Improving Methodologies for Cumulative Risk Assessment: A Case Study of Noncarcinogenic Health Risks from Volatile Organic Compounds in Fenceline Communities in Southeastern Pennsylvania
Publication: Environmental Health Perspectives
Volume 133, Issue 5
CID: 057004
Abstract
Background:
Cumulative risk assessment (CRA) is key to characterizing health risks in fenceline and disadvantaged communities, which face environmental pollution and challenging socioeconomic conditions. Traditional approaches for inclusion of mixtures in CRA are limited and only assess the most sensitive target organ system for each chemical.
Methods:
We developed an expanded approach to cumulative risk assessment that considers all known target organ systems associated with a chemical. Specifically, we created a multi-effects toxicity database by a) compiling toxicological and epidemiological data from the Agency for Toxic Substances and Disease Registry’s (ATSDR) Toxicological Profiles and the Environmental Protection Agency (US EPA) CompTox Chemicals Dashboard; b) developing a tiering system to prioritize identified data for use in developing toxicity values; and c) accounting for uncertainty to create toxicity values for additional target organ systems. We demonstrated differences between the traditional approach and our expanded approach by using state-of-the-art mobile monitoring data from our Southeastern Pennsylvania Hazardous Air Pollutant Monitoring and Assessment Project (SEPA HAP-MAP) to conduct a cumulative risk assessment.
Results:
Of the 32 chemicals quantified in SEPA HAP-MAP, 28 were represented in our multi-effects toxicity database, whereas only 16 were included using a traditional approach. In total, we derived toxicity values for 172 chemical-target organ system combinations. Our expanded approach found neurological, renal, respiratory, endocrine, and systemic risks (hazard index ) in SEPA HAP-MAP fenceline communities, whereas no risks were identified using a traditional approach limited to the most sensitive target organ systems only.
Conclusion:
Our results suggest that traditional approaches to CRA underestimate health risks in fenceline and other highly exposed communities and highlight the need for improved methods to inform health-protective and just risk management decisions. https://doi.org/10.1289/EHP14696
Introduction
Cumulative risk assessment—defined as “analysis, characterization, and possible quantification of the combined risks to health or the environment from multiple agents or stressors”— is key to understanding risks faced by fenceline and disadvantaged populations, who typically experience multiple exposures to chemicals in their communities.1,2 Numerous US statutes contain language authorizing (and in some cases specifically calling for) consideration of cumulative risks in regulatory rulemaking, including the Food Quality Protection Act, the Federal Food, Drug, and Cosmetic Act, the Federal Insecticide, Fungicide, and Rodenticide Act, and the Lautenberg amendment to the Toxic Substances Control Act. In addition, authoritative bodies have voiced support for the development and use of cumulative risk assessment approaches in regulatory rulemaking has been voiced by authoritative bodies.1,3,4
Recognizing the multifactorial nature of many disease processes, US regulatory and public health agencies have attempted to develop frameworks to aid in assessing risks from mixtures of chemicals and other stressors. Many of these efforts, however, have been conceptual and lack specific methodological recommendations2,5–8 or have limited focus on chemicals that act through a common mechanism.9,10 In 2023, the US Environmental Protection Agency (US EPA) released draft guidance for conducting the planning and scoping activities involved in the problem formulation step of a cumulative risk assessment,11 but specific steps for implementation and quantitative methodologies for analysis remain to be determined. As currently practiced, cumulative risk assessment remains limited in scope and may not provide accurate conclusions regarding risks, thereby leading to risk management decisions that do not sufficiently protect public health.1
Current cumulative approaches to noncancer risk assessment for inhalation exposures typically consider mixtures by estimating the hazard quotients (HQs) for each chemical (i.e., the exposure concentration divided by the reference concentration; see Table S1 for definitions of terms) and summing them up across chemicals that act on the same organ systems to derive a hazard index (HI). However, because reference concentrations (RfCs) are based only on the most sensitive target organ systems of chemicals (i.e., the systems affected at the lowest concentrations of exposure), this approach fails to consider the potential for effects on additional target organ systems at higher levels of exposure.12,13 Though perhaps negligible if considered in isolation, these additional effects may be significant when considered in the context of coexposures to numerous chemicals. In addition, because of the integrated function of organs across physiological systems, effects of exposure on target organs may impact the function of other organs in that system.14 Indeed, Fox et al.13 conducted a risk assessment that considered the effects of 40 hazardous air pollutants beyond their most sensitive target organ systems and, in comparison with an approach using US EPA RfCs only, found hazards associated with three additional target organ systems. In the Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures, the US EPA describes the possibility of RfCs based on additional target organ systems (termed “target organ toxicity concentrations”) but did not note active work at the agency to derive them at that time.15 In recent IRIS assessments, however, the agency has included candidate organ system reference values that it indicates may be useful in cumulative risk assessments.16,17
The purpose of this paper is to demonstrate improved methods for cumulative risk assessment of noncarcinogenic health effects using case study data from the Southeastern Pennsylvania Hazardous Air Pollutant Monitoring and Assessment Project (SEPA HAP-MAP), which examined fenceline communities in southeastern Pennsylvania along the Delaware River.18 Our objective was to develop and pilot a reproducible approach for consideration of noncancer risks associated with multiple target organ systems simultaneously for complex mixtures.
Methods
SEPA HAP-MAP Case Study Context
Fenceline communities along the Delaware River in southeastern Pennsylvania have long faced high levels of pollution from petrochemical refineries, municipal waste incinerators, and several other industrial facilities.19,20 In 1996, the US EPA acknowledged the disproportionate siting of harmful facilities in the city of Chester, Pennsylvania, and noted elevated rates of infant mortality, cancer mortality, and asthma among residents.21 The SEPA HAP-MAP was designed to improve neighborhood-level characterization of ongoing chemical exposures and associated health risks in Chester and surrounding boroughs. We generated Figure 1 in R (version 4.4.1; R Development Core Team) using data from OpenStreetMap,22,23 Google Earth,24 and the US Census Bureau25 to display the study area of the SEPA HAP-MAP campaign.
To demonstrate our multi-effects toxicity database (METDB) approach, we leveraged data from the SEPA HAP-MAP study, which employed state-of-the-art analytical instrumentation to measure 32 volatile organic compounds (VOCs), including formaldehyde, benzene, toluene, ethylbenzene, and xylenes (as o-xylene alone and m- and p-xylene combined), using a mobile laboratory operated by Aerodyne Research Inc. (Table S2).26–32 We collected mobile measurements were over a 3-wk period between the 8th and 29th of September of 2021. These intensive measurements occurred on a near daily basis and entailed a repeated 8-h driving route in the area between Widener University and the far western side of the Marcus Hook industrial complex (Figure 1). We designed the driving route with the intent of characterizing concentrations both at facility fencelines and in adjacent communities. Although the mobile laboratory typically drove the fixed route each day, there were occasional deviations from this preplanned route to follow identified chemical plumes or to investigate a potential emission source.
The concentrations analyzed were measured by three instruments in the mobile laboratory: a proton-transfer reaction mass spectrometer; a fast gas chromatograph with mass spectrometry detection; and laser spectrometer(s).26–28,33 The proton-transfer reaction mass spectrometer and laser spectrometers measured at 1-s time resolution, and the fast gas chromatograph measured at 10-min resolution. A subset of chemicals was measured by two instruments: in these cases, we preferentially selected measurements from the instrument with the highest time resolution (Table S2). We obtained concentrations of measured VOCs in blocks of our study area, limiting our analyses to blocks with a minimum of 10 unique visits by the mobile laboratory.34 This 10-visit minimum over the 3-wk study was established to reduce the influence of variations by time of day and weather conditions and excluded areas that were only visited a few times from the larger analysis.
Instrumentation used in the SEPA HAP-MAP mobile monitoring campaign.
Limits of detection on spatial measurements.
Measurements made by instrumentation on Aerodyne Mobile Laboratory (AML) have low limits of detection (LODs) better than 100 parts per trillion by volume (pptv) for 1-s measurements.26 Averaging multiple measurements for spatial analysis of air toxic concentration improves the LOD to lower levels. The improvement and consideration of nuances of LOD and spatial averaging for individual species such as ethylene oxide are discussed in more detail in Robinson et al.34
TILDAS.
A single-laser tunable infrared laser direct absorption spectrometer (TILDAS; Aerodyne Research, Inc.)35 was used during the HAP-MAP to measure formaldehyde (HCHO) and formic acid (HCOOH).33 TILDAS instruments use the HITRAN database of spectral lines, supplemented as needed with additional high resolution spectral data.36,37 For uncalibrated TILDAS species, raw concentrations are expected to be within 5%–10% of reported values. During the campaign, ultrazero air was delivered to the gas-phase inlet every 15 min in excess of the inlet flow to zero the instrument.
HCHO is calibrated indirectly via the water measurement from the HCHO-TILDAS and compared with two other water measurements available on the CS--CO- TILDAS and two additional redundant measurements from instruments (LI-6262, LI-COR Environmental). These comparisons indicate that the calibration for (and by extension for HCHO) is within 2% of true concentration.
Vocus.
