Sarnat Jeremy A, Wilson William E, Strand Matthew, Brook Jeff, Wyzga Ron, Lumley Thomas
Department of Environmental and Occupational Health, Rollins School of Public Health of Emory University, Atlanta, Georgia 30322, USA.
J Expo Sci Environ Epidemiol. 2007 Dec;17 Suppl 2:S75-82. doi: 10.1038/sj.jes.7500621.
Examining the validity of exposure metrics used in air pollution epidemiologic models has been a key focus of recent exposure assessment studies. The objective of this work has been, largely, to determine what a given exposure metric represents and to quantify and reduce any potential errors resulting from using these metrics in lieu of true exposure measurements. The current manuscript summarizes the presentations of the co-authors from a recent EPA workshop, held in December 2006, dealing with the role and contributions of exposure assessment in addressing these issues. Results are presented from US and Canadian exposure and pollutant measurement studies as well as theoretical simulations to investigate what both particulate and gaseous pollutant concentrations represent and the potential errors resulting from their use in air pollution epidemiologic studies. Quantifying the association between ambient pollutant concentrations and corresponding personal exposures has led to the concept of defining attenuation factors, or alpha. Specifically, characterizing pollutant-specific estimates for alpha was shown to be useful in developing regression calibration methods involving PM epidemiologic risk estimates. For some gaseous pollutants such as NO2 and SO2, the associations between ambient concentrations and personal exposures were shown to be complex and still poorly understood. Results from recent panel studies suggest that ambient NO2 measurements may, in some locations, be serving as surrogates to traffic pollutants, including traffic-related PM2.5, hopanes, steranes, and oxidized nitrogen compounds (rather than NO2).
检验空气污染流行病学模型中所使用暴露指标的有效性，一直是近期暴露评估研究的重点。这项工作的主要目标，很大程度上是确定给定的暴露指标代表什么，并量化和减少因使用这些指标替代真实暴露测量而产生的任何潜在误差。本手稿总结了共同作者在2006年12月举行的美国环境保护局（EPA）近期研讨会上的发言，该研讨会探讨了暴露评估在解决这些问题中的作用和贡献。文中呈现了美国和加拿大的暴露与污染物测量研究结果以及理论模拟结果，以研究颗粒物和气态污染物浓度分别代表什么，以及它们在空气污染流行病学研究中使用时可能产生的误差。量化环境污染物浓度与相应个人暴露之间的关联，引出了定义衰减因子（或α）的概念。具体而言，表征特定污染物的α估计值，在开发涉及PM流行病学风险估计的回归校准方法中被证明是有用的。对于一些气态污染物，如二氧化氮（NO2）和二氧化硫（SO2），环境浓度与个人暴露之间的关联显示出很复杂，且仍未得到充分理解。近期专题研究的结果表明，在某些地点，环境NO2测量值可能正在充当交通污染物的替代指标，这些交通污染物包括与交通相关的PM2.5、藿烷、甾烷和氧化氮化合物（而非NO2）。
