中国花生黄曲霉毒素风险评估中膳食暴露非参数概率评估方法

基于2009年、2010年对全国13个花生主产省155个县市2 571份代表性花生样品检测分析获得的12 855个黄曲霉毒素检测数据,以及卫生部2002年、2004年中国居民营养与健康状况调查报告中膳食消费数据,采用Monte Carlo模拟和Bootstrap抽样方法量化人群花生黄曲霉毒素膳食暴露量的变异度和不确定度,构建了中国花生黄曲霉毒素风险评估中膳食暴露非参数概率评估模型,依据年龄段、地区间和收入水平差异的人群花生黄曲霉毒素膳食暴露量,分析了各暴露量分布的指标统计量和90%置信区间。评估结果显示,2～5岁儿童为花生黄曲霉毒素膳食摄入的高暴露人群,其高端暴露量（P97.5）中位值为每千克体重每天1.342ng（表示为ng.kg-1.d-1,下同）,四分位间距为0.151ng.kg-1.d-1;18～45岁成人中,城市人群花生黄曲霉毒素膳食暴露量高于农村,其高端暴露量（P97.5）中位值为0.695ng.kg-1.d-1,四分位间距为0.078ng.kg-1.d-1;中等收入水平人群花生黄曲霉毒素膳食暴露量最高,其高端暴露量（P97.5）中位值为0.874ng.kg-1.d-1,四分位间距为0.086ng.kg-1.d-1。花生黄曲霉毒素膳食暴露非参数概率评估方法克服了点评估方法不能定量评估结果变异度和不确定度的缺点,结果更为精确。

Non-parametric probability assessment of risk evaluation on dietary exposure to Chinese peanut aflatoxins

Based on 12 855 aflatoxins data from 2 571 peanut samples of 155 counties in 13 provinces in 2009 and 2010,and according to the dietary consumption data in the Chinese resident nutrition and health survey reports in 2002 and 2004,a non-parametric probability assessment model was built using Monte Carlo simulation and Bootstrap sampling methods.In the model,dietary aflatoxins exposure distributions were constructed and quartiles of corresponding statistics were calculated.Results showed that children aged 2 to 5,adults aged 18 to 45 from cities or the middle-income aged 18 to 45 were the high aflatoxins exposure populations.Their median values of high exposure（P97.5） were 1.342ng·kg-1·d-1,0.695ng·kg-1·d-1 and 0.874 and with the quartiles of P97.5 as 0.151ng·kg-1·d-1,0.078ng·kg-1·d-1 and 0.086ng·kg-1·d-1,respectively.Compared to point estimate,present assessment was more accurate on variability and uncertainty in quality risk assessments.