Objectives:Insulin infusion therapy is commonly used in the hospital setting to manage diabetic ketoacidosis and hyperosmolar hyperglycemic state. Clinical evidence suggests both hypoglycemia and glycemic variability negatively impact patient outcomes. The hypothesis of this study was that moderate-intensity insulin therapy decreases hospital length of stay and prevalence of hypoglycemia in patients with diabetic ketoacidosis and hyperosmolar hyperglycemic state.Design:Pre-post study.
Setting:Large academic medical center in the United States.Patients:Two-hundred one consecutive, nonpregnant, adult patients admitted for diabetic ketoacidosis and hyperosmolar hyperglycemic state between October 2010 and December 2014.Interventions:High-intensity insulin therapy versus moderate-intensity insulin therapy. High-intensity insulin therapy was designed to rapidly normalize blood glucose levels with bolus doses of insulin and rapid insulin titration. Moderate-intensity insulin therapy was designed to mitigate glycemic variability and hypoglycemia through avoidance of bolus dosing, a liberalized blood glucose target, and gradual insulin titration.Measurements and Main
Results:Hospital and ICU length of stay were reduced by 23.6% and 38%, respectively. The relative risk of remaining in the hospital at day 7 (0.51; p = 0.022) and day 14 (0.28; p = 0.044) were significantly reduced by the moderate-intensity insulin therapy strategy. The relative risk of remaining in the ICU at 48 hours was significantly lower in the moderate-intensity insulin therapy cohort (0.34; p = 0.0048). The prevalence (35% vs 1%; p = 0.0003) and relative risk (0.028; p = 0.0004) of hypoglycemia were significantly lower in the moderate-intensity insulin therapy cohort. Glycemic variability decreased by 28.6% (p < 0.0001). There was no difference in the time to anion gap closure (p = 0.123).
Conclusions:Moderate-intensity insulin therapy for diabetic ketoacidosis and hyperosmolar hyperglycemic state resulted in improvements in hospital and ICU length of stay, which appeared to be associated with decreased glycemic variability.