BACKGROUND: Portable glucometers are often utilized at the patient's bedside in the ICU or operating room for frequent measurements of the blood glucose concentration. Many of these devices are based on a glucose oxidase method that may be influenced by PO2. The aim of this study was to evaluate the influence of a high PO2 of arterial blood on measured glucose values compared with venous blood. METHODS: Forty adult patients who underwent surgery with general anesthesia were included in this study. Each patient had a cannula inserted into the radial artery and a central venous catheter through the right internal jugular vein. Two hours after the induction of anesthesia, we drew arterial and venous blood and measured the blood glucose concentration using both a bedside glucometer based on a glucose oxidase method and a laboratory glucometer based on a hexokinase method. We also measured blood gas, electrolyte, and hematocrit values. Statistical analyses were performed with repeated measure ANOVA, multiple linear regression, and Bland-Altman's analysis. Data is expressed as mean +/- SD. RESULTS: The arterial blood glucose concentration measured by the glucose oxidase method (119.5 +/- 25.0 mg/dl) was significantly lower than the venous blood (133.5 +/- 24.8 mg/dl) and hexokinase method (134.2 +/- 27.1 mg/dl). There was no significant difference between the venous blood glucose concentration by the glucose oxidase method and hexokinase method. When we used the correction formula: corrected glucose value = arterial glucose value by glucose oxidase method + 0.1053 X PaO2 - 5.414, the bias improved from - 14.6 mg/dl to 1.0 mg/dl. CONCLUSIONS: The blood glucose concentration measured by the glucose oxidase method is more accurate in venous blood than oxygenated arterial blood. When we measure the blood glucose level using the glucose oxidase method, we should consider the influence of high oxygen tension.