Research On Accuracy And Clinical Influence Of Bedside Glucometry In Critically Ill Patients

Objective To determine the accuracy and clinical impact of two common methods of bedside point-of-care testing(POCT) for glucose measurements. The aim is to make glucometry used safely and normatively in ICU and radically reduce incidence rates of hypoglycemia in critical ill patients during tight glucose control period.Methods ICU patients aged≥18 years were enrolled in this study. During blood glucose monitoring, the patient's blood samples of capillary, vein and artery were drawn simultaneously. The glucose value of blood samples was detected respectively by glucometry of glucose oxidase(GOD) and glucose dehydrogenase(GDH) method. The patients'real-time laboratory parameters (hematocrit, lactic acid etc.) and assessment of organ function, critical degree(APACHEⅡscore) were recorded. Compared with reference standred of central laboratory, the accuracy of bedside POCT for glucose measurements among different methods, different blood samples was analysed.The influence factors of glucometry were analysed statisticly by the logistic regression method.Results 240 patients were enrolled in this study. (1) A strong correlation was found between the laboratory method and GOD or GDH method (r 0.953～0.980, P<0.05). (2) The average blood glucose level (BGL) of capillary, venous and arterial samples measured by GOD method (150.5±64.8 mg/dL, 152.2±59.5mg/dL, 151.6±61.9mg/dL) was respectively higher than laboratory reference (135.0±59.1mg/dL, P<0.01). (3) As to laboratory reference, glucose values of capillary, venous and arterial samples measured by GOD method were inaccurate in 21.7%, 27.9%, 28.7% of samples. Glucose values measured by GDH method were respectively inaccurate in 5.4%, 13.3%, 11.2% of samples. (4) During laboratory reference value≤80mg/dL, glucose values of capillary, venous and arterial samples measured by GOD method conflicted with laboratory reference values in 59.4%, 71.9%, 71.9% of samples. Glucose values measured by GDH method were respectively inaccurate in 25.0%, 40.6%, 40.6% of samples. (5) The occurance rate of error measurements by GOD or GDH method in MODS patients was significantly higher than non-MODS patients (P<0.05). With the patients'APACHEⅡscore increased, the inaccurate rates of measurement for three kinds of samples measured by GOD or GDH method were gradually increasing. (6) The accuracy of POCT for capillary glucose measurements was significantly effected by insulin application, fingertip skin changes and arterial pressure (P<0.05). (7) On the impact factors of accuracy of POCT for glucose measurements, BGL, hematocrit level and APACHEⅡscores were into the logistic regression equation.Conclusions (1) Bedside POCT methods for glucose measurements had a strong correlation with laboratory method, but they tended to yield higher glucose values. (2) The risk of inaccurate measurements by GOD method in critically ill patients (CRRT excluded) was higher than GDH method. (3) During hypoglycemia, clinical agreement between laboratory method and POCT methods for glucose measurements was obviously fallen. (4) The accuracy of POCT for glucose measurements was significantly effected by critical degree, hematocrit level and BGL in ICU. (5) The accuracy of POCT for capillary glucose measurements was significantly effected by insulin application fingertip, skin changes and mean arterial pressure.