1. Casarett & Doull’s Toxicology: The Basic Science of Poisons (Curtis D. Klaassen ed., 9th ed. 2019) [hereinafter Casarett & Doull’s].

2. Elaine M. Faustman, Risk Assessment, in Casarett & Doull’s, supra note 1, at ch. 4.

3. For a discussion of more modern formulations of this principle, which was articulated by Paracelsus in the sixteenth century, see David L. Eaton, Scientific Judgment and Toxic Torts—A Primer in Toxicology for Judges and Lawyers, 12 J.L. & Pol’y 5, 15 (2003); Lauren M. Aleksunes & David L. Eaton, Principles of Toxicology, in Casarett & Doull’s, supra note 1, at ch. 2. A short review of the field of toxicology can be found in Curtis D. Klaassen, Principles of Toxicology and Treatment of Poisoning, in Goodman and Gilman’s The Pharmacological Basis of Therapeutics 1739 (11th ed. 2008).

4. Some substances, such as central nervous system toxicants, can produce complex and nonspecific symptoms, such as headaches, nausea, and fatigue.

5. Forensic toxicology, a subset of toxicology generally concerned with criminal matters, is not addressed in this reference guide, because it is a highly specialized field with its own literature and methodologies that do not relate directly to toxic tort or regulatory issues.

6. In standard risk assessment terminology, hazard—the ability of a chemical to cause (a) specific type(s) of toxic effect(s)—is an intrinsic property of a chemical or physical agent, while risk—the likelihood/probability that the effect will occur in an individual or population under a given set of circumstances—is dependent both upon hazard and on the extent of exposure (both the concentration or amount and the duration/frequency of exposure).

including industrial, contract, and in-house laboratories.⁷ Known as good laboratory practices (GLPs), these codes govern many aspects of laboratory standards, including such details as the number of animals per cage, dose and chemical verification, and the handling of tissue specimens. GLPs are remarkably similar across agencies, but the tests called for differ depending on the mission. For example, there are major differences between the Food and Drug Administration’s (FDA) and the Environmental Protection Agency’s (EPA) required procedures for testing drugs and environmental chemicals.⁸ The FDA requires and specifies both efficacy and safety testing of drugs in humans and animals. Carefully controlled clinical trials using doses within the expected therapeutic range are required for premarket testing of drugs, because exposures to prescription drugs are carefully controlled and should not exceed specified ranges or uses.

Although new chemicals manufactured for use as food additives, pharmaceuticals, pesticides, and cosmetics have required substantial toxicity testing prior to regulatory approval since at least the 1950s, the vast majority of industrial chemicals had little, if any, premarket toxicity testing until the mid-1970s. In 1976, Congress established new legislation, called the Toxic Substances Control Act (TSCA), to require chemical manufacturers of new chemical entities (chemicals not on an established list of the then approximately 60,000 chemicals in commerce) to submit premarket notifications that would require EPA evaluation and approval prior to the manufacture of significant quantities of the chemical. Although well intended, the legislation “lacked teeth” and over the years was widely viewed as ineffectual. In 2016, the TSCA was extensively revised under the Frank R. Lautenberg Chemical Safety for the 21st Century Act.⁹ This revised TSCA legislation (the Lautenberg Act) increases substantially the expectations for toxicity evaluation for existing chemicals, with a focus on those that the EPA believes pose the highest risk to human health and the environment, as well as giving the EPA more oversight prior to approval for the manufacturing of new chemicals. The EPA now provides extensive guidance to industry on the types of information and data

7. A dramatic case of fraud involving a pharmaceutical company reporting toxicology research to support the safety of a consumer product is described in Schueneman v. Arena Pharmaceuticals, Inc., 840 F.3d 698 (9th Cir. 2016). The defendants made positive public statements to investors about the safety of its weight loss drug, specifically indicating that the drug was not carcinogenic and referred to supporting “animal studies.” Id. at 700. However, it was discovered that the defendants were conducting a study in which the drug was given to lab rats, and the drug was in fact causing cancer in the rats. Id. at 701. The Ninth Circuit found that the plaintiff, one of the defendant’s investors, successfully showed that the defendants intentionally and deliberately did not disclose the negative findings of the rat study to investors. Id. at 707–08.

8. Comparison of FDA, EPA, OECD GLP (Apr. 7, 2015), https://www.fda.gov/inspections-compliance-enforcement-and-criminal-investigations/fda-bioresearch-monitoring-information/articles.

9. See Frank R. Lautenberg Chemical Safety for the 21st Century Act, Pub. L.114-182,130 Stat. 448 (2016), https://perma.cc/B4T2-9PV7.

required under the TSCA Premanufacture Notification guidelines.¹⁰ While the law and regulatory guidance often do not explicitly specify which of the numerous available toxicity testing procedures must be performed, they usually do provide a framework. Such frameworks, which may vary between different regulations, enable both industry and the EPA to assess the potential harm a new chemical may pose to people or the environment.¹¹ Much of the evaluation relies upon chemical structure and comparison with other similar chemicals for which substantial toxicological and environmental data may be available.

In a similar manner, the European Union Regulation on Registration, Evaluation, Authorization and Restriction of Chemicals (REACH) requires extensive testing of both new chemicals and chemicals already in commerce.¹² Moreover, because exposures are less predictable, doses usually are given in a wider range in animal tests for nonpharmaceutical agents.¹³

10. See EPA, Points to Consider When Preparing TSCA New Chemical Notifications (June 2018), https://perma.cc/H9G5-7SAT.

11. The 2016 revisions of the TSCA set forth five possible EPA determinations on new chemical notices:

• The chemical or significant new use presents an unreasonable risk of injury to health or the environment;
• Available information is insufficient to allow the EPA to make a reasoned evaluation of the health and environmental effects associated with the chemical or significant new use;
• In the absence of sufficient information, the chemical or significant new use may present an unreasonable risk of injury to health or the environment;
• The chemical is or will be produced in substantial quantities and either enters or may enter the environment in substantial quantities or there is or may be significant or substantial exposure to the chemical; or
• The chemical or significant new use is not likely to present an unreasonable risk of injury to health or the environment.¹⁴

The lack of toxicity data for new chemicals (or new uses of existing chemicals) introduced into commerce has led the EPA to propose methods of evaluation using (1) in vitro toxicity pathway testing, (2) computational toxicology approaches using structural analogs for which toxicity information is available, (3) followed by whole-animal testing where warranted. See EPA’s Review Process for New Chemicals, https://perma.cc/U65D-WANT.

12. REACH is a regulation of the European Union, adopted in 2007 to “improve the protection of human health and the environment from the risks that can be posed by chemicals, while enhancing the competitiveness of the EU chemicals industry. It also promotes alternative methods for the hazard assessment of substances in order to reduce the number of tests on animals.” See European Chemicals Agency, Understanding REACH, https://perma.cc/AC3M-UTYF.
The legislation was updated in 2022. See European Chemicals Agency, Upcoming Changes to REACH Information Requirements, https://perma.cc/A6WV-L4GZ.

13. The development of a new drug inherently requires searching for an agent that at useful doses has a biological effect (e.g., decreasing blood pressure), whereas those developing a new chemical for consumer use (e.g., a house paint) hope that at usual doses no biological effects will occur. There are other compounds, such as pesticides and antibacterial agents, for which a biological effect is desired, but it is intended that at usual doses humans will not be affected. These different expectations

are part of the rationale for the differences in testing information available for assessing toxicological effects. Under FDA rules, approval of a new drug usually will require extensive animal and human testing, including a randomized double-blind clinical trial for efficacy and toxicity. In contrast, under the TSCA, the only requirement before a new chemical can be marketed is that a premanufacturing notice be filed with the EPA, including any toxicity data in the company’s possession.

14. See EPA, Regulatory Determinations made under Section 5 of the Toxic Substances Control Act (TSCA), https://perma.cc/6FNB-R74U

15. For a comparison of European REACH legislation and the U.S. EPA’s 2016 TSCA provisions, see Ágnes Botos, John D. Graham & Zoltán Illés, Industrial Chemical Regulation in the European Union and the United States: a Comparison of REACH and the Amended TSCA, 22 J. Risk Rsch. 1187, https://doi.org/10.1080/13669877.2018.1454495. For an overview of current trends in toxicity testing, see Ida Fischer, Catherine Milton & Heather Wallace, Toxicity Testing Is Evolving!, 9 Toxicology Rsch. 67 (2020), https://doi.org/10.1093/toxres/tfaa011; see also Daniel Krewski et al., Toxicity Testing in the 21st Century: Progress in the Past Decade and Future Perspectives, 94 Archives Toxicology 1 (2020), https://doi.org/10.1007/s00204-019-02613-4.

16. Nat’l Rsch. Council, Science and Decisions: Advancing Risk Assessment (2009). See also In re Johnson & Johnson Talcum Powder Prods. Mktg., Sales Pracs. & Prods. Litig., 509 F. Supp. 3d 116 (D.N.J. 2020) (expert opinion that talcum powder may cause ovarian cancer based on risk assessment and analysis of Bradford Hill factors considerations is admissible). Recently, a National Academy of Sciences panel has discussed potential approaches using systematic review to improve the EPA’s risk paradigm. See Nat’l Acads. of Scis., Eng’g & Med., The Use of Systematic Review in EPA’s Toxic Substances Control Act Risk Evaluations (2021).

public from adverse effects.¹⁷ Acceptable risk levels—for example, 1 in 1,000 to 1 in 1,000,000—are usually well below what can be measured through epidemiological study. Inevitably, this means that risk assessment is usually based solely on toxicological data—or, if epidemiological findings of an adverse effect are observed, then toxicological reasoning must be used to extrapolate to the appropriate lower exposure or dose standard aimed at protecting the public.

The four-part paradigm of risk assessment is heavily based on toxicological precepts. First, hazard identification, which is roughly equivalent to the legal term general causation, reflects the toxicological rule of specificity of effects, and the second step, dose–response assessment, is based upon “the dose makes the poison” and is fundamental to specific causation. The hazard identification process often uses weight-of-evidence approaches in which the toxicological, mechanistic, and epidemiological data are rigorously assessed to form a judgment regarding the likelihood that the agent produces a specific effect. Second, establishing the appropriate dose–response curve, threshold, or “one-hit” model is an exercise in toxicological reasoning. Even for those chemicals known to be carcinogens, a threshold model is appropriate if the toxicological mechanism of action can be demonstrated to depend upon a threshold. The third step, exposure assessment, requires knowledge of specific toxicological dynamics; for example, the impact on the lung of an air pollutant varies by factors such as: inhalation rate per unit body mass, which is affected by exercise and by age; the size of a particle or the solubility of a gas, either of which will affect the depth of penetration into the more sensitive parts of the airways; the competence of the usual airway defense mechanisms, such as mucus flow and macrophage function; and the ability of the lung to metabolize the agent.¹⁸ The fourth and final step is the characterization of risk.¹⁹

Risk assessment is not an exact science. It should be viewed as a useful framework to organize and synthesize information and to provide estimates on which policy making can be based. In recent years, codification of the methodology used to assess risk has increased confidence that the process can be reasonably free of bias; however, significant controversy remains, particularly when actual data are limited and generally conservative default assumptions are used.

17. Pharmaceuticals intended for human use are an exception in that a tradeoff between desired and adverse effects may be acceptable, and human data are available prior to, and as a result of, the marketing of the agent.

18. Some toxic agents pass through the lung without producing any direct effects on this organ. For example, inhaled carbon monoxide produces its toxicity in essence by being treated by the body as if it is oxygen. Carbon monoxide readily combines with the oxygen-combining site of hemoglobin, the molecule in red blood cells that is responsible for transporting oxygen from the lung to the tissues. In this way, the effective transport and tissue utilization of oxygen is blocked.

19. See, e.g., Nat. Res. Defense Council v. U.S. EPA, 38 F.4th 34 (9th Cir. 2022) (according to the EPA’s Office of Research and Development, the EPA did not follow its 2005 Guidelines for Carcinogen Risk Assessment, which lay out four steps for conducting risk assessments of chemicals’ carcinogenic potential).

20. See Nat’l Rsch. Council & Comm. on Improving Risk Analysis Approaches Used by the U.S. EPA, Toward a Unified Approach to Dose–Response Assessment, in Science and Decisions: Advancing Risk Assessment 127 (2009), https://perma.cc/47RG-NKMG. See also Yates v. Ford Motor Co., No. 5:12-CV-752-FL, 2015 WL 2189774, at *17, *27 (E.D.N.C. May 11, 2015) (holding that an EPA pamphlet that includes regulatory risk assessment can be relied on by the plaintiff’s experts only to “support the reliability of studies the experts themselves rely upon”).

21. It is also claimed that standard risk assessment will underestimate true risks, particularly for sensitive populations exposed to multiple stressors, an issue of particular pertinence to discussions of environmental justice. Comm. on Env’t Just., Institute of Med., Toward Environmental Justice: Research, Education, and Health Policy Needs (1999), https://doi.org/10.17226/6034. In 2003, the EPA developed formal guidance for cumulative risk assessment, wherein it defined aggregate exposure as the “combined exposure of an individual (or defined population) to a specific agent or stressor via relevant routes, pathways, and sources.” EPA, Framework for Cumulative Risk Assessment (2003), https://perma.cc/D7CT-Z2RT. In 2022, the EPA released a draft update to this document, with an initial emphasis on identifying research needs. See EPA, Cumulative Impacts: Recommendations for ORD Research (2022), https://perma.cc/B6SL-EHUT.

