通过不同检测系统评价降钙素原对血流感染的临床诊断价值

目的通过溯源及检测原理不同的两种仪器评价降钙素原对血流感染的临床诊断价值。方法选取北京协和医院门诊及住院血培养报警阳性的患者标本91份及体检健康标本100份,分别使用VIDAS 30全自动免疫分析仪(简称VIDAS 30)及iFlash 3000化学发光免疫分析仪(简称iFlash 3000)检测降钙素原,分析其变化与血流感染发生的相关性及其对血流感染的诊断效能。结果两种方法拟合的线性回归方程为Y=0.823X-0.374(r=0.987,P<0.01);以血培养为“金标准”,通过VIDAS 30及iFlash 3000检测降钙素原诊断血流感染的灵敏度分别为64.7%、61.2%,特异度均为100.0%,准确度分别为84.3%、82.7%,阳性预测值均为100.0%,阴性预测值分别为77.9%、76.3%;两种方法诊断血流感染的受试者工作特征曲线下面积(AUC)分别为0.913、0.876(P=0.001、0.002),最佳截断值分别为0.29、0.27 ng/mL,此时的灵敏度分别为75.3%、71.8%,特异度均为100.0%。结论溯源及原理不同的两种仪器检测降钙素原对血流感染的诊断均有较好的临床应用价值,合理调整最佳截断值可以提高血流感染诊断的灵敏度。

The evaluation of the clinical diagnostic value of procalcitonin in blood infection by different detection systems

Objective To evaluate the clinical diagnostic value of procalcitonin in blood infection by two kinds of instruments.Methods 91 samples with positive blood culture alarm and 100 samples from healthy persons were selected from Beijing Union Hospital.The samples were detected procalcitonin by Automatic Immune Analyzer VIDAS 30(VIDAS 30)and iFlash 3000 Chemiluminescence Immunoassay Analyzer(iFlash 3000).The correlation between the changes and the occurrence of blood infection and its diagnostic efficacy were analyzed.Results The linear regression equation fitted by the two methods is Y=0.823X-0.374(r=0.987,P<0.01).When blood culture was taken as the"gold standard",the sensitivity of VIDAS 30 and iFlash 3000 method was 64.7%and 61.2%,specificity was 100.0%,accuracy was 84.3%and 82.7%,positive predictive value was 100.0%,negative predictive value was 77.9%and 76.3%,respectively.The area under receiver operating characteristic curve(AUC)for diagnosing the blood infection was 0.913 and 0.876(P=0.001,0.002),the cutoff value was 0.29 and 0.27 ng/mL,and then the sensitivity of the two methods was 75.3%and 71.8%respectively,the specificity was both 100.0%.Conclusion The procalcitonin detected by two different instruments has good clinical value in the diagnosis of blood infection,and the sensitivity of blood infection diagnosis can be improved by adjusting the best cutoff value reasonably.