Computerized intensive insulin dosing can mitigate hypoglycemia and achieve tight glycemic control when glucose measurement is performed frequently and on time
1 Division of Endocrinology, Indiana University School of Medicine, 545 Barnhill Drive, EH 421, Indianapolis, IN 46202, USA
2 Department of Medicine and Clarian Health, Indiana University School of Medicine, 545 Barnhill Drive, EH 421, Indianapolis, IN 46202, USA
3 Department of Surgery, Tufts Medical Center, Tufts University School of Medicine, 750 Washington Street, NEMC Box 4630, Boston, MA 02111, USA
4 Methodist Hospital/Clarian Health, 1701 N. Senate Blvd., Indianapolis, IN 46202, USA
5 Medical Quality, Clarian Health, ERC 6102, 1701 N. Senate Blvd., Indianapolis, IN 46202, USA
6 The Epsilon Group Virginia LLC, 615 Woodbrook Drive, Charlottesville, VA 22901, USA
7 William Beaumont Hospital, 3601 W 13 Mile Road, Royal Oak, MI 48073-9952, USA
Critical Care 2009, 13:R163 doi:10.1186/cc8129
See related letter by Hoekstra et al., http://ccforum.com/content/14/1/404Published: 12 October 2009
Control of blood glucose (BG) in critically ill patients is considered important, but is difficult to achieve, and often associated with increased risk of hypoglycemia. We examined the use of a computerized insulin dosing algorithm to manage hyperglycemia with particular attention to frequency and conditions surrounding hypoglycemic events.
This is a retrospective analysis of adult patients with hyperglycemia receiving intravenous (IV) insulin therapy from March 2006 to December 2007 in the intensive care units of 2 tertiary care teaching hospitals. Patients placed on a glycemic control protocol using the Clarian GlucoStabilizer™ IV insulin dosing calculator with a target range of 4.4-6.1 mmol/L were analyzed. Metrics included time to target, time in target, mean blood glucose ± standard deviation, % measures in hypoglycemic ranges <3.9 mmol/L, per-patient hypoglycemia, and BG testing interval.
4,588 ICU patients were treated with the GlucoStabilizer to a BG target range of 4.4-6.1 mmol/L. We observed 254 severe hypoglycemia episodes (BG <2.2 mmol/L) in 195 patients, representing 0.1% of all measurements, and in 4.25% of patients or 0.6 episodes per 1000 hours on insulin infusion. The most common contributing cause for hypoglycemia was measurement delay (n = 170, 66.9%). The median (interquartile range) time to achieve the target range was 5.9 (3.8 - 8.9) hours. Nearly all (97.5%) of patients achieved target and remained in target 73.4% of the time. The mean BG (± SD) after achieving target was 5.4 (± 0.52) mmol/L. Targeted blood glucose levels were achieved at similar rates with low incidence of severe hypoglycemia in patients with and without diabetes, sepsis, renal, and cardiovascular disease.
Glycemic control to a lower glucose target range can be achieved using a computerized insulin dosing protocol. With particular attention to timely measurement and adjustment of insulin doses the risk of hypoglycemia experienced can be minimized.