Effect of Household Air Pollution on the Gut Microbiome and Virome of Adult Women Living in Uganda

BACKGROUND: Emerging observational studies suggest
that air pollution can influence the gut microbiome. However, this
association is often highly confounded by factors, such as diet and
poverty. The gut virome may influence respiratory health independent
of the gut microbiome. We recently demonstrated in a randomized waitlist-controlled
trial (ClinicalTrials.gov NCT03351504) that a clean lighting intervention
reduced the level of personal exposure to air pollution among adult
women in rural Uganda. OBJECTIVES: To determine the effect
of a solar lighting intervention on changes to the gut microbiome
and virome and secondarily to determine the association between these
changes on lung health. METHODS: Between 2018 and 2019,
we collected stool samples and assessed respiratory symptoms and spirometry
from 80 adult women living in rural Uganda at baseline and 12 and
18 months postrandomization. The intervention group received a solar
lighting system after randomization, while the waitlist-controlled
group received one at 12 months. Deep metagenomics sequencing of stool
was performed and profiled for nonviral and viral taxonomic composition.
The primary analysis focused on pre- vs postintervention changes due
to
power considerations, adjusting for potential confounding by age,
diet, antibiotic use, and season. A sensitivity analysis was conducted
using intention-to-treat principles. When comparing pre- vs postintervention
periods, we used sparse partial least-squares models to identify nonviral
and viral signatures of reduced air pollution exposure. Mixed effects
models were used to evaluate changes in health outcomes as well as
associations between microbial signatures of reduced air pollution
exposure and health. RESULTS: The average age was 39.2
years. The solar lighting intervention led to larger changes in viral
compared to nonviral microbial community structure and differential
abundance of bacteria, eukaryotes, and viruses. Provision of solar
lighting systems was associated with a reduction in the presence of
respiratory symptoms from 57.1% to 36.1% (p = 0.002),
while there was no impact on lung function. Microbiome and virome
signatures had AUCs of 0.74 and 0.76, respectively, in predicting
pre- vs postintervention stool samples. Microbiome signatures were
associated with a lower risk of respiratory symptoms (OR = 0.68 (0.49
– 0.94), p = 0.020). CONCLUSION: Among adult women living in rural Uganda, both nonviral and viral
components of the gut microbial community changed after a clean lighting
intervention. Microbiome signatures reflective of lower air pollution
exposures were associated with improved respiratory symptoms. These
observations suggest that air pollution may influence lung health
through the gut-lung axis, warranting further exploration in future
intervention studies.

背景：新兴观察性研究表明，空气污染可影响肠道微生物组（gut microbiome）。然而，该关联常受到饮食、贫困等因素的强烈混杂。肠道病毒组（gut virome）可独立于肠道微生物组影响呼吸系统健康。我们近期在一项随机等待名单对照试验（ClinicalTrials.gov NCT03351504）中证实，清洁照明干预可降低乌干达农村成年女性的个人空气污染暴露水平。
研究目的：明确太阳能照明干预对肠道微生物组与肠道病毒组变化的影响，并次要探究上述变化与肺部健康之间的关联。
研究方法：2018至2019年间，我们于基线、随机分组后12个月及18个月时，从乌干达农村的80名成年女性中收集粪便样本，评估其呼吸症状与肺量测定结果。干预组在随机分组后即刻获得太阳能照明系统，等待名单对照组则于12个月时获得该系统。我们对粪便样本开展深度宏基因组测序（deep metagenomics sequencing），并对其非病毒与病毒分类组成进行分析。基于检验效能考量，主要分析聚焦于干预前后的变化，并校正年龄、饮食、抗生素使用及季节等潜在混杂因素。我们采用意向性治疗原则开展敏感性分析。在比较干预前后阶段时，我们使用稀疏偏最小二乘模型（sparse partial least-squares models）识别与空气污染暴露降低相关的非病毒与病毒特征。我们采用混合效应模型（mixed effects models）评估健康结局的变化，以及与空气污染暴露降低相关的微生物特征与健康之间的关联。
研究结果：受试者平均年龄为39.2岁。相较于非病毒微生物群落结构，太阳能照明干预对病毒群落结构的改变更为显著，同时可引起细菌、真核生物与病毒的丰度差异。提供太阳能照明系统与呼吸症状患病率从57.1%降至36.1%相关（p=0.002），但对肺功能无显著影响。在预测干预前后粪便样本时，微生物组与病毒组特征的曲线下面积（AUC）分别为0.74与0.76。微生物组特征与呼吸症状风险降低相关（比值比OR=0.68，95%置信区间0.49~0.94，p=0.020）。
研究结论：在乌干达农村成年女性群体中，清洁照明干预后肠道微生物群落的非病毒与病毒组分均发生显著变化。反映空气污染暴露水平降低的微生物组特征与呼吸症状改善相关。上述结果提示，空气污染或可通过肠-肺轴（gut-lung axis）影响肺部健康，未来需开展更多干预研究进一步探索该方向。