Objectives: To determine the association of pre-existing diabetes, hyperglycemia, and hypoglycemia during the first 24 hours of ICU admissions with 90-day mortality in patients with sepsis admitted to the ICU.
Design: We used mixed effects logistic regression to analyze the association of diabetes, hyperglycemia, and hypoglycemia with 90-day mortality (n = 128,222).
Setting: All ICUs in the Netherlands between January 2009 and 2014 that participated in the Dutch National Intensive Care Evaluation registry.
Patients: All unplanned ICU admissions in patients with sepsis.
Interventions: The association between 90-day mortality and pre-existing diabetes, hyperglycemia, and hypoglycemia, corrected for other factors, was analyzed using a generalized linear mixed effect model.
Measurements and main results: In a multivariable analysis, diabetes was not associated with increased 90-day mortality. In diabetes patients, only severe hypoglycemia in the absence of hyperglycemia was associated with increased 90-day mortality (odds ratio, 2.95; 95% CI, 1.19-7.32), whereas in patients without pre-existing diabetes, several combinations of abnormal glucose levels were associated with increased 90-day mortality.
Conclusions: In the current retrospective large database review, diabetes was not associated with adjusted 90-day mortality risk in critically ill patients admitted with sepsis.