ExpoKids: An R-based tool for characterizing aggregate chemical exposure during childhood

Aggregate exposure assessments consider combined exposures to a single chemical across multiple routes and media, which is especially important for children's health due to unique behaviors, physiology, and potential increased susceptibility during development. Children's intake rates per unit body weight can be higher than adults. Regulatory contexts, such as the US Food Quality Protection Act and amendments to the Toxic Substances Control Act, emphasize considering aggregate exposures in risk assessments. While frameworks exist across US and international organizations, practical implementation is often limited by data availability for individual chemicals. To address visualization and communication needs for aggregate oral exposures across lifestages, the authors developed ExpoKids to depict contributions of multiple oral media within and across lifestages using ADD and LADD metrics, facilitating comparisons among seven postnatal childhood lifestages and adulthood.

ExpoKids Version 1.0 was developed in R (Version 3.4.0) to visualize aggregate oral exposure estimates and can be used as an extension of ExpoFIRST (Version 2.0), which applies exposure factor data from the US EPA Exposure Factors Handbook (EFH, 2011) to estimate deterministic potential dose for user-defined scenarios. Central tendency oral ADDs (mg/kg-day) were estimated in ExpoFIRST for ten media (soil, dust, water, breast milk, dairy, meat, fish, vegetables, fruit, grains) across EFH children's age groups and an adult group. ADD was calculated as: ADD = (C × IR × EF × ED) / (AT × BW), where C is concentration, IR intake rate, EF exposure frequency, ED exposure duration, AT averaging time, and BW body weight. EFH-derived age-specific central tendency values (means or medians) were used for IR, EF, ED, AT, and BW; chemical concentrations (C) were drawn from media-specific literature sources. General population scenarios were selected: average wet weight concentrations representative of entire media (even if subset-sampled), per capita food group estimates (combined sexes) for total fruits/vegetables/etc., and community drinking water (tap water) only. ExpoFIRST produced medium-specific ADDs, which were exported to Excel and reorganized by lifestage for input to ExpoKids. Data gaps and assumptions: For children under 1 month, EFH provided no soil or dust ingestion values; both intake rates were assumed 0 mg/day. Due to limited data, ExpoFIRST v2.0 included only breastfeeding data for young infants (no formula-fed data). ExpoKids processing and visualization: EFH/ExpoFIRST age groups were recategorized into seven lifestages (young infant; infant; young child; child; young youth; youth; adult) following EPA guidance on age grouping. ExpoKids generates five graph types: (1) ADD by lifestage (stacked media contributions), (2) LADD by lifestage (time-weighted ADD by lifestage), (3) Percent ADD by lifestage (relative media contributions within a lifestage), (4) medium-specific ADD across lifestages, and (5) medium-specific LADD across lifestages. Data were ingested using readxl, reshaped with reshape2 (melt), and plotted as stacked bar charts using ggplot2/tidyverse. LADDs were computed within ExpoKids by time-weighting each lifestage ADD by the fraction of a 70-year lifespan spent in that lifestage. Lifestage aggregation formulas: For a given medium, lifestage-specific ADD was computed as a duration-weighted average of EFH age-group ADDs: ADD_lifestage = Σ(ADD_i × Y_i) / ΣY_i, where Y_i is years in age group i within the lifestage. Lifestage LADD was computed as LADD_lifestage = ADD_lifestage × (Y_lifestage / 70 years). Percent contributions within a lifestage were calculated as % contribution = (ADD_medium,lifestage / Σ ADD_all media,lifestage) × 100%. Case example selection: To evaluate ExpoKids, the authors developed illustrative scenarios for three chemicals—di(2-ethylhexyl) phthalate (DEHP), manganese (an essential nutrient with potential developmental toxicity at elevated exposure), and endosulfan—guided by criteria: evidence of developmental toxicity; at least one postnatal sensitive window; and available concentration data in at least four oral media in ExpoFIRST. Literature searches (PubMed, Sept 2018) targeted US data (2000–2018), quantitative media measurements, and general population contexts. Additional large US databases with multi-year monitored/measured data were consulted to gauge concentration accuracy. Mean media concentrations were, when possible, taken from a single source per chemical.

ExpoKids provides a data visualization framework to illustrate aggregate oral exposure across seven childhood lifestages and adulthood for ten media, producing ADD and LADD views and relative contributions by medium and lifestage. The tool effectively conveys ExpoFIRST-derived estimates to highlight relative media and lifestage contributions to aggregate exposure. Its application to three illustrative chemicals (DEHP, manganese, endosulfan) demonstrated use in exploring typical exposure assessment questions about media contributions across lifestages, differences across lifestages for an essential nutrient, and the influence of regulation on aggregate exposure patterns. Specific quantitative exposure levels are not detailed in the provided text, but the framework enables identification of lifestages and media that dominate aggregate exposure profiles.

The development and application of ExpoKids directly address the need to understand and communicate how multiple oral media contribute to aggregate chemical exposure across childhood lifestages relative to adulthood. By organizing EFH- and ExpoFIRST-based exposure estimates into lifestage-focused, time-weighted visualizations, the tool supports stakeholders in identifying critical lifestages and dominant media contributing to aggregate exposure. The case examples illustrate the tool’s versatility for assessing chemicals with developmental toxicity concerns, essential nutrients with potential toxicity at elevated exposures, and chemicals affected by regulatory actions. ExpoKids thereby enhances the translation of existing exposure factor data into insights relevant to children’s health risk assessment and management.

ExpoKids, an R-based extension to ExpoFIRST, was developed to visualize aggregate oral exposure as ADD and LADD across seven childhood lifestages and adulthood for ten media. The tool facilitates comparison of media contributions within and across lifestages and supports answering common exposure assessment questions, as illustrated with DEHP, manganese, and endosulfan. The approach improves communication of aggregate exposure patterns for children and can be applied to available exposure data to inform children's environmental health assessments.

Implementation of aggregate exposure assessments is constrained by data availability for individual chemicals and media. EFH lacked ingestion values for soil and dust for children under 1 month (assumed 0 mg/day), and ExpoFIRST v2.0 had breastfeeding data only for young infants (no formula-fed infant data). Scenario assumptions included general population parameters, per capita total food groups, and community tap water only, which may not reflect all subpopulations or sources (e.g., bottled water). ExpoFIRST estimates medium-specific ADDs and does not itself compute aggregate exposures; ExpoKids integrates and visualizes only oral media and relies on available central tendency inputs, which may limit generalizability or omit high-end exposures.