Multisite Evaluation of a New Diabetes Self-Test for Glucose and Glycated Protein (Fructosamine)

Abstract

Background: In diabetes management, the true average blood glucose is best obtained using glycated protein tests that give the average blood glucose over a previous time window of either weeks (fructosamine tests) or months (glycated hemoglobin tests). Until now, glycated protein tests have only been available as laboratory tests and have therefore been underutilized in diabetes management. Recently, a fructosamine self-test for use by diabetes patients at home was cleared for marketing by the U.S. Food and Drug Administration (FDA). We have studied the performance of this test in three geographically distinct diabetes clinics to confirm the performance and accuracy of both glucose and fructosamine testing with this device. This new self-testing system has the potential to improve glycemic control dramatically in patients with diabetes, including those patients with type 2 diabetes using oral drug therapy. Methods: Three geographically different sites (San Diego, CA, Tallahassee, FL, and Minneapolis, MN) were selected for the study. Sixty male and 56 female adult patients, with both type 1 (59) and type 2 (57) diabetes, were selected for participation in the study (total patients = 116). Fingerstick puncture capillary blood glucose was tested using the YSI Model 1500 and the Duet^® Glucose test. A fingerstick puncture capillary blood test was also tested with the Duet GlucoProtein™ (fructosamine) test strip in duplicate. For fructosamine comparison, a venipuncture blood sample of ethylene diaminetetraacetic acid (EDTA) plasma was collected and tested using the Roche Unimate^® laboratory test. Results : The glucose test gave excellent correlation to the reference laboratory method (r = 0.98) and the GlucoProtein test gave a correlation of 0.72 compared to the laboratory method. The bias of both tests compared to the laboratory tests was 10% or less at all concentrations. Error grid analysis of the glucose test showed that 97.5% of test results were in the accurate zone and 2.5% were in the clinically neutral or benign errors zone. Analysis of fructosamine test results using a two-by-two grid yielded sensitivity of 100%, specificity of 92% and accuracy of 94%. Conclusions: The Duet Glucose Control System is accurate for both measuring glucose and GlucoProtein (fructosamine) using a fingerstick blood sample. This new testing system has the potential to provide useful information to both healthcare specialists in their office and also to patients at home to help them achieve better long-term glucose control and avoid the potential acute and chronic complications of diabetes.