Optimal Glycemic Targets in Critically Ill Patients with Diabetes

Abstract

Dysglycemia occurs frequently during critical illness and is associated with increased mortality. There are phenotypical associations between the individual components of acute dysglycemia—hyperglycemia, glycemic variability, and hypoglycemia—and outcomes that appear modulated by the presence or absence of pre-existing diabetes. Observational studies suggest harm from hyperglycemia is attenuated in patients with pre-existing diabetes, but that the associations of glycemic variability and hypoglycemia with harm are as strong or stronger in those with pre-existing diabetes. These observations provided the rationale for studies to evaluate whether personalized strategies allowing mild to moderate hyperglycemia to diminish glycemic variability and/or hypoglycemia in certain critically ill patients is beneficial. The CONTROLING trial evaluated the effect of altering the blood glucose target range based on admission glycated hemoglobin (HbA1c) measurements regardless of whether a patient had pre-existing diabetes. Frequency of hypoglycemic episodes increased in the intervention group, particularly in those without diabetes who were assigned lower targets for their blood glucose range with no statistical differences in clinical outcomes. The LUCID trial restricted enrolment to critically ill patients with pre-existing diabetes and the intervention was a more liberal target range of 10–14 mmol/L, which reduced the incidence of hypoglycemia (<4.0 mmol/L). This did not translate into improvement in patient centered outcomes, although underpowered to do so. Further exploration of the most appropriate target blood glucose range in critically ill patients with diabetes, and the utilization of novel technology or therapeutics is indicated.