Two-Stage Machine Learning-Based Approach to Predict Points of Departure for Human Noncancer and Developmental/Reproductive Effects.

Chemical points of departure (PODs) for critical health effects are crucial for evaluating and managing human health risks and impacts from exposure. However, PODs are unavailable for most chemicals in commerce due to a lack of in vivo toxicity data. We therefore developed a two-stage machine learning (ML) framework to predict human-equivalent PODs for oral exposure to organic chemicals based on chemical structure. Utilizing ML-based predictions for structural/physical/chemical/toxicological properties from OPERA 2.9 as features (Stage 1), ML models using random forest regression were trained with human-equivalent PODs derived from in vivo data sets for general noncancer effects (n = 1,791) and reproductive/developmental effects (n = 2,228), with robust cross-validation for feature selection and estimating generalization errors (Stage 2). These two-stage models accurately predicted PODs for both effect categories with cross-validation-based root-mean-squared errors less than an order of magnitude. We then applied one or both models to 34,046 chemicals expected to be in the environment, revealing several thousand chemicals of moderate concern and several hundred chemicals of high concern for health effects at estimated median population exposure levels. Further application can expand by orders of magnitude the coverage of organic chemicals that can be evaluated for their human health risks and impacts.

Modeling Clothing as a Vector for Transporting Airborne Particles and Pathogens across Indoor Microenvironments.

Evidence suggests that human exposure to airborne particles and associated contaminants, including respiratory pathogens, can persist beyond a single microenvironment. By accumulating such contaminants from air, clothing may function as a transport vector and source of "secondary exposure". To investigate this function, a novel microenvironmental exposure modeling framework (ABICAM) was developed. This framework was applied to a para-occupational exposure scenario involving the deposition of viable SARS-CoV-2 in respiratory particles (0.5-20 µm) from a primary source onto clothing in a nonhealthcare setting and subsequent resuspension and secondary exposure in a car and home. Variability was assessed through Monte Carlo simulations. The total volume of infectious particles on the occupant's clothing immediately after work was 4800 µm3 (5th-95th percentiles: 870-32 000 µm3). This value was 61% (5-95%: 17-300%) of the occupant's primary inhalation exposure in the workplace while unmasked. By arrival at the occupant's home after a car commute, relatively rapid viral inactivation on cotton clothing had reduced the infectious volume on clothing by 80% (5-95%: 26-99%). Secondary inhalation exposure (after work) was low in the absence of close proximity and physical contact with contaminated clothing. In comparison, the average primary inhalation exposure in the workplace was higher by about 2-3 orders of magnitude. It remains theoretically possible that resuspension and physical contact with contaminated clothing can occasionally transmit SARS-CoV-2 between humans.

Textile Washing Conveys SVOCs from Indoors to Outdoors: Application and Evaluation of a Residential Multimedia Model.

Indoor environments have elevated concentrations of numerous semivolatile organic compounds (SVOCs). Textiles provide a large surface area for accumulating SVOCs, which can be transported to outdoors through washing. A multimedia model was developed to estimate advective transport rates (fluxes) of 14 SVOCs from indoors to outdoors by textile washing, ventilation, and dust removal/disposal. Most predicted concentrations were within 1 order of magnitude of measurements from a study of 26 Canadian homes. Median fluxes to outdoors [µg·(year·home)-1] spanned approximately 4 orders of magnitude across compounds, according to the variability in estimated aggregate emissions to indoor air. These fluxes ranged from 2 (2,4,4'-tribromodiphenyl ether, BDE-28) to 30 200 (diethyl phthalate, DEP) for textile washing, 12 (BDE-28) to 123 200 (DEP) for ventilation, and 0.1 (BDE-28) to 4200 (bis(2-ethylhexyl) phthalate, DEHP) for dust removal. Relative contributions of these pathways to the total flux to outdoors strongly depended on physical-chemical properties. Textile washing contributed 20% tris-(2-chloroisopropyl)phosphate (TCPP) to 62% tris(2-butoxyethyl)phosphate (TBOEP) on average. These results suggest that residential textile washing can be an important transport pathway to outdoors for SVOCs emitted to indoor air, with implications for human and ecological exposure. Interventions should try to balance the complex tradeoff of textile washing by minimizing exposures for both human occupants and aquatic ecosystems.

Transient Multimedia Model for Investigating the Influence of Indoor Human Activities on Exposure to SVOCs.

