Combinations of Air Pollutants Linked With Adverse Childhood Asthma Outcomes

Published on: 

Twenty combinations of toxic pollutants were identified by a machine learning algorithm as being tied with asthma outcomes.

The idea that air pollutants are associated with childhood asthma is well established, but a new study suggests particular mixtures of pollutants can further exacerbate asthma risk.

The study was published in The Journal of Clinical Investigation. It was based on a novel approach to asthma epidemiology—using a machine learning algorithm to draw connections between real-world asthma outcomes and the pollutants to which the patients might have been exposed.

“Asthma is one the most prevalent diseases affecting children in the United States,” said senior author Supinda Bunyavanich, MD, MPH, MPhil, of the Icahn School of Medicine at Mount Sinai.

Bunyavanich and colleagues suspected that combinations of air toxics—hazardous pollutants that can lead to health effects—might increase the risk of asthma in children. To test their hypothesis, they used geocoding levels of 125 air toxics from the US Environmental Protection Agency’s National Air Toxic Assessment. They then mapped those data to the addresses of 151 children who participated in a study called the Airway in Asthma study.


The goal of the model was to examine potential links between toxic exposures early in life and particular asthma outcomes around age 12, such as the need for daily asthma medication, visits to emergency departments, and hospitalizations.

The model identified 18 individual pollutants that correlated with adverse asthma outcomes. Yet, it also found another 20 combinations of pollutants that appeared to increase the risk of negative asthma outcomes.

“Traditionally, for technical reasons, it has been difficult to study the health effects of more than one toxic at a time,” said Gaurav Pandey, PhD, another senior author of the study. “We overcame this by tapping into the power of machine learning algorithms.”

Many of those 20 combinations had not previously been linked with asthma risk.

The investigators noted that a significant number of the toxics in problematic combinations “are similar in structure, and have analogous formation, production, chemical fate, and chemical transport properties.” Chlorine was commonly found in dangerous air toxic combinations, they wrote. A few combinations had heavy metal compounds, and many were acidic chemicals.

“This aligns with prior literature implicating acidic chemicals, chlorinated chemicals, and heavy metal compounds as risk factors for asthma and asthma severity,” the authors wrote. “However, an understanding of the biological mechanisms through which these combinations of air toxics can jointly affect respiratory health and asthma merits further study.”

The authors noted that exposure to acrylic acid by itself increases a child’s risk of needing daily asthma medication, but when acrylic acid is mixed with other toxics, not only does the risk of needing daily medication increase, so does the risk of emergency department visits and hospitalizations.

“As a physician who treats children with asthma, I was struck by how many potential air toxics are not on our radar,” Bunyavanich added. “These results changed my view of the heightened risk some children face.”

The investigators also found that socioeconomic factors may be linked with exposure to certain toxic combinations. For instance, the combination of hydroquinone and ethylidene dichloride was the strongest predictor of overnight hospitalizations. Children from low-income families were more likely to be exposed to the combination, they noted.


Li Y, Hsu HL, Chun Y, et al. Machine learning-driven identification of early-life air toxic combinations associated with childhood asthma outcomes. J Clin Invest. Published online October 5, 2021. doi:10.1172/JCI152088