COMPARISON OF ACCURACY GLUCOMETER USED IN OUR HOSPITAL WITH HEXOKINASE METHOD.

Autor: Çolak, Emel, Çalık, Gizem Yılmaz, Ibiş, Oğulcan, Haskılıç, Yunus Emre, Şeneş, Mehmet, Yücel, Doğan
Předmět:
Zdroj: Turkish Journal of Biochemistry / Turk Biyokimya Dergisi; 2019 Supplement, Vol. 44, p46-46, 1/3p
Abstrakt: Objectives: The aim of this study is to compare the results of the glucose meter device used in our hospital with the glucose measurement results of an automated analyser. Materials and Methods: 61 patients who were admitted to the emergency department of our hospital and arterial blood gases requested were included in the study. One sample below 50 mg/dL glucose concentration, eight samples in the 50-80 mg/dL, twenty-three samples in the 80-120 mg/dL,eleven samples in the 120-200 mg/dL, eight samples in the 200-300 mg/dL, five samples in the 300-400 mg/dL and five samples above 400 mg/dL were included in the study. A sample was stored at room temperature to obtain a concentration below 50 mg/dL. Three samples were stored at room temperature for 50-80 mg/dL concentration. 10% Dextrose was added to two samples to obtain a concentration of 300-400 mg/dL. To obtain values above 400 mg/dL, 10% Dextrose was added to four samples. Each sample was analyzed in duplicate with the Abbott Freestyle Precision Pro glucose meter (glucose dehydrogenase) and the Roche cobas 6000 automated analyser (hexokinase). Automated analyser was accepted as a reference method. Whole blood samples admitted to the laboratory were studied in glucose meter device within five minutes, and then the plasma samples were separated by centrifugation at 1300 g for 10 minutes and analyzed in the automated analyzer. Data were evaluated with Deming regression analysis and Clarke Error Grid analysis in EP Evaluator. Results: We calculated CV values as 4. 94% for glucose meter and 1. 85% for automated analyser from data obtained from duplicate measurements. Correlation between automated analyser and glucose meter data were found as R: 0. 9950 by the Pearson correlation analysis. In the Deming regression analysis the coefficient of determination (R2) = 0. 99, Slope = 0. 9659 (95% CI = 0. 9288 to1. 0031); intercept = -1. 9663 (95% CI = -6. 9379 to 3. 0053) were found. In the Clarke Error Grid analysis, 98% of the samples were in area A and 2% in area B. Conclusions: Glucose meter and automated analyser (hexokınase) results were found to be consistent in Deming regression analysis. With the Clarke Error Grid analysis, 98% of the data were in the A area. Glucose meter results are compatible with the reference method. [ABSTRACT FROM AUTHOR]
Databáze: Complementary Index