The Accuracy of Home Glucose Meters for the Glucose Range Anticipated in Pregnancy

EDITORIAL COMMENT: All readers who manage patients with diabetes will be interested in this paper. As the authors indicate home glucose meters are an essential component in assessing the metabolic control in patients with diabetes and in making management decisions the most important of which is whether or not to commence insulin. The reviewer of this paper made the comment that especially in the management of gestational diabetes that unless a patient's sugar levels are extremely high, insulin therapy should not be commenced until the patient's glucose levels are checked by the reference laboratory. Furthermore, as this paper shows, home glucose meters tend to give lower readings than laboratory measurements. Therefore, if home glucose meters alone are used, women who should be on insulin will not receive it. Some centres use the 2‐hour postprandial level and others the I‐hour postprandial level to decide whether insulin therapy is indicated when dietary measures have failed. There is discussion at meetings of the Australian Diabetes in Pregnancy Group indicating that different clinics have quite different criteria for the blood sugar level that is needed prior to commencement of insulin. The figures quoted vary from a 2‐hour postprandial level of 6.5 mmol/L on 2 or more occasions to 8.5 mmol/L This accounts for the difference in insulin usage in different clinics managing women with gestational diabetes. The management of women with gestational diabetes from the point of view of metabolic control is not straightforward. There are different definitions used for the diagnosis of gestational diabetes, and different criteria used for when the patient should be asked to commence insulin therapy. This paper highlights the additional difficulty of interpretation of results obtained by home glucose meters. Authors'reply 1. I do not necessarily agree that because home glucose meters read lower than the laboratory glucose that some women are not receiving insulin treatment. In clinical practice the criteria which should be used are appropriate to the measuring instrument. 2. I would consider it likely that most home glucose meters and reagent strips are calibrated for maximum accuracy at the range where most of the results in people with diabetes are to be found, e.g. 8–12 mmol/L. It should be reasonably simple for a manufacturer to prepare a system for use in pregnancy which is accurate below a level of 8.0 mmol/L. Perhaps your Editorial Comment could throw some weight behind the plea for manufacturers to do so. 3. While waiting for the above to happen, clinicians should try and ensure that their patients only use the best performing meters. Summary: Home glucose meters are used almost universally for the management of pregnant women with diabetes. However the degree of accuracy of meters in general, and particularly for the lower glucose range, is questionable. Six of the most commonly used meters in Australia were evaluated for analytical accuracy for glucose levels <8.0mmol/L ‐ a level below which many important clinical management decisions are made. Two meters, Elite and Precision QID, cannot be recommended. The other 4 meters, Glucometer 3, Reflolux S, Alpha and Advantage produced results with a higher degree of analytical accuracy. Of the meters tested, the Accutrend Alpha had the highest proportion of results within 5% and 10% of a laboratory reference method, the lowest bias with the Altaian Bland method of comparison, and is recommended for use in pregnancy. No meter approached the analytical accuracy goal recommended by the American Diabetes Association. Manufacturers should be encouraged to develop or adapt systems for use in pregnancy.