A Vocus proton-transfer reaction time-of-flight mass spectrometer (Vocus PTR-ToF-MS, referred to as “Vocus”) was deployed on the AML. This instrument measured the time of flight of several gas-phase VOCs. The observed time-of-flight spectra were mass calibrated to yield mass spectra (ions signals as a function of ion mass-to-charge). Each of the signals was detected as voltages and converted to ions per second.
During this campaign, the Vocus was operated predominantly in and occasionally in chemical ionization modes. The raw data collected from the Vocus when using chemical ionization were processed using custom software based on Igor Pro (Tofware; TOFWERK and Aerodyne Research, Inc.) and used to quantify the concentrations of the species designated as Vocus in Table S2. The time series of the mass concentrations of each species was obtained via high-resolution fitting of the observed mass spectra. Instrument backgrounds were corrected for by periodically sampling zero air. Similarly, changes in sensitivity were corrected for by measuring known concentrations of calibrant species every 3.5 h. The reported data is corrected for both instrument background and instrument sensitivity as a function of sampling time. Aromatics compounds or group and (their uncertainties) are as follows: benzene (30%), furan (100%), C8 aromatics (30%), toluene (30%), and styrene (70%). Note: Furan uncertainty is due to high signal from adjacent isoprene ion in the mass spectra.
During the campaign, instabilities in the behavior of the ion source were observed. Thus, the real-time species concentrations obtained from the Vocus PTR data were obtained by scaling the Vocus PTR measurements to the collocated gas chromatography measurements on the AML. Scaling was performed by comparing the ratio of the Vocus data (averaged appropriately over the gas chromatography sampling time) in units of ions/s to the gas chromatography data in units of parts per billion (ppb). The Hz/ppb values obtained in this way were then applied to convert the time-resolved Vocus data to ppb values. Four different time periods with different apparent sensitivities and ion source operating performance were identified, and the scaling was done separately for each measurement time period. The 1-sigma estimated uncertainties in the reported Vocus data are estimated from the Gaussian fits of histograms of the ratios between the Vocus data and the gas chromatography data over the entire campaign.
GC-EI-ToF.
A 2-channel gas chromatograph (GC) with electron ionization time-of-flight mass spectrometer (GC-EI-ToF) was deployed on the AML. This instrument quantifies speciated gas-phase VOCs for those species listed in Table S2. The GC-EI-ToF operates on a 30-min duty cycle with 10-min sample collection per half-hour and a 20-min sample prep, separation and detection. The gas chromatograph is equipped with two separation channels, each optimized to resolve different volatility rages. Channel 1 is optimized to resolve hydrocarbons, oxygenates, and halocarbons in the -alkane volatility range, whereas Channel 2 is optimized to resolve compounds in the -alkane volatility range. One chromatogram from each separation channel is acquired every 30 min (4 chromatograms per hour, 96 chromatograms per day).
The GC operates with thermal desorption preconcentration, where sample air passes through a multistage preconcentration system for sample collection and focusing before being injected on to the GC column for separation. For this campaign, each GC channel used multibed adsorbents from Markes International (Tenax TA/Graphitized Carbon/Carboxen 1000), which are optimized for species. Before the sample was collected onto the adsorbent, it passed through an oxidant trap equipped with activated to mitigate artifacts that are known to occur with adsorbent sampling. During this campaign the adsorbent tubes and oxidant traps were replaced approximately every 4 d (192 GC cycles). After preconcentration, the analyte is injected onto the GC column. For this campaign, Channel 1 was equipped with a Restek PLOT MXT-Q-Bond nonpolar column for the high-volatility species (C2–C5 hydrocarbons, C1–C3 oxygenated VOCs) with of column guard (MXT-1) on each end of the column, and a Restek Rxi-624MS midpolarity column was used on Channel 2 for lower volatility VOCs (C5–C12 hydrocarbons, C2–C10 oxygenated VOCs). After separation, each column effluent is directed to the detector via a two-way selector valve.
The data reported for the GC-EI ToF listed in Table S2 have undergone high-resolution mass spectral fitting followed by chromatographic peak fitting to generate VOC time series. The data were then normalized to correct for instrument drift due to changes in detector performance or environmental changes, and, finally, measured sensitivities from pre- and postcampaign calibrations were applied to make mixing ratio time series. The mixing ratios reported are in units of ppb. All data reported from the GC-EI-ToF should be treated with 30% uncertainty.
Traditional Cumulative Risk Assessment
The assessment phase of a cumulative risk assessment investigating the noncarcinogenic effects of multiple chemicals often follows a four-step paradigm: hazard identification, dose–response assessment, exposure assessment, and risk characterization.1 Hazard identification involves determining the health effects associated with each chemical of interest. In the dose–response assessment, metrics are selected to describe the quantitative relationship between exposure to each included chemical and its most sensitive target organ system (i.e., the target organ system that shows adverse effects at the lowest level of exposure). Exposure assessment entails characterizing the frequency, magnitude, and duration of exposures to included chemicals. During the risk characterization phase, estimated exposures are compared to dose–response metrics to generate conclusions on the likelihood and magnitude of expected risks.
In a traditional cumulative risk assessment of inhalation exposures to chemicals eliciting noncarcinogenic effects, the hazard identification and dose–response assessment steps typically rely on assessments and reference concentrations (RfCs) from the US EPA’s Integrated Risk Information System (IRIS). RfCs are defined as an “estimate (with uncertainty spanning perhaps an order of magnitude) of a daily exposure to the human population (including sensitive subgroups) that is likely to be without an appreciable risk of deleterious effects during a lifetime.”38 RfCs are based on the most sensitive target organ system (e.g., the neurological system or developmental system), which is the target organ system that has been shown to have adverse health effects at the lowest concentration of exposure to the chemical. The US EPA develops IRIS RfCs based on review of the literature and rigorous intra-agency, interagency, and external peer review processes.39,40
Following the traditional cumulative risk assessment approach, we searched the US EPA IRIS database for available RfCs for each of the volatile organic compounds measured in the SEPA HAP-MAP. To assess inhalation exposure to each chemical, we first calculated the 95th percentile upper confidence limit of the mean air concentration in each included 50- by block of our study area. Then, we used an average of those values to derive exposure concentrations for the study area. For the purposes of our case study, we assumed that residents living within our study area were continuously exposed to this concentration over the course of a lifetime.
To characterize risks, we divided chemical-specific exposure concentrations from the SEPA HAP-MAP campaign by their corresponding RfCs to generate hazard quotients (HQs) as follows:where is the hazard quotient for chemical i, is the exposure concentration for chemical i, and is the reference concentration for chemical i. Hazard indices (HIs) were calculated for each target organ system by summing the HQs from all chemicals with an RfC for that target organ system as follows:where is the HI for target organ system TOS, and is the HQ for chemical i for target organ system TOS.
(1)
(2)
Expanded Cumulative Risk Assessment: Multi-Effects Toxicity Database
To expand consideration of adverse effects associated with exposure to our measured chemicals beyond the most sensitive target organ system (as in the traditional approach to cumulative risk assessment), we developed an expanded approach that considered additional target organ systems for each chemical, thereby generating more comprehensive information from which to estimate risks of exposures to chemical mixtures (Figure 2). Our expanded approach, as described in more detail below, involved creating a multi-effects toxicity database (METDB) by identifying and categorizing toxicological and epidemiological information, prioritizing identified data, and developing values akin to RfCs for additional target organ systems (Figure 3); in accordance with US EPA recommendations, we refer to these values as target organ toxicity concentrations rather than RfCs.15
Data sources and identification and categorization of potential points of departure.
We compiled points of departure (PODs) from Agency for Toxic Substances and Disease Registry (ATSDR) Toxicological Profiles and the US EPA CompTox Chemicals Dashboard Version 2.1.1 (hereafter referred to as ToxProfiles and CompTox, respectively). ToxProfiles are peer-reviewed documents that present a narrative review of the available toxicological and epidemiological data available for a given chemical.41 Developed for chemicals of concern based on toxicity, occurrence, and likelihood of human exposure, ToxProfiles exist for just over 300 chemicals.41 ToxProfiles also include minimal risk levels (MRLs), which are akin to the US EPA’s RfCs or RfDs and developed using a comprehensive internal and external review process.41 On the other hand, CompTox includes physicochemical, exposure, hazard, and risk data for more than 1,200,000 chemicals.42 It has been compiled from a variety of data sources, including the US EPA IRIS and Provisional Peer-Reviewed Toxicity Values, the National Library of Medicine’s Hazardous Substances Databank, the Organization for Economic Co-operation and Development’s eChem Portal, and several databases published by the European Union and the European Chemicals Agency (ECHA).43
We searched CompTox and ToxProfiles for all human or animal PODs, which are intended to be low- or no-effect levels, for noncancer end points associated with chronic or subchronic inhalation exposure to the VOCs measured in the SEPA HAP-MAP. Because our measurements could not differentiate between all xylene isomers, we included PODs for the isomers separately, for mixed xylenes, and for xylenes not otherwise specified. PODs were categorized as benchmark concentration levels (BMCLs; lower confidence limit of the concentration corresponding to a specific change in adverse response in exposed in comparison with unexposed subjects), no-observed-adverse-effect concentrations (NOAECs; the highest tested concentration at which no adverse effect on the target organ system is observed), or lowest-observed-adverse-effect concentrations (LOAECs; the lowest tested concentration at which an adverse effect on the target organ system is observed).
