Fixed-rate insulin for adult diabetic ketoacidosis is associated with more frequent hypoglycaemia than rate-reduction method: a retrospective cohort study.

Abstract

To assess whether hypoglycaemia incidence during management of adult diabetic ketoacidosis (DKA) differed following transition from a fixed-rate insulin protocol to a protocol using an empiric insulin rate reduction after normoglycaemia. We retrospectively reviewed charts from adult patients managed with a DKA order set before and after order set revision. In cohort 1 (n=77), insulin rate was 0.1unit/kg/h with no adjustments and dextrose was infused at 12.5g/h after glucose reached 250mg/dl. In cohort 2 (n=78), insulin was reduced to 0.05unit/kg/h concurrent with dextrose initiation at 12.5g/h after glucose reached 200mg/dl. The primary outcome was hypoglycaemia (glucose<70mg/dl) within 24h of the first order for insulin. The 24-h incidence of hypoglycaemia was 19.2% in cohort 2 versus 32.5% in cohort 1; the adjusted odds ratio was 0.46 (95% confidence interval (CI) [0.21, 0.98]; P=0.047). The 24-h use of dextrose 50% in water (D50W) was also reduced in cohort 2. No differences were seen in anion gap or bicarbonate normalization, rebound hyperglycaemia or ICU length of stay. In most patients who became hypoglycaemic, the preceding glucose value was below 100mg/dl. The insulin rate-reduction protocol was associated with less hypoglycaemia and no obvious disadvantage. Robust intervention for low-normal glucose values could plausibly achieve low hypoglycaemia rates with either approach.