22. A public health standard to protect against the lifetime risk of inhaling a known carcinogen will usually be based on lifetime exposure calculations of twenty-four hours a day, every day for seventy years. This is more than 25,000 days and 600,000 hours. Exceeding this standard for a few hours would presumably have little impact on cancer risk. In contrast, for a short-term standard set to avoid a threshold-based risk, exceeding the standard for this short time may make a major difference—for example, an asthma attack caused by being outdoors on a day that the ozone standard is exceeded. Regulatory statutes sometimes limit consideration of multiple pathways of exposure to the same chemical, potentially underestimating overall exposure and risk. See, e.g., Swati D. G. Rayasam et al., Toxic Substances Control Act (TSCA) Implementation: How the Amended Law Has Failed to Protect Vulnerable Populations from Toxic Chemicals in the United States, 56 Env’t Sci. & Tech. 11969 (2022), https://doi.org/10.1021/acs.est.2c02079. However, the interpretation of how such evaluations under the revised TSCA law are done by the EPA has been criticized. See Anthony C. Tweedale, Correspondence on “Toxic Substances Control Act (TSCA) Implementation: How the Amended Law has Failed to Protect Vulnerable Populations from Toxic Chemicals in the United States,” 56 Env’t Sci. & Tech. 16533 (2022), https://doi.org/10.1021/acs.est.2c06032.

23. See, e.g., Hardeman v. Monsanto Co., 997 F.3d 941 (9th Cir. 2021) (discussion of appropriate use in an expert opinion of the IARC’s classification of glyphosate as a “probable carcinogen”).

24. There are exceptions—for example, when measurements of levels in the blood or other body constituents of the potentially offending agent are at a high enough level to be consistent with reasonably specific health impacts, such as in carbon monoxide poisoning and lead poisoning.

25. Direct-acting toxic agents are those whose toxicity is due to the parent chemical entering the body. A change in chemical structure through metabolism usually results in detoxification. Indirect-acting chemicals are those that must first be metabolized to a harmful intermediate for toxicity to occur. For an overview of metabolism in toxicology, see David L. Eaton & Julia Cui, Biotransformation, in Patty’s Toxicology (James Klaunig ed., 7th ed. 2023), https://doi.org/10.1002/0471125474.tox122.

26. Grapefruit juices and certain other fruit juices at sufficient dose have the potential to inhibit the metabolism of a wide variety of drugs and chemicals. See Michael J. Hanley et al., The Effect of Grapefruit Juice on Drug Disposition, 7 Expert Op. Drug Metabolism & Toxicology 267 (2011), https://doi.org/10.1517/17425255.2011.553189; Meng Chen et al., Food-Drug Interactions

Precipitated by Fruit Juices Other than Grapefruit Juice: An Update Review, 26 J. Food & Drug Analysis S61 (2018), https://doi.org/10.1016/j.jfda.2018.01.009.

27. See Steve C. Gold et al., Reference Guide on Epidemiology, section titled “General Causation,” in this manual. See also Nat’l Acads. Sci., Eng’g & Med, Assessing Causality from a Multidisciplinary Evidence Base for National Ambient Air Quality Standards (2022), https://perma.cc/6L9G-NBCN (discussing the strengths and limitations of epidemiological studies in establishing “causation” between exposure to a hazardous chemical (in this case, various air pollutants) and the development of specific diseases). For example, in In re Nexium Esomeprazole, 662 F. App’x 528 (9th Cir. 2016), testimony on general causation was excluded as unreliable in which the expert did not adequately explain how he inferred a causal relationship from epidemiological studies that did not come to the same conclusions. However, epidemiological studies are not always necessary. McMunn v. Babcock & Wilcox Power Generation Grp., Inc., No. 2:10CV143, 2014 WL 814878, at *15 (W.D. Pa. Feb. 27, 2014) (citing Glastetter v. Novartis Pharms. Corp., 252 F.3d 986, 992 (8th Cir. 2001); Rider v. Sandoz Pharms. Corp., 295 F.3d 1194, 1198 (11th Cir. 2002)).

artificial distinctions being drawn between them. However, although courts generally rule epidemiological expert opinion admissible, the admissibility of toxicological expert opinion has been more controversial because of uncertainties regarding extrapolation from animal and in vitro data to humans. This has been particularly true in cases in which relevant epidemiological research data exist. However, the methodological weaknesses of some epidemiological studies, including their inability to accurately measure exposure, the often small numbers of subjects, and issues such as recall bias when individuals with specific diseases are asked to recall their past exposure to chemicals in comparison with a control population, often render these studies difficult to interpret. As detailed in the Reference Guide on Epidemiology, an epidemiological study by itself can only show an association, which may or may not be causal.

The Role of Toxicology in Establishing Causation

The most well-known guidelines for considering causality of a specific effect by a specific cause are the Bradford Hill considerations.²⁸ Austin Bradford Hill was a British epidemiologist who contributed greatly to studies linking cigarette smoking to lung cancer. As discussed in detail in the Reference Guide on Epidemiology, Hill described nine points of consideration to facilitate the determination of a causal association seen in an epidemiology study. A modified list of the Bradford Hill considerations that inform epidemiologists in making judgments about causation includes:

1. Temporal relationship
2. Strength of the association
3. Dose–response relationship
4. Replication of the findings
5. Biological plausibility
6. Consideration of alternative explanations
7. Cessation of exposure
8. Specificity of the association
9. Consistency with other knowledge

28. Austin Bradford Hill, The Environment and Disease: Association or Causation?, 108 J. Royal Soc’y Med. 32 (2015) (reprint of 1965 article), https://doi.org/10.1177/0141076814562718. For a discussion of the Bradford Hill considerations, see Steve C. Gold et al., Reference Guide on Epidemiology, section titled “General Causation,” in this manual. Although the points raised by Bradford Hill in his 1965 article are often referred to as “criteria,” the author made it clear that none were absolutes. As discussed in the Reference Guide on Epidemiology in this manual, the term “consideration,” rather than “criteria” or even “guidelines,” is perhaps more appropriate to represent these ideas.

toxicological data where relevant. These were discontinued as no longer necessary because many other organizations had developed similar approaches.²⁹

A more recent example of a common approach to developing a systematic review process for evaluating the medical science literature has been that of the Cochrane reviews.³⁰ These are systematic approaches that analyze the existing literature and focus almost exclusively on epidemiological findings, primarily related to diagnosis and treatment of disease. In part building on the Cochrane review model, systematic reviews have been suggested for toxicology³¹ and for public health.³²

In contrast to epidemiology, because animal and cell studies permit researchers to isolate the effects of exposure to a single chemical or to mixtures, toxicological findings offer unique information concerning dose–response relationships, mechanisms of action, specificity of response, and other information relevant to the assessment of causation—although these can present issues related to extrapolation to humans.³³

The gold standard in clinical epidemiology and in the testing of pharmaceutical agents is the randomized double-blind placebo control study in which the control and intervention groups are usually well matched. Although appropriate and very informative for the testing of pharmaceutical agents, it is generally unethical for chemicals used for other purposes. The randomized control design in essence is what is used in a classic toxicological study in laboratory animals,

29. For a review of consensus approaches in general and the importance of unbiased selection of members of review committees that reflect the broad disciplines involved, see Bernard D. Goldstein, What the Trump Administration Taught Us About the Vulnerabilities of EPA’s Science-Based Regulatory Processes: Changing the Consensus Processes of Science into the Confrontational Processes of Law, 31 Health Matrix 299 (2021), https://perma.cc/YJ3H-TPPR. For a discussion of general issues regarding medical practice and testimony, see John B. Wong et al., Reference Guide on Medical Testimony, in this manual.

30. Cochrane Library, About Cochrane Reviews, https://perma.cc/UXR9-4HPL.

31. Lena Smirnova et al., 3S—Systematic, Systemic, and Systems Biology and Toxicology, 35 ALTEX 139 (2018) https://doi.org/10.14573/altex.1804051. See also Thomas Hartung et al., Systems Toxicology II: A Special Issue, 30 Chem. Rsch. Toxicology 869 (2017), https://doi.org/10.1021/acs.chemrestox.7b00038.

32. See also Steve C. Gold et al., Reference Guide on Epidemiology, section titled “Methods for Synthesizing or Combining the Results of Multiple Studies,” in this manual.

33. Courts have found animal studies to be relevant evidence of causation where there is a sound basis for extrapolating conclusions from those studies to humans in real-world conditions. See Hardeman v. Monsanto Co., 997 F.3d 941 (9th Cir. 2021); see also Milana v. Eisai, Inc., No. 8:21-cv-831-CEH-AEP, 2022 WL 846933 (M.D. Fla. Mar. 22, 2022) (increase in tumors in mammary tissue of experimental animals compared to higher rates of cancer in patients taking weight loss drug); In re Levaquin Prods. Liab. Litig., No. 08–1943 (JRT), 2010 WL 8400514, at *2–4, *8 (D. Minn. Nov. 8, 2010); Brandeis Univ. v. Keebler Co., No. 1:12–cv–01508, 2013 WL 5911233 (N.D. Ill. Jan. 18, 2013); In re Abilify (Aripiprazole) Prods. Liab. Litig., 299 F. Supp. 3d 1291, 1310 (N.D. Fla. 2018).

although matching is more readily achieved because the animals are bred to be genetically similar and have identical environmental histories.

Dose issues are at the intersection between toxicology and epidemiology. Many epidemiological studies of the potential risk of chemicals do not have direct information about dose, although qualitative differences among subgroups or in comparison with other studies can be inferred. The epidemiology database includes many studies that are probing for the potential for an association between a cause and an effect. Thus, a study asking all those suffering from a specific disease a multiplicity of questions related to potential exposures is bound to find some statistical association between the disease and one or more exposure conditions due to chance alone. Such studies generate hypotheses that can then be evaluated by subsequent studies that more narrowly focus on the potential cause-and-effect relationship. One way to evaluate the strength of the association is to assess whether those epidemiological studies evaluating cohorts with relatively high exposure support the association.³⁴

The requirement in certain jurisdictions for epidemiological evidence of a relative risk greater than two (RR > 2) for general causation also has limited the utilization of toxicological evidence.³⁵ A firm legal requirement for such evidence means that if the epidemiological database only showed statistically significant evidence that cohorts exposed to ten parts per million of an agent for twenty years produced an 80% increase in risk, the court could not hear the case of a plaintiff alleging that exposure to fifty parts per million for twenty years of the same agent caused the adverse outcome. Yet to a toxicologist, there would be little question that exposure to the fivefold-higher dose of an agent producing an 80% increase in risk would lead to more than a 100% increase in risk, i.e., more than doubling of the risk.

34. As an example of considering dose and exposure issues across epidemiological studies, see Luoping Zhang et al., Formaldehyde Exposure and Leukemia: A New Meta-Analysis and Potential Mechanisms, 681 Mutation Rsch. 150 (2008), https://doi.org/10.1016/j.mrrev.2008.07.002. The subject of the strength of an epidemiological association and its relation to causality is considered in Steve C. Gold et al., Reference Guide on Epidemiology, in this manual; see also note 79, infra.

35. The basis for the use of RR > 2 is the translation of the preponderance of evidence rule, or “more likely than not” legal standard, into the epidemiological concept of at least a doubling of risk. An example is the Havner rule in Texas, which for general causation requires that there be reliable epidemiological studies with a statistically significant RR > 2 associating a putative cause with an effect. See Bostic v. Ga. Pac. Corp., 439 S.W.3d 332 (Tex. 2014). But see Hardeman v Monsanto Co., 997 F.3d 941, 966 (9th Cir. 2021) (“[W]e have never suggested that a hardline increase in a risk statistic, or even an adjusted odds ratio above 2.0, is necessary for finding a strong association. To the contrary, flexibility is warranted considering the contextual nature of the Daubert inquiry.”) (citations omitted). For a discussion of the use by courts in various jurisdictions of relative risk > 2 for general and specific causation, see Russellyn S. Carruth & Bernard D. Goldstein, Relative Risk Greater Than Two in Proof of Causation in Toxic Tort Litigation, 41 Jurimetrics 195 (2001); for the toxicological issues, see Bernard D. Goldstein, Toxic Torts: The Devil Is in the Dose, 16 J.L. & Pol’y 551–85 (2008).

36. See, for example, Nat’l Toxicology Prog., 15th Rep. on Carcinogens (2021), https://perma.cc/V4DP-3YNZ, for a discussion of how epidemiology and animal toxicity studies are used to assess the potential cancer risk of different chemicals.

