A wide variety of laws designed to protect human health and the environment do so through setting some type of limit on specific chemicals. For example, worker safety laws include occupational exposure limits. Drinking water regulations specify maximum contaminant levels. Food safety laws contain migration limits for materials intended to contact food. Although other factors (e.g., analytical capabilities, technology limitations, and cost) do affect these limits, typically each limit is based on a risk assessment that has concluded that a particular amount or “dose” of the regulated chemical will not present an appreciable risk to the health of the relevant population or environment.
With a few exceptions, human health risk analyses (e.g., occupational, general population and specific subpopulations, bystanders, and consumers/residential) for chemicals underpinning all of these laws have been done on a single-substance basis. Yet we know that people are often exposed to trace levels of many different chemicals through a variety of exposure routes.
Globally, there are growing concerns that limits based on single-substance risk assessments are not adequately protective, especially for compromised or sensitive populations. The European Union Chemicals Strategy for Sustainability Towards a Toxic-Free Environment calls for the development of consistent legal requirements to “ensure that risks from simultaneous exposure to multiple chemicals are effectively and systematically taken into account across chemicals-related policy areas.” EU Commission Communication Regarding Chemicals Strategy for Sustainability: Towards a Toxic-Free Environment at 12, COM (2003) 667 final (Oct. 14, 2020). In response to President Biden’s executive orders regarding environmental justice and climate change, the U.S. Environmental Protection Agency (EPA) Office of Research and Development (ORD) published a Report on Cumulative Impacts and Recommendations for ORD Research, which states: “For EPA to fulfill its mission to protect human health and the environment, the Agency needs to address the cumulative impacts of chemical and non-chemical stressors with the best available science. . . .” EPA, Cumulative Impacts Research: Recommendations for EPA’s Office of Research and Development at 1 (2022). So, policy makers are prioritizing the development of different approaches to regulate co-exposure, including the use of cumulative chemical impact assessments and mixture assessment factors (MAF).
The “cumulative impacts” affecting an individual, community, or the environment at a point in time or over a period of time can include both chemical and nonchemical stressors and the interactions between them. “Cumulative impact assessments” are used to account for the effects of all these stressors—positive, neutral, or negative—in the context of decision-making. A “cumulative chemical impact assessment” (alternatively called a “combined chemical impact assessment” or “multiple chemical impact assessment”) is a narrower assessment and evaluates the combined risk to an individual, population, or ecosystem from exposures to multiple chemicals. Policy makers are particularly concerned about the combination (or cocktail) effect of the “known unknowns” and the possible cumulative exposure scenarios that unintentionally result from mixtures of different chemical substances from different sources.
For intentional mixtures, which are typically manufactured and formulated products like shampoos, detergents, and pesticides, the composition and hazardous properties of the mixture are typically known and regulated (e.g., rules on the composition of cosmetics and plant protection products). For unintentional mixtures that originate from a single source, such as a wastewater effluent, the composition of the mixture can be assessed and, where relevant, regulated (e.g., water quality standards).
However, for mixtures of chemicals resulting from multiple sources and multiple pathways, assessing composition and hazardous properties and characterizing cumulative exposure are difficult. Regulations requiring cumulative chemical assessments do exist—for example, maximum residue levels for pesticides and occupational exposure levels for workers exposed to chemicals. However, there is no systematic approach to regulating mixture effects resulting unintentionally from different products and different routes of exposure. Whether an approach to addressing potential mixtures will further improve safety and health outcomes is the subject of important scientific research and dialogue.
While policy makers are demanding increased use of cumulative chemical impact assessments, these assessments are complex and the resources required to complete them can exceed those available to agencies. This article summarizes the key concepts and methodologies used in conducting risk assessments for exposures to multiple chemicals, discusses the common themes and challenges that are emerging from the European Union and the United States, and notes some of the collaborations underway to develop relevant laws and methodologies to manage cumulative chemical exposure scenarios.
In assessing the risk posed by a chemical to an individual, a population, or the environment, there are two essential considerations: the hazards posed by the chemical and the exposure pathways available for the chemical to interact with the community of interest.
A “chemical hazard assessment” identifies and attempts to quantify the intrinsic dangers posed by a chemical. Physical dangers such as flammability, explosivity, and corrosiveness can usually be assessed with relatively few standard laboratory tests. Health effects, such as carcinogenicity, neurotoxicity, and endocrine disruption, are more complicated to evaluate. This work typically relies on the weight of evidence of toxicity testing on animals (in vivo), laboratory analyses of cell cultures and tissues (in vitro), and computer modeling (in silico). Chemical hazard assessments typically provide an estimated dose or range of doses that are without harmful effects. A “dose” is the amount of a chemical that a person (or other living organism) receives over a given period of time.
