剂量-效应关系两种置信区间的比较

毒性实验存在随机误差,同时剂量-效应曲线(DRC)拟合也会产生误差.因此,合理表征剂量-效应关系及其置信区间非常重要.本研究在剂量-效应曲线非线性拟合基础上,阐述了构建拟合函数置信区间(FCI)与实验观测置信区间(OCI)的基本原理.并以对氯苯酚及等效应浓度比混合物为例,说明目前多数文献使用的FCI不能有效地描述真实毒性实验的不确定度,提出以OCI表征DRC,以求真实反映观测毒性效应的不确定度.利用剂量加和与独立作用模型预测了混合物毒性,并与OCI比较,得到混合物毒性作用类型.建议在分析混合物协同、拮抗、独立与加和作用时,应充分考虑OCI.

Comparision between two confidence intervals of dose-response relationships

Random error exists not only in toxicity experiments but also in the fitting of dose-response curves (DRC). It is, therefore, very important to rationally characterize the dose-response relationship and its confidence interval (CI). In the present research, the basic principles of the function-based confidence interval (FCI) and the observation-based confidence interval (OCI) were explained on the basis of the nonlinear least square fitting of DRC. p-Chlorophenol and an equipotent concentration ratio mixture were taken as example to illustrate why the FCI used in most literatures can not describe the uncertainty of experimental data efficiently. Therefore, to really reflect the uncertainty in toxicity test, the OCI should be selected as an alternative to characterize DRC. Thus, patterns of mixture toxic action (PMTA) might be induced from the comparison of the experimental OCI with the toxicity values predicted by dose addition (DA) and independent action (IA). It is suggested that the OCI should be sufficiently taken into account in the analysis of PMTA such as synergism, antagonism, IA and DA.