37. See, e.g., Int’l Agency for Rsch. on Cancer, IARC Monographs on the Identification of Carcinogenic Hazards to Humans, https://perma.cc/R56M-EQMM; Nat’l Toxicology Prog., supra note 36; Waite v. AII Acquisition Corp., 194 F. Supp. 3d 1298 (S.D. Fla. 2016) (court remains “unconvinced” that lack of statistically significant epidemiological studies properly overcomes all other types of scientific evidence including animal, cellular, and molecular studies). The absence of epidemiological data is due, in part, to the difficulties in conducting cancer epidemiology studies, including the lack of suitably large groups of individuals exposed for a sufficient period of time, long latency periods between exposure and manifestation of disease, the high variability in the background incidence of many cancers in the general population, and the inability to measure actual exposure levels. See also Parker v. Mobil Oil Corp., 857 N.E.2d 1114 (N.Y. 2006), in which New York’s highest court reviewed an appeal from a summary judgment of a lower court that had dismissed a plaintiff’s case because of the lack of a specific exposure estimate by the plaintiff’s experts. The plaintiff’s experts claimed that unusual work practices during his seventeen years employed at a gasoline station led to higher exposure to benzene, a known cause of acute myelogenous leukemia (AML) from which the patient suffered, than would have occurred in the work practices of a usual gasoline station worker. The higher court found for the plaintiff on the lack of a need for a specific exposure assessment, citing other cases finding that an expert does not need to establish the amount of exposure the plaintiff experienced. Id. at 1121 (citing McClain v. Metabolife Int’l, Inc., 401 F.3d 1233, 1241 n.6 (11th Cir. 2005)). However, the same court ruled for the defense based on the absence of epidemiological evidence that gasoline caused AML, despite the fact that benzene is a component of the gasoline mixture. Court preference for epidemiology is also evident in the Havner rule (see supra note 35). The requirement for epidemiological evidence of a doubling of risk precludes toxicological reasoning related to dose. For example, suppose there were two epidemiological studies showing a statistically significant 80% increase of an adverse endpoint in large workforces in well-regulated industries that were exposed to ten parts per million (ppm) of a chemical for thirty years. Suppose also that there was a plaintiff with the same endpoint whose experts presented evidence that the plaintiff was exposed to fifty ppm for thirty years to the same chemical. Based on dose considerations, a toxicologist could argue that the plaintiff’s level of exposure would exceed 100%, i.e., more than double the plaintiff’s risk. But under a strict interpretation of the Havner rule, the plaintiff’s case would be dismissed.

studies in which animals deliberately are exposed to the chemical. Such deliberate exposure is not possible in humans. As a general rule, unequivocally positive epidemiological studies reflect prior workplace practices that led to relatively high levels of chemical exposure for a limited number of individuals and that, fortunately, in most cases no longer occur now. Thus, an additional prospective epidemiological study often is not possible, and even the ability to do retrospective studies is constrained by the passage of time.

In essence, epidemiological findings consistently demonstrating an adverse effect in humans of a manufactured chemical represent a failure of toxicology as a preventive science. A corollary of the tenet that, depending upon dose, all chemical and physical agents are harmful is that society depends upon toxicological science to discover and anticipate these harmful effects, and we also rely on regulators and other responsible parties to prevent human exposure to a harmful level or to ensure that the agent is not manufactured and distributed. In this regard epidemiology is a valuable backup approach that functions to detect failures of primary prevention. The two disciplines complement each other, particularly when the approaches are iterative.

Toxicologists may use epidemiological findings in their expert opinions concerning general causation and specific causation. This is particularly true for well-studied compounds that have been thoroughly evaluated by respected organizations, such as the International Agency for Research on Cancer (IARC, affiliated with the World Health Organization (WHO)) or the National Toxicology Program (NTP, affiliated with the U.S. Department of Health and Human Services), which include toxicologists and epidemiologists in their review of the evidence. For example, in mortality studies of large cohorts of workers exposed to high levels of a carcinogenic agent in which all types of cancer are reported, but only an increase in brain cancer is noted, the toxicologist may consider this information in determining if a pancreatic cancer was caused by exposure to a lesser dose of the agent. In terms of assessing specific causation for exposure to this known carcinogen, there may be the equivalent of a No Observed Adverse Effect Level (NOAEL; see section titled “Toxicological Study Design” below) in the reviewed and accepted epidemiological data in which an increase in brain cancer risk is seen only in the more highly exposed populations. Whether a toxicologist is qualified to offer an opinion based on both toxicology and epidemiology data depends on the particular background of the proffered expert and the legal standard applied in the jurisdiction.

Toxicology and Exposure Science

In recent decades, exposure assessment has developed into a scientific field (exposure science) with journals, learned societies, and research funding processes.

Exposure science methodologies include mathematical models used to predict exposure resulting from an emission source, which might be a long distance upwind; chemical or physical measurements of media such as air, food, and water; and biological monitoring within humans, including measurements of blood and urine specimens if available, for the chemical of concern and/or its metabolites. Extrapolation back to the level of external exposure depends upon toxicokinetic analyses (see section titled “Safety and Risk Assessment” above). In this continuum of exposure metrics, the closer to the human body, the greater the overlap with toxicology.³⁸ An exposure assessment should also look for competing exposures from other sources of the chemical of concern.

Exposure assessment is central to epidemiology as well and is often the limiting factor in using an epidemiological study to infer causation. Many of the causal associations between chemicals and human disease have been developed from epidemiological studies relating a workplace chemical to an increased risk of the specific disease in cohorts of workers, often with only a qualitative assessment of exposure. More sophisticated modern approaches to chemical exposures at complex workplaces include development of a job exposure matrix based on specific exposure levels throughout the working lifetime. An improved quantitative understanding of such exposures enhances the likelihood of observing causal relations.³⁹ It also can support the opinion of the expert toxicologist on the likelihood that a specific exposure was responsible for an adverse outcome.

38. Toxicologists also have indirect means of approaching exposure through symptoms— e.g., for many agents, there is a known threshold for smell and a reasonable range of levels that might cause symptoms. For example, the use of toxicological expertise is appropriate in a situation in which chronic exposure to a volatile hydrocarbon is alleged to have occurred at levels at which acute exposure would be expected to render the individual unconscious. Toxicologists may also contribute knowledge of the extent of individual exposure based upon appropriate assumptions concerning inhalation rate or water use; for example, children inhale more per body mass than do adults, and outdoor workers in hot climates will drink more fluids than those in cool climates. For a discussion of general issues regarding assessment of exposure of toxic substances, see M. Elizabeth Marder & Joseph V. Rodricks, Reference Guide on Exposure Science and Exposure Assessment, in this manual.

39. In terms of general causation, accurate exposure assessment is important because a true effect can be missed owing to confounding caused by cohorts that often include workers with little exposure to the putative offending agent, thereby diluting the actual effect. See Peter F. Infante, Benzene Exposure and Multiple Myeloma: A Detailed Meta-analysis of Benzene Cohort Studies, 1076 Annals N.Y. Acad. Scis. 90 (2006), https://doi.org/10.1196/annals.1371.081, for a discussion of this issue in relation to a meta-analysis of the potential causative role of benzene in multiple myeloma. On the other hand, an association between exposure and effect occurring solely by chance is more likely if the effect does not meet the expected standard of being more pronounced in those receiving the highest dose. See Goldstein, supra note 35. Setting regulatory standards based on the observed effect in a cohort often requires a risk assessment, which in turn is dependent on understanding the extent of the exposure. This has led to extensive retrospective reconstruction of exposure in key cohorts.

40. See Nat’l Rsch. Council, Human Biomonitoring for Environmental Chemicals (2006), https://perma.cc/9AQ2-QXUQ; see also Lesa L. Aylward, Integration of Biomonitoring Data into Risk Assessment, 9 Current Op. Toxicology 14 (2018), https://doi.org/10.1016/j.cotox.2018.05.001.

41. Sometimes also referred to in the popular press as “forever chemicals.” For an EPA summary of research on PFAS, see https://perma.cc/JLV4-54XX.

have half-lives measured in years, and thus slow accumulation from environmental exposures can occur over many years. For these types of chemicals, blood concentrations are indicative of lifetime exposures—not just recent exposure—and can be useful to establish whether an unusual exposure to a specific agent has occurred in the past.

The Centers for Disease Control and Prevention’s (CDC) National Health and Nutrition Examination Survey (NHANES) monitors a broad level of biological markers of health—e.g., blood cholesterol levels—on a U.S. population sample. This survey now includes measurements of a wide variety of environmental and dietary substances. For environmental agents, the CDC considers blood levels above the statistical ninety-fifth percentile level in the NHANES samples as an indication of an “unusual exposure,” although it does not state whether exposures above, or below, the ninety-fifth percentile are necessarily harmful or safe, respectively.⁴² However, other regulatory agencies might use such levels to indicate that an exposure at that level is associated with an unacceptable risk.

Biomarkers of Effect

Research on certain toxic chemicals has established the presence of biological changes indicative of the effects of that specific chemical and, thus, can also provide evidence of exposure, at least within the time period that the biological changes remain within the body. For example, when carbon monoxide enters the bloodstream, it binds to hemoglobin, the molecule responsible for delivering oxygen from the lungs to the rest of the body. Once bound, it prevents oxygen from binding, which results in oxygen deprivation that can be rapidly fatal. For sublethal doses of carbon monoxide, the time to full equilibrium with blood hemoglobin is eight to twelve hours, depending largely on breathing rate. The measurement of the amount of this carboxyhemoglobin in blood is therefore both a biomarker of exposure—reflecting carbon monoxide exposure during the past eight to twelve hours—as well as a biomarker of effect, as it is on the direct pathway to toxicity. Carbon monoxide can also be measured in expired air, in which case it primarily reflects exposure. Another example of a biomarker of exposure and effect is provided by a broad class of insecticides known as organophosphorus chemicals (OPs). As a chemical class, OPs bind to and inhibit an important

42. For more details on the NHANES program, see https://perma.cc/UY4L-S8TY. One objective of the NHANES program was to “establish reference ranges that physicians and scientists can use to determine whether a person or group has an unusually high exposure. . . . The 95th percentile is helpful when determining whether levels observed in separate public health investigations or other studies are unusual.” CDC, Fourth National Report on Human Exposure to Environmental Chemicals (Volume Three: Analysis of Pooled Serum Samples for Select Chemicals, NHANES 2005–2016), https://perma.cc/E28W-X5HL.

enzyme in nerve function called cholinesterase and can be rapidly lethal. Cholinesterase activity unrelated to its role in the nervous system is also present in red blood cells and in plasma. Measurements of cholinesterase in the blood can thus be an effects-related biomarker of exposure to many OP pesticides. This test is used frequently in the occupational environment and clinically to diagnose OP poisoning. Where available, and when testing is done within the required time frames, such effects-related biomarkers provide information that is directly on the pathway between external exposure and internal effects. Although currently there are only a relatively few chemicals for which specific biomarkers are available, it is predicted that the increasing understanding of molecular effects of toxic substances will lead to more effects-related biomarkers of exposure.

Biomarkers of Susceptibility

Just as biological measurement of environmental exposures is an area of research interest, another area of research interest is DNA analyses to identify individuals who may be more susceptible to exposures, such as genetically determined variability in enzymes responsible for detoxifying chemicals of concern. Given the complexities of environmental and genetic interactions, much discussion in this area remains theoretical.

Toxicological Study Design

Toxicological studies usually involve exposing laboratory animals (in vivo research) or cells or tissues (in vitro research) to chemical or physical agents, monitoring the outcomes (such as cellular abnormalities, tissue damage, organ toxicity, or tumor formation), and comparing the outcomes with those for unexposed control groups. As explained below,⁴³ the extent to which animal and cell experiments accurately predict human responses to chemical exposures is subject to debate. However, because it is generally unethical to experiment on humans by exposing them to known doses of chemical agents, animal toxicological evidence often provides the best scientific information about the risk of disease from a chemical exposure.⁴⁴

43. See section titled “Extrapolation from Animal (In Vivo) and Cell (In Vitro) Research to Humans” below.

44. Clinical drug trials for therapeutic drugs are an example of experimental human exposure, where as a policy matter potential benefit is considered greater than the risk of adverse events. Notably, information from animal research is used to determine risk profiles, prior to human trials. The human trials are highly regulated, and subject to ethical requirements, including informed

consent. See, e.g., FDA, Clinical Trials Guidance Documents, https://www.fda.gov/science-research/clinical-trials-and-human-subject-protection/clinical-trials-guidance-documents.

45. However, it is from drug studies in which multiple animal species are compared directly with humans that many of the principles of toxicology have been developed.

46. See, e.g., OECD, Considerations for Assessing the Risks of Combined Exposure to Multiple Chemicals (2018), https://perma.cc/L48K-W79J.

47. See, e.g., Casarett & Doull’s, supra note 1.

48. Robert J. Kavlock et al., Accelerating the Pace of Chemical Risk Assessment, 31 Chem. Rsch. Toxicology 287 (2018), https://doi.org/10.1021/acs.chemrestox.7b00339.

49. See, e.g., Michael Dorato et al., The Toxicologic Assessment of Pharmaceutical and Biotechnology Products, in Hayes’ Principles and Methods of Toxicology 325 (A. Wallace Hayes & Claire L. Kruger eds., 6th ed. 2014), https://doi.org/10.1201/b17359.

50. The terms mechanism of action and mode of action are conceptually similar and are used to describe the specific molecular, biochemical, and cellular changes that are responsible for the observed toxic effects following the administration of a chemical. Although not identical, the two terms have similar meaning. Generally, mode of action is somewhat broader and refers to the generalized sequence of molecular events that lead to toxicity, whereas mechanism of action refers to a specific cellular, biochemical, or molecular target of the substance (e.g., a specific receptor).

including diligently searching for a dose that has no measurable adverse effect. This information is useful in understanding the mechanisms of toxicity and in extrapolating data from animals to humans.⁵¹

Acute Toxicity Testing and the Lethal Dose 50% (LD₅₀)

To determine the dose–response relationship for a compound, a short-term lethal dose 50% (LD₅₀) may be derived experimentally. The LD₅₀ is the dose at which a compound kills 50% of laboratory animals and usually is determined from a single exposure, although the observation may extend for a period of days to weeks. The use of this easily measured endpoint for acute toxicity largely has been replaced, in part because recent advances in toxicology have provided other pertinent endpoints, and in part because of animal welfare concerns that have focused on reduction, refinement, or replacement of the use of animals in laboratory research.⁵² Although determining the LD₅₀ was frequently the primary goal of acute toxicity testing, much additional useful information is obtained, such as the identification of the target organ(s) of effect, overall characterization of the symptoms, and progression of toxicity following single doses. Acute toxicity studies are also necessary to establish the dose range necessary for subsequent repeated-dose (subacute, usually fourteen days; and subchronic, usually ninety days) toxicity studies.