Empirical evidence suggests that human occupants indoors, through their presence and activities, can influence the dynamics of semivolatile organic compounds (SVOCs). To better understand these dynamics, a transient multimedia human exposure model was developed (Activity-Based Indoor Chemical Assessment Model (ABICAM)). This model parametrizes mass-balance equations as functions of time-dependent human activities. As a case study, ABICAM simulated exposures of an archetypal adult and toddler over 24 h to diethyl phthalate (DEP), butyl benzyl phthalate (BBzP), and di-2-ethylhexyl phthalate (DEHP) that span a wide range of gas-particle partitioning tendencies. Under baseline (no activities beyond respiration), the toddler's time-average internal doses were three to four times higher than the adult's, due to differences in physical human attributes (e.g., inhalation rate). When time-dependent activities were considered, interindividual (e.g., adult vs toddler) variability was accentuated by up to a factor of 3 for BBzP. Activities with the greatest influence on time-average internal dose were showering (-71% for BBzP), cooking (+27% for DEHP), and sleeping (-26% for DEHP). Overall, the results support the hypotheses that (1) transient indoor activities can give rise to intraindividual variability in estimated internal doses of SVOCs, and (2) interindividual variability in such exposure can result from differences in activity patterns and physical human attributes, according to a compound's physical-chemical properties.

Phthalates: Relationships between Air, Dust, Electronic Devices, and Hands with Implications for Exposure.

Exposure to phthalates is pervasive and is of concern due to associations with adverse health effects. Exposures and exposure pathways of six phthalates were investigated for 51 women aged 18-44 years in Ontario, Canada, based on measured phthalate concentrations in hand wipes and indoor media in their residences. All six phthalates had detection frequencies of 100% in air (∑6670 ng m-3 geomean) and floor dust (∑6630 µg g-1), nearly 100% detection frequencies for hand palms and backs that were significantly correlated and concentrations were repeatable over a 3 week interval. Phthalates on hands were significantly correlated with levels in air and dust, as expected according to partitioning theory. Total exposure was estimated as 4860 ng kg bw-1 day-1 (5th and 95th percentiles 1980-16 950 ng kg bw-1 day-1), with dust ingestion, followed by hand-to-mouth transfer, as the dominant pathways. With the exception of diethyl phthalate (DEP), phthalates had over 50% detection frequencies in surface wipes of most electronic devices sampled, including devices in which the use of phthalates was not expected. Phthalate concentrations on surfaces of hand-held devices were ∼10 times higher than on non-hand-held devices and were correlated with levels on hands. The data are consistent with phthalate emissions from sources such as laminate flooring and personal care products (e.g., scented candles), followed by partitioning among air, dust, and surface films that accumulate on electronic devices and skin, including hands. We hypothesize that hands transfer phthalates from emission sources and dust to hand-held electronic devices, which accumulate phthalates due to infrequent washing and which act as a sink and then a secondary source of exposure. The findings support those of others that exposure can be mitigated by increasing ventilation, damp cloth cleaning, and minimizing the use of phthalate-containing products and materials.

Modeling clothing as a secondary source of exposure to SVOCs across indoor microenvironments.

BACKGROUND: Evidence suggests that clothing can influence human exposure to semi-volatile organic compounds (SVOCs) through transdermal uptake and inhalation. OBJECTIVES: The objectives of this study were [1] to investigate the potential for clothing to function as a transport vector and secondary source of gas-phase SVOCs across indoor microenvironments, [2] to elucidate how clothing storage, wear, and laundering can influence the dynamics of transdermal uptake, and [3] to assess the potential for multiple human occupants to influence the multimedia dynamics of SVOCs indoors. METHODS: A computational modeling framework (ABICAM) was expanded, applied, and evaluated by simulating and augmenting two "real-world" chamber experiments. A primary strength of this framework was its representation of occupants and their clothing as unique entities with the potential for location changes. RESULTS: Estimates of transdermal uptake of diethyl phthalate (DEP) and di(n-butyl) phthalate (DnBP) were generally consistent with those extrapolated from measured concentrations of urinary metabolites, and those predicted by two other mechanistic models. ABICAM predicted that clean clothing (long sleeves, long pants, and socks, 100% cotton, 1 mm thick) readily accumulated DEP (6900-9700 µg) and DnBP (4500-4800 µg) from the surrounding chamber air over 6 h of exposure to average concentrations of 233 (DEP) and 114 (DnBP) µg·m-3. Because of their high capacity, clean clothing also effectively minimized transdermal uptake. In addition, ABICAM predicted that contaminated clothing functioned as a vector for transporting DEP and DnBP across indoor microenvironments and reemitted 13-80% (DEP) and 3-27% (DnBP) of the accumulated masses over 48 h. SIGNIFICANCE: Though the estimated secondary inhalation exposures from contaminated clothing were low compared to the corresponding primary exposures, these secondary exposures could be accentuated in other contexts, for example, involving longer timeframes of clothing storage, multiple occupants wearing contaminated clothing, and/or repeated instances of clothing-mediated transport of contaminants (e.g., from an occupational setting). IMPACT: This modeling study reaffirms the effectiveness of clean clothing in reducing transdermal uptake of airborne SVOCs and conversely, that contaminated clothing could be a source of SVOC exposure via transdermal uptake and by acting as a vector for transporting those contaminants to other locations.

