中文摘要：

目的评价葡萄糖氧化酶法（GOD法）和葡萄糖脱氢酶法（GDH法）检测重症患者床旁即时血糖的准确性。方法危重症患者240例，年龄20～88岁，性别不限，急性生理学与慢性健康状况评分系统Ⅱ评分1～45分。床旁同步采集患者静脉血、动脉血、指端血，分别采用GOD法和GDH法检测即时血糖，同时以中心实验室己糖激酶法测定血糖作为标准水平，即时血糖检测水平与标准水平进行clarke误差表格分析和Bland—Ahman一致性检验，并评价位于Bland—Altmen一致性界限内的即时血糖检测水平的准确性。结果Clarke误差表格分析显示：静脉血、动脉血、指端血样即时血糖检测水平与标准水平的差异率分别有98．7％、98．3％、98．3％（GDH法）及96．2％、96．6％、96．7％（GOD法）位于临床准确区（A区）与临床可接受区（B区），1．2％、1．7％、1．7％（GDH法）及2．9％、3．3％、3．3％（GOD法）位于数据错误区（D区），GDH法没有位于数据误差区（c区）和严重错误区（E区）的情况，GOD法0．8％静脉血即时血糖检测水平与标准水平的差异率位于数据误差区（C区），没有位于严重错误区（E区）的情况。Bland．Altman一致性检验结果：标准水平与静脉血、动脉血、指端血即时血糖检测水平的差值依次为-0．1、-0．3和-0．2mmol／L（GDH法）及-0．9、-1．0和-0．9mmol／L（GOD法），超上、下限范围发生率为4．6％、6．7％、6．6％（GDH法）及4．5％、5．0％、7．1％（GOD法）。位于Bland-Altman一致性界限内的静脉血、动脉血、指端血样即时血糖检测水平的准确率分别为94．3％、92．1％、93．7％（GOD法）和96．6％、95．1％、95．5％（GDH法）。结论GOD法和GDH法检测重症患者床旁即时血糖总体上具有良好的准确性，但有高估实际血糖水平的可能。

英文摘要：

Objective To evaluate the accuracy of point-of-care testing （POCT） for blood glucose monitoring in critically ill patients. Methods Two hundred and forty critically ill patients, of both sexes, aged 20- 88 yr, with Acute Physiology and Chronic Health Evaluation II score of 1-45, were enrolled. The venous, arterial and capillary blood samples were collected to determine the real-time blood glucose level using glucose oxidase （GOD） and glucose dehydrogenase （GDH） methods. The blood glucose level measured by central laboratory hexokinase method simultaneously was served as standard level. Error Grid analysis （EGA） and Bland-Ahman analysis were used to determine accuracy and consistency, respectively. The accuracy of real-time blood glucose levels within the consistent limits was evaluated. Results 1. The results of EGA showed that 98.7 %, 98.3 %, 98.3% （ GDH method） and 96.2%, 96.6%, 96.7% （GOD method） of the difference between venous, arterial and capillary blood glucose levels measured and the standard level were located in the A and B zones, respectively, and 1.2%, 1.7%, 1.7% （GDH method） and 2.9% , 3.3% , 3.3% （GOD method） in the D zone. 0.8% （GOD method） of the difference between venous blood glucose levels and the standard level were located in the C zone. 2. Bland-Altman analysis showed that the difference between the standard level and glucose level measured in blood samples from the vein, artery and capillary was - 0.1, - 0.3, - 0.2 mmol/L （ GDH method） and - 0.9, - 1.0, -0.9 mmol/L （GOD method）, respectively, and the incidence beyond the upper and lower limits of consistency zone was 4.5%, 6.7% , 6.6% （GDH method） and 4.6% , 5.0%, 7.1% （GOD method）, respectively. The accuracy of venous, arterial and capillary blood glucose levels within the consistent limits was 94.3 %, 92. 1 %, 93.7%（GOD method） and 96.6%,95.1% ,95.5%（GDH method）, respetively. Conclusion The accuracy of POCT for blood glucose monitored by GOD and GDH methods is good in critically