A “chemical exposure assessment” identifies all of the pathways by which a particular organism or a population may be exposed to a chemical. Examples of pathways include ingesting the chemical from food, water, or toys; inhaling the chemical because it is present in indoor or outdoor air; and absorbing the chemical through the skin from clothing, toys, or tools. The output of an exposure assessment is typically a range of potential doses of a chemical estimated to be received by the population of interest.
A “chemical risk assessment” compares the outputs from a chemical hazard assessment and chemical exposure assessment to determine if the dose of chemical received by the population being considered is “without appreciable risk” or more commonly understood as “safe.” The results of a chemical risk assessment are typically expressed as margins of exposure or risk quotients. For example, if the dose to be received by the population of interest is 100 times smaller than the dose that is considered safe, this would be expressed as a margin of exposure of 100. Thus, the higher the margin of exposure, the lower the estimated risk.
A recent survey of chemical inventories around the world suggests that more than 350,000 different chemicals and chemical mixtures have been identified.
The Food Quality Protection Act of 1996 requires the U.S. EPA to conduct cumulative risk assessments for pesticides that share a common mechanism of toxicity. 21 U.S.C. § 346a (2022). A common mechanism of toxicity is assumed when two or more chemicals cause the same health effect by the same or very similar sequence of biochemical events. To consider potential human health risks from all pathways of dietary and nondietary exposures to a single chemical (aggregate assessments) and/or more than one pesticide acting through a common mechanism of toxicity (cumulative assessments), researchers undertook considerable efforts to develop data, data analytics, and modeling frameworks. This foundational work was necessary to conduct, evaluate, and provide tiered guidance. Thus, a substantial “learning curve” was established that led to valuable case studies and scientific review. See EPA, Pesticide Science and Assessing Pesticide Risk, at epa.gov. This provided a basis for cumulative risk analysis and its consideration in other regulatory settings.
A recent survey of chemical inventories around the world suggests that more than 350,000 different chemicals and chemical mixtures have been identified. Zhanyun Wang et al., Toward a Global Understanding of Chemical Pollution: A First Comprehensive Analysis of National and Regional Chemical Inventories, 54 Env’t Sci. Tech. 2575 (Jan. 22, 2020). Even if a population is exposed to only a small fraction of these chemicals, the number of potentially different mixtures is overwhelming. To help identify where further alignment in scientific considerations could be made, the Organization for Economic Co-operation and Development (OECD) published an overview of the technical aspects of various approaches used in risk assessment of multiple chemicals. OECD, Considerations for Assessing the Risks of Combined Exposure to Multiple Chemicals, Series on Testing & Assessment No. 296, Env’t, Health & Safety Div., Env’t Directorate (2018).
The potential for a population to be exposed to multiple chemicals is a threshold question in deciding whether to conduct a cumulative chemical impact assessment. Examples of data to consider in making this decision include analytical measurements of multiple chemicals in environmental samples such as drinking water, the composition of intentionally produced mixtures such as pesticide formulations and cosmetic products, information of intended uses for regulated chemicals, and market data on chemical sales. Id. at 24–25.
In formulating problem statements for the assessment, it is beneficial for researchers to consider a variety of factors, including the questions that they are being asked to answer, the applicable legal frameworks for any actions, key stakeholders, and any community or site-specific concerns. Defining the scope of the assessment early is also important, and key considerations include which exposure pathways and exposed populations will be considered and the types of effects to be evaluated. Researchers should also consider the implications and value of additional data collection. Id. at 22–23.
The potential for a population to be exposed to multiple chemicals is a threshold question in deciding whether to conduct a cumulative chemical impact assessment.
A “fit for purpose” assessment uses only the resources needed to support the decisions that will rely on the assessment. The OECD paper summarizes a key initiative conducted by the World Health Organization and the International Program on Chemical Safety to create a model that uses a “fit for purpose” strategy to conduct risk assessments for multiple chemical exposures. The model employs parallel tiered hazard assessments and exposure assessments to support the risk assessment. The tiers range from zero to three, and lower tiers require less data and employ conservative assumptions. For example, the Tier 0 hazard assessment uses a default dose addition for all chemicals. The Tier 3 assessments for hazard and exposure employ probabilistic estimates, which require a significant amount of data for each chemical. The assessment progresses from lower tiers to higher tiers until the assessment answers the regulatory question, or until the absence of data limits further refinement. Id. at 13–17.