In addition to the POD for each chemical–target organ system combination, we also extracted information on the source of data (i.e., CompTox or ToxProfiles), study duration (e.g., 90 d), exposure regimen (e.g., 6 h per day, 5 d per week) and specific health end point analyzed. Because this information was not always reported in a consistent manner in CompTox (e.g., information on study duration was found in multiple data columns), we manually checked each entry to ensure that ineligible (e.g., acute) PODs were excluded. Target organ systems were manually assigned to each CompTox POD based on guidance from ATSDR and professional judgment.44 Table 1 presents definitions for each target organ system and examples of included end points. PODs missing sufficient details in CompTox to enable categorization of target organ systems were not included for further consideration, but PODs of unknown duration were retained. After we compiled and categorized all eligible PODs, we excluded chemical–end point combinations without any evidence of effect (i.e., those for which only NOAECs were identified) to avoid overestimating risks.
| Target organ system | Definition | Examples of corresponding health effects |
|---|---|---|
| Systemic | Effects that may occur at multiple sites in the body, such as changes in body weight or body weight gain (in a non-reproductive study), with or without changes in food consumption | Decreased body weight |
| Cardiovascular | Effects related to the heart and circulatory system and its functioning | Altered blood pressure; bradycardia; myocarditis |
| Dermal | Effects related to the skin and its functioning | Dermatitis; edema; hyperkeratosis |
| Developmental | Effects on the offspring resulting from exposures to parental germ cells (formed when the parents were in utero), the conceptus through the preimplantation blastocyst stage, and all subsequent developmental stages up through 18 y of age in humans or sexual maturity in animals | Delayed ossification, alteration in offspring organ weight; visceral anomalies (heart defects) |
| Endocrine | Effects involving ductless hormone-secreting glands, including the hypothalamus, pituitary gland, adrenal glands (including the adrenal cortex and medulla), thyroid glands, parathyroid glands, and the pancreatic islets | Thyroid hyperplasia; adrenal cortical atrophy; decreased thyroid, pituitary, or adrenal function |
| Gastrointestinal | Effects related to the digestive system, including effects of the esophagus, stomach, and small and large intestines | Diarrhea; ulceration; nausea |
| Hematological | Effects related to blood chemistry and hematology | Anemia; leukopenia; erythrocytopenia |
| Hepatic | Effects related to the liver and gallbladder and their functioning | Altered liver enzymes; hepatomegaly; cirrhosis |
| Immunological | Morphological effects involving lymphatic tissues such as the lymph nodes, spleen, and thymus | Altered T-cell activity; thymus or spleen lymphoid atrophy; histiocytosis of lymph node or spleen |
| Metabolic | Effects involving disturbances in acid–base balance | Acidosis or alkalosis; increased osmolal gap; altered metabolic rate |
| Musculoskeletal | Effects related to the muscles and skeletal system and its functioning | Muscular atrophy; altered bone density; muscular rigidity |
| Neurological | Effects occurring from a change in the structure or function of the central or peripheral nervous system by a biological, chemical, or physical agent | Behavioral changes; decreased locomotor activity, altered EEG |
| Ocular | Effects are those related to the eyes and their functioning | Cataracts; blindness; conjunctivitis |
| Renal | Effects related to the kidneys and urinary bladder and their functioning | Decreased urine volume; hematuria; renal tubular degeneration |
| Reproductive | Effects resulting from exposures during the interval from the generation of parental germ cells to conception up through implantation of the offspring | Abnormal sperm; decreased fertility; tubular degeneration |
| Respiratory | Effects related to the respiratory system and its functioning | Bronchitis; lung or nasal irritation; pulmonary edema |
Note: Definitions and examples adapted from Agency for Toxic Substances and Disease Registry (2018).44 EEG, electroencephalogram; METDB, multi-effects toxicity database.
Selection of chemical–target organ system points of departure.
Our second step was to select one POD for each identified chemical–target organ system combination with evidence of effect for inclusion in the METDB. Where they were available, we preferentially selected IRIS RfCs (including candidate RfCs, such as those for trichloroethylene and formaldehyde)45,46 for use in our METDB, followed by ATSDR MRLs. However, because our approach expands consideration of health risks beyond those associated with the most sensitive end point for a chemical, IRIS RfCs and ATSDR MRLs were not available for most chemical–target organ system combinations in our METDB. If epidemiological or occupational studies were available, the most sensitive (i.e., lowest) POD in humans was selected for inclusion (case reports and case studies were not considered). In the absence of human data, the most sensitive POD from rats or mice was used; per US EPA guidance under the Toxic Substances Control Act, these are the preferred animal models for health effects testing for inhalation studies.47 If data on humans, rats, and mice were unavailable, the most sensitive POD from other available animal models (e.g., guinea pigs or dogs) was selected for inclusion. Within each of these tiers of preferred data, the most sensitive BMCL was preferentially selected over the most sensitive NOAEC, which in turn was preferred over the most sensitive LOAEC.
Application of uncertainty factors and development of target organ toxicity concentrations.
To develop target organ toxicity concentrations (TTCs) for these chemical–target organ system combinations, we used a standardized approach to account for sources of uncertainty in the PODs from our METDB. PODs based on animal models were first converted to human equivalents to account for uncertainty in animal–human toxicokinetic differences. This conversion was calculated by adjusting the POD for duration of exposure regimen (i.e., for less-than-continuous exposure during the study period) and applying a dosimetric adjustment factor (DAF):where is the POD from an animal study adjusted to represent a human equivalent concentration, POD is the unadjusted POD from experimental animal study, D is the number of hours exposed per 24 h, and W is the number of days of exposure per 7 d.
(3)
Study-specific values for D and W were used for PODs obtained from ToxProfiles. Because these data were not available in CompTox, all studies from CompTox were assumed to follow the Organization for Economic Co-operation and Development (OECD) inhalation test guidelines for exposure regimens: 6 h per day for 5 d per week for studies of subchronic or unknown duration and 6 h per day for 7 d per week for chronic studies.48,49 DAFs, which are chemical-specific and effect-dependent (i.e., in the portal of entry or in a location remote from the portal of entry) values representing interspecies differences in delivered dose when extrapolating from animal studies to humans,50 were obtained from IRIS where possible. When chemical-specific DAFs were not available and the POD corresponded to a remote effect (i.e., not a portal-of-entry effect), we used a DAF of 1, per US EPA guidance.51 One chemical (isopentane) elicited a portal-of-entry effect and did not have an established DAF; for this chemical only, we did not apply Equation 3 to animal PODs and instead used an animal-to-human uncertainty factor () of 10.
Target organ toxicity concentrations were calculated for each chemical–target organ system combination by dividing human-derived or human-equivalent PODs by the product of all uncertainty factors (UFs) as follows:where TTC is the target organ toxicity concentration (in ppb); is the POD from a human study, or POD from an animal study adjusted to represent a human equivalent concentration (with the exception of isopentane, as above); is the interspecies UF, is the intraspecies UF; is the subchronic-to-chronic uncertainty factor; is the LOAEL to NOAEL UF; and is the database UF.
(4)
Human-equivalent PODs were assigned a of 3 to account for animal–human differences in toxicodynamics only (as differences in toxicokinetics were accounted for using Equation 3). PODs from animal studies of isopentane were assigned a of 10 to account for animal–human differences in toxicokinetics and toxicodynamics. Human-derived PODs were assigned a of 1. All studies were assigned a full intraspecies () of 10 to reflect differences in vulnerability among humans. For PODs derived from subchronic studies, we assigned a full subchronic-to-chronic UF () of 10 to incorporate uncertainty resulting from studies based on less-than-lifetime exposures. Chronic studies were assigned a of 1. Studies of unknown duration were assigned a of 10 to be conservative. If the POD was a NOAEC or BMCL, a LOAEC-to-NOAEC UF () of 1 was assigned. PODs that were LOAECs were assigned a of 10 to reflect the uncertainty in NOAEC-to-LOAEC extrapolation. All PODs were assigned a database UF () of 1, because this UF is typically assigned when there are substantial gaps in available data (e.g., no developmental studies) and it is unclear whether a more sensitive end point exists for the chemical. Because we are considering nonsensitive target organ systems in our analysis, the does not apply.
Demonstration of the METDB Method for Identifying and Selecting PODs
Using the chemical 1,2,4-trimethylbenzene, we identified 25 potential PODs in the CompTox Chemicals Dashboard42; 1,2,4-trimethylbenzene does not have an ATSDR Toxicological Profile (so no PODs were identified there). From the list of PODs, we were able to identify two target organ systems (neurological and developmental) with evidence of effect.
Neurological.
POD identification and selection: For 1,2,4-trimethylbenzene, a POD [ (human equivalent concentration) of , or ] corresponding to an IRIS RfC (, or ) was identified, with a critical effect of “decreased pain sensitivity in male Wistar rats.” Because an established RfC was available, we selected this value (and thus did not need to establish our own TTC).
