These data support the clinical observation that decreasing EOSM during treatment of DKA is associated with increased ICP and suggest that DKA can be treated effectively with i.v. fluids and insulin without increasing ICP.