The OECD paper also summarizes the two fundamentally different approaches used for assessing the risks associated with exposures to multiple chemicals: “whole mixture” approaches and “component-based” approaches. The choice of approach will greatly influence how the assessment is conducted, and its limitations.
In “whole mixture” approaches, testing is conducted on the mixture of chemicals that are the subject of the study, and the mixture is then assessed as if it were a single chemical. The benefit of this approach is it considers any unidentified chemicals as well as any interactions among the chemicals. One limitation is that whole mixture approaches assume that the composition of the mixture does not vary over time or exposure routes, which may not be realistic. Nor do they indicate which of the chemicals in the mixture are responsible for any adverse effects, which can hinder the development of mitigation strategies. Id. at 19.
For “component-based” approaches, the estimated effects of the mixture are based on the toxicities of the individual chemicals and on any interactions between the chemicals and the organisms that influence the toxicities of the individual chemicals. The strength of component-based approaches is that they can be performed prospectively, without the need to conduct experimental assessments on the mixture. However, they are highly dependent upon having sufficient data for each chemical in the mixture and on the predictive model used to estimate the interactions. There are three predominant predictive model types that can be used to estimate the effects of the exposure to the mixture of chemicals. A dose addition/concentration addition model is used when the chemicals in the mixture are expected to have the same or similar modes of action or adverse outcome pathways (e.g., all of the chemicals in the mixture have the potential to cause liver cancer). A response addition/independent action model is used when the chemicals are expected to have independent modes of action (e.g., one chemical has the potential to cause liver cancer; another has the potential to cause thyroid disorders). Third, customized models are developed to address chemical interactions when it is expected that the combined effect of two or more chemicals will be greater (synergistic) or less (antagonistic) than what would be predicted by the other two models. Id. at 19–20. Drug interactions provide common examples of synergistic and antagonistic effects. The model used will influence the conclusion of what doses are safe. When data regarding the mode of action of the chemicals are limited, experts may have different opinions about the appropriate model to use. In an effort to harmonize the model selection process, scientific communities and agencies within the EU and the United States have developed decision trees to guide choices on what type of assessment approaches to use.
With limited exceptions, most laws contain no requirement to conduct cumulative chemical impact assessments. In 2020, the EU Commission staff stated: “[N]o systematic identification, of (unintentional) real-life priority mixtures is currently performed. The regulatory risk assessment or risk management of real-life unintentional mixtures is accordingly performed only rarely and on an ad hoc basis, while available scientific case studies are isolated examples.” EU Commission Staff Progress Report on the Assessment and Management of Combined Exposures to Multiple Chemicals (Chemical Mixtures) and Associated Risks, at 24, SWD (2020) 250 final (Oct. 14, 2020). Moreover, no one agency has a comprehensive view of the chemicals encountered by a population of interest. So, to develop a comprehensive picture of all the chemicals a population of interest is exposed to from food, drinking water, product use, and the environment requires a significant level of interagency coordination. Completing a single cumulative chemical impact assessment can take years, and agencies rarely have the resources to engage in multiyear collaborative efforts.
Insufficient data are likely the biggest challenge to broader implementation of cumulative chemical impact assessments. Toxicity data are needed for each chemical included in a component-based assessment and often from multiple animal species and exposure routes. If multiple health effects are considered in an assessment, data are needed for each of the health effects. To choose the correct interaction model when conducting component-based assessments, data on the mode of action of each chemical to produce a given health effect and data about the synergistic and antagonistic effects of particular chemical mixtures are critical. When considering exposure pathways to chemicals found in the environment, researchers seek information regarding the types and numbers of sources of release to the environment, whether uses are widespread or dispersive, the location of the sources compared to critical resources (such as a drinking water aquifer or sensitive communities), waste disposal methods, existing concentrations of the chemicals in the environment, and whether or not the chemicals occur naturally in the environment. When considering exposure pathways from the use of products, important information includes the use pattern of the product, the duration of the exposure, the concentration of the chemicals in the products, other products used simultaneously or within a brief time interval, the percentage of the population of interest that uses the product, and the routes of exposure (such as inhalation, skin contact, and oral exposure). Collecting the existing data and developing new data for a higher-tier cumulative chemical impact assessment can consume a prohibitive amount of resources.
