RA. Fisken (1999). Severe diabetic ketoacidosis: the need for large doses of insulin. Diabetic medicine : a journal of the British Diabetic Association, 16(4), 347-50. https://doi.org/10.1046/j.1464-5491.1999.00054.x

PubMed 10220211

[Abstract]
A 21-year-old female with Type 1 diabetes mellitus (DM) presented in ketoacidosis. She received intravenous normal saline and insulin at 6 U/h and 1.26% sodium bicarbonate solution. After the blood glucose had fallen to 9.5 mmol/l, the saline infusion was changed to 5% glucose solution and the insulin infusion rate to 2 to 3 U/h. The next day the patient became more drowsy (Glasgow coma scale 13/15, later falling to 4/15). Computed tomography (CT) scan suggested cerebral oedema and the patient was treated with dexamethasone and mannitol. She remained critically ill for 48 h, eventually making a full recovery. Insulin was given at rates of 8 to 14 U/h, with 10% or 20% glucose infusion to maintain the blood glucose above 5 mmol/l; despite this it was not until the fifth day that her serum bicarbonate became normal. Textbooks usually advise starting insulin at 6 U/h and reducing the infusion rate to 1-4 U/h when the blood glucose falls below a certain level. In this case, even with high rates of insulin infusion, it took 5 days before the patient's serum bicarbonate returned to normal. Thus, in severe diabetic ketoacidosis (DKA), protocols should advise that the insulin infusion be continued at high dose (4 to 6 U/h or more), with appropriate glucose infusion to prevent hypoglycaemia, until the serum bicarbonate is normal or nearly so.