Department of Environmental Science, Institute for Water and Wetland Research, Radboud University Nijmegen, P.O. Box 9010, 6500, GL, Nijmegen, the Netherlands.
Department of Environmental Science, Institute for Water and Wetland Research, Radboud University Nijmegen, P.O. Box 9010, 6500, GL, Nijmegen, the Netherlands; Environment Department, University of York, York, United Kingdom.
Environ Int. 2019 May;126:37-45. doi: 10.1016/j.envint.2019.01.080. Epub 2019 Feb 15.
The use of down-the-drain products and the resultant release of chemicals may lead to pressures on the freshwater environment. Ecotoxicological impact assessment is a commonly used approach to assess chemical products but is still influenced by several uncertainty and variability sources. As a result, the robustness and reliability of such assessments can be questioned. A comprehensive and systematic assessment of these sources is, therefore, needed to increase their utility and credibility. In this study, we present a method to systematically analyse the uncertainty and variability of the potential ecotoxicological impact (PEI) of chemicals using a portfolio of 54 shampoo products. We separately quantified the influence of the statistical uncertainty in the prediction of physicochemical properties and freshwater toxicity as predicted from Quantitative Structure-Property Relationships (QSPRs) and Quantitative Structure-Activity Relationships (QSARs) respectively, and of various sources of spatial and technological variability as well as variability in consumer habits via 2D Monte Carlo simulations. Overall, the variation in the PEIs of shampoo use was mainly the result of uncertainty due to the use of toxicity data from three species only. All uncertainty sources combined resulted in PEIs ranging on average over seven orders of magnitude (ratio of the 90th to the 10th percentile) so that an absolute quantification of the ecological risk would not be meaningful. In comparison, variation in shampoo composition was the most influential source of variability, although less than compared to uncertainty, leading to PEIs ranging over three orders of magnitude. Increasing the number of toxicity data by increasing the number of species, either through additional measurements or ecotoxicological modelling (e.g. using Interspecies Correlation Equations), should get priority to improve the reliability of PEIs. These conclusions are not limited to shampoos but are applicable more generally to the down-the-drain products since they all have similar data limitations and associated uncertainties relating to the availability of ecotoxicity data and variability in consumer habits and use.
将产品直接排入下水道，导致其中的化学物质释放，可能会对淡水环境造成压力。生态毒理学影响评估是评估化学产品的常用方法，但仍受到多种不确定性和变异性来源的影响。因此，需要对这些来源进行全面系统的评估，以提高其效用和可信度。在这项研究中，我们提出了一种方法，使用 54 种洗发水产品的组合，系统地分析化学物质潜在生态毒性（PEI）的不确定性和变异性。我们分别量化了预测物理化学性质和淡水毒性的定量结构-性质关系（QSPR）和定量结构-活性关系（QSAR）预测中统计不确定性的影响，以及各种空间和技术变异性来源以及通过 2D 蒙特卡罗模拟的消费者习惯变异性的影响。总体而言，洗发水使用的 PEI 变化主要是由于仅使用三种物种的毒性数据导致的不确定性所致。所有不确定性来源的综合结果导致 PEI 平均范围超过七个数量级（第 90 百分位与第 10 百分位的比值），因此对生态风险进行绝对量化没有意义。相比之下，洗发水成分的变化是变异性的最主要来源，尽管其影响小于不确定性，导致 PEI 范围超过三个数量级。通过增加物种数量（无论是通过额外的测量还是生态毒理学建模，例如使用种间相关方程）来增加毒性数据数量，应优先考虑提高 PEI 的可靠性。这些结论不仅限于洗发水，而且更普遍适用于直接排入下水道的产品，因为它们都具有类似的数据限制以及与生态毒性数据的可用性和消费者习惯和使用的变异性相关的不确定性。