Harmonized US National Health and Nutrition Examination Survey 1988-2018 for high throughput exposome-health discovery.

The National Health and Nutrition Examination Survey (NHANES) provides data on the health and environmental exposure of the non-institutionalized US population. Such data have considerable potential to understand how the environment and behaviors impact human health. These data are also currently leveraged to answer public health questions such as prevalence of disease. However, these data need to first be processed before new insights can be derived through large-scale analyses. NHANES data are stored across hundreds of files with multiple inconsistencies. Correcting such inconsistencies takes systematic cross examination and considerable efforts but is required for accurately and reproducibly characterizing the associations between the exposome and diseases. Thus, we developed a set of curated and unified datasets and accompanied code by merging 614 separate files and harmonizing unrestricted data across NHANES III (1988-1994) and Continuous (1999-2018), totaling 134,310 participants and 4,740 variables. The variables convey 1) demographic information, 2) dietary consumption, 3) physical examination results, 4) occupation, 5) questionnaire items (e.g., physical activity, general health status, medical conditions), 6) medications, 7) mortality status linked from the National Death Index, 8) survey weights, 9) environmental exposure biomarker measurements, and 10) chemical comments that indicate which measurements are below or above the lower limit of detection. We also provide a data dictionary listing the variables and their descriptions to help researchers browse the data. We also provide R markdown files to show example codes on calculating summary statistics and running regression models to help accelerate high-throughput analysis and secular trends of the exposome. [Table: see text].

A comprehensive analysis of racial disparities in chemical biomarker concentrations in United States women, 1999-2014.

BACKGROUND: Stark racial disparities in disease incidence among American women remain a persistent public health challenge. These disparities likely result from complex interactions between genetic, social, lifestyle, and environmental risk factors. The influence of environmental risk factors, such as chemical exposure, however, may be substantial and is poorly understood. OBJECTIVES: We quantitatively evaluated chemical-exposure disparities by race/ethnicity, life stage, and time in United States (US) women (n = 38,080) by using biomarker data for 143 chemicals from the National Health and Nutrition Examination Survey (NHANES) 1999-2014. METHODS: We applied a series of survey-weighted, generalized linear models using data from the entire NHANES women population along with cycle and age-group stratified subpopulations. The outcome was chemical biomarker concentration, and the main predictor was race/ethnicity with adjustment for age, socioeconomic status, smoking habits, and NHANES cycle. RESULTS: Compared to non-Hispanic White women, the highest disparities were observed for non-Hispanic Black, Mexican American, Other Hispanic, and Other Race/Multi-Racial women with higher levels of pesticides and their metabolites, including 2,5-dichlorophenol, o,p'-DDE, beta-hexachlorocyclohexane, and 2,4-dichlorophenol, along with personal care and consumer product compounds, including parabens and monoethyl phthalate, as well as several metals, such as mercury and arsenic. Moreover, for Mexican American, Other Hispanic, and non-Hispanic black women, there were several exposure disparities that persisted across age groups, such as higher 2,4- and 2,5-dichlorophenol concentrations. Exposure levels for methyl and propyl parabens, however, were the highest in non-Hispanic black compared to non-Hispanic white children with average differences exceeding 4-fold. Exposure disparities for methyl and propyl parabens are increasing over time in Other Race/Multi-Racial women while fluctuating for non-Hispanic Black, Mexican American, and Other Hispanic. Cotinine levels are among the highest in Non-Hispanic White women compared to Mexican American and Other Hispanic women with disparities plateauing and increasing, respectively. DISCUSSION: We systematically evaluated differences in chemical exposures across women of various race/ethnic groups and across age groups and time. Our findings could help inform chemical prioritization in designing epidemiological and toxicological studies. In addition, they could help guide public health interventions to reduce environmental and health disparities across populations.