No Observed Adverse Effect Level (NOAEL)

A dose–response study also permits the determination of another important characteristic of the biological action of a chemical—the no observed adverse effect level (NOAEL; Figure 1).⁵³ The NOAEL sometimes is called a threshold, because it is the level above which observed adverse or toxic effects in test

51. See sections titled “Benchmark Dose (BMD),” “Linearized Non-Threshold (LNT) Model and Determination of Cancer Risk,” and “Maximum Tolerated Dose (MTD) and Chronic Toxicity Tests” below.

52. Dale M. Cooper et al., The Humane Use and Care of Laboratory Animals in Toxicology Research, in Hayes’ Principles and Methods of Toxicology 1023, supra note 49.

53. For example, undiluted acid on the skin can cause a horrible burn. As the acid is diluted to lower and lower concentrations, less and less of an effect occurs until there is a concentration sufficiently low (e.g., one drop in a bathtub of water, or a sample with less than the acidity of vinegar) that no effect occurs. This “no observed effect” concentration differs from person to person. For example, a baby’s skin is more sensitive than that of an adult, and skin that is irritated or broken responds to the effects of an acid at a lower concentration. However, the key point is that there is some concentration that is completely harmless to the skin.

animals are believed to occur and below which no toxicity is observed.⁵⁴ Of course, because the NOAEL is dependent on the ability to observe an effect, the level is sometimes lowered after more sophisticated methods for detecting effects are developed. Typically, NOAEL studies involve the administration of the test chemical at varying different doses, daily over a period of ninety days. At least three doses (plus the “zero dose” control) are used, with doses generally spanning a dose range of thirty- to one hundred-fold, with the highest dose causing clear evidence of toxicity, and the lowest dose showing no evidence of toxicity. Typically, rats are used, with at least ten males and ten females for each dose group. The ultimate goal is to establish a dose–response curve, with at least one dose that does not show any evidence of toxicity. In circumstances where the lowest dose tested resulted in an adverse effect, the terminology for that dose is the lowest observed adverse effect level (LOAEL), which may then be used in place of a NOAEL, with additional caveats (e.g., uncertainty factors) if used in risk assessment.

Benchmark Dose (BMD)

For regulatory toxicology, the NOAEL has been replaced by a more statistically robust approach known as the benchmark dose (BMD; Figure 1). The BMD is determined based on dose–response modeling and is defined as the exposure associated with a specified low incidence of risk, generally in the range of 1% to 10%, of a health effect, or the dose associated with a specified measure or change of a biological effect (thus, if the lower range on the BMD is set at a 10% response, it is referred to as a BMD₁₀). To model the BMD, sufficient data must exist, such as at least a statistically or biologically significant dose–related trend in the

54. The significance of the NOAEL was relied on in an action alleging various health issues caused by hazardous levels of wind-borne thallium dust that escaped a military base landfill and entered plaintiff’s living environment. Myers v. United States, No. 02cv1349–BEN, 2014 WL 6611398 (S.D. Cal. Nov. 20, 2014), aff’d, 673 F. App’x 749 (9th Cir. 2016). The court ruled in favor of defendants, in part from finding persuasive evidence that the low amount of thallium found in soil at the landfill was significantly below the NOAEL. 2014 WL 6611398, at *31. See also Env’t L. Found. v. Beech-Nut Nutrition Corp., 235 Cal. App. 4th 307, 311, 317–18 (2015), as modified on denial of reh’g (Apr. 16, 2015) (affirming judgment in favor of defendants because it found the trial court did not err in relying on the defendant’s toxicology expert testimony that the average and single-exposure levels of lead in the manufacturer’s baby food does not exceed the NOAEL (i.e., the regulatory safe harbor exposure level)). But see Chiaracane v. Port Auth. Trans-Hudson Corp., No. 18-CV-2995, 2020 WL 906268, at *4, *8 (S.D.N.Y. Feb. 25, 2020) (holding that plaintiffs’ expert testimony by a neurotoxicologist on the issue of causation of respiratory symptoms by a cleaning chemical cannot be admitted because, although the expert spoke to NOAEL, the expert did not quantify the plaintiffs’ dose or exposure to the chemical in his analysis).

Data reflect hypothetical experiment with twenty animals at each of five dose groups and a control group:

NOAEL, No Observed Adverse Effect Level (dose; mg/kg-d)
LOAEL, Lowest Observed Adverse Effect Level (dose; mg/kg-d)
BMD₁₀, Benchmark Dose at 10% response level (mg/kg-d)
BMDL₁₀, 95th percentile Lower Confidence Limit on the BMD10 (mg/kg-d)

selected endpoint.⁵⁵ Although the NOAEL and BMD are conceptually similar, the BMD is generally considered to be a more scientifically robust value because it uses all of the dose–response information from a multidose study, rather than just the single dose point that was associated with no observable response. In the regulatory arena, the upper and lower 95% confidence limits on the BMD are often statistically determined. The lower confidence limit, called the BMDL, is often used as the point of departure (POD) for extrapolating to lower doses or determining reference doses (Rf D) or acceptable daily intakes (ADIs). If the BMD₁₀ is used, then the lower bounds response is called the BMDL₁₀ (Figure 1). Rf Ds and ADIs are then determined by dividing the point of departure by uncertainty or safety

55. Courts also recognize the benchmark dose. See, e.g., Wyman v. U.S. Surgical Corp., No. 1:18-cv-00095-JAW, 2020 WL 1932338, at *21 (D. Me. Apr. 21, 2020) (explaining that the EPA bases its estimations of daily exposure levels of methylmercury a human can have without “appreciable risk of deleterious effects during a lifetime” on benchmark doses, which are amounts of methylmercury that are “associated with the threshold for a health effect on sensitive populations according to epidemiological studies”); Kolakowski v. Sec’y of Health & Hum. Servs., No. 99-0625V, 2010 WL 5672753, at *11 (Fed. Cl. Nov. 23, 2010) (explaining that the National Academy of Sciences establishes its risk assessment benchmark dose for mercury “from a percentage of the population with a physical response to the dose”).

factors. Many chemicals have multiple different toxic effects, and the dose–response relationship for each adverse outcome (toxic effect) may be different. Typically, regulatory agencies use the most sensitive toxic effect (the effect associated with the lowest dose or threshold) to establish regulatory values, such as Maximum Contaminant Levels (MCLs) under the Clean Water Act, or National Ambient Air Quality Standards (NAAQS) under the Clean Air Act. Thus, in the absence of robust epidemiological data, the NOAEL or BMDL for the most sensitive toxic effect are frequently used to establish “safe” or “acceptable” levels of exposure, which are obtained by simply dividing the POD (identified as the BMDL₁₀ or NOAEL) by safety or uncertainty factors (SF or UF). In the early days of regulation of food additives and contaminants, NOAELs were established as discussed above, and typically divided by a factor of one hundred to account for: (1) potential species differences, with a default assumption that humans might be tenfold more sensitive than the rat or mouse used to derive the NOAEL, and (2) interindividual variability within the human population, with a default assumption of another factor of ten. Today, regulatory agencies such as the EPA and the FDA often still use these basic assumptions or uncertainty factors but may modify overall uncertainty by adding additional uncertainty factors based on the quality of the study and the target population of concern. For example, the Food Quality Protection Act of 1996⁵⁶ generally requires an additional uncertainty factor (again, usually a factor of ten) if potentially susceptible populations, such as children and pregnant women, are likely to be exposed.

In toxic tort litigation related to a specific effect (i.e., a particular type of birth defect, liver or kidney disease, neurological damage, etc.), experts may argue that the dose of the agent was below the threshold and thus insufficient to have caused or substantially contributed to the disease. An individual plaintiff’s levels of exposure to a toxic substance can sometimes be estimated using exposure science techniques (see Reference Guide on Exposure Science and Exposure Assessement) and are usually expressed in units of milligrams (mg) of chemical per kilogram (kg) of body weight, per day. These are the same units used in determining NOAEL and BMDL levels from experimental animal studies, and sometimes from human epidemiology studies where airborne, dietary, or drinking water concentrations have been measured. Comparison of the estimated NOAEL or BMDL to the estimated dose of an individual can provide a useful approximation of whether the estimated exposure is likely to have exceeded a level known to be toxic. The ratio of the human-exposed dose to that of an estimate of a potential “threshold” (e.g., BMDL or NOAEL with appropriate safety factor adjustments) is called the Margin of Exposure (MOE). The larger the MOE, the less likely it is that an individual’s actual exposure could have caused or contributed to the disease. Thus, an MOE greater than 1,000 is unlikely to be sufficient to have caused or

56. See EPA, Summary of the Food Quality Protection Act, https://perma.cc/5G7J-KYLS.

substantially contributed to a toxic effect, whereas an MOE that is less than 1.0 might reasonably be anticipated to have caused or contributed to the toxic effect in question. However, there are many assumptions that go into such estimates and comparisons, and each assumption should be justified with data. It is important to note that this approach to estimation of a potential adverse outcome (risk) is generally not used for estimating the probability of cancer risk from chemicals that are thought to cause cancer by damaging DNA—so-called genotoxic carcinogens. The reason for this is that regulatory policies for genotoxic carcinogens have generally assumed that risk is proportionately related to dose and thus there is some risk at any dose, as discussed below.

Linearized Non-Threshold (LNT) Model and Determination of Cancer Risk

Certain genetic mutations, such as those leading to cancer and some inherited disorders, have been widely hypothesized to occur without any threshold. In theory, the cancer-causing mutation to the genetic material of the cell can be produced by any one molecule of certain chemicals. The LNT model led to the development of the one-hit hypothesis of cancer risk, in which each molecule of a cancer-causing chemical has some finite possibility of producing the mutation that leads to cancer. (See Figure 2 for an idealized comparison of an LNT and threshold dose–response.) This risk is very small, because it is unlikely that any one molecule of a potentially cancer-causing agent will reach that one particular spot in a specific cell’s DNA and result in the specific change that then eludes the body’s defenses and leads to a clinical case of cancer. However, the LNT hypothesis holds that the risk is not zero. The same model also can be used to predict the risk of inheritable mutational events. There is also considerable scientific controversy over the origins of the LNT model and whether it is scientifically appropriate to use for risk assessment of chemical carcinogens.⁵⁷

57. For further discussion of the LNT model of carcinogenesis, see James E. Klaunig & Zemin Wang, Chemical Carcinogenesis, in Casarett & Doull’s, supra note 1, at 329. See also Edward J. Calabrese et al., Dose Response and Risk Assessment: Evolutionary Foundations, 309 Env’t Pollution 119787 (2022), https://doi.org/10.1016/j.envpol.2022.119787; Rebecca A. Clewell et al., Dose–Dependence of Chemical Carcinogenicity: Biological Mechanisms for Thresholds and Implications for Risk Assessment, 301 Chemico-Biologocal Interactions 112 (2019), https://doi.org/10.1016/j.cbi.2019.01.025. Although the one-hit model may explain the response to some carcinogens, there is accumulating evidence that for most cancers there is in fact a multistage process and that some cancer-causing agents, so-called epigenetic or nongenotoxic agents, act through non-mutational processes. For example, the multistage cancer process may explain the carcinogenicity of benzo[a]pyrene (produced by the combustion of hydrocarbons such as oil) and aflatoxin (a mold toxin that contaminates corn and peanuts). However, non-mutational responses to asbestos, dioxin, and estradiol cause their carcinogenic effects. The appropriate mathematical model to use to depict the dose–response relationship for such carcinogens is still a matter of debate. See John E. Doe et al., The Codification of Hazard and Its Impact on the Hazard Versus Risk Controversy, 95 Archives Toxicology 3611 (2021), https://doi.org/10.1007/s00204-021-03145-6.

58. Even the determination of the MTD can be fraught with controversy. See Environmental Science Deskbook § 4:5 (2022) (illustrating problems with making inferences about human toxicity based on MTD established by animal studies).

must trade statistical power for extrapolation from higher doses to lower doses. However, recent research indicates that the MTD used in studies of some chemicals—and perhaps even the lower doses—may overwhelm biologically important protective pathways, thereby causing cellular damage integral to the disease process that would not occur at lower, more realistic doses to which humans may be exposed. Because of these concerns, there has been increasing research interest among toxicologists to redefine the MTD to include assessment of repair response pathways, where data are available, such as DNA repair processes, to ensure that the highest dose used does not overwhelm protective pathways that are operable at lower doses.⁵⁹

Accordingly, proffered toxicological expert opinion on potentially disease-causing chemicals almost always is based on evaluation of the research studies that extrapolate from animal experiments involving doses significantly higher than that to which humans are exposed.⁶⁰ In both regulatory and toxic tort scenarios, such extrapolation is backed up by other types of data, where available, such as epidemiological studies.

In Vitro Research

In vitro research concerns the effects of a chemical on human or animal cells, bacteria, yeast, isolated tissues, or embryos. Thousands of in vitro toxicological tests have been described in the scientific literature. Many tests are for chemicals that cause changes in DNA (mutations) in bacterial or mammalian systems. There are short-term in vitro tests for just about every physiological response and every organ system, such as perfusion tests and DNA studies. Many of these short-term tests have been thoroughly validated, and standardized protocols for conducting such studies are widely available.⁶¹ However, in vitro studies often do not

59. Yu-Mei Tan et al., Opportunities and Challenges Related to Saturation of Toxicokinetic Processes: Implications for Risk Assessment, 127 Regul. Toxicology & Pharmacology 105070 (2021), https://doi.org/10.1016/j.yrtph.2021.105070.

60. See, e.g., Int’l Agency for Research on Cancer, World Health Organization, Preamble, Amended 2019, https://perma.cc/A3L5-NLN7; Joseph V. Rodricks, Evaluating Disease Causation in Humans Exposed to Toxic Substances, 14 J.L. & Pol’y 39 (2006).