Source of the POD: In the IRIS Toxicological Review of Trimethylbenzenes: Executive Summary, the principal study supporting the POD and RfC is identified as Korsak and Rydzyński.52
Developmental.
POD identification and selection: For 1,2,4-trimethylbenzene, four potential PODs were identified using a developmental study design (and having relevant end points). All four PODs were NOAECs derived from studies in rats with no listed study durations. Two PODs had the same (and lowest) value (, or ).
Source of the POD: The source for the POD is listed as the ECHA IUCLID database, but the link provided in the Dashboard connects to the ECHACHEM Database.53 Once on this site, by clicking the magnifying glass icon and entering “1,2,4-trimethylbenzene” a record for the chemical is returned and can be selected. Once the “Substance Infocard” opens, it is necessary to scroll down to a section called “Key datasets” and select the tab “REACH registered substance factsheets.” It will open a new record for 1,2,4-trimethylbenzene, with a series of options on the left-hand side of the screen; selecting the box that says, “Toxicological information” will open another menu of options, which will show the drop-down menu for “Toxicity to reproduction.” Selecting this category will open yet another menu of options to locate and select “Developmental toxicity/teratogenicity.” Scrolling to the bottom of this record will yield the “Applicant’s summary and conclusion,” which indicates that the NOAEC for maternal and developmental toxicity was (or ).
The ECHA record provides many details about the underlying study to accompany the presentation of the NOAEC. ECHA rates the study as “reliable with restrictions” (ECHA notes that the study follows the OECD Guideline 414 for Prenatal Developmental Toxicity Studies and was published in a peer-reviewed journal, but GLP status for the study was unknown) and declares it “fully adequate for assessment.” The ECHA record provides details about test animals, dose administration and groups, statistical methods, outcome assessment, etc. It is noted in the record that the dosing regimen is noncontinuous; on gestation days 6 through 20, dams were exposed to 1,2,4-trimethylbenzene for 6 h per day. The ECHA record does not name the study or provide any identifying information that can be used to locate it.
Derivation of the TTC: To derive a developmental TTC, we first adjust the NOAEC from an animal study using a noncontinuous dosing regimen to a converted to continuous exposure, as follows:
We then set the dosimetric adjustment factors and UFs as DAF: 0.98; : 3 (for toxicodynamic differences between rats and humans); : 10 (for interindividual variation in vulnerability among humans); : 10 (for extrapolation from a subchronic study to a chronic duration); : 1 (no adjustment needed, because POD is a NOAEC); and : 1 (no adjustment needed, as this is not intended to be the most sensitive end point).
Then, we calculated the TTC by dividing the product of the and the DAF by the product of the UFs as follows:and
Noncancer risk characterization for VOCs.
To characterize noncancer risk using the METDB, we calculated HQs and HIs.54 For each chemical–target organ system combination, we extended the traditional HQ equation (Equation 2) by dividing the chemical-specific exposure concentrations from the SEPA HAP-MAP campaign by all available MRLs and TTCs in addition to RfCs. Similarly, HIs were calculated for each target organ system by summing the HQs from all chemicals with an RfC, MRL, or TTC for that target organ system. All calculations were performed using R (version 4.4.1). As above, HQs and HIs above 1 indicated the potential for adverse health effects. We compared results of our risk characterization using the traditional and expanded approaches to cumulative risk assessment.
Sensitivity Analyses
Acknowledging that the studies that comprise CompTox have not been evaluated by the US EPA and may be of varying data quality, we also conducted a sensitivity analysis using a version of our METDB that only included TTCs derived from ToxProfiles and RfCs from IRIS. We used the same approach to developing target organ toxicity concentrations as described above for the ToxProfile-only sensitivity analysis. As an additional sensitivity analysis, we assumed that PODs with missing information on study duration were chronic instead of subchronic. This approach resulted in higher target organ toxicity concentrations because a of 10 for subchronic-to-chronic extrapolation was eliminated for PODs of unknown study duration.
In its 2002 report, the US EPA’s Reference Dose/Reference Concentration Technical Panel recommended against deriving a reference value when the total UF exceeds 3,000.55 In the METDB, the total UF cannot exceed 3,000 because the UFD is always set at 1, so it could be considered that a total UF of 3,000 is too uncertain. To assess the impact of highly uncertain target organ toxicity concentrations on the characterization of risks, we performed an additional sensitivity analysis in which all TTCs with UF were excluded from the METDB (and thus did not contribute to organ-specific HIs).
Results
Multi-Effects Toxicity Database
The METDB contains points of departure for 28 of the 32 VOCs quantified in the SEPA HAP-MAP study area. The remaining four VOCs either did not have PODs for subchronic or chronic inhalation in ToxProfiles or CompTox (furan and ) or did not contain sufficient information in the CompTox database to allow for inclusion in our database (propene and n-propane, which were missing target organ systems for inhalation PODs).42 Of the VOCs included in the METDB, 19 chemicals had available ToxProfiles56–75 and 16 had available RfCs in IRIS.76 In total, PODs for 172 chemical–target organ systems combinations are included in the METDB (Supplemental Excel Table S1).
For all chemicals included in the METDB, our expanded approach identified additional adverse target organ systems not captured by IRIS RfCs or ATSDR MRLs (Figure 4). For example, the IRIS RfC only considers respiratory effects for formaldehyde, whereas the METDB identified 10 additional target organ systems (neurological, developmental, reproductive, systemic, hepatic, renal, hematological, ocular, immunological, and dermal) associated with formaldehyde exposure. PODs were identified in the METDB for a mean of six target organ systems per chemical (range from 1 to 13). IRIS RfCs captured two target organ systems for 1,4-dioxane, methyl tert-butyl-ether, and trichloroethylene; all other chemical RfCs were based on a single target organ system.
The most commonly identified target organ systems in the METDB were developmental and reproductive effects (20 and 19 chemicals, respectively); only 3 chemicals each were identified as reproductive or developmental toxicants using IRIS RfCs (Figure 4). Neurological effects were better represented than other target organ systems in the IRIS database: 9 of the 21 chemicals identified as neurotoxicants in the METDB had available RfCs for this end point in IRIS.
Risk Characterization
As shown in Figure 5 and Table 2, our expanded approach based on toxicity data from the METDB identified potential risks to the neurological, renal, respiratory, and endocrine systems, as well as for systemic effects (HIs of 9.21, 4.16, 1.54, 1.43, and 1.09, respectively) in the overall SEPA HAP-MAP study area. These risks were primarily driven by high mean concentrations of formaldehyde, which had an HQ above 1 for both neurological and renal effects (Figure 5). In contrast, no potential health risks were identified under the traditional approach of comparing SEPA HAP-MAP concentrations to available RfCs in IRIS (Table 2). The largest HI in our analysis based only on IRIS RfCs was 0.38 for respiratory effects (Table 2). Though many chemicals in the METDB had available TTCs for developmental and reproductive effects, as noted above, many of these values were negligible (), and thus we did not identify the overall SEPA HAP-MAP study area to be at risk for either effect.
| End point | Hazard indicesa | ||||
|---|---|---|---|---|---|
| Traditional approach (IRIS RfCsc only) | Expanded approach (METDBb) | CompTox excluded | Chronic assumption | Total uncertainty factor | |
| Neurological | 0.04 | 9.22a | 0.12 | 1.06a | 9.22a |
| Renal | — | 4.16a | 4.16a | 4.16a | 4.14a |
| Respiratory | 0.38 | 1.54a | 0.40 | 0.52 | 1.52a |
| Endocrine | — | 1.43a | 1.43a | 1.43a | 0.00 |
| Systemic | — | 1.09a | 0.40 | 0.47 | 1.09a |
| Reproductive | 0.09 | 0.74 | 0.66 | 0.67 | 0.72 |
| Ocular | 0.02 | 0.64 | 0.64 | 0.64 | 0.64 |
| Dermal | — | 0.60 | 0.60 | 0.60 | 0.60 |
| Immunological | 0.10 | 0.58 | 0.58 | 0.58 | 0.58 |
| Hematological | — | 0.46 | 0.26 | 0.28 | 0.46 |
| Developmental | 0.05 | 0.44 | 0.00 | 0.48 | 0.44 |
| Hepatic | 0.00 | 0.40 | 0.40 | 0.40 | 0.40 |
| Musculoskeletal | — | 0.10 | 0.10 | 0.10 | 0.00 |
| Cardiovascular | — | 0.02 | 0.02 | 0.02 | 0.02 |
| Gastrointestinal | — | 0.01 | 0.01 | 0.01 | 0.01 |
| Metabolic | — | 0.01 | 0.01 | 0.00 | 0.01 |
Note:
—, no data; IRIS, Integrated Risk Information System; RfC, reference concentration; METDB, Mult-Effect Toxicity Database; US EPA, US Environmental Protection Agency.
a
A Hazard index above 1 indicates adverse effects are possible.
b
METDB, multi-effects toxicity database.
c
IRIS RfC, Reference concentrations from the US EPA’s IRIS.
