NEW PORTABLE DEVICES FOR BLOOD-GLUCOSE ME ASUREMENT - COMPARISON OF MEDISENSE CARD, GLUCOMETER ELITE, HYPOCOUNT SUPREME AND OMNICAN CONTROLWITH THE REFERENCE GLUCOSE-OXIDASE METHOD

Patients as well as medical personnel are increasingly confronted with new devices for blood glucose measurement. Devices smaller in size, f aster in action and easier to handle are being developed by various co mpanies. We investigated whether precision and accuracy of recently ma rketed blood glucose meters would be efficient and safe enough for cli nical use, and compared the meters MediSense Card, Glucometer Elite, H ypocount Supreme, and Omnican Control with our reference glucose oxida se method. On average 200 pairs of blood glucose values from capillary blood samples of type 1 and 2 diabetic patients were determined using two meters of each make. The measurements were performed by one exper ienced technician using blood from the same sample for the meter and t he Beckman Analyzer 2. For evaluation a linear regression analysis, th e percentage of values within a maximum deviation of less than 10% fro m the reference value, and clinically relevant models such as Clarke's error-grid analysis and Koschinsky's acceptance analysis were used. C ombining all analyses, MediSense Card and Glucometer Elite showed the least deviations in all blood glucose ranges. For MediSense Card 56%, for Glucometer Elite 72%, but for Hypocount Supreme and Omnican Contro l less than 50% of all values were within a +/- 10% range compared wit h the reference method. Clarke's error-grid analysis reflected the bes t results for MediSense Card and Glucometer Elite with 96% and 97% res pectively of all measurements within the ideal error-grid zone A, wher eas this was the case for Hypocount Supreme in 85% and for Omnican Con trol in only 71%. The coefficient of variance for measurements in seri es using 3 different blood glucose ranges was acceptable for all devic es besides the Hypocount Supreme, but overall performance was poor com pared to our reference method. Nevertheless, according to Koschinsky's clinical acceptance scale even Hypocount Supreme and Omnican Control can be classified as clinically acceptable (range 2 from a three point scale). The 1994 criteria of the American Diabetes Association, a max imum deviation of less than 5% in 100% of measurements, were not met b y any device, but after training and with proper handling all devices work efficiently enough and are safe for clinical use. Nevertheless, c ompanies are urged to seek new technologies for blood glucose measurem ent, as all recently developed devices tested in our study did not per form better than known older glucose meters.