摘要
目的:探讨诊断2型糖尿病时糖化血红蛋白(HbA1c)的最佳切点,并对比分析空腹血糖(FPG)和HbA1c对于诊断的敏感性和特异性。方法:117名2型糖尿病高危人群,男67名,女50名,以1999年美国糖尿病学会(ADA)标准为诊断标准,通过绘制受试者工作特征曲线(ROC曲线),分析比较FPG及HbA1c诊断的敏感性和特异性。结果:①HbA1c的ROC曲线的最佳切点为6.45%,灵敏度74.6%,特异性90.1%,曲线下面积0.91(95%CI 0.83-0.95),阳性似然比9.6,阴性似然比0.27,阳性预测值93.0%,阴性预测值70.4%。②FPG ROC曲线的最佳切点为6.45 mmol/L,灵敏度87.3%,特异性85.9%,曲线下面积0.92(95%CI 0.86-0.96),阳性似然比7.0,阴性似然比0.15。③当HbA1c≥6.45%时,诊断的准确性为93.0%,而FPG≥6.45 mmol/L时诊断的准确性为84.1%。结论:HbA1c≥6.45%可作为诊断2型糖尿病的一种可靠方法,且诊断的准确性较FPG高,敏感性较FPG低,若同时联合检测分别以HbA1c 6.45%和FPG6.45 mmol/L切点,诊断DM的敏感性和特异性分别为90.1%和95.8%。
Objective:Explore diagnosed type 2 diabetes glycated hemoglobin(HbA1c) the best non-oil imports,and comparative analysis of fasting plasma glucose(FPG) and HbA1c to the diagnosis of sensitivity and specificity.Methods: 117 in type 2 diabetes high-risk groups,male 67,female 50 to 1999 American diabetes association(ADA) criteria for diagnosis criteria,draw a receiver-operating characteristic curve(ROC curves),analysis and comparison FPG HbA1c diagnosis and the sensitivity and specificity.Results:(1) the best HbA1c the ROC curves for 6.45% non-oil imports,sensitivity and specificity,curve 90.1% 74.6% in the area under 0.91(95% CI 0.83-0.95),positive,negative likelihood ratio 9.6 likelihood ratio,0.27 positie predictie 70.4% 93.0%,negative predictive value.2 the best FPG ROC curves for 6.45 mmol/l non-oil imports 87.3%,specificity and sensitivity,curve 85.9% area under the following CI 0.86-0.92(95%),positive,negative likelihood ratio 7.0 likelihood ratio 0.15.3 when HbA1c p 6.45%,the diagnostic accuracy of 93.0%,and for mmol/l p FPG 6.45 when the diagnostic accuracy of is 84.1%.Conclusion: HbA1c p 6.45% can serve as a diagnosis of type 2 diabetes a reliable method,and the diagnostic accuracy of high sensitivity is relatively FPG FPG low,if also joint test respectively by 45% HbA1c6.FPG6.45 mmol/l and non-oil imports,diagnostic DM sensitivity and specificity 90.1% and localizing respectively.
出处
《中国卫生检验杂志》
CAS
2011年第5期1226-1227,共2页
Chinese Journal of Health Laboratory Technology