Clinical question: Is glucose variability associated with increased mortality independent of mean glucose values in intensive-care-unit (ICU) patients on a strict glucose-control algorithm?
Background: Initial studies demonstrating that strict glycemic control in the ICU improves mortality have not been reproduced in more recent trials and meta-analyses. This inconsistency may be due to unstudied aspects of glycemic control, such as glucose variability.
Study design: Retrospective cohort.
Setting: Eighteen-bed medical/surgical ICU in a teaching hospital in Amsterdam, Netherlands.
Synopsis: Data were collected on 5,728 patients admitted to the ICU from January 2004 to December 2007, all of whom were treated with a computerized intensive insulin protocol. Mean glucose, standard deviation in glucose, and mean absolute glucose change per hour (glucose variability) were calculated for each patient stay in the ICU. The results from these three calculated values were divided into quartiles and evaluated for their predictive value of ICU death and in-hospital death.
Within each mean glucose quartile, the uppermost glucose variability quartile was associated with increased risk of death. Compared with the lowest glucose variability quartile, the highest quartile had a 3.3-fold increased risk of ICU death and a 2.8-fold increased risk of in-hospital death. Patients in the highest glucose quartile with the highest glucose variability had a 12.4-fold increased risk of ICU death.
Bottom line: High glucose variability is associated with increased ICU and in-hospital mortality independent of mean glucose values in patients on a strict glucose control algorithm.
Reviewed for TH eWire by Dimitriy Levin, MD, Jeffrey Carter, MD, Erin Egan, MD, JD, Jonathan Pell, MD, Laura Rosenthal, MSN, ACNP, Nichole Zehnder, MD, Hospital Medicine Group, University of Colorado Denver
For more reviews of HM-related literature, visit our website.