Glucometers are used to quickly obtain glucose concentrations in the hospital and at home. For quality assurance, the American Diabetes Association recommends periodically comparing the results from glucometers with those from referenced laboratories (1). Usually, measurement using glucometers is done one-by-one by technicians, which is costly regarding manpower. In contrast, measurement using blood analyzers in laboratories saves manpower by analyzing after sufficient samples are collected. Tubes containing glycolytic inhibitors such as sodium fluoride are used to keep the glucose level constant (2). If analyses by glucometers could be done similarly, the laboratory could also save manpower. However, the Food and Drug Administration (FDA) suggests that fluoride should not be used as a preservative for venous specimens when using glucometers (3). Since there was not much explanation in the FDA's statement, we conducted a study to assess if blood stored in fluoride-containing tubes could be used to measure glucose concentration by glucometers.
Six glucometers that allow venous blood to be used were selected: Glucometer Elite (Bayer), Accu-Chek Advantage (Roche), Accu-Chek Go (Roche), Accu-Chek Active (Roche), Optium (Abbott), and Sure Step (Johnson & Johnson). One new meter was randomly selected from each manufacturer. Venous blood was collected in fluoride-containing tubes (sodium fluoride/K3EDTA; Greiner Bio-One VACUETTE North America) from 62 consecutive subjects who had diabetes and were not receiving dialysis. The stored blood was applied to the reagent strip of glucometers in random order. The reference measurements were performed by blood analyzer (Toshiba Medical Systems, Tokyo, Japan) in the laboratory.
The regression and correlation coefficients between the values from the reference measurement and each glucometer ranged from 0.898 to 1.366 and from 0.938 to 0.993, respectively. The mean ± SD of the bias of glucometers, which was calculated as (glucometer − reference) × 100%/reference, ranged from 5.3 ± 3.7% to 31.5 ± 7.5%. The percentage of bias <5%, as recommended by the American Diabetes Association (1), was low among all glucometers. The most accurate was the Accu-Chek Advantage meter (54.8%), and the others were all under 20%. Alternatively, using criteria by the FDA (bias <20%) (3), the Accu-Chek Advantage, Elite, and Optium meters had 100, 83.9, and 83.9% of values met, respectively, while the others were all under 70%. The error grid analysis showed that all of the measurements from the Accu-Chek Advantage, Elite, and Optium meters lay in zones A and B. Interestingly, these three devices were all amperometric systems, while the others were all photometric systems. However, we could not find supporting evidence from a review of the literature. In conclusion, our study demonstrated that when manpower is a concern, fluoride-containing tubes may be used with certain glucometers with acceptable results.
