El-Masri Hisham A
Experimental Toxicology Division, National Health and Environmental Effects Research Laboratory, U.S. Environmental Protection Agency, Research Triangle Park, NC 27711, USA.
Toxicol Appl Pharmacol. 2007 Sep 1;223(2):148-54. doi: 10.1016/j.taap.2006.07.009.
While procedures have been developed and used for many years to assess risk and determine acceptable exposure levels to individual chemicals, most cases of environmental contamination can result in concurrent or sequential exposure to more than one chemical. Toxicological predictions of such combinations must be based on an understanding of the mechanisms of action and interaction of the components of the mixtures. Statistical and experimental methods test the existence of toxicological interactions in a mixture. However, these methods are limited to experimental data ranges for which they are derived, in addition to limitations caused by response differences from experimental animals to humans. Empirical methods such as isobolograms, median-effect principle and response surface methodology (RSM) are based on statistical experimental design and regression of data. For that reason, the predicted response surfaces can be used for extrapolation across dose regions where interaction mechanisms are not anticipated to change. In general, using these methods for predictions can be problematic without including biologically based mechanistic descriptions that can account for dose and species differences. Mechanistically based models, such as physiologically based pharmacokinetic/pharmacodynamic (PBPK/PD) models, include explicit descriptions of interaction mechanisms which are related to target tissues levels. These models include dose-dependent mechanistic hypotheses of toxicological interactions which can be tested by model-directed experimental design and used to identify dose regions where interactions are not significant.
虽然用于评估风险和确定个体化学品可接受暴露水平的程序已经开发并使用多年，但大多数环境污染情况可能导致同时或相继接触多种化学品。对此类混合物的毒理学预测必须基于对混合物成分的作用机制和相互作用的理解。统计和实验方法用于测试混合物中毒理学相互作用的存在。然而，这些方法仅限于其推导所基于的实验数据范围，此外还受到实验动物与人类反应差异所造成的限制。诸如等效线图、中位效应原理和响应面方法（RSM）等经验方法基于统计实验设计和数据回归。因此，预测的响应面可用于在预计相互作用机制不会改变的剂量区域进行外推。一般来说，如果不包括能够解释剂量和物种差异的基于生物学的机制描述，使用这些方法进行预测可能会有问题。基于机制的模型，如基于生理学的药代动力学/药效学（PBPK/PD）模型，包括与靶组织水平相关的相互作用机制的明确描述。这些模型包括毒理学相互作用的剂量依赖性机制假设，可通过模型指导的实验设计进行测试，并用于识别相互作用不显著的剂量区域。
