The U.S. Environmental Protection Agency (EPA) has been working on its Integrated Risk Information System (IRIS) assessment of inorganic arsenic (iAs) for many years, and recently released its plans for completing it in the Updated Problem Formulation and Protocol for the Inorganic Arsenic IRIS Assessment (EPA 2019a). Much of the update was made in response to recommendations in a 2013 report made by a previous committee of the National Academies of Sciences, Engineering, and Medicine (the National Academies). At the request of EPA, the National Academies convened another ad hoc committee to evaluate whether the various elements of the IRIS iAs assessment plan are appropriate to synthesize the scientific evidence and quantitate estimates of iAs toxicity. Specific areas the committee was asked to comment on included the prioritization of health outcomes; the systematic review methods; consideration of early life exposures; and dose-response analysis techniques (e.g., consideration of mode-of-action information, a hierarchical, Bayesian meta-analysis approach, conversion of exposure metrics to a common intake value, and modeling approaches).
Overall, 10 of the 11 committee members found that the approaches outlined in EPA’s IRIS iAs assessment plan and supplemental materials address many of the recommendations made by the 2013 committee. EPA has implemented systematic review approaches to evaluate the scientific literature and has appropriately prioritized the health outcomes for dose-response assessment. EPA has begun to implement several analytical approaches to utilize human data in the low-dose range, and has introduced more advanced statistical methods in its dose-response evaluations that take into consideration human data from multiple studies. Some of these techniques are novel for the IRIS program and were introduced because the database of epidemiological studies with exposures at lower doses has grown since 2013, making human study–based estimates for the dose-response curve in the lower range feasible. The committee was unable to reach consensus about the adequacy of these dose-response methods; the majority of the committee supported EPA’s approach and one member objected to this conclusion (his arguments are presented in a dissenting statement in Appendix A, along with the committee rebuttal). The bullets below highlight the committee’s major findings and recommendations. EPA’s efforts should be directed toward ensuring that the following issues are addressed in completing the iAs IRIS assessment, as opposed to revising the assessment plan.
- Transparency and communication: The committee found that the IRIS iAs assessment plan alone did not provide sufficient detail on all aspects of how the assessment would be performed, but many of the committee’s questions and concerns were addressed in its interactions with EPA during a public meeting. EPA indicated that more detail about the methods will be available in other documents that it is preparing for publication. These materials should be in the public domain as soon as possible or otherwise included in the documentation when the iAs IRIS assessment is completed.
- Prioritizing health outcomes: The committee agrees with EPA’s approach to prioritizing the health outcomes for further analysis as those that have robust or moderate evidence of a causal association. Recommendations for streamlining the presentation of the prioritization scheme are made.
- Systematic review methods: The committee commends EPA’s use of systematic review methods to identify evidence, evaluate evidence for hazard identification, and for subsequent use in dose-response analyses. Certain standard elements of systematic reviews are missing from the IRIS iAs assessment plan, but adding more contextual information and detail about the process would alleviate these concerns.
- Early life exposures: The committee evaluated EPA’s handling of this issue in terms of (1) the role in children’s health end points (e.g., neurodevelopmental toxicity and birth outcomes) and (2) the role of exposure timing in relation to health outcomes later in life. On the first issue, the committee noted that the majority of the childhood studies rely on continuous health outcome measures, and an approach to handle these types of data in a dose-response meta-analysis is still in development