Impact of long-term temporal trends in fine particulate matter (PM2.5) on associations of annual PM2.5 exposure and mortality: An analysis of over 20 million Medicare beneficiaries.
Impact of long-term temporal trends in fine particulate matter (PM2.5) on associations of annual PM2.5 exposure and mortality: An analysis of over 20 million Medicare beneficiaries.
复制标题
细颗粒物 (PM2.5) 的长期时间趋势对年度 PM2.5 暴露与死亡率之间关系的影响：对超过 2000 万医疗保险受益人的分析。
DOI：
10.1097/ee9.0000000000000009
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Manjourides,Justin
中科院分区：
文献类型：
--
作者：
Eum,Ki-Do;Suh,HelenH;Pun,VivianChit;Manjourides,Justin
随着时间的推移，环境细颗粒物 (PM 2.5) 浓度不断降低，加上预期寿命不断增加，引发了人们对 PM 2.5 与死亡率之间关联的时间混杂的担忧。为了解决这个问题，我们研究了 2000 年 12 月至 2012 年 12 月期间居住在距环境保护局空气质量监测点 6 英里范围内的约 20,000,000 名美国医疗保险受益人的 PM 2.5 相关死亡风险比 (MRR)。我们通过研究 PM 2.5 相关 MRR 是否随研究周期长度的不同而变化，从而评估了时间混杂因素。然后，我们评估了控制时间混杂的三种方法：(1) 使用按时间回归的 PM 2.5 残差评估暴露；(2) 在健康模型中添加时间惩罚样条项； (3) 将描述 PM 2.5 时间变化的术语纳入健康模型，并使用分解方法估计该术语。我们发现，在 13 年的研究期间，PM 2.5 暴露量增加 10 μg/m 3 与死亡风险增加 1.20 倍（95% 置信区间 [CI]= 1.20, 1.21）相关，且相关程度随着研究周期的缩短而降低。 MRR 仍然具有统计显着性，但当模型根据 PM 2.5 的长期时间趋势进行调整时，MRR 会减弱。在 13 年的研究期间，基于残留、时间调整的 MRR 等于每 10 μg/m 3 1.12 (95% CI= 1.11, 1.12)，并且在检查较短的研究期间时没有变化。基于样条和分解的方法产生类似但不太稳定的 MRR。我们的研究结果表明，长期 PM 2.5 的流行病学研究可能会被长期时间趋势所混淆，并且可以使用 PM 2.5 按时间回归的残差来控制这种混淆。
Decreasing ambient fine particulate matter (PM 2.5) concentrations over time together with increasing life expectancy raise concerns about temporal confounding of associations between PM 2.5 and mortality. To address this issue, we examined PM 2.5-associated mortality risk ratios (MRRs) estimated for approximately 20,000,000 US Medicare beneficiaries, who lived within six miles of an Environmental Protection Agency air quality monitoring site, between December 2000 and December 2012. We assessed temporal confounding by examining whether PM 2.5-associated MRRs vary by study period length. We then evaluated three approaches to control for temporal confounding:(1) assessing exposures using the residual of PM 2.5 regressed on time;(2) adding a penalized spline term for time to the health model; and (3) including a term that describes temporal variability in PM 2.5 into the health model, with this term estimated using decomposition approaches. We found a 10 μg/m 3 increase in PM 2.5 exposure to be associated with a 1.20 times (95% confidence interval [CI]= 1.20, 1.21) higher risk of mortality across the 13-year study period, with the magnitude of the association decreasing with shorter study periods. MRRs remained statistically significant but were attenuated when models adjusted for long-term time trends in PM 2.5. The residual-based, time-adjusted MRR equaled 1.12 (95% CI= 1.11, 1.12) per 10 μg/m 3 for the 13-year study period and did not change when shorter study periods were examined. Spline-and decomposition-based approaches produced similar but less-stable MRRs. Our findings suggest that epidemiological studies of long-term PM 2.5 can be confounded by long-term time trends, and this confounding can be controlled using the residuals of PM 2.5 regressed on time.