Association between Diabetes Mellitus and Sepsis for the Glycemic Control Outcome of Two Intensive Care Units in Malaysia

Abstract

Close monitoring and tight glycemic control are required among critically ill patients as they have dynamic metabolism which may precipitate stress-induced hyperglycemia. Clinically, diabetes mellitus (DM) patient with sepsis indicated a high mortality rate. This study investigates the association between DM and non-DM related to sepsis and non-sepsis patients from different insulin infusion therapy management. This study used 128 retrospective data from Hospital A, and 37 retrospective data from Hospital B. ICU patients who received insulin infusion therapy during their stay in the ICU were selected. Both centres implement the sliding scale-based insulin infusion therapy with the target range for blood glucose (BG) level within 6.0 - 10.0 mmol/L. The retrospective clinical data were compared among cohorts for DM and non-DM associated with sepsis and non-sepsis conditions. Findings showed that the DM group had higher insulin sensitivity than non-DM for both cohorts. Meanwhile, cohort B had higher insulin sensitivity than cohort A for all classes. Cohort A (DM+Sepsis) had low insulin sensitivity (66.7 L/(mU.min) and worst condition with sepsis which resulted from the lowest percentage (30.81%) of BG measurement within the target range. The (nonDM+nonSepsis) class had the tightest glycemic control for cohort A (3.4 mmol/L) and cohort B (2.2 mmol/L), as observed by the BG interquartile range. Furthermore, cohort A (nonDM+nonSepsis) had a 41.55% of severe hyperglycemia and 0.12% for severe hypoglycemia. Contrary, cohort B (nonDM+nonSepsis) had the highest percentage within the target range (74.31%) and the lowest percentage of hyperglycemia (18.78%). There was significantly different (p-values <0.05) between cohort A and cohort B in BG level and glucose intake, likewise between sepsis and non-sepsis of non-DM for both cohorts. The findings indicate that a successful glycemic control protocol is much influenced by insulin sensitivity, patient variability, diabetes condition, and patient sepsis status.