We conducted sensitivity analyses to assess the extent to which PODs from CompTox influence our results, in recognition that PODs from that source may be more variable in quality than PODs from ATSDR and IRIS. In these analyses, the METDB restricted to ATSDR and IRIS sources contained only 116 PODs. When PODs from CompTox were excluded, potential health risks were observed for renal and endocrine effects only (Figure S1). HIs for neurological effects decreased from 9.22 to 0.12, respiratory effects from 1.54 to 0.30, and systemic effects from 1.09 to 0.40, though values for other end points remained similar (Table 2). In our sensitivity analysis assuming PODs of unknown study duration were chronic rather than subchronic (therefore eliminating an UF of 10 for subchronic-to-chronic extrapolation), the HIs for neurological, respiratory, and systemic effects also decreased (Table 2). However, renal, endocrine, and neurological effects remained of concern () (Table 2). In our sensitivity analysis excluding the most uncertain TTCs (i.e., those calculated using a total UF of 3,000), we found similar HIs for all examined effects except for endocrine effects, where the HI decreased from 1.43 to 0.00.
Discussion
The METDB is a novel cumulative risk assessment methodology that can be used to improve the characterization of noncancer hazards by expanding consideration of the health effects of chemicals beyond the most sensitive end point. It has many potential applications but is especially useful for fenceline communities and other disadvantaged populations experiencing concurrent exposures to chemical mixtures. We piloted the METDB with air monitoring data from fenceline communities along the Delaware River and found the potential for neurological, renal, and respiratory, endocrine, and systemic risks () associated with measured VOCs. In contrast, using a traditional cumulative risk assessment approach incorporating only IRIS RfCs resulted in the conclusion that there were no potential adverse health effects. Our results highlight the importance of moving beyond toxicity information restricted to the most sensitive target organ system. In the context of mixtures, including the full range of potential health effects for each component may reveal additional effects of concern; the exposure due to each component may result in an but in combination results in an , leading to a different conclusion regarding the acceptability of noncancer risks and potentially different risk management decisions.
This methodology is intended to be a simple, reproducible approach for accounting for the health effects of chemical exposures more holistically. To facilitate the use of our methodology by others, we provide the METDB for the assessed chemicals from our case study (Supplemental Excel Table S1). Nevertheless, we acknowledge that our approach may still be time-intensive when applied to other chemicals, given that PODs derived from CompTox data require manual review and classification.
Although our method facilitates a more complete consideration of the noncancer toxicological burden resulting from chemical mixture exposures, there are several associated limitations. Our approach leverages PODs acquired from large, curated databases of toxicological information that are regularly updated. These databases are powerful, in that they facilitate the rapid identification and filtering of PODs and contain information needed to identify the documents from which the PODs originate. Without these databases, efforts to compile PODs to use in our method would have been prohibitively labor-intensive. Given the number of PODs required to implement this approach, however, we could not evaluate the underlying toxicological studies supporting the PODs selected for use in the METDB.
Although identifying and locating information about the underlying toxicological studies is possible, it can be challenging, because the PODs aggregated in CompTox are sourced from an array of government and other toxicological repositories that are not navigated in the same way. In some cases, the source material may be a well-described study published in a scientific journal, but in others, it could be part of a brief paragraph written by agency scientists summarizing the results of multiple unpublished studies. When original studies were not available, a lack of information about study length and dosing regimen contributed to uncertainty in our estimated TTCs. Thirty-five of our identified PODs lacked detail on study length and were treated as less than chronic duration (i.e., assigned a of 10) in our main analyses. We conducted a sensitivity analysis assuming that PODs of unknown duration were from chronic studies, which still found potential adverse health effects associated with our SEPA HAP-MAP case study measurements. Similarly, we assumed a dosing regimen consistent with OECD guidelines for PODs derived from animal studies in CompTox; deviations from this regimen may result in a misspecification of a continuous duration-adjusted POD, which could be a limitation to our approach. However, when we examined data on dosing regimens for PODs from ToxProfiles, we found that 58% used the OECD regimen and a further 23% used a similar regimen (e.g., 7 h instead of 6 h per day).
Consequently, a related limitation of our method is that each POD cannot be evaluated with the same rigor that would be applied in current regulatory approaches to deriving toxicology values (i.e., IRIS RfCs or ATSDR MRLs), which introduces uncertainty to our characterizations of noncancer hazards. We did not assess risk of bias for the studies underlying the identified PODs; therefore, it is possible that some of the PODs came from poor-quality studies that would have been excluded under a formal risk of bias assessment. However, when we limited our METDB to studies included in ToxProfiles (i.e., those that have been subjected to the peer-review process), we found results similar to those in our main analyses for most end points: our METDB approach still highlighted potential adverse health effects that were not shown in an approach limited to RfCs only. In addition, our results align with prior analyses that used different data sources to expand consideration of chemical hazards beyond the most sensitive end point.12,13
There are also limitations regarding availability of PODs for each potential target organ system. Our assessment of POD availability for the included chemicals found that a disproportionate number of PODs were available for reproductive and developmental end points in comparison with other target organ systems. This finding does not necessarily mean that the included chemicals were less likely to target other organ systems; it may instead reflect an uneven research focus on end points in other organ systems. The absence of a POD for a given chemical–target organ system combination when such an effect truly exists would result in the underestimation of corresponding HIs for that organ system.
An additional source of uncertainty in estimating noncancer burdens is our pragmatic assumption that mixture chemicals act additively to elicit effects. Although the mixtures evidence base is relatively limited (and populated primarily with studies of pairwise interactions and select groups of well-studied chemical classes),77,78 research on synergisms and antagonisms suggests that departures from additivity may occur in –25% of tested combinations, depending on the criteria selected to determine departures from additivity.77 It is difficult to predict the frequency, nature, and magnitude of these departures from additivity, and thus it is challenging to forecast the resulting influence on our estimation of noncancer burdens.
We used an approach for assigning UFs that mirrors their application by US agencies in the development of inhalation toxicology values. Recent research, however, suggests that this approach may not reflect the most current science and thus may inadequately protect public health. In particular, some have argued that differences in vulnerability among human populations are inadequately considered when the intraspecies UF () is assigned a value of 10.79,80 For example, women who smoked and had low iodide levels were up to 100 times more sensitive to disruptions in thyroid hormones following perchlorate exposure than women who did not have these risk factors.81 There is even precedent for using a in developing health-based reference values: the California Environmental Protection Agency’s Office of Environmental Health Hazard Assessment (OEHHA) set a standard for benzene using a of 60 for human variability, taking into account data on increased susceptibility to this chemical among people with specific genetic polymorphisms.82 As a result, the OEHHA standard was 70% lower than the US EPA RfC based on the same health end point.80 Varshavsky et al.80 note that exposures to extrinsic factors such as systemic racism can similarly act to increase susceptibility.
Although exposure concentrations derived from the SEPA HAP-MAP study VOC measurements were useful in demonstrating our approach, they reflect a 3-wk sampling campaign and are not equivalent to the longer-term measurements or modeled concentrations typically employed in regulatory risk assessments. Given the potential for seasonal variation in ambient VOC concentrations,83–85 there is uncertainty in generalizing these measurements to year-long or chronic exposure levels. This uncertainty can be reduced in future work by extending the measurement duration or by developing scaling factors to account for seasonal or year-to-year variation. In addition, our mobile monitoring quantified only a subset of hazardous air pollutants. Unmeasured pollutants may also contribute to noncancer health burdens.
Our SEPA HAP-MAP case example focuses on a geographically specific region; despite this, we believe the METDB approach could be applicable on a broader geographic scale, were credible exposure data for the affected population available. Other effect-based approaches have been proposed, in contrast with the stressor-based approach we demonstrate. These approaches involve using specific diseases as starting points, and the subsequent identification of relevant chemicals and other stressors as a basis for risk management.86 An alternative approach designed to account for mixture risks in chemical regulations called the mixtures assessment or allocation factor (MAF) has been proposed as part of the European Chemicals Strategy for Sustainability.87 Different permutations of the approach have been proposed, and their potential value in the regulation of mixtures is currently the subject of much stakeholder discourse in the European regulatory community.88 In the absence of widespread agreement or adoption of a specific approach, attempting different methodologies may be warranted, given the circumstances and available data.
There are several potential applications for the METDB approach in ongoing US EPA risk assessment and risk management activities. One opportunity lies in potentially augmenting the Air Toxics Screening Assessment (AirToxScreen), a mapping tool used by state, local, and Tribal air agencies to estimate noncancer hazards posed by ambient air toxics.89 Though the AirToxScreen has the capability to report HIs for target organ systems, it only uses the most sensitive end point for each chemical; the METDB approach would improve the AirToxScreen’s consideration of cumulative hazards from air toxics and potentially lead to better-justified public health interventions. Our approach may also have value in Superfund decision-making and may have the potential to improve siting and permitting decisions by better characterizing population noncancer hazards under different exposure scenarios.
