Researchers have analyzed the chemical composition of 81 everyday household items and assessed the potential risks for users.
In a joint effort, the Southwest Research Institute (SwRI) worked alongside the Environmental Protection Agency (EPA) to investigate the chemical composition of 81 commonly used household products. Furthermore, they assessed the possible risks these items might pose to users.
According to the Centers for Disease Control and Prevention, exposure to certain chemicals can have harmful health effects. Building on prior studies that identified chemicals in consumer products, SwRI and the EPA also looked at how materials such as rubber, plastic, clothing, upholstery, and fabric reacted to environmental conditions, like high temperatures in a car or during wear.
The findings, featured in the Environmental Science & Technology journal, are based on four years of data collected using advanced chromatography techniques, suspect screening, non-targeted analysis, and a machine-learning method called Highlightâ„¢, developed by SwRI. Unlike traditional methods that screen a sample against known compounds, Highlightâ„¢ enables scientists to identify and evaluate an extensive array of chemicals through suspect-screening analysis. This approach uncovered 88,795 unique chemical features and 1,883 compound groups across 13 analytical batches.
“Highlight uses machine learning algorithms for quick pattern matching, speeding up our workflow,” remarked William Watson, a research engineer from SwRI’s Intelligent Systems Division and the study’s lead author.
The research also aimed to enhance the field of exposomics, which looks into the impact of lifelong chemical exposure from various sources—environmental, dietary, and lifestyle—on human health. By identifying chemicals in household goods and frequent exposure sources, this study may aid future biomonitoring initiatives.
“Consumer products aren’t comprised of just one chemical. They are a mixture of related chemicals,” explained Dr. Kristin Favela, a staff scientist in SwRI’s Chemistry and Chemical Engineering Division. “We wanted to find out whether the chemicals in our samples could be ’emitted’ or ‘extracted’ to gauge the extent and probability of human exposure.”
To conduct the analysis, SwRI subjected samples of clothing, upholstery, fabrics, rubber, and plastics to two different heat conditions and solvent strengths. The goal was to see if these materials would release chemical vapors that people might inhale in indoor settings, such as a hot car, or while being worn. The study also examined the potential for chemicals to be extracted, which would provide insights into real-world exposure risks, like when a child bites on a household item.
“In addition to enhancing our understanding of the risks of chemical exposure for the public, this research showcases our ability to utilize machine learning alongside Highlightâ„¢ findings to retrospectively analyze and interpret older datasets,” said Watson.
Utilizing the EPA’s Toxicity Forecasting program, ToxCast, the team conducted a further analysis to gauge risk based on human exposure and available biological activity data. Out of the 88 verified chemicals that were both extractable and emittable, 66 had ToxCast data available, and most of the ToxCast in vitro assay data (92%, averaging 441 assays per chemical) indicated no concentration-dependent activity. Data for the remaining 22 chemicals was lacking. Nevertheless, at elevated concentrations, the synthetic antioxidant BKF—commonly used to stabilize plastics and rubbers—showed negative effects at levels reaching 42.3 mg/kg/day. This research could pave the way for better screening models that predict emission activity for household products.