61. See Joanna Klapacz & B. Bhaskar Gollapudi, Genetic Toxicology, in Casarett & Doull’s, supra note 1, at 497–54. Use of in vitro data for evaluating human mutagenicity and teratogenicity is described in John M. Rogers, Developmental Toxicology, in Casarett & Doull’s, supra note 1, at 547–92. For a critique of expert testimony using in vitro data, see In re Zoloft (Sertraline Hydrochloride) Prods. Liab. Litig., 26 F. Supp. 3d 466, 477–82 (E.D. Pa. 2014); In re Mirena IUS Levonorgestrel-Related Prods. Liab. Litig. (No. II), 341 F. Supp. 3d 213, 294–95 (S.D.N.Y. 2018), aff’d, 982 F.3d 113 (2d Cir. 2020); Davis v. McKesson Corp., No. CV-18-1157-PHX-DGC, 2019 WL 3532179, at *17 (D. Ariz. Aug. 2, 2019); In re Incretin-Based Therapies Prods. Liab. Litig., 524 F. Supp. 3d 1007 (S.D. Cal. 2021), aff’d, No. 21-55342, 2022 WL 898595 (9th Cir. Mar. 28, 2022).

accurately reflect the complexity of in vivo biological responses to chemicals, and thus can be misleading. A particular challenge with interpreting in vitro studies is how to compare the concentration of the test chemical used in the test tube or petri dish with the actual concentration of the same chemical in target tissues of the individual exposed to the chemical (the in vivo situation). As such, in vitro studies are often “hypothesis generating,” rather than providing definitive scientific proof that a chemical can cause a particular type of toxic response in vivo.

The criteria of reliability for an in vitro test include the following: (1) whether the test has come through a published protocol in which many laboratories used the same in vitro method on a series of unknown compounds prepared by a reputable organization (such as the NIH or the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use [ICH]) to determine if the test consistently and accurately measures toxicity; (2) whether the test has been adopted by a U.S. or international regulatory body; and (3) whether the test is predictive of in vivo outcomes related to the same cell or target organ system.

Understanding the molecular pathways that are responsible for the ability of a chemical to cause a particular toxic effect, such as a specific type of cancer, is of value to a toxicologist charged with assessing the likelihood that a specific chemical caused a specific cancer in an individual. Identification of biological pathways is an important component of specific causation in most toxic tort cases. Thus, in vitro toxicology studies that investigate exactly how it is that a chemical causes its adverse biological effects are relevant to toxic tort litigation and can contribute significant support to a toxicologist’s expert opinions on specific causation.⁶²

New Approach Methodologies (NAMs)

Advances in analyzing the impact of external chemical agents are dependent primarily on advances in biology and in medicine leading to a greater understanding of normal body functions and of the changes observed in disease states. Integrating advances in biological and medical understanding of the cellular, subcellular, and molecular basis for normal function is reflected in the continued incorporation of new multidisciplinary science into toxicology.

62. See, e.g., In re Johnson & Johnson Talcum Powder Prods. Mktg., Sales Pracs. & Prods. Liab. Litig., 509 F. Supp. 3d 116 (D.N.J. 2020) (expert may opine that his in vitro studies show that talc causes cellular inflammation and oxidative stress but may not offer an opinion on a causal link between talc and ovarian cancer); see also In re Denture Cream Prods. Liab. Litig., No. 09-2051-MD, 2015 WL 392021 (S.D. Fla. Jan. 28, 2015) (expert’s failure to explain the relevancy of in vitro studies to humans or to account for factors needed to make a proper extrapolation renders expert’s opinion unreliable); Hardeman v. Monsanto Co., 997 F.3d 941, 963 (9th Cir. 2021) (“[C]ell studies can support more substantial evidence of causation.”).

63. Nat’l Rsch. Council, Toxicity Testing in the 21st Century: A Vision and a Strategy (2007), https://doi.org/10.17226/11970. See also Kavlock et al., supra note 48; Sebastian Schmeisser et al., New Approach Methodologies in Human Regulatory Toxicology—Not If, but How and When!, 178 Env’t Int’l 108082 (2023), https://doi.org/10.1016/j.envint.2023.108082.

64. See EPA, Guidance for Applying Quantitative Data to Develop Data-Derived Extrapolation Factors for Interspecies and Intraspecies Extrapolation (2014), https://perma.cc/JJ2K-SD8C.

receptor binding affinity) between humans and experimental animals. If a heavy metal, such as mercury, causes kidney toxicity in laboratory animals, it is highly likely to do so at some dose in humans. However, the dose at which mercury causes this effect in laboratory animals is modified by many internal factors, and the dose–response curve may be different from that for humans. Through the study of factors that modify the toxic effects of chemicals, including absorption, distribution, metabolism, and excretion, researchers can improve the ability to extrapolate from laboratory animals to humans and from higher to lower doses.⁶⁵ The mathematical depiction of the process by which an external exposure is absorbed into the bloodstream and then moves through various compartments in the body until it reaches the target organ is often called physiologically based pharmacokinetic or toxicokinetic (PBPK/PBPT) modeling.⁶⁶

Extrapolation from studies in nonmammalian species to humans is much more difficult but can be done if there is sufficient information on similarities in absorption, distribution, metabolism, and excretion. Advances in computational toxicology have increased the ability of toxicologists to make such extrapolations.⁶⁷ Quantitative determinations of human toxicity based on in vitro studies usually are not considered appropriate. In vitro or in vivo animal data for elucidating the mechanisms of toxicity are more persuasive when positive human epidemiological data or human tissue–based toxicological information also exists.

65. For example, benzene undergoes a complex metabolic sequence that results in toxicity to the bone marrow in all species, including humans. Martyn T. Smith, Advances in Understanding Benzene Health Effects and Susceptibility, 31 Annual Rev. Public Health 133 (2010), https://doi.org/10.1146/annurev.publhealth.012809.103646. The exact metabolites responsible for this bone marrow toxicity are the subject of much interest but remain unknown. Mice are more susceptible to benzene than are rats. If researchers could determine the differences between mice and rats in their metabolism of benzene, they would have a useful clue about which portion of the metabolic scheme is responsible for benzene toxicity to the bone marrow. See, e.g., Angela L. Slitt, Absorption, Distribution, and Excretion of Toxicants, in Casarett & Doull’s, supra note 1, at 159–92; Andrew Parkinson et al., Biotransformation of Xenobiotics, in Casarett & Doull’s, supra note 1, at 193–400.

66. For a discussion of methods used to extrapolate from animal toxicity data to human health effects, see text and references cited in the sections titled “Benchmark Dose (BMD)” and “Linearized Non-Threshold (LNT) Model and Determination of Cancer Risk” above.

67. See, e.g., Nicole C. Kleinstreuer et al., Introduction to Special Issue: Computational Toxicology, 34 Chem. Rsch. Toxicology 171 (2021), https://doi.org/10.1021/acs.chemrestox.1c00032. For discussion of the use of artificial intelligence and machine learning in toxicology, see Zhoumeng Lin & Wei-Chun Chou, Machine Learning and Artificial Intelligence in Toxicological Sciences, 189 Toxicological Scis. 7 (2022), https://doi.org/10.1093/toxsci/kfac075. For an overview of new approaches in structure-activity relationships, see Mark T.D. Cronin et al., A Scheme to Evaluate Structural Alerts to Predict Toxicity—Assessing Confidence by Characterising Uncertainties, 135 Regul. Toxicology & Pharmacology 105249 (2022), https://doi.org/10.1016/j.yrtph.2022.105249.

68. Note that the diffuse toxicity of cyanide also reflects its ability to spread widely in the body. Certain mitochondrial poisons primarily affect the brain and active muscles, including the heart, which are particularly oxygen dependent. Others, unable to penetrate the blood–brain barrier, will primarily affect peripheral muscles including the heart.

69. Acetaminophen (Tylenol) is among the most widely used analgesic/non-steroidal anti-inflammatory drugs, and is used safely every day by millions of people at normal therapeutic doses (less than five grams per day). However, doses in excess of ten to fifteen grams in a single day deplete an intracellular antioxidant called glutathione (GSH), and the excess dose can cause serious damage to the liver. Acetaminophen overdose is among the leading causes of acute liver failure in the United States. To produce toxicity it must first be metabolized to a reactive form by a specific enzyme that is only present in significant amounts in the liver. This is why acetaminophen is only toxic to the liver.

70. Lung fibrosis is a key pathological finding in a group of diseases known as pneumoconiosis that includes coal miners’ black lung disease, silicosis, asbestosis, and other conditions usually caused by occupational exposures to small airborne particles.

71. As a simplification, agent–receptor interactions often are described as a key in a lock, with the key needing to be able to both fit into the lock and turn the mechanism. An example from the

activity include daily cycles controlled by internal “biological clocks,” and feedback loops that allow cyclic variation, such as in the menstrual cycle or in the variations of hormone and receptor levels that are linked to normal functions such as sleeping and sexual activity. These complex normal up-and-down variations produce conceptual difficulties when attempting to extrapolate the results from model systems to the functioning human.⁷²

The understanding of the relationship between mutation and cancer discussed previously led to some of the first toxicological tests to determine whether an external agent could cause cancer. Such tests have grown in sophistication because of the advances in molecular biology and computational toxicology that have occurred concomitantly with an increased understanding of the variety of potential pathways that lead to mutagenesis.⁷³

Toxicological testing for chemical carcinogens ranges from relatively simple studies to determine whether the substance is capable of producing bacterial mutations to observation of cancer incidence as a result of long-term administration of the substance to laboratory animals. Between these two extremes are a multiplicity of tests that build upon the understanding of the mechanism of cancer causation. In vitro or in vivo tests may focus on the evidence of effects in DNA, such as the presence of adducts of the chemical or its metabolites bound to the DNA molecule or the cross-linking of the DNA molecule to protein. Researchers may look for changes in the nucleus of the cell suggestive of DNA damage that could result in mutagenesis and carcinogenesis, for example, the micronucleus test or the comet assay. Certain mutagens cause an increase in the normal exchange of nuclear material among DNA components during normal cell division, which gives rise to a test known as the sister chromatid exchange.⁷⁴ The direct observation of chromosomes

nervous system is the use in treating a heroin overdose of another opiate that has a much higher affinity for the receptor site but produces little effect once bound. When given to a normal person, this second opiate would have a mild depressant effect, but it can reverse a near fatal overdose of heroin by displacing the heroin from the receptor site. Thus, the directionality of opiate effect depends on the interaction of the components of the mixture. This interaction is even more complex when dealing with estrogenic agents that are naturally occurring as well as made within the body at different levels in response to different external and internal stimuli and at different time intervals.

72. See, e.g., In re Denture Cream Prods. Liab. Litig., No. 09-2051-MD, 2015 WL 392021 (S.D. Fla. Jan. 28, 2015) (expert’s failure to correct for circadian rhythm fluctuations in zinc plasma levels rendered his opinion inadmissible). In another example, the complexity of the interaction of a mixture of dioxins with receptors governing the endocrine system can be contrasted with that of the reaction of carbon monoxide with the hemoglobin oxygen receptor discussed supra, note 18. The latter is unidirectional in that any additional carbon monoxide will interfere with oxygen delivery, of which there cannot be too much under normal physiological conditions.

73. See, e.g., James E. Klaunig, Carcinogenesis, in An Introduction to Interdisciplinary Toxicology: From Molecules to Man 87 (2020), https://doi.org/10.1016/B978-0-12-813602-7.00008-9.

74. All of these tests require validation regarding their relevance to predicting human carcinogenesis, as well as to their technical reproducibility. See Klapacz & Gollapudi, supra note 61, for a discussion of the wide variety of short-term mutagenesis assays for predicting mutagenic and potentially carcinogenic potential of chemicals.

to look for specific abnormalities, known as cytogenetic analysis, is providing more information about the pathways of carcinogenesis. For cancers such as acute myelogenous leukemia, it has long been recognized that those individuals who present with recognizable chromosomal abnormalities are more likely to have been exposed to a known human chemical leukemogen such as benzene.⁷⁵ These and other tests provide information that can be used in evaluating whether a chemical is a potential human carcinogen.

The recent almost explosive growth in the ability to measure minuscule changes in the complex genome, a process known as next-generation sequencing, is rapidly being applied to a broad range of human diseases. The emphasis thus far has been on the promise of developing new therapies appropriate to personalized medicine, but the same techniques may in the future provide a fingerprint of the interaction of a chemical with DNA. However, there are very few chromosomal abnormalities or specific mutations that are unequivocally linked to a specific chemical or physical carcinogen.⁷⁶

Assessing whether a chemical or physical agent produces human cancer requires careful evaluation. The World Health Organization’s IARC and the U.S. National Toxicology Program (NTP) have formal processes to evaluate the weight of evidence that a chemical causes cancer.⁷⁷ Each classifies chemicals on the basis of epidemiological evidence, toxicological findings in laboratory animals, and mechanistic considerations, and then assigns a specific category of carcinogenic potential to the individual chemical or exposure situation (e.g., employment as a painter).⁷⁸ Only a small percentage of the total chemicals in commerce

75. Patrick J. Kerzic & Richard D. Irons, Distribution of Chromosome Breakpoints in Benzene-Exposed and Unexposed AML Patients, 55 Env’t Toxicology & Pharmacology 212 (2017), https://doi.org/10.1016/j.etap.2017.08.033.

