Objectives: Although glucose meters are well-established instruments for self-monitoring blood glucose levels, diagnostic and screening procedures should be performed using standard laboratory methods. In addition to standard laboratory methods, HemoCue is authorized for screening and diagnostic purposes in Germany. The rapid development of other glucose meters makes it necessary to re-evaluate this recommendation. Our objective was to test the usefulness of glucose meters in screening pregnant patients for gestational diabetes.
Methods: The 50-g glucose challenge test was administered to one hundred and ninety-three pregnant patients whose blood glucose levels were then simultaneously measured with five portable meters and the HemoCue. The results were compared to our standard method (Hexokinase). A cut-off of 7.8 mmol/L was used and sensitivity, specificity, accuracy, the Youden index, and the Kappa index were calculated. The tests were performed by well-trained personnel (C.D. and U.S.).
Results: 1212 measurements were performed on 193 patients. All glucose meters showed a very good correlation (r > 0.90). None of the measurements showed an extreme deviation necessitating the error grid analysis. The GlucoTouch (5.93% +/- 7.4) and the HemoCue (-9.04% +/- 5.9) showed a mean deviation greater than 5%. None of the meters had a mean deviation greater than 10%. The accuracy fluctuated between 0.85 and 0.94. The Kappa index was between 66 to 85. In our clinical trial, the Accu-Chek, Glucotouch, OneTouch, and Precision demonstrated greater accuracy and a higher Kappa index than the HemoCue.
Conclusions: Our data showed good concordance in statistical and clinical parameters for most of the six glucose meters. The HemoCue, recommended as a standard method in several countries, did not show better concordance than most of the tested glucose meters. When used by well-trained personnel, the accuracy of the Accu-Chek, Glucotouch, One-Touch, and Precision was acceptable for use in gestational diabetes screening.