Coyne S, Lacour B, Hennequin-Le Meur C
Laboratoire de biochimie générale, Hôpital Necker-Enfants Malades, APHP, Paris, France.
Ann Biol Clin (Paris). 2008 May-Jun;66(3):249-54. doi: 10.1684/abc.2008.0223.
In order to build a continuous quality improvement approach for control of glucose meters in clinical divisions at Necker-Enfants Malades hospital, the analytical performances (precision and accuracy) of 2 glucose meters have been evaluated in our laboratory according to SFBC recommendations. Fifty-six heparinized whole blood specimens from patients and thirty-nine from healthy volunteers were analyzed on each of the two meters and compared to plasma glucose measurement on the Roche Hitachi 917 system. The correlation coefficient was 0.938 for Optium Xceed and 0.911 for One Touch Ultra. However, 14.7% and 18.9% of the results (n = 95) for respectively Optium Xceed and One Touch Ultra were discordant, i.e. higher than a 20% difference compared to reference blood glucose concentrations. Inaccuracy was more important for low glucose concentrations (< 5 mmol/L; 12/14 discrepant samples for Optium Xceed and 16/19 for One Touch Ultra). This data suggests a lack of accuracy, particularly for low glucose concentrations. Capillary blood glucose concentrations must therefore be interpreted with caution concerning the diagnosis of hypoglycemia and treatment of unstable patients. Moreover, quality control of glucose meters (blood glucose determinations concurrently at bedside and in the laboratory) is difficult to perform. It also raises questions about the responsibility of "point-of-care testing", an area still subject to discussion.
为了在巴黎内克尔儿童医院的临床科室建立一种持续质量改进方法来控制血糖仪，我们实验室根据法国生物临床分析协会（SFBC）的建议，对两款血糖仪的分析性能（精密度和准确度）进行了评估。在这两款血糖仪上分别分析了来自患者的56份肝素化全血标本和来自健康志愿者的39份标本，并与罗氏日立917系统上的血浆葡萄糖测量结果进行比较。Optium Xceed血糖仪的相关系数为0.938，One Touch Ultra血糖仪的相关系数为0.911。然而，Optium Xceed和One Touch Ultra血糖仪分别有14.7%和18.9%的结果（n = 95）不一致，即与参考血糖浓度相比差异高于20%。对于低葡萄糖浓度（< 5 mmol/L），不准确度更为明显（Optium Xceed血糖仪有12/14份差异样本，One Touch Ultra血糖仪有16/19份）。该数据表明存在准确度不足的问题，尤其是对于低葡萄糖浓度。因此，在诊断低血糖和治疗不稳定患者时，必须谨慎解读毛细血管血糖浓度。此外，血糖仪的质量控制（同时在床边和实验室进行血糖测定）难以实施。这也引发了关于“即时检验”责任的问题，这一领域仍有待讨论。
