Cardiovascular health and exposure to PM2.5 constituents: a multi-cohort study

项目摘要 心血管疾病（CVD）是美国人死亡的主要原因，也是导致寿命损失的主要原因 也是世界上最好的在北美和欧洲进行的队列研究提供了关键信息， 确定暴露于环境细颗粒物（PM2.5）是CVD风险增加的原因。他们有 并指出了未来研究的重要方向，包括PM2.5化学品的差异毒性 成分对CVD发病率的影响，浓度-反应（C-R）函数的形状，暴露时间的影响 窗户和室内空气污染暴露。我们研究的目的是评估 长期暴露于PM2.5成分和CVD事件，并调查暴露时间的影响 窗户和室内空气污染对这些协会。Approach.我们将进行回顾性多队列研究， 使用5个国家队列的个人健康记录（具有裁定的CVD终点）的中国研究 由阜外医院维持。为约120，000个队列提供个人水平的暴露轨迹 参与者在他们的后续阶段，我们将开发一个集成机器学习模型，使用地面 空气质量测量，卫星数据和空气质量模型模拟，以产生1公里分辨率的国家 PM2.5质量和主要成分（硫酸盐、硝酸盐、元素碳、有机碳、铅和镍） 1998年至2020年的暴露估计（目标1）。使用2007年至2020年的队列数据， 分析中，我们将使用考克斯比例风险模型来估计长期暴露与 PM2.5成分和不良心血管结局，包括总CVD、急性心肌梗死和 中风（目标2）。该模型的变体将用于研究C-R函数的形状（线性与线性）。 非线性）以及使用1998年至2020年期间的暴露估计值（Aim 2a）。我们还将使用时变和时- 不变暴露度量（目标2b）。最后，我们将评估室内空气污染的混杂效应， C-R函数使用一个子队列，估计个人PM2.5暴露估计值基于阜外 医院正在进行的多地区小组研究（Aim 2c）。预期成果。考虑到 我国PM2.5组分浓度随时间和空间的变化，我们的研究结果将为PM2.5 在美国和世界范围内，成分特异性C-R在CVD负担计算中发挥作用。我们的研究还将 扩大我们对影响这些关联的重要因素的知识，例如暴露窗口和室内 大气污染.

Project Summary Cardiovascular diseases (CVD) are the primary cause of death and a major cause of years of life lost in the US and in the world. Cohort studies conducted in North America and Europe have provided crucial information in identifying exposure to ambient fine particulate matter (PM2.5) as a cause of increased CVD risks. They have also pointed out important directions for future research, including the differential toxicity of PM2.5 chemical constituents to CVD incidence, the shape of concentration-response (C-R) functions, impact of exposure time windows and indoor air pollution exposure. The objectives of our study are to assess the associations between long-term exposure to PM2.5 constituents and incident CVD, and investigate the impact of exposure time window and indoor air pollution on these associations. Approach. We will conduct a retrospective multi-cohort study in China using personal health records from five national cohorts with adjudicated CVD endpoints maintained by the Fuwai Hospital. To provide individual-level exposure trajectories for ~120,000 cohort participants during their follow-up period, we will develop an ensemble machine learning model using ground air quality measurements, satellite data, and air quality model simulations to generate 1 km resolution national PM2.5 mass and major constituent (sulfates, nitrates, elemental carbon, organic carbon, lead, and nickel) exposure estimates between 1998 and 2020 (Aim 1). Using cohort data from 2007 to 2020 for our main analyses, we will use Cox proportional hazard models to estimate the associations of long-term exposures to PM2.5 constituents and adverse cardiovascular outcomes including total CVD, acute myocardial infarction, and stroke (Aim 2). Variations of this model will be used to investigate the shape of the C-R functions (linear vs. non-linear) as well as the impact of exposure errors using exposure estimates between 1998 and 2020 (Aim 2a). We will also examine windows of susceptibility for CVD endpoints using both time-variant and time- invariant exposure metrics (Aim 2b). Finally, we will evaluate the confounding effect of indoor air pollution on the C-R functions using a sub-cohort with estimated personal PM2.5 exposure estimates based on Fuwai Hospital's ongoing Multi-Region Panel Study (Aim 2c). Expected Results. Given the tremendous variability in PM2.5 constituent concentrations in China over time and space, our findings will contribute new PM2.5 constituent-specific C-R functions in CVD burden calculation in both the US and the world. Our study will also expand our knowledge on important factors affecting these associations such as exposure window and indoor air pollution.