Blood Glucose Variability as a Predictor of Mortality for Septic Patients in Intensive Care Unit

Abstract

Background: One of the ways of modulation of septic response is glycaemic control. Increased Bloodglucose variability rather than mean glucose level in critically ill patients is an important factorassociated with sepsis.

Objective: Objective of this study was to find out whether the increased blood glucose variabilitypredicts mortality for septic patients in Intensive Care Unit.

Methods: Purposively allocated 40 septic patients in Intensive Care Unit of Dhaka Medical CollegeHospital, Bangladesh were observed prospectively from January 2012 to July 2013. Glucose variabilitywas assessed within first twenty four hours of recruitment of septic patients. By considering a targetblood glucose level (5 - 8 m mol/L), the assigned patients were separated into three groups (accordingto blood glucose variability). Group I (mild variable group) included those patients having less than twoblood glucose values not within the target blood glucose level. Group II (moderate variable group)included those patients having more than two to four blood glucose values not within the target range.Group III (more variable group) included those patients having more than four blood glucose values outof range of target glucose level. Chi- square test along with a p-value (< 0.05) was done to assess whichgroup of blood glucose variability best reflects the association of mortality. Logistic regression was usedto determine the odds ratio of ICU death in relation to blood glucose variability.

Results: Out of all the septic patients, the more blood glucose variability group (37.5%) more reflectedthe blood glucose variability than the others. Using the Chi-square, it was found that a highlysignificant difference (x2-14.56, p-value 0.001) was existed between three blood glucose variabilitygroups with respect to mortality. Logistic regression analysis demonstrated that more glucosevariability group had predicted higher mortality rate with a p- value of 0.007 and an odds ratio of 16.0.Result is significant. On the other hand, significant effect of moderate glucose variability group onmortality was not found with a p-value of 0.665 and an odds ratio of 0.667.

Conclusions: The septic patients having more blood glucose variability had predicted higher mortalityrate than that of moderate and less blood glucose variability in Intensive Care Unit. This observationindicates that blood glucose variability should be included as a future approach to glucose managementof septic patients as a target for therapeutic intervention.

JBSA 2022; 35 (2) : 43-50