There are also steps that regulatory agencies could take to expand the applicability of the METDB approach and to make its use easier and more rigorous. One opportunity is in the process for derivation of new toxicity values; some newer toxicological reviews from the US EPA’s IRIS program have provided candidate reference values alongside final selections of reference values.90,91 These candidates can be used for consideration of target organ systems beyond the one corresponding to the critical effect, allowing for the chemical to be readily implemented into a cumulative risk assessment using METDB approach. Further, although toxicological databases like the US EPA’s CompTox Chemicals Dashboard facilitate rapid aggregation of PODs for chemicals, additional resources dedicated to simplifying location of the underlying studies supporting each POD would reduce the resources needed to evaluate their quality; similarly, efforts to fill in details that are missing from POD records (e.g., study duration) would aid in bolstering confidence in applying the METDB approach. One other potential innovation would come from a shift toward the use of the probabilistic risk-specific dose methodology for noncancer effects1,92 (as the US EPA has demonstrated with acrolein)93; this shift would help move beyond limitations of the HI methodology toward a more quantitative characterization of population incidence of noncancer effects.
Understanding risks from multiple stressor exposures is the objective of cumulative risk assessment.2,94 Our work demonstrates an enhanced method to better understand the potential health risks stemming from exposures to multiple chemicals. Although this is an important advance built with existing data sources that can guide risk management actions, it does not yet account for other essential elements of a holistic cumulative risk assessment. Stressors are not limited to chemicals; biological, physical, psychological and economic stressors are recognized as exposures of concern that can co-occur with chemicals.2,94–97 Developing practical risk assessment methods that include these and other stressors will further advance our understanding of the full range of exposures and risks in fenceline communities to guide risk management toward environmental justice.
Acknowledgments
The authors thank all operators of the Aerodyne Mobile Laboratory who worked on the sampling campaign.
The authors acknowledge support from Bloomberg Philanthropies (Grant ID 2021-100480). A.A.C. was supported by NIEHS Training Grant T32ES007141 and the 21st Century Cities Initiative at Johns Hopkins University. A.A.C. and K.E.N. were supported by the Environmental Defense Fund (Grant ID 141857). C.G. was supported by the Johns Hopkins University Education and Research Center for Occupational Safety and Health (ERC) from the National Institute for Occupational Safety and Health (NIOSH) under Grant No. 5 T42 OH 008428.
Article Notes
M. Claflin, E.F., H.S., J.K., M. Canagaratna, S.H., and T.I.Y. are current or former employees of Aerodyne Research, Inc., which developed and commercialized the PTR-MS, gas chromatography–mass spectrometry, and TILDAS instruments used in this study. The remaining authors declare they have no conflicts of interest related to this work to disclose.
Supplementary Material
References
1.
National Research Council. 2009. Science and Decisions: Advancing Risk Assessment. Washington, DC: The National Academies Press. https://doi.org/10.17226/12209.
2.
US EPA (US. Environmental Protection Agency). Framework for Cumulative Risk Assessment. https://www.epa.gov/sites/default/files/2014-11/documents/frmwrk_cum_risk_assmnt.pdf [accessed 27 August 2024].
3.
International Programme on Chemical Safety & Inter-Organization Programme for the Sound Management of Chemicals. Assessment of Combined Exposures to Multiple Chemicals: Report of a WHO/IPCS International Workshop on Aggregate/Cumulative Risk Assessment. https://iris.who.int/handle/10665/44113 [accessed 5 August 2024].
4.
National Research Council. Phthalates and Cumulative Risk Assessment: The Tasks Ahead. https://nap.nationalacademies.org/catalog/12528/phthalates-and-cumulative-risk-assessment-the-tasks-ahead [accessed 27 August 2024].
5.
ATSDR (Agency for Toxic Substances and Disease Registry). Framework for Assessing Health Impacts of Multiple Chemicals and Other Stressors. https://www.atsdr.cdc.gov/interactionprofiles/ip-ga/ipga.pdf [accessed 5 August 2024].
6.
US EPA. Guidelines for the Health Risk Assessment of Chemical Mixtures. https://www.epa.gov/sites/default/files/2014-11/documents/chem_mix_1986.pdf [accessed 5 August 2024].
7.
US EPA. Concepts, Methods and Data Sources for Cumulative Health Risk Assessment of Multiple Chemicals, Exposures and Effects: a Resource Document. https://ordspub.epa.gov/ords/eims/eimscomm.getfile?p_download_id=474337 [accessed 5 August 2024].
8.
US EPA. Cumulative Risk Assessment Under the Toxic Substances Control Act. https://www.epa.gov/system/files/documents/2023-02/Draft%20Principles%20of%20CRA%20under%20TCSA_0.pdf [accessed 5 August 2024].
9.
US EPA. Guidance on Cumulative Risk Assessment of Pesticide Chemicals That Have a Common Mechanism of Toxicity. https://www.epa.gov/sites/default/files/2015-07/documents/guidance_on_common_mechanism.pdf [accessed 5 August 2024].
10.
US EPA. Draft Proposed Approach for Cumulative Risk Assessment of High-Priority Phthalates and a Manufacturer-Requested Phthalate under the Toxic Substances Control Act. https://www.epa.gov/system/files/documents/2023-02/Draft%20Phthalate%20CRA%20Approach.pdf [accessed 5 August 2024].
11.
US EPA. Guidelines for Cumulative Risk Assessment: Planning and Problem Formulation. https://ordspub.epa.gov/ords/eims/eimscomm.getfile?p_download_id=546986 [accessed 5 August 2024].
12.
Fox MA. 2002. Evaluating cumulative risk assessment for environmental justice: a community case study. Environ Health Perspect 110 Suppl 2(suppl 2):203–209. https://pubmed.ncbi.nlm.nih.gov/11929729/, https://doi.org/10.1289/ehp.02110s2203.
13.
Fox MA, Tran NL, Groopman JD, Burke TA. 2004. Toxicological resources for cumulative risk: an example with hazardous air pollutants. Regul Toxicol Pharmacol 40(3):305–311. https://pubmed.ncbi.nlm.nih.gov/15546684/, https://doi.org/10.1016/j.yrtph.2004.07.008.
14.
Carpenter DO, Arcaro K, Spink DC. 2002. Understanding the human health effects of chemical mixtures. Environ Health Perspect 110 Suppl 1(suppl 1):25–42. https://pubmed.ncbi.nlm.nih.gov/11834461/, https://doi.org/10.1289/ehp.02110s125.
15.
US EPA. 2000. Supplementary Guidance for Conducting Health Risk Assessment of Chemical Mixtures. Washington, DC: US EPA.
16.
US EPA. 2025. IRIS Toxicological Review of Perfluorohexanesulfonic Acid (PFHxS, CASRN 335-46-4) and Related Salts. https://iris.epa.gov/static/pdfs/0705tr.pdf [accessed 19 January 2025].
17.
US EPA. 2025. IRIS Toxicological Review of Inorganic Arsenic. https://iris.epa.gov/static/pdfs/0278tr.pdf [accessed 19 January 2025].
18.
Tehrani MW, Fortner EC, Robinson ES, Chiger AA, Sheu R, Werden BS, et al. 2023. Characterizing metals in particulate pollution in communities at the fenceline of heavy industry: combining mobile monitoring and size-resolved filter measurements. Environ Sci Process Impacts 25(9):1491–1504. https://pubmed.ncbi.nlm.nih.gov/37584085/, https://doi.org/10.1039/d3em00142c.
19.
US EPA. EasyRSEI Dashboard Version 2.3.11. https://edap.epa.gov/public/extensions/EasyRSEI/EasyRSEI.html [accessed 19 January 2025].
20.
Whitehead R Jr, Block W. 2000. Environmental justice risks in the petroleum industry. Wm & Mary Envtl L & Pol’y Rev 24:67. https://scholarship.law.wm.edu/wmelpr/vol24/iss1/4 [accessed 19 January 2025].
21.
Chester Residents Concerned for Quality Living v. Seif. 1996. 944 F. Supp. 413 (United States District Court, E.D. Pennsylvania). https://law.justia.com/cases/federal/district-courts/FSupp/944/413/1485694/ [accessed 4 August 2024].
22.
OpenStreetMap Contributors. 2025. OpenStreetMap. https://www.openstreetmap.org/copyright [accessed 19 January 2025].
23.
Fellows I. 2025. OpenStreetMap: Access to Open Street Map Raster Images. R package version 0.3.4. https://cran.r-project.org/package=OpenStreetMap [accessed 16 February 2025].
24.
Google. 2025. Google Earth. https://earth.google.com/ [accessed 16 February 2025].
25.
United States Census Bureau. 2025. TIGER/Line Shapefile, 2020, Census Tracts, Pennsylvania. https://www.census.gov/geo/tiger/ [accessed 16 February 2025].
26.
Krechmer J, Lopez-Hilfiker F, Koss A, Hutterli M, Stoermer C, Deming B, et al. 2018. Evaluation of a new reagent-ion source and focusing ion–molecule reactor for use in proton-transfer-reaction mass spectrometry. Anal Chem 90(20):12011–12018. https://pubmed.ncbi.nlm.nih.gov/30220198/, https://doi.org/10.1021/acs.analchem.8b02641.
27.
Claflin MS, Pagonis D, Finewax Z, Handschy AV, Day DA, Brown WL, et al. 2021. An in situ gas chromatograph with automatic detector switching between PTR-and EI-TOF-MS: isomer-resolved measurements of indoor air. Atmos Meas Tech 14(1):133–152, https://doi.org/10.5194/amt-14-133-2021.
