中文摘要：

目的比较排序计数法和LMS（偏度系数-中位数-变异系数）方法计算江苏省居民膳食镉暴露量百分位数结果，探讨LMS法在处理膳食暴露年龄协变量上的优势。方法采用分布点评估模型，通过Monte Carlo模拟获得人群膳食镉暴露量分布，使用排序计数法和LMS方法分别计算镉暴露量年龄别百分位数。结果排序计数法和LMS法镉暴露量年龄组别百分位数差值均数的95％可信区问包含零，差异无统计学意义。LMS法提示儿童、青少年暴露量明显高于成人，18岁之后暴露量基本稳定。结论排序计数法和LMS法对江苏省居民膳食镉暴露年龄组别百分位数的评估结果基本一致，但LMS法在利用年龄协变量信息上较排序计数法有明显优势。

英文摘要：

Objective To compare the percentiles of dietary exposure of the population to cadmium in Jiangsu province calculated by LMS （Box-Cox power-median-coefficient of variation） and sort counting methods, and to explore the merit of LMS method in dealing with age covariant of dietary exposure. Methods Use distribution point model to estimate dietary exposure distribution through Monte Carlo simulation, calculate percentiles of dietary exposure of the population to cadmium by LMS and sort counting methods respectively. Results 95 % confidence interval of difference value means of percentiles of dietary exposure in different age groups included 0, there were no significant differences. LMS method showed that children and young person under 18 risked a higher cadmium exposure than adults. The exposure percentile curves of the above 18 age group stayed basicly stable. Conclusion Percentiles of dietary exposure of the population to cadmium in Jiangsu province estimated by LMS and sort counting methods were basicly the same but LMS method had the obvious merit to use age covariant.