Comparison of bedside and laboratory blood glucose estimations in critically ill children with shock

Bedside glucometers are often used for frequent glucose measurements on capillary blood in critically ill children. However, there are concerns that capillary blood glucose estimations may not be accurate in children with shock and peripheral edema. The objective of this study was to compare simultaneously obtained laboratory values of arterial or central venous blood glucose with capillary blood glucose estimation using a glucometer in children with shock.

Prospective observational study.

Tertiary pediatric intensive care unit in a children's hospital in Chennai.

All children admitted between July 2007 and September 2008 with shock as defined by the American College of Critical Care Medicine criteria were eligible for inclusion.

None.

Three hundred thirty-seven sets of simultaneous measurements were obtained from 52 children (age range, 3 months to 18 yrs; average Pediatric Risk of Mortality III score 9.6). The mean blood glucose measurement using capillary blood on a bedside glucometer was 135 ± 67 mg/dL (7.5 ± 3.7 mmol/L). The mean laboratory glucose was 130 ± 67 mg/dL (7.2 ± 3.7 mmol/L). The correlation coefficient between the measurements was 0.94. There were no differences between those with and without peripheral edema. Using Bland-Altman plots, the mean difference between capillary samples vs. laboratory glucose was 6 mg/dL (0.3 mmol/L). The spread was wider at the higher ends of blood glucose values.

Capillary blood glucose estimation in children with shock was similar to the laboratory measurement in the midranges of glucose values.