Accuracy of glucose meter use in gestational diabetes

Glucose monitoring is essential for the successful management of gestational diabetes. The accuracy of glucose meters is typically determined over a much wider range of glucose values than that commonly encountered in gestational diabetes. The objective of our study was to look at the accuracy of self-monitoring glucose meters in a clinic setting over a range of glucose values seen in gestational diabetes. We retrospectively analyzed 107 case records of subjects with gestational diabetes, each of whom had three simultaneous laboratory and glucose meter glucose tests. The results were compared using the performance goals that (1) all of glucose meters should have readings within 10% of the reference value and (2) the error grid analysis in the standard format and a modified version suitable for gestational diabetes. We also examined the range of the differences from the reference value. Nearly half of the values (47%) were in excess of 10% of the reference range (either above or below). Close to 15% were in excess of 20% difference from the reference range. Standard error grid analysis showed that 96% of the values fell within sections A of the error grid which are considered acceptable, and 100% fell within sections A and B, differences which are generally considered to have no major impact on care. The modified version of the error grid analysis demonstrated that 39% of the values were outside the acceptable range. Within subjects, a substantial number (26%) had a range of differences that exceeded 20% difference between each other. Although the meters give reasonable results that might be acceptable for general diabetes care, the results provide some cause for concern in the management of gestational diabetes. Given the need for precision in the setting of pregnancy particularly in making the decision of whether to start or withhold insulin therapy, caregivers need to be cognizant of these inaccuracies.