76. See Luoping Zhang et al., The Nature of Chromosomal Aberrations Detected in Humans Exposed to Benzene, 32 Critical Revs. Toxicology 1 (2002), https://doi.org/10.1080/20024091064165. However, studies on liver cancer patients exposed to the potent dietary carcinogen aflatoxin B1 (AFB) have demonstrated that a specific mutation in codon 249 in the important cancer gene called TP53 is uniquely associated with exposure to AFB. Only liver tumors from patients with a history of AFB exposure showed this specific mutation. S. P. Hussain et al., TP53 Mutations and Hepatocellular Carcinoma: Insights into the Etiology and Pathogenesis of Liver Cancer, 26 Oncogene 2166 (2007), https://doi.org/10.1038/sj.onc.1210279.

77. The U.S. National Toxicology Program issues a congressionally mandated report on carcinogens. Nat’l Toxicology Prog., 15th Rep. on Carcinogens (2021), https://perma.cc/V4DP-3YNZ. IARC produces its reports through a monograph series that provides detailed description of the agents or processes under consideration as well as the findings of the IARC expert working group. See the IARC website for a list of these monographs, https://perma.cc/UEP2-5L9A.

78. IARC uses the following classifications:
Group 1, The agent (mixture) is carcinogenic to humans;
Group 2A, The agent (mixture) is probably carcinogenic to humans;
Group 2B, The agent (mixture) is possibly carcinogenic to humans;
Group 3, The agent (mixture) is not classifiable as to its carcinogenicity to humans.

are considered to be known human carcinogens, although this must be considered in the context of how many of these chemicals have been rigorously tested (see discussion of REACH and the Lautenberg Act above). In the past, when chemicals were evaluated for carcinogenicity, assignment to the highest category was dependent almost totally on whether epidemiological research had been conducted, although animal data and mechanistic information were also considered. In recent years, with improved understanding of the mechanism of action of chemical carcinogens, there has been increased use of mechanistic data.⁷⁹ For example, higher credence is given to the likelihood that a chemical is a human carcinogen if the metabolite found to be responsible for carcinogenesis in a laboratory animal is also found in the blood or urine of humans exposed to this chemical, or if there is evidence of the same type of DNA damage in humans as there is in laboratory animals in which the agent does cause cancer.⁸⁰

Inherent in putting chemicals into distinct categories when there is a continuum for the strength of the evidence is that some chemicals will be very close to the dividing line between the discrete categories. Inevitably, small differences in the interpretation of the evidence for such chemicals will lead to disagreement regarding categorization.

79. See Daniele S. Wikoff et al., A Framework for Systematic Evaluation and Quantitative Integration of Mechanistic Data in Assessments of Potential Human Carcinogens, 167 Toxicological Scis. 322 (2019), https://doi.org/10.1093/toxsci/kfy279, for a discussion and for specific examples of the use of mechanistic data in evaluating carcinogens. The evolution in the approach to determining cancer causality is evident from reviewing the guidelines used to assemble the weight of evidence for causality by IARC and NTP, two of the organizations that have the lengthiest track record of responsibility for the hazard identification of carcinogens. Both have increased the weight given to mechanistic evidence in characterizing the overall strength of the total evidence used to classify the potential for a chemical or an exposure to be causal. IARC now permits classification in Group 1 when there is less than sufficient evidence in humans but sufficient evidence in animals and “strong evidence in exposed humans that the agent exhibits key characteristics of carcinogens and sufficient evidence of carcinogenicity in experimental animals.” IARC Preamble, supra note 60, at 23. The criterion used by NTP for listing a chemical as a known human carcinogen in its biannual Report on Carcinogens is: “There is sufficient evidence of carcinogenicity from studies in humans,* which indicates a causal relationship between exposure to the agent, substance, or mixture, and human cancer.” The asterisk is particularly notable in that it specifies that the evidence need not be solely epidemiological: “*This evidence can include traditional cancer epidemiology studies, data from clinical studies, and/or data derived from the study of tissues or cells from humans exposed to the substance in question, which can be useful for evaluating whether a relevant cancer mechanism is operating in humans.” See Nat’l Toxicology Prog., supra note 77, at 6. For a recent discussion on using mechanism/mode of action data in assessment of genotoxic carcinogens, see also Andrea Hartwig et al., Mode of Action-Based Risk Assessment of Genotoxic Carcinogens, 94 Archives Toxicology 1787 (2020), https://doi.org/10.1007/s00204-020-02733-2. The EPA also considers mechanism of action in its regulatory approaches and distinguishes further between mechanism of action and mode of action. See Kathryn Z. Guyton et al., Improving Prediction of Chemical Carcinogenicity by Considering Multiple Mechanisms and Applying Toxicogenomic Approaches, 681 Mutation Rsch. 230, 240 (2009), https://doi.org/10.1016/j.mrrev.2008.10.001.

80. An example is the IARC evaluation of formaldehyde that upgraded the categorization from 2A to 1 based on epidemiological data that were strongly supported by the finding of nasal cancer in laboratory animals and by the presence of DNA-protein cross-links in the nasal tissue of the

laboratory animals and of humans inhaling formaldehyde. However, epidemiological evidence associating formaldehyde with human acute myelogenous leukemia was questioned on the basis of the lack of mechanistic evidence, including questions about how such a highly reactive agent could reach the bone marrow following inhalation. See Formaldehyde, 2-Butoxyethanol and 1-tert-Butoxypropan-2-ol, in 88 IARC Monographs on the Evaluation of Carcinogenic Risks to Humans (2006).

absorption, distribution, metabolism, and excretion, enhances the expert’s ability to extrapolate from laboratory animals to humans.⁸¹

The expert should review similarities and differences between the animal species in which the compound has been tested and humans. This analysis should form the basis of the expert’s opinion regarding whether extrapolation from animals to humans is warranted.

In general, an overwhelming similarity is apparent in the biology of all living things, and there is a particularly strong similarity among mammals. Of course, laboratory animals differ from humans in many ways. For example, rats do not have gallbladders. Thus, rat data would not be pertinent to the possibility that a compound produces human gallbladder toxicity.⁸² The life stage at which exposure occurs can also be an important determinant of toxicity. It is generally recognized that the developing embryo and fetus are unusually susceptible to many toxic substances, and thus exposure of a pregnant animal requires special consideration for toxic effects that might occur to the developing offspring. In a similar manner, it is also generally recognized that early life exposure (from infancy through adolescence) is also a period of enhanced susceptibility, and effects not seen in adults may occur. This is especially true for chemicals that affect the nervous system, since the development of the nervous system is not complete until early adulthood. Thus, for example, the effects of lead poisoning in children are manifest mostly through impacts on the central nervous system that lead to impairment of learning and development, whereas adults exposed to the same level of lead have little or no detectable effects on the central nervous system, although at high doses lead can cause effects on the peripheral nervous system in adults (so-called “lead palsy”).

Note that, in animal studies, many subjective symptoms are poorly modeled regardless of the age of exposure. Thus, complaints that a chemical has caused nonspecific symptoms, such as nausea, headache, and weakness, for which there may be no objective manifestations in humans, are difficult to test in laboratory animals.

81. See generally supra note 3 and references therein.

82. See, e.g., Hardeman v. Monsanto Co., 997 F.3d 941, 963 (9th Cir. 2021) (“Animal studies are relevant evidence of causation where there is a sound basis for extrapolating conclusions from those studies to humans in real-world conditions.”). See generally David Spurgeon et al., Species Sensitivity to Toxic Substances: Evolution, Ecology and Applications, 8 Frontiers Env’t Sci. 1 (2020), https://doi.org/10.3389/fenvs.2020.588380. Species differences that produce a qualitative difference in response to xenobiotics are well known. Sometimes understanding the mechanism underlying the species difference can allow one to predict whether the effect will occur in humans. Thus, carbaryl, an insecticide commonly used for gypsy moth control, produces fetal abnormalities in dogs but not in hamsters, mice, rats, and monkeys. Dogs lack the specific enzyme involved in metabolizing carbaryl; the other species tested all have this enzyme, as do humans. Therefore, it has been assumed that humans are not at risk for fetal malformations produced by carbaryl.

83. Certain chemicals act directly to produce toxicity, whereas others require the formation of a toxic metabolite. For example, the potent rat liver carcinogen, N-nitroso-dimethylamine (NDMA, present at trace levels in a typical diet and in some pharmaceutical preparations), requires metabolic activation in order to bind to DNA and cause mutations that lead to cancer in experimental animals. In humans, there appears to be only one enzyme capable of activating NDMA to its reactive intermediate, and this enzyme (called CYP2E1) is only expressed in the liver, suggesting that liver cancer would be the only kind of cancer expected in humans exposed to NDMA. However, other carcinogenic nitrosamines can be activated by enzymes present in other human tissues, as well as the liver. See Yupeng Li & Stephen S. Hecht, Metabolic Activation and DNA Interactions of Carcinogenic N-Nitrosamines to Which Humans Are Commonly Exposed, 23 Int’l J. Molecular Scis. 4559 (2022), https://doi.org/10.3390/ijms23094559.

84. Brian J. Day et al., Potentiation of Carbon Tetrachloride-Induced Hepatotoxicity and Pneumotoxicity by Pyridine, 8 J. Biochem. Toxicology 11 (1993), https://doi.org/10.1002/jbt.2570080104.

85. Nat’l Rsch. Council, Applications of Toxicogenomic Technologies to Predictive Toxicology and Risk Assessment (2007), https://doi.org/10.17226/12037. Genomics can also be misinterpreted. An example is the use of white blood cell gene expression to determine whether benzene was a cause of acute myelogenous leukemia (AML) in individual workers. Martyn T. Smith, Misuse of Genomics in Assigning Causation in Relation to Benzene Exposure, 14 Int’l J. Occupational & Env’t Health 144 (2008), https://doi.org/10.1179/oeh.2008.14.2.144, describes why the failure to match a pattern of DNA expression in workers with AML who were previously exposed to benzene is not scientifically defensible as a means to establish the lack of causation, as said to have been done in workers’ compensation cases in California. The wide range in the rate of metabolism of chemicals is at least partly under genetic control. A study of Chinese workers exposed to benzene found approximately a doubling of risk in people with high levels of either an enzyme that increased the rate of formation of a toxic metabolite or an enzyme that decreased the rate of detoxification of this metabolite. There was a sevenfold increase in risk for those who had both genetically determined variants. Nathan Rothman et al., Benzene Poisoning, A Risk Factor for Hematological Malignancy, Is Associated with the NQO1 609C→T Mutation and Rapid Fractional Excretion of Chlorzoxazone, 57 Cancer Rsch. 2839 (1997), https://perma.cc/4BDM-JXNZ. See also Lucio G. Costa & David L. Eaton, Gene Environment Interactions: Fundamentals of Ecogenetics (2006).

extensively by the EPA in evaluating many new chemicals required to be tested under the registration requirements of the TSCA, its reliability has a number of limitations.⁸⁶

Has the Compound Been the Subject of In Vitro Research, and If So, Can the Findings Be Related to What Occurs In Vivo?

Cellular and tissue culture research can be particularly helpful in identifying mechanisms of toxic action and potential target organ toxicity. The major barrier to the use of in vitro results is the frequent inability to relate doses that cause cellular toxicity to doses that cause whole-animal toxicity. In many critical areas, knowledge that permits such quantitative extrapolation is lacking.⁸⁷ Nevertheless, the ability to quickly test new products through in vitro tests, using

86. For example, benzene and the alkyl benzenes (which include toluene, xylene, and ethyl benzene) share a similar chemical structure. SAR works exceptionally well in predicting the acute central nervous system anesthetic-like effects of both benzene and the alkyl benzenes. Although there are slight differences in dose–response relationships, they are readily explained by the interrelated factors of chemical structure, vapor pressure, and lipid solubility (the brain is highly lipid). Nat’l Rsch. Council, The Alkyl Benzenes (1981). However, only benzene produces damage to the bone marrow and leukemia; the alkyl benzenes do not have this effect. This difference is the result of specific toxic metabolic products of benzene in comparison with the alkyl benzenes. Thus, SAR is predictive of neurotoxic effects but not bone marrow effects. See David A. Eastmond et al., Lymphohematopoietic Cancers Induced by Chemicals and Other Agents and Their Implications for Risk Evaluation: An Overview, 761 Mutation Rsch. 40 (2014), https://doi.org/10.1016/j.mrrev.2014.04.001. Advances in computational approaches also show promise in improving SAR. See Nat’l Rsch. Council, supra note 63. For an example of how gains in computational sciences and “artificial intelligence” (e.g., “Deep Learning”) are being applied to improve SAR analysis for predicting chemical toxicity, see Gabriel Idakwo et al., Deep Learning-Based Structure-Activity Relationship Modeling for Multi-Category Toxicity Classification: A Case Study of 10K Tox21 Chemicals With High-Throughput Cell-Based Androgen Receptor Bioassay Data, 10 Frontiers Physiology 2044 (2019), https://doi.org/10.3389/fphys.2019.01044. In Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), the Court rejected a per se exclusion of SAR, animal data, and reanalysis of previously published epidemiological data where there were negative epidemiological data. Most cases involving SAR expert opinion involve patent litigation or analogous drugs under the Controlled Substance Analogue Enforcement Act. See, e.g., Takeda Pharm. Co. v. Torrent Pharms. Ltd., No. 17-3186 (SRC)(CLW), 2020 WL 549594, at *26 (D.N.J. Feb. 4, 2020), aff’d, 844 F. App’x 339 (Fed. Cir. 2021); United States v. Cooper, No. 3:14-cr-014-J-20MCR, 2015 WL 13850123, at *6, *9 (M.D. Fla. Jan. 14, 2015).

