Intensity of early correction of hyperglycaemia and outcome of critically ill patients with diabetic ketoacidosis.

OBJECTIVES: To determine the impact of the intensity of early correction of hyperglycaemia on outcomes in patients with diabetic ketoacidosis (DKA) admitted to the intensive care unit. METHODS: We studied adult patients with DKA admitted to 171 ICUs in Australia and New Zealand from 2000 to 2013. We used their blood glucose levels (BGLs) in the first 24 hours after ICU admission to determine whether intensive early correction of hyperglycemia to <= 180 mg/dL was independently associated with hypoglycaemia, hypokalaemia, hypo-osmolarity or mortality, compared with partial early correction to > 180 mg/dL as recommended by DKA-specific guidelines. RESULTS: Among 8553 patients, intensive early correction of BGL was applied to 605 patients (7.1%). A greater proportion of these patients experienced hypoglycaemia (20.2% v 9.1%; P < 0.001) and/or hypo-osmolarity (29.4% v 22.0%; P < 0.001), but not hypokalaemia (16.7% v 15.6%; P = 0.47). Overall, 11 patients (1.8%) in the intensive correction group and 112 patients (1.4%) in the partial correction group died (P = 0.42). However, after adjustment for illness severity, partial early correction of BGL was independently associated with a lower risk of hypoglycaemia (odds ratio [OR], 0.38; 95% CI, 0.30-0.48; P < 0.001), lower risk of hypo-osmolarity (OR, 0.80; 95% CI, 0.65-0.98; P < 0.03) and lower risk of death (OR, 0.44; 95% CI, 0.22-0.86; P = 0.02). CONCLUSIONS: In a large cohort of patients with DKA, partial early correction of BGL according to DKA-specific guidelines, when compared with intensive early correction of BGL, was independently associated with a lower risk of hypoglycaemia, hypo-osmolarity and death.