Characterization of age-based trends to identify chemical biomarkers of higher levels in children.

BACKGROUND: Chemical biomarker concentrations are driven by complex interactions between chemical use patterns, exposure pathways, and toxicokinetic parameters such as biological half-lives. Criteria to differentiate legacy from current exposures are helpful for interpreting variation in age-based and time trends of chemical exposure and identifying chemicals to which children are highly exposed. A systematic approach is needed to study temporal trends for a wide range of chemicals in the US population. OBJECTIVES: Using National Health and Nutrition Examination Survey (NHANES) data on measured biomarker concentrations for 141 chemicals from 1999 to 2014, we aim to 1) understand the influence of temporal determinants, in particular time trends, biological half-lives, and restriction dates on age-based trends, 2) systematically define an age-based pattern to identify chemicals with ongoing and high exposure in children, and 3) characterize how age-based trends for six Per- and Polyfluoroalkyl Substances (PFASs) are changing over time. METHODS: We performed an integrated analysis of biological half-lives and restriction dates, compared distributions of chemical biomarker concentrations by age group, and then applied a series of regression models to evaluate the linear (ßage) and nonlinear (ßage2) relationships between age and chemical biomarker levels. RESULTS: For restricted chemicals, a minimum persistence of 1 year in the human body is needed to observe substantial differences between the less exposed young population and historically exposed adults. We define a metric ( [Formula: see text] ) that identifies several phthalates, brominated flame retardants, pesticides, and metals such as lead and tungsten as elevated and ongoing exposures in children. While a substantial reduction in children's exposures was reflected in PFOS and PFOA, levels of PFNA and PFHxS in children were higher in 2013-2014 compared to those in 1999-2000. CONCLUSIONS: Integrating a series of regression models with systemized stratified analyses by age group enabled us to define an age-based pattern to identify chemicals that are of higher levels in children.

Human Health Benefits from Fish Consumption vs. Risks from Inhalation Exposures Associated with Contaminated Sediment Remediation: Dredging of the Hudson River.

BACKGROUND: Billions of dollars are spent on environmental dredging (ED) to remediate contaminated sediments, with one goal being reduced human health risks. However, ED may increase health risks in unanticipated ways, thus potentially reducing net benefits. OBJECTIVES: To assess the ways that ED may increase health risks in unanticipated ways, thus potentially reducing net benefits, we quantitatively assessed a subset of population health benefits and risks of ED, using the 2009-2015 remediation of the Hudson River Polychlorinated Biphenyls (PCBs) Superfund Site as a case study. Three remediation scenarios were evaluated: No Action (NA), Source Control (SC), and ED. METHODS: We quantified health benefits for each scenario from reduced PCB levels in Hudson River fish, and health risks from ED operations due to increased inhalation exposures to PCBs and fine particulate matter (PM2.5), using disability-adjusted life years (DALYs) as a common metric. Occupational health risks were also considered in a separate sensitivity analysis. Estimates of population-level benefits and risks included Monte Carlo simulation-based uncertainty analysis. RESULTS: Under NA, fish consumption would result in an estimated health burden of 112 DALYs, and ED would lead to a reduction of 15 DALYs in excess of SC. ED operations were estimated to induce a total burden of 33 DALYs, dominated by PM2.5 impacts from rail transport emissions (32 DALYs). Including uncertainty, the net health benefit of ED ranged from -138 to +1,326 avoided DALYs (90% confidence), with a median of -11 avoided DALYs. CONCLUSIONS: For the considered impacts, ED in the Hudson River might not have led to an overall net positive human health impact. The benefits and risks of ED, however, have different degrees of uncertainty and involve different populations. Reducing long-distance transport of dredged sediment is a priority. This comparative approach could be used prospectively to better determine trade-offs involved in different remediation scenarios and to improve remediation design to maximize benefits while minimizing risks. https://doi.org/10.1289/EHP5034.