87. See, e.g., In re Denture Cream Prods. Liab. Litig., No. 09-2051-MD, 2015 WL 392021 (S.D. Fla. Jan. 28, 2015) (expert failed to explain relevancy of in vitro studies to humans or account for factors needed to make a proper extrapolation).

human cells, provides invaluable “early warning systems” for toxicity.⁸⁸ For example, screening of new chemicals with a simple bacterial mutagenicity assay such as the Ames test will identify if the new chemical is likely to be mutagenic, and thus carcinogenic. The pharmaceutical industry routinely evaluates new chemicals for mutagenic/carcinogenic potential, and generally will not spend large amounts of time and money in further development if the candidate drug is identified through such simple tests as likely to be mutagenic and thus carcinogenic.

Is the Association Between Exposure and Disease Biologically Plausible?

No matter how strong the temporal relationship between exposure and the development of disease, or the supporting epidemiological evidence, it is difficult to accept an association between a compound and a health effect when no plausible mechanism can be identified by which the chemical or physical exposure leads to the putative effect.⁸⁹

Specific Causal Association Between an Individual’s Exposure and Onset of Disease

An expert who opines that exposure to a compound caused a person’s disease engages in deductive clinical reasoning.⁹⁰ In most instances, cancers and other diseases do not wear labels documenting their causation. The opinion is based on an assessment of the individual’s exposure, including the amount, the temporal relationship between the exposure and disease, and other disease-causing factors. This information is then compared with scientific data on the relationship

88. Despite its limitations, in vitro research can strengthen inferences drawn from whole-animal bioassays and can support opinions regarding whether the association between exposure and disease is biologically plausible. See Klapacz & Gollapudi, supra note 61; Rogers, supra note 61.

89. See, e.g., Sarkees v. E.I. DuPont de Nemours & Co., No. 17-CV-651, 2020 WL 906331 (W.D.N.Y. Feb. 25, 2020) (expert reviewed animal studies, in vitro studies, and human studies and the inferential steps utilized by researchers to propose a likely mechanism for toxic effects in humans).

90. For an example of deductive clinical reasoning based on known facts about the toxic effects of a chemical and the individual’s pattern of exposure, see Bernard D. Goldstein, Is Exposure to Benzene a Cause of Human Multiple Myeloma?, 609 Annals N.Y. Acad. Scis. 225 (1990), https://doi.org/10.1111/j.1749-6632.1990.tb32070.x.

between exposure and disease. The certainty of the expert’s opinion depends on the strength of the research data demonstrating a relationship between exposure and the disease at the dose in question and the presence or absence of other disease-causing factors (also known as confounding factors).⁹¹

Particularly problematic are generalizations made in personal injury litigation from regulatory positions. Regulatory standards are set for purposes far different than determining the preponderance of evidence in a toxic tort case. Further, if regulatory standards are discussed in toxic tort cases to provide a reference point for assessing exposure levels, it must be recognized that there is a great deal of variability in the extent of evidence required to support different regulations.⁹² The extent of evidence required to support regulations depends on

1. the law (e.g., the Clean Air Act National Ambient Air Quality Standard provisions have language focusing regulatory activity for primary pollutants on adverse health consequences to sensitive populations with an adequate margin of safety and with no consideration of economic consequences, while regulatory activity under the TSCA and its revision, the Lautenberg Act, clearly asks for some balance between the societal benefits and risks of new chemicals⁹³);
2. the specific endpoint of concern (e.g., consider the concern caused by cancer and adverse reproductive outcomes versus almost anything else); and

91. Causation issues are discussed in Steve C. Gold et al., Reference Guide on Epidemiology, section titled “General Causation and Specific Causation,” and John B. Wong et al., Reference Guide on Medical Testimony, section titled “Medical Decision-Making,” both in this manual. For a detailed analysis of the challenges of assessing epidemiological and toxicological data to establish causal relationships between exposure and disease, see Nat’l Acads. Scis., Eng’g, & Med., Advancing the Framework for Assessing Causality of Health and Welfare Effects to Inform National Ambient Air Quality Standard Reviews (2022).

92. See, e.g., Kirk v. Schaeffler Grp. USA, Inc., 887 F.3d 376, 392 (8th Cir. 2018) (noting that regulatory standards are “designed to protect public health,” not to be a measure of causation, and that regulatory agencies may set standards “without having precise data on the question of how much harm, or what kind of harm, some specific amount of . . . substance might reasonably be expected to cause” (quoting Wright v. Willamette Indus., Inc., 91 F.3d 1105, 1107 (8th Cir. 1996)); Williams v. Mosaic Fertilizer, LLC, 889 F.3d 1239, 1247 (11th Cir. 2018) (stating that regulatory standards are designed to be protective whereas “dose–response calculations aim to identify the exposure levels that actually cause harm”).

93. Cmty. Voice v. U.S. Env’t Prot. Agency, 997 F.3d 983, 990–91 (9th Cir. 2021) (citing Whitman v. Am. Trucking Ass’ns, 531 U.S. 457, 467–68 (2001)) (explaining TSCA’s consideration of “non-health factors such as achievability and cost” and the opposite for the Clean Air Act, which the Supreme Court has held “does not” permit “the EPA to consider costs in setting clean air standards”). See also Nat’l Res. Defense Council v. U.S. Env’t Prot. Agency, 38 F.4th 34 (9th Cir. 2022) (challenge to the EPA’s conclusions on human health risks and cost-benefit analysis for glyphosate under FIFRA, based on the EPA’s failure to follow its Cancer Guidelines, resulting in court vacating human health portion of the EPA’s Interim Decision and remanding).

94. These concerns are discussed in Stephen Breyer, Breaking the Vicious Circle: Toward Effective Risk Regulation (1993).

95. “‘[S]cientific knowledge of the harmful level of exposure to a chemical’ is a ‘minimal fact’ necessary for establishing causation.” Cooper v. Meritor, Inc., No. 4:16-CV-52-DMB-JMV, 2019 WL 545271, at *5 (N.D. Miss. Feb. 11, 2019) (quoting Allen v. Pa. Eng’g Corp., 102 F.3d 194, 199 (5th Cir. 1996)). In toxic tort cases, the plaintiff bears the burden of demonstrating the level of exposure. Zellars v. NexTech Ne. LLC, 895 F. Supp. 2d 734, 741 (E.D. Va. 2012) (citing Westberry v. Gislaved Gummi AB, 178 F.3d 257, 260 (4th Cir. 1999)). The court in Zellars discussed cases from the Fourth, Fifth, and Seventh Circuit U.S. Courts of Appeal where expert opinions were or were not found to be reliable, noting that where “substantial exposure” is established by the expert, “more detailed quantitative data about [the plaintiff’s] level of exposure” may not be necessary. 895 F. Supp. 2d at 739 (citing Westberry, 178 F.3d at 264). On the other hand, where there is “no direct evidence of the plaintiff’s level of exposure to the chemical at issue,” the expert’s opinion may be excluded as unreliable. Zellars, 895 F. Supp. 2d at 740 (citing Allen, 102 F.3d at 199). Likewise, where the expert did not review medical records, perform any tests or calculations, or seek to learn any specifics about the environment in which the alleged exposure took place, the expert’s opinion as to causation based on exposure level was excluded as being “not based on sufficient information about the level of the plaintiff’s exposure to the alleged toxin.” 895 F. Supp. 2d at 740 (citing Wintz v. Northrop Corp., 110 F.3d 508, 510–13 (7th Cir. 1997)). In Zellars, the expert opinions were ultimately excluded because the experts “did not consider the extent of Plaintiffs’ exposure,” rendering their opinions “speculative at best.” 895 F. Supp. 2d at 741, aff’d sub nom. Zellers v. NexTech Ne., LLC, 533 F. App’x 192, 198 (4th Cir. 2013). For a discussion of general issues

mathematical modeling, in which one uses a variety of physical factors to estimate the transport of the pollutant from the source to the individual. For example, mathematical models take into account such factors as wind variations to allow calculation of the transport of radioactive iodine from a federal atomic research facility to nearby residential areas. Second, exposure can be directly measured in the medium in question—air, water, food, or soil. When the medium of exposure is water, soil, or air, hydrologists or meteorologists may be called upon to contribute their expertise to measuring exposure. The third approach directly measures human individuals through some form of biological monitoring, such as blood tests to determine blood lead levels or urinalyses to check for a urinary metabolite indicative of pollutant exposure. Ideally, both environmental testing and biological monitoring are performed; however, this is not always possible, particularly in instances of past exposure (unless the chemical(s) had very long half-lives (e.g., greater than three or four years), in which case blood levels may be useful to demonstrate significant exposures that occurred many years earlier).⁹⁶

The toxicologist must go beyond understanding exposure to determine if the individual was exposed to the compound in a manner that can result in absorption into the body. The absorption of the compound is a function of its physiochemical properties, its concentration, and the presence of other agents or conditions that assist or interfere with its uptake. For example, inhaled lead is absorbed almost totally, whereas ingested lead is taken up only partially into the body. Of toxicological relevance is the finding that children absorb ingested lead substantially more efficiently than adults, which may contribute to the enhanced susceptibility of children to the neurotoxic effects of lead.⁹⁷ Iron deficiency and low nutritional calcium intake, both common conditions of children living in impoverished urban environments, increase the amount of ingested lead that is absorbed in the gastrointestinal tract and passes into the bloodstream.⁹⁸

regarding assessment of exposure of toxic substances see M. Elizabeth Marder & Joseph V. Rodricks, Reference Guide on Exposure Science and Exposure Assessment, in this manual.

96. See sections titled “Toxicology and the Law” and “Use of Biomarkers in Toxicology Exposure Assessment” above.

97. Alan R. Abelsohn & Margaret Sanborn, Lead and Children: Clinical Management for Family Physicians, 56 Canadian Fam. Physician 531 (2010), https://perma.cc/M6S4-P2L7.

98. The term bioavailability is used to describe the extent to which a compound, such as lead, is taken up into the body. In essence, bioavailability is at the interface between exposure and absorption into the organism. See, e.g., In re Denture Cream Prods. Liab. Litig., No. 09-2051-MD, 2015 WL 392021 (S.D. Fla. Jan. 28, 2015) (expert’s analysis of in vitro studies failed to demonstrate bioavailability of zinc in denture cream); Ridings v. Maurice, No. 15-00020-CV-W-JTM, 2019 WL 13159921, at *4 (W.D. Mo. Aug. 12, 2019) (permitting expert to testify regarding bioavailability of drug alleged to have caused, at least in part, the plaintiff’s injuries). Under certain circumstances bioavailability is not a necessary condition to establish general causation, however. Adkisson v. Jacobs Eng’g Grp., Inc., 342 F. Supp. 3d 791, 805–07 (E.D. Tenn. 2018) (“Biological plausibility and bioavailability are important scientific concepts. But it does not appear that either is strictly

necessary for an association between a particular toxic agent and a particular disease to be considered causal. Accordingly, neither is required to establish proof of general causation.”).

99. Courts have explored the relationship between metabolic transformation and carcinogenesis. See, e.g., In re E.I. Du Pont De Nemours & Co. C-8 Personal Inj. Litig., No. 2:13-md-2433, 2016 WL 3064124, at *4 (S.D. Ohio May 30, 2016) (“For example, one pound of a toxic, cancer-causing chemical that biopersists in humans for decades and bioaccumulates in the environment for millennia may cause more harm than 100 pounds of a toxic, cancer-causing chemical that metabolizes and/or degrades quickly.”).

compounds are often excreted through the biliary tract into the feces. Certain fat-soluble, poorly metabolized compounds, such as PCBs, may persist in the body for decades, although they can be excreted in the milk fat of lactating women, with potential effects in their infants.

Does the Temporal Relationship Between Exposure and the Onset of Disease Support or Contradict Causation?

In acute toxicity, there is usually a short time period between cause and effect. However, in some situations, the length of basic biological processes necessitates a longer period of time between initial exposure and the onset of observable disease. For example, in AML (acute myelogenous leukemia), the adult form of acute leukemia, at least one to two years must elapse from initial exposure to radiation, benzene, or cancer chemotherapy before the manifestation of a clinically recognizable case of leukemia, and the period of significantly higher risk from the last exposure usually persists for no more than about fifteen years. A toxic tort claim alleging a shorter or longer time period between cause and effect is scientifically highly debatable. Much longer latency periods are necessary for the manifestation of solid tumors caused by agents such as asbestos and arsenic.¹⁰⁰

100. The temporal relationship between exposure and causation is discussed in many cases. See, e.g., Johnson v. Arkema, Inc., 685 F.3d 452, 466–67 (5th Cir. 2012) (quoting Curtis v. M&S Petroleum, Inc., 174 F.3d 661, 670 (5th Cir. 1999)) (“[T]emporal connection standing alone is entitled to little weight in determining causation[,]” but may be “entitled to greater weight when there is an established scientific connection between exposure and illness or other circumstantial evidence supporting the causal link.”) (internal citations omitted); In re Paulsboro Derailment Cases, 746 F. App’x 94, 99 (3d Cir. 2018) (citing Kannankeril v. Terminix Int’l, Inc., 128 F.3d 802, 808–09 (3d Cir. 1997)) (noting that an expert may only rely on a temporal relationship analysis if the expert uses that relationship within the greater methodology of a differential diagnosis); C.W. ex rel. Wood v. Textron, Inc., 807 F.3d 827, 838 (7th Cir. 2015) (quoting Ervin v. Johnson & Johnson, 492 F.3d 901, 904–05 (7th Cir. 2007)) (rejecting expert methodology relying on the relationship between the time of exposure and the onset of injury, explaining that “[t]he mere existence of a temporal relationship between taking a medication and the onset of symptoms does not show a sufficient causal relationship”); Arias v. DynCorp, 928 F. Supp. 2d 10, 8 (D.D.C. 2013), aff’d, 752 F.3d 1011 (D.C. Cir. 2014) (discussing the relationship between temporal evidence and specific causation); Leake v. United States, 843 F. Supp. 2d 554, 561–62 (E.D. Pa. 2011) (noting that “even a strong temporal relationship between exposure and injury, in and of itself,” may not be “sufficient to establish a reliable general causation opinion”).