To address the intense need for data, researchers are leveraging existing databases and creating new ones. Some examples are the OECD eChemPortal, European Chemical Agency databases for REACH and biocides, European Food Safety Authority data warehouse, EU Commission Information Platform for Chemical Monitoring, U.S. EPA Integrated Risk Information System, and U.S. Geological Survey monitoring of pesticides in surface and groundwater. Significant gaps still remain, however, such as data on impurities and by-products and environmental monitoring data at local levels. Additionally, there are challenges for agencies in relying on data that the agency did not develop, including questions about whether the data were developed with methods consistent with agency protocols and legal requirements, whether the programs for collecting the data will continue to be supported in the future, and whether the agency will continue to have access to the data over time.
Completing a single cumulative chemical impact assessment can take years, and agencies rarely have the resources to engage in multiyear collaborative efforts.
When toxicological data are unavailable or limited for chemicals that are the subject of the cumulative chemical risk assessment, researchers may create small groups of chemicals based on similar structures and physical properties. Then, using the known toxicological information from some of the chemicals in the group, “quantitative structure–activity relationship” (QSAR) models are used to predict the effects of the chemicals of interest. QSAR models reduce the need for animal testing, and many agencies around the world use them. Deciding which chemicals can be grouped together and which data can be “read across” from one chemical to another involves significant judgment and can be the subject of scientific debate.
Developing broadly accepted modeling tools is essential to increasing the use of cumulative chemical impact assessment. The Dutch National Institute of Public Health and the Environment is leading the EuroMix project, which aims to develop an experimentally verified, tiered test strategy for mixtures of multiple chemicals.
Cumulative chemical impact assessments include many assumptions and variables. It is important that these choices and uncertainties be documented and communicated to encourage one agency to rely on the work done by another and for the public to have confidence in the results. To meet transparency requirements and to aid assessors in communicating uncertainties, the European Food Safety Authority (EFSA) published guidance on uncertainty analysis. EFSA Sci. Comm., Guidance on Uncertainty Analysis in Scientific Assessments, 16 EFSA J. 5123 (2018). This guidance describes uncertainty analysis as the process of identifying and characterizing uncertainty about questions of interest and/or quantities of interest in a scientific assessment. Uncertainties can be associated with assessment inputs or with assessment methodology. The guidance also encourages quantitative reporting of the effect of uncertainties whenever possible and suggests that expert assistance may be required from statisticians and other professionals to conduct the analysis. A rigorous uncertainty analysis can identify priority areas for further research. As prioritization is a significant challenge in conducting fit for purpose cumulative chemical impact assessments, it is expected that the use of uncertainty analysis will increase.
Environmental justice and green chemistry are two important trends asking questions about the effect of cumulative chemical exposures. Cumulative chemical impact assessment is a powerful method that can address these concerns. Policy makers wishing to expand the use of these assessments are likely to amend laws to explicitly require that these assessments be conducted. To manage the virtually infinite number of potentially different chemical mixtures and the increased resource demand, international efforts to develop shared data platforms and widely accepted modeling tools will continue.
With mounting public concern about the risk of combined chemical effects and in the face of insufficient data to inform systematic cumulative impact assessment, policy makers are also considering alternative regulatory approaches. For example, under the upcoming revision of EU regulation for the Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH), the European Commission is proposing to introduce a “mixture assessment factor” (MAF). REACH currently requires that the safety of substances be assessed prior to placing them on the market, but the assessment does not need to consider the possibility of co-exposure to other substances. Registrants of substances typically do not have information on the presence or use of other substances that could result in a cumulative exposure scenario. The introduction of MAF is being proposed as a pragmatic approach to managing this absence of information and the potential of an unknown, unintentional cumulative exposure. In practice, MAF is a factor by which the safety threshold of a given substance is reduced (by dividing the safety threshold by the MAF) to ensure a level of protection against unintended mixture effects that is similar to the level of protection for a single-substance assessment. Whether the revision of REACH sets a single MAF to be used for all mixtures of chemicals or multiple MAFs for different mixtures of chemicals remains to be determined. However, different MAF values could apply to different chemicals and different exposure scenarios.
To fill data gaps, crowdsourcing and citizen science will likely play an increasing role, especially when environmental data are required at a local level. In the United States, CitzenScience.gov is designed to accelerate the use of crowdsourcing and citizen science across federal agencies. The site provides access to a catalog of hundreds of federally supported citizen science projects. Funded by Horizon 2020, the EU has launched the EU-Citizen.Science platform for sharing knowledge tools training and resources for citizen science.
At the agency level, the development of additional capabilities in interagency collaboration, community engagement, uncertainty analyses, and communication of risk and uncertainty to diverse stakeholders will be important to support broader application of these assessments.