28.
McManus JB, Zahniser MS, Nelson DD, Williams LR, Kolb CE. 2002. Infrared laser spectrometer with balanced absorption for measurement of isotopic ratios of carbon gases. Spectrochim Acta A Mol Biomol Spectrosc 58(11):2465–2479. https://pubmed.ncbi.nlm.nih.gov/12353697/, https://doi.org/10.1016/s1386-1425(02)00064-1.
29.
Yacovitch TI, Herndon SC, Roscioli JR, Floerchinger C, Knighton WB, Kolb CE. 2015. Air pollutant mapping with a mobile laboratory during the BEE-TEX field study. Environ Health Insights 9(suppl 4):7–13. https://pubmed.ncbi.nlm.nih.gov/26819556/, https://doi.org/10.4137/EHI.S15660.
30.
Yacovitch TI, Lerner BM, Canagaratna MR, Daube C, Healy RM, Wang JM, et al. 2023. Mobile laboratory investigations of industrial point source emissions during the MOOSE field campaign. Atmosphere 14(11):1632, https://doi.org/10.3390/atmos14111632.
31.
Herndon SC, Jayne JT, Zahniser MS, Worsnop DR, Knighton B, Alwine E, et al. 2005. Characterization of urban pollutant emission fluxes and ambient concentration distributions using a mobile laboratory with rapid response instrumentation. Faraday Discuss 130:327–339. https://pubmed.ncbi.nlm.nih.gov/16161792/, https://doi.org/10.1039/b500411j.
32.
Kolb CE, Herndon SC, McManus JB, Shorter JH, Zahniser MS, Nelson DD, et al. 2004. Mobile laboratory with rapid response instruments for real-time measurements of urban and regional trace gas and particulate distributions and emission source characteristics. Environ Sci Technol 38(21):5694–5703. https://pubmed.ncbi.nlm.nih.gov/15575289/, https://doi.org/10.1021/es030718p.
33.
Spinei E, Whitehill A, Fried A, Tiefengraber M, Knepp TN, Herndon S, et al. 2018. The first evaluation of formaldehyde column observations by improved Pandora spectrometers during the KORUS-AQ field study. Atmos Chem Phys 11(9):4943–4961. https://pubmed.ncbi.nlm.nih.gov/33424951/, https://doi.org/10.5194/amt-11-4943-2018.
34.
Robinson ES, Tehrani MW, Yassine A, Agarwal S, Nault BA, Gigot C, et al. 2024. Ethylene oxide in southeastern Louisiana’s petrochemical corridor: high spatial resolution mobile monitoring during HAP-MAP. Environ Sci Technol 58(25):11084–11095. https://pubmed.ncbi.nlm.nih.gov/38860676/, https://doi.org/10.1021/acs.est.3c10579.
35.
McManus JB, Zahniser MS, Nelson DD, Shorter JH, Herndon SC, Jervis D, et al. 2015. Recent progress in laser-based trace gas instruments: performance and noise analysis. Appl Phys B 119(1):203–218, https://doi.org/10.1007/s00340-015-6033-0.
36.
Yacovitch TI, Herndon SC, Roscioli JR, Floerchinger C, McGovern RM, Agnese M, et al. 2014. Demonstration of an ethane spectrometer for methane source identification. Environ Sci Technol 48(14):8028–8034. https://pubmed.ncbi.nlm.nih.gov/24945706/, https://doi.org/10.1021/es501475q.
37.
Yacovitch TI, Dyroff C, Roscioli JR, Daube C, McManus JB, Herndon SC. 2023. Ethylene oxide monitor with part-per-trillion precision for in situ measurements. Atmos Meas Tech 16(7):1915–1921, https://doi.org/10.5194/amt-16-1915-2023.
38.
US EPA. 1993. Reference Dose (RfD): Description and Use in Health Risk Assessments, Background Document 1A. Washington, DC: US EPA.
39.
US EPA. Basic Information about the Integrated Risk Information System. https://www.epa.gov/iris/basic-information-about-integrated-risk-information-system#process [accessed 4 August 2024].
40.
US EPA. ORD Staff Handbook for Developing IRIS Assessments. https://ordspub.epa.gov/ords/eims/eimscomm.getfile?p_download_id=545991 [accessed 9 August 2024].
41.
ATSDR. 10 Steps to making an ATSDR ToxProfile. https://www.atsdr.cdc.gov/toxprofiles/docs/FINAL_508-ready_ToxProfile-Infographic_6.10.14.pdf [accessed 9 August 2024].
42.
US EPA. CompTox Chemicals Dashboard v2.2.1. https://comptox.epa.gov/dashboard/ [accessed 9 August 2024].
43.
Williams AJ, Grulke CM, Edwards J, McEachran AD, Mansouri K, Baker NC, et al. 2017. The CompTox Chemistry Dashboard: a community data resource for environmental chemistry. J Cheminform 9(1):61–27. https://pubmed.ncbi.nlm.nih.gov/29185060/, https://doi.org/10.1186/s13321-017-0247-6.
44.
ATSDR. 2018. Draft Guidance for the Preparation of Toxicological Profiles. https://www.atsdr.cdc.gov/toxicological-profiles/media/pdfs/2024/10/tox-profile-development-guidance.pdf [accessed 9 August 2024].
45.
US EPA. Toxicological Review of Trichloroethylene. https://iris.epa.gov/static/pdfs/0199tr.pdf [accessed 9 August 2024].
46.
US EPA. IRIS Toxicological Review of Formaldehyde (Inhalation). https://iris.epa.gov/static/pdfs/0419tr.pdf [accessed 9 August 2024].
47.
US EPA. 1998. Health Effects Test Guidelines OPPTS 870.3465 90-Day Inhalation Toxicity. https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=P100IRTA.TXT [accessed 3 September 2024].
48.
OECD (Organisation for Economic Co-operation and Development). Test No. 452: chronic Toxicity Studies. https://www.oecd-ilibrary.org/content/publication/9789264071209-en [accessed 3 September 2024].
49.
OECD. OECD Guideline For the Testing of Chemicals, Subchronic Inhalation Toxicity: 90-Day Study. https://ntp.niehs.nih.gov/sites/default/files/iccvam/suppdocs/feddocs/oecd/oecd-tg413.pdf [accessed 3 September 2024].
50.
Jarabek AM, Asgharian B, Miller FJ. 2005. Dosimetric adjustments for interspecies extrapolation of inhaled poorly soluble particles (PSP). Inhal Toxicol 17(7–8):317–334. https://pubmed.ncbi.nlm.nih.gov/16020031/, https://doi.org/10.1080/08958370590929394.
51.
US EPA. Risk Assessment Guidance for Superfund Volume I: Human Health Evaluation Manual (Part F, Supplemental Guidance for Inhalation Risk Assessment). https://semspub.epa.gov/src/document/HQ/140530 [accessed 8 August 2024].
52.
Korsak Z, Rydzyński K. 1996. Neurotoxic effects of acute and subchronic inhalation exposure to trimethylbenzene isomers (pseudocumene, mesitylene, hemimellitene) in rats. Int J Occup Med Environ Health 9(4):341–349. https://pubmed.ncbi.nlm.nih.gov/9117193/.
53.
European Chemicals Agency. 2025. ECHA Chemicals Database (ECHACHEM). https://echa.europa.eu/information-on-chemicals/registered-substances [accessed 9 August 2024].
54.
US EPA. Risk Assessment Guidance for Superfund Volume I Human Health Evaluation Manual (Part A). https://www.epa.gov/sites/default/files/2015-09/documents/rags_a.pdf [accessed 9 August 2024].
55.
US EPA. 2025. A Review of the Reference Dose and Reference Concentration Processes. https://www.epa.gov/sites/default/files/2014-12/documents/rfd-final.pdf [accessed 9 August 2024].
56.
Public Health Service, US Department of Health and Human Services. 2022. Toxicological profile for Chloromethane (Draft for Public Comment). Washington, DC: US Department of Health and Human Services, Public Health Service.
57.
Public Health Service, US Department of Health and Human Services. 1997. Toxicological Profile for Chloroform. Washington, DC: US Department of Health and Human Services, Public Health Service.
58.
Public Health Service, US Department of Health and Human Services. 1999. Toxicological Profile for Formaldehyde. Washington, DC: US Department of Health and Human Services, Public Health Service.
59.
Public Health Service, US Department of Health and Human Services. 1999. Toxicological Profile for n-Hexane. Washington, DC: US Department of Health and Human Services, Public Health Service.
60.
Public Health Service, US Department of Health and Human Services. 2006. Toxicological Profile for Vinyl Chloride. Washington, DC: US Department of Health and Human Services, Public Health Service.
61.
Public Health Service, US Department of Health and Human Services. 2007. Toxicological Profile for Xylenes. Washington, DC: US Department of Health and Human Services, Public Health Service.
62.
Public Health Service, US Department of Health and Human Services. 2010. Addendum to the Toxicological Profile for Formaldehyde. Washington, DC: US Department of Health and Human Services, Public Health Service.
63.
Public Health Service, US Department of Health and Human Services. 2010. Toxicological Profile for Ethylbenzene. Washington, DC: US Department of Health and Human Services, Public Health Service.
64.
Public Health Service, US Department of Health and Human Services. 2010. Toxicological Profile for Styrene. Washington, DC: US Department of Health and Human Services, Public Health Service.
65.
Public Health Service, US Department of Health and Human Services. 2012. Toxicological Profile for 1,3-Butadiene. Washington, DC: US Department of Health and Human Services, Public Health Service.
66.
Public Health Service, US Department of Health and Human Services. 2012. Toxicological Profile for 1,4-Dioxane. Washington, DC: US Department of Health and Human Services, Public Health Service.
67.
Public Health Service, US Department of Health and Human Services. 2016. Addendum to the Toxicological Profile for Vinyl Chloride. Washington, DC: US Department of Health and Human Services, Public Health Service.
68.
Public Health Service, US Department of Health and Human Services. 2017. Toxicological Profile for Benzene. Washington, DC: US Department of Health and Human Services, Public Health Service.
69.
Public Health Service, US Department of Health and Human Services. 2017. Toxicological Profile for Toluene. Washington, DC: US Department of Health and Human Services, Public Health Service.
70.
Public Health Service, US Department of Health and Human Services. 2019. Toxicological Profile for Tetrachloroethylene. Washington, DC: US Department of Health and Human Services, Public Health Service.
71.
Public Health Service, US Department of Health and Human Services. 2019. Toxicological Profile for Trichloroethylene. Washington, DC: US Department of Health and Human Services, Public Health Service.
72.
Public Health Service, US Department of Health and Human Services. 2020. Toxicological Profile for Chlorobenzene. Washington, DC: US Department of Health and Human Services, Public Health Service.
73.
Public Health Service, US Department of Health and Human Services. 2020. Toxicological Profile for Ethylbenzene (Draft for Public Comment). Washington, DC: US Department of Health and Human Services, Public Health Service.
74.
Public Health Service, US Department of Health and Human Services. 2022. Toxicological Profile for Methyl-tert-butyl Ether (MTBE) (Draft for Public Comment). Washington, DC: US Department of Health and Human Services, Public Health Service.
75.
Public Health Service, US Department of Health and Human Services. 2000. Toxicological Profile for Methylene Chloride. Washington, DC: US Department of Health and Human Services, Public Health Service.
76.
US EPA. IRIS Assessments. https://iris.epa.gov/AtoZ/?list_type=alpha [accessed 9 August 2024].
77.
Martin O, Scholze M, Ermler S, McPhie J, Bopp SK, Kienzler A, et al. 2021. Ten years of research on synergisms and antagonisms in chemical mixtures: a systematic review and quantitative reappraisal of mixture studies. Environ Int 146:106206. https://pubmed.ncbi.nlm.nih.gov/33120228/, https://doi.org/10.1016/j.envint.2020.106206.
78.
Maffini MV, Rayasam SDG, Axelrad DA, Birnbaum LS, Cooper C, Franjevic S, et al. 2023. Advancing the science on chemical classes. Environ Health 21(S1):1–14, https://doi.org/10.1186/s12940-022-00919-y.
79.
Chiu WA. 2024. Invited perspective: uneven progress addressing population variability in human health risk assessment. Environ Health Perspect 132(3):031305. https://pubmed.ncbi.nlm.nih.gov/38498339/, https://doi.org/10.1289/EHP13461.
80.
Varshavsky JR, Rayasam SDG, Sass JB, Axelrad DA, Cranor CF, Hattis D, et al. 2023. Current practice and recommendations for advancing how human variability and susceptibility are considered in chemical risk assessment. Environ Health 21(S1):1–20, https://doi.org/10.1186/s12940-022-00940-1.
81.
Blount BC, Pirkle JL, Osterloh JD, Valentin-Blasini L, Caldwell KL. 2006. Urinary perchlorate and thyroid hormone levels in adolescent and adult men and women living in the United States. Environ Health Perspect 114(12):1865–1871. https://pubmed.ncbi.nlm.nih.gov/17185277/, https://doi.org/10.1289/ehp.9466.
82.
Air Toxicology and Epidemiology Branch, Office of Environmental Health Hazard Assessment California Environmental Protection Agency. 2008. Technical Support Document for the Derivation of Noncancer Reference Exposure Levels. Oakland, California: California Environmental Protection Agency.
83.
Gentner DR, Harley RA, Miller AM, Goldstein AH. 2009. Diurnal and seasonal variability of gasoline-related volatile organic compound emissions in riverside, California. Environ Sci Technol 43(12):4247–4252. https://pubmed.ncbi.nlm.nih.gov/19603630/, https://doi.org/10.1021/es9006228.
84.
Sadeghi B, Pouyaei A, Choi Y, Rappenglueck B. 2022. Influence of seasonal variability on source characteristics of VOCs at Houston industrial area. Atmos Environ 277:119077, https://doi.org/10.1016/j.atmosenv.2022.119077.
85.
Dumanoglu Y, Kara M, Altiok H, Odabasi M, Elbir T, Bayram A. 2014. Spatial and seasonal variation and source apportionment of volatile organic compounds (VOCs) in a heavily industrialized region. Atmos Environ 98:168–178, https://doi.org/10.1016/j.atmosenv.2014.08.048.
86.
Menzie CA, MacDonell MM, Mumtaz M. 2007. A phased approach for assessing combined effects from multiple stressors. Environ Health Perspect 115(5):807–816. https://pubmed.ncbi.nlm.nih.gov/17520072/, https://doi.org/10.1289/ehp.9331.
87.
Backhaus T. 2023. The mixture assessment or allocation factor: conceptual background, estimation algorithms and a case study example. Environ Sci Eur 35(1):55, https://doi.org/10.1186/s12302-023-00757-w.
88.
Backhaus T. 2024. Exploring the mixture assessment or allocation factor (MAF): a brief overview of the current discourse. Curr Opin Toxicol 37:100460, https://doi.org/10.1016/j.cotox.2024.100460.
89.
US EPA. Air Toxics Screening Assessment. https://www.epa.gov/AirToxScreen [accessed 10 August 2024].
90.
US EPA. IRIS Toxicological Review of Inorganic Arsenic (External Peer Review Draft). https://ordspub.epa.gov/ords/eims/eimscomm.getfile?p_download_id=547771[accessed 10 August 2024].
91.
US EPA. IRIS Toxicological Review of Perfluorodecanoic Acid (PFDA) and Related Salts. https://ordspub.epa.gov/ords/eims/eimscomm.getfile?p_download_id=549465 [accessed 10 August 2024].
92.
Chiu WA, Slob W. 2015. A unified probabilistic framework for dose–response assessment of human health effects. Environ Health Perspect 123(12):1241–1254. https://pubmed.ncbi.nlm.nih.gov/26006063/, https://doi.org/10.1289/ehp.1409385.
93.
Blessinger T, Davis A, Chiu WA, Stanek J, Woodall GM, Gift J, et al. 2020. Application of a unified probabilistic framework to the dose–response assessment of acrolein. Environ Int 143:105953. https://pubmed.ncbi.nlm.nih.gov/32768806/, https://doi.org/10.1016/j.envint.2020.105953.
94.
US EPA. Guidance on Cumulative Risk Assessment. Part 1. Planning and Scoping. https://www.epa.gov/sites/default/files/2015-01/documents/cumrisk2_0.pdf [accessed 27 August 2024].
95.
Young GS, Fox MA, Trush M, Kanarek N, Glass TA, Curriero FC. 2012. Differential exposure to hazardous air pollution in the United States: a multilevel analysis of urbanization and neighborhood socioeconomic deprivation. Int J Environ Res Public Health 9(6):2204–2225. https://pubmed.ncbi.nlm.nih.gov/22829799/, https://doi.org/10.3390/ijerph9062204.
96.
Fox MA, Niemeier RT, Hudson N, Siegel MR, Dotson GS. 2021. Cumulative risks from stressor exposures and personal risk factors in the workplace: examples from a scoping review. Int J Environ Res Public Health 18(11):5850. https://pubmed.ncbi.nlm.nih.gov/34072475/, https://doi.org/10.3390/ijerph18115850.
97.
Chari R, Burke TA, White RH, Fox MA. 2012. Integrating susceptibility into environmental policy: an analysis of the national ambient air quality standard for lead. Int J Environ Res Public Health 9(4):1077–1096. https://pubmed.ncbi.nlm.nih.gov/22690184/, https://doi.org/10.3390/ijerph9041077.
Information & Authors
Information
Published In
License Information
EHP is a Diamond Open Access journal published with support from the NIEHS, NIH. All content is public domain unless otherwise noted. Contact the corresponding author for permission before any reuse of content. Full licensing information is available online.
History
Received: 25 January 2024
Revision received: 18 February 2025
Accepted: 20 March 2025
Published ahead of print: 24 March 2025
Published online: 8 May 2025
Notes
Conclusions and opinions are those of the individual authors and do not necessarily reflect the policies or views of EHP Publishing or the National Institute of Environmental Health Sciences.
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click DOWNLOAD.