101. See, e.g., Wyman v. U.S. Surgical Corp., No. 1:18-cv-00095-JAW, 2020 WL 1932338, at *21 (D. Me. Apr. 21, 2020) (In the context of mercury, benchmark dose is “the amount . . . associated with the threshold for a health effect on sensitive populations according to epidemiological studies.”); Kolakowski v. Sec’y of Health & Hum. Servs., No. 99-0625V, 2010 WL 5672753, at *11 (Fed. Cl. Nov. 23, 2010) (Benchmark dose “is a dose at which some statistical interval, measured from a percentage of the population with a physical response to the dose.”); In re Valsartan, Losartan, and Irbesartan Prods. Liab. Litig., No. 2022 WL 807343, at *2 (D.N.J. Mar. 17, 2022) (permitting an expert to testify using benchmark-dose methodology, acknowledging that while “this is not the methodology used by many scientists,” it “is nonetheless used by some in the relevant scientific community”).

102. See, e.g., supra note 54 & accompanying text; Robert G. Tardiff & Joseph V. Rodricks, Toxic Substances and Human Risk: Principles of Data Interpretation 391 (1987); Joseph V. Rodricks, Calculated Risks 230–39 (2d ed. 2006). For regulatory toxicology, NOAEL is being replaced by a more statistically robust approach known as the benchmark dose. See supra note 55 & accompanying text. For example, the EPA’s use of the benchmark dose takes into account comprehensive dose–response information, unlike NOAEL.

the BMDLow, or BMDL, or BMDL₁₀) can also be calculated from animal data or from human toxicity data if they exist. The estimation of a threshold dose using either NOAEL or BMD data, however, is generally not applied to substances that exert toxicity by causing DNA damage and mutations that may lead to cancer. Theoretically, any exposure at all to mutagens may increase the risk of cancer, although the risk may be very small and not achieve statistical significance.¹⁰³

Although the concept of the non-threshold response is generally used only for genotoxic carcinogens or other outcomes associated with gene mutations (e.g., development of some types of birth defects), it is sometimes claimed that other toxic responses, such as the effects of lead on children’s IQ, also follow a linear response at low doses. However, it is biologically likely that there are thresholds for such responses, but if the threshold is below background levels of exposure (such as may be the case with lead), the dose–response would appear to be linear since a threshold dose could not be established and background exposures exceed the putative threshold level.

What Is the Likelihood That the Disease Would Have Occurred Without the Specific Exposure?

In coming to an expert opinion as to whether the exposure to a known cause of a disease is likely to be responsible for the case observed in the plaintiff, the expert must consider other potential causes of the same disease. Some may be

103. See sources cited supra notes 57 and 65. See also Myers v. United States, No. 02cv1349–BEN, 2014 WL 6611398, at *39 (S.D. Cal. Nov. 20, 2014) (explaining that although thallium may be harmful at certain levels, “[s]ince we are exposed to thallium on a daily basis, there must be a level that does not cause adverse health effects”). U.S. regulatory approaches aimed at protecting the general population tend to avoid setting a standard for a known human carcinogen, because any allowable level below the standard is at least theoretically capable of causing cancer. However, exposure to many chemical carcinogens, including benzene and arsenic, cannot be eliminated. Thus, agencies and Congress have developed a number of ingenious means to regulate carcinogens while not seeming to acquiesce in exposure of the general population to a carcinogen. These include the FDA’s approach to de minimis risk and the EPA’s setting of a zero maximum contaminant level goal for carcinogens in drinking water while setting a maximum contaminant level above zero that is set as closely as possible to the MCLG, taking technology and cost data into account. In contrast, occupational standards, which also take into account feasibility, permit exposure to known human carcinogens. A generally outmoded approach for environmental or indoor air guidelines has been to divide the permissible OSHA standard by a factor accounting for the presumed lifetime exposure to the environmental chemical compared with forty-five years at a forty-hour workweek. Plaintiff’s experts in a toxic court case must rely on more than exposure levels established by regulatory agencies, as those levels “often build in considerable cushion to account for the most sensitive members of the population.” Williams v. Mosaic Fertilizer, LLC, 889 F.3d 1239, 1246–47 (11th Cir. 2018) (affirming exclusion of expert opinion for, inter alia, relying on regulatory standards of exposure levels).

due to other known causal factors, such as lung cancer in a uranium miner who is also a cigarette smoker. In addition, with rare exceptions, almost all diseases have some anticipated background causes that appear related to human biology, particularly aging. An additional challenge comes in unraveling the contribution of a specific substance when exposure may have been to multiple agents. There are numerous examples where the effect of two combined exposures to two substances both known to cause the same disease is greater than the sum of the individual effects (synergism). The example above of lung cancer in uranium miners who also smoke is a good illustration. Another example of such synergism is the incidence of liver cancer in people who are exposed to the dietary contaminant aflatoxin B1, and also are infected with hepatitis B virus. Both are known causes of liver cancer (relative risk for hepatitis B virus antigen positivity is approximately tenfold; relative risk for aflatoxin B1 alone is approximately threefold) but the presence of both risk factors increased the risk of developing liver cancer by approximately sixty times.¹⁰⁴

Not surprisingly, cancer has received the most attention in research aimed at disentangling external and internal causes. For some cancers, such as lung cancer, smoking is an obvious predominant external cause, and the extent to which other known causes contribute to overall lung cancer incidence also can be estimated. But these known external causes do not currently account for all cases of lung cancer. For other cancers, such as pancreatic cancers, it has been very difficult to find any external cause. There are, however, a few instances in which the exposure to a specific chemical is nearly uniquely associated with a particular (usually rare) type of cancer. For example, the industrial chemical vinyl chloride, widely used in the synthesis of certain plastics, is associated with a rare form of cancer of the blood vessels in the liver (angiosarcoma).¹⁰⁵ Likewise, certain types of mesothelioma (a cancer of the lining of the chest cavity or stomach/intestine) are associated with prior exposure to asbestos, especially among smokers, suggesting a synergistic interaction between smoking and asbestos exposure.¹⁰⁶

That mutations in the DNA of stem cells of virtually every tissue accumulate with age is now generally recognized. Upwards of thousands of mutations

104. John D. Groopman et al., Aflatoxin and Hepatitis B Virus Biomarkers: a Paradigm for Complex Environmental Exposures and Cancer Risk, 1 Cancer Biomarkers 5 (2005), https://doi.org/10.3233/cbm-2005-1103. See also Joshua Jin et al., Synergism in Actions of HBV with Aflatoxin in Cancer Development, 499 Toxicology 153652 (2023), https://doi.org/10.1016/j.tox.2023.153652.

105. The relative risk (RR) of angiosarcoma of the liver was ninety-seven in one of the first occupational cohort studies to demonstrate the association between occupational exposure to vinyl chloride and this very rare form of liver cancer. There is also suggestive evidence that vinyl chloride can contribute to the more common form of liver cancer (hepatocellular carcinoma), although the relative risks are much lower. See Ugo Fedeli et al., Occupational Exposure to Vinyl Chloride and Liver Diseases, 25 World J. Gastroenterology 4885 (2019), https://doi.org/10.3748/wjg.v25.i33.4885.

106. Sonja Klebe et al., Asbestos, Smoking and Lung Cancer: An Update, 17 Int’l J. Env’t Rsch. & Pub. Health 258 (2019), https://doi.org/10.3390/ijerph17010258.

per cell are not uncommon in the cells of the elderly. Certain of these mutations, often called driver mutations, are more commonly associated with cancer.

A major cause for the accumulation of mutations as we age is related to the frequency of mistakes made in copying DNA. These reflect the fallibility of normal cell DNA replication as well as the increasing failure with age of DNA repair enzymes to successfully carry out appropriate copy editing of the replicated DNA.

There is increasing evidence that many cancers are associated with heritable genetic differences among individuals. For example, women who inherit one mutated copy of the so-called “breast cancer genes,” BRCA1 and BRCA2, are at substantially greater risk of developing breast cancer when compared to women who have the common (“wild type”) genetic form.¹⁰⁷ The study of gene–environment interactions is of growing importance in the field of toxicology as well as medicine, as there is now substantial evidence that many diseases result from a combination of certain genetic variants and exposure to specific chemicals.¹⁰⁸ For many cancers, genetic variants that decrease the efficiency and accuracy of DNA repair are widely recognized as “susceptibility genes.” The evidence suggests that individuals with genetic deficiency in a DNA repair gene are likely to be more susceptible to the mutagenic effects of chemicals that cause the specific type of mutation repaired by the deficient DNA repair gene.

Estimates of what percent of individual cancer types are due to external versus internal factors are beginning to appear and can be expected to improve with advances in stem cell biology. (Unfortunately, cancers due to intrinsic factors, such as the failure of DNA repair, have been described as “bad luck” to distinguish them from cancers due to external factors.) This information will likely become increasingly important in considering the differential causation of individual cancer types.¹⁰⁹ It is likely that the introduction of potential genetic

107. Women who have mutated copies of BRCA genes have a lifetime risk of developing breast cancer of 45–75%, compared with a “background” lifetime risk of ~10–15%. About 5–10% of all breast cancers can be attributed to genetic risk. See Zora Baretta et al., Effect of BRCA Germline Mutations on Breast Cancer Prognosis: A Systematic Review and Meta-Analysis, 95 Medicine e4975 (2016), https://doi.org/10.1097/MD.0000000000004975.

108. See Nat’l Inst. Env’t Health Scis., Gene and Environment Interaction, https://perma.cc/X5HB-39TV, for a general discussion of how genes and the environment can interact to produce certain diseases.

109. A readily understandable account of the chemistry and biological implications of DNA repair is in the Popular Science Background piece accompanying the award of the 2015 Nobel Prize for Chemistry. DNA Repair—Providing Chemical Stability for Life, https://perma.cc/S9VL-9KRH. See also McManaway v. KBR, Inc., No. H-10-1044, 2012 WL 13059744 (S.D. Tex. Aug. 22, 2012) (expert failed to take into account whether alleged genetic transformation injuries caused by sodium dichromate were repaired, rendering his opinion on persistence of these injuries unreliable). A brief overview of the literature on the extent of estimated external and internal causes in different human cancer types is in Bernard D. Goldstein & Varun Patel, Controversy About the “Bad Luck” Cancer Hypothesis Could Lead to a Useful Tool for Planning Primary Prevention Cancer

susceptibility factors to environmental/occupational exposures to certain chemicals will become more common in toxic tort litigation, potentially used by plaintiff and defense lawyers alike to address the likelihood of a causal association between an exposure and a disease.

A relatively new and important area of genetic research, called epigenetics, has identified that there can be heritable changes (passed on from one generation to the next) in the functions of DNA that are not associated with a change in the DNA sequence (i.e., not the result of a mutation). Epigenetics (literally translated as “above genetics”) describes how the expression of DNA is regulated by several different biological processes.¹¹⁰ Importantly, there are a growing number of examples where chemicals found in the diet, workplace, and general environment can cause epigenetic changes that may contribute to a wide variety of adverse effects, including many types of cancers, diabetes and metabolic syndrome, neurological and cognitive dysfunction, as well as alterations in the functions of important organs such as the lung, the heart, the immune system, and reproductive organs. Indeed, a search of the scientific literature today shows over 127,000 scientific papers published in the general field of epigenetics. Although much of this scientific literature is focused on the normal (endogenous) functioning of epigenetic phenomena in chronic disease processes, there is growing recognition that many environmental and dietary factors can alter epigenetic processes, including some heavy metals, certain pesticides, some chemicals used in plastics, diesel exhaust, tobacco

Research, 32 Chem. Rsch. Toxicology, 949 (2019), https://doi.org/10.1021/acs.chemrestox.8b00390. For a discussion of methodologies for ruling out idiopathic (unknown) causes of a disease, see Hardeman v. Monsanto Co., 997 F.3d 941, 953 (9th Cir. 2021) (quoting Wendell v. GlaxoSmithKline LLC, 858 F.3d 1227, 1233–34 (9th Cir. 2017)) (ruling out idiopathy for disease with 70% idiopathy rate where expert relied on clinical experience, literature, and medical records).

110. For example, many of the nucleotide bases that are strung together to make up DNA can have methyl (-CH₃) groups attached to them in specific positions by enzymes known as DNA methyl transferases. DNA methylation functions somewhat like a switch, providing critical information on when a gene is “turned on” or “turned off.” In a similar manner, certain enzymes can modify proteins called chromatin that are intimately associated with DNA and help to regulate when specific genes are turned on or off (gene expression). One group of these chromatin proteins, called histones, can be modified by attaching, or detaching, acetyl groups on the histone proteins (histone acetylases and histone deacetylases), which then modulate whether the gene is expressed or not. This epigenetic process is known as histone modification (histone acetylation or deacetylation). Several other biological processes that also modify the expression of genes via epigenetic processes include protein phosphorylation, ubiquination, and sumolyation of chromatin proteins. Similar to histone acetylation, adding these different chemical entities to proteins that are involved in regulation of DNA function can modify how genes are expressed, without changing the actual sequence of the DNA, and thus are referred to as epigenetic effects. Other forms of RNA called short non-coding RNA and microRNAs have been found to be important epigenetic regulators of expression of DNA. See Lian Zhang et al., