Glucose meters in the intensive care unit

Intensive intravenous insulin therapy directed at maintaining blood glucose concentration within tight normal limits (4.4–6.1 mmol/L) is now a standard of care for the critically ill. This requires frequent point of care blood glucose testing.  

For appropriate intravenous insulin dosing it is vital that the chosen glucose measuring system is precise, accurate and not prone to analytical interference likely to be encountered among critically ill patients. A recent study sought to compare the suitability of four blood glucose meters for use in a critical care setting. 

The four meters were chosen on the basis that each was representative of a different glucose measuring technology. All four meters were found to be equally and sufficiently precise (within and between batch CV < 5 %). However accuracy, judged by correlation of results obtained by meters with a reference laboratory based hexokinase method, was found to vary between the four meters. 

The best performing meter in this regard returned a result within 10 % of the reference value for 170 of 185 samples tested. The comparable figure for the least accurate meter was 79 out of 185 samples tested. A previous study, cited by the authors, has demonstrated that as long as measured glucose is within 10 % of its true value, no significant insulin dosing errors are likely to occur. 

Ascorbic acid was shown to interfere with analysis for three out of the four meters, but acetaminophen, maltose, lactate, b-hydroxybutyrate and epinephrine were shown not to interfere with any glucose meter measurement, even when present in a very high concentration. Finally, hematocrit of the sample was shown to affect glucose results for two of the four meters. 

For these two meters there was a significant negative bias in glucose that correlated with increasing hematocrit. It is clear that glucose meters are not all equally suited for use in the intensive care unit.