ROC曲线在评价铅中毒指标中的应用

Application of Receiver Operating Characteristic Curve in Assessment of Indicators for Lead Poisoning

目的评价几种铅中毒指标筛选不同水平职业性铅接触的准确性。方法连续性收集１５７名铅接触工人的血标本，同时测定血铅、红细胞游离原卟啉（ＦＥＰ）、锌原卟啉（ＺＰＰ）、尿铅、尿δ氨基酮戊酸（δＡＬＡ），以血铅≥４０、≥６０μｇ／ｄｌ的铅接触水平为金标准，用ＲｅｃｅｉｖｅｒＯｐｅｒａｔｉｎｇＣｈａｒａｃｔｅｒｉｓｔｉｃ（ＲＯＣ受试者工作特征）曲线分析软件评价所测定指标筛选血铅≥４０、≥６０μｇ／ｄｌ职业性铅接触的价值，在ＲＯＣ曲线上制定各测定指标筛选上述２个铅接触水平的最佳临界点，用ＥＰＩＮＦＯＲ统计软件比较筛选血铅≥４０、≥６０μｇ／ｄｌ的平行试验的敏感度。结果筛选血铅≥４０、≥６０μｇ／ｄｌ的铅接触水平，ＲＯＣ曲线下的面积（ＡＵＣＲＯＣ）最大的是ＺＰＰ，其他依次为ＦＥＰ、尿铅、尿δＡＬＡ；联合ＺＰＰ与ＦＥＰ的平行试验的敏感度高于联合尿铅与尿δＡＬＡ的平行试验。结论筛选血铅≥４０、≥６０μｇ／ｄｌ水平的职业性铅接触，ＺＰＰ的准确性最高，并且联合ＺＰＰ与ＦＥＰ的平行试验的敏感度高于联合尿铅与尿δＡＬＡ的平行试验。

Objective To evaluate the accuracy of several indicators for lead poisoning in screening for occupational lead exposure to varied extent.Methods Blood specimens from 157 workers exposed to lead were collected and determined for levels of blood lead,free erythrocyte protoporphyrin(FEP),zinc protoporphyrin(ZPP),urine lead and δ aminolevulinic acid(δ ALA),Golden standards for lead exposure were defined as blood lead levels greater than 40 and 60 μg/dl,respectively.Diagnostic value for occupational exposure to lead with blood lead levels of greater than 40 and 60 μg/dl was evaluated by a receiver operating characteristic(ROC)curve software,and optimal cut off points for the above mentioned indicators for screening those with blood lead levels greater than 40 and 60 μg/dl were determined on the ROC curve.Sensitivity of biphasic screening in series with blood lead levels of greater than 40 and 60 μg/dl was compared with EPINFOR software.Results For screening those with blood lead levels greater than 40 and 60 μg/dl,ZPP covered the maximal area under the ROC curve(AUC ROC ),and the next was FEP,urine lead,and δ ALA.Sensitivity of screening with both ZPP and FEP in parallel was higher than that with both urine lead and δ ALA.Conclusion ZPP was the most accurate indicator for screening those occupational exposed with blood lead levels greater than 40 and 60μg/dl,and sensitivity of both ZPP with FEP was higher than that of both urine lead with δ ALA.