Cerebral oedema which develops during the treatment of diabetic ketoacidosis is an important cause of mortality and morbidity in children. We examined 25 management protocols from throughout the UK and related variations in fluid, bicarbonate, insulin, and potassium regimens to the incidence of cerebral oedema recalled in each centre. Treatment of shock ranged from 5 to 25 (median 20) mi kg(-1) plasma (5 recommended 0.9 % saline only) over 10-60 min. Subsequent fluid regimens used 0.9 % saline in 24 (0.45 % saline in 1); 8 used 0.45 % saline if hypernatraemia was present. The rehydration period ranged from 24 h (n = 20) to 48 h (n = 1) and was based on steady (n = 12) or irregular (n = 13) replacement. The quantity of potassium added to fluids was 20-80 (median 30) mmol l(-1). Eight centres recalled having seen 1-5 (median 2) cases of cerebral oedema in the past 5 yr, 10 centres recalled none. Compared with the 10 centres without cerebral oedema, protocols from the 8 with cerebral oedema used more plasma to resuscitate (22 +/- 3 (mean +/- SD) vs 18 +/- 4 ml kg(-1); p < 0.025), suggested larger maintenance fluid volumes for ages 6-9 yr (81 +/- 2 vs 70 +/- 11 ml kg(-1) day(-1); p < 0.005) and were more likely to change to 0.18 % saline when blood glucose had fallen (8/8 vs 5/10) than 0.45 % saline (0/8 vs. 5/10; p < 0.05). Free water overload may contribute to cerebral oedema. These data strengthen the case for a nationally agreed protocol and a prospective study of the antecedents to the development of cerebral oedema.
