Accuracy of bedside glucose measurement from three glucometers in critically ill patients

Objective: Implementation of strict glucose control in most intensive care units has resulted in increased use of point-of- care glucose devices in the intensive care unit. The aim of this study was to determine the reliability of point-of-care testing glucose meters among critically ill patients under intensive insulin treatment. Design: Prospective observational study. Patients. Intensive care unit and non-intensive care unit patients in a tertiary care teaching hospital. Measurements. A glucose oxidase method was used to validate the point-of-care testing devices. Three different point-of-care testing devices, Accu-Chek Sensor (Roche Diagnostics), Precision (Abbott Diagnostics), and HemoCue were tested. Glucose measurements were performed in duplicate by an experienced technician under standardized conditions in the hospital's laboratory, using arterial (intensive care unit patients) and arterial or venous (non-intensive care unit patients) heparinized whole blood samples. Main Results: A strong correlation was found between the glucose oxidase method and the Accu-Chek device (r(2) = .9596, p < 0.001). Mean absolute difference between the glucose oxidase and Accu-Chek was -0.32 mmol/L (95% confidence interval -0.84 to 1.48 mmol/L). Using the International Organization for Standardization (ISO) criteria, 27 of 197 samples (13.7%) were inaccurate. In all samples that failed to meet the ISO criteria, glucose values measured by the Accu-Chek device were higher compared with the glucose oxidase method. In another set of experiments among intensive care unit patients, strong positive correlations were also found between the other point-of-care testing devices and the glucose oxidase method. However, paired samples from Accu-Chek, HemoCue, and Precision failed the ISO criteria in 9 of 82 (11.0%), 4 of 82 (4.9%), and 11 of 82 (13.4%) of cases, respectively. In non-intensive care unit patients paired samples from Accu-Chek, HemoCue, and Precision failed the ISO criteria in 3 of 120 (2.5%), 11 of 120 (9.2%), and 16 of 120 (13.3%) cases, respectively. Conclusions. Under standardized conditions, glucose results from three point-of-care testing devices were inaccurate in both intensive care unit and non-intensive care unit patients. Among intensive care unit patients, inaccurate glucose readings were most frequently falsely elevated, resulting in misinterpretation of high glucose values with subsequent inappropriate insulin administration or masking of true hypoglycemia. (Crit Care Med 2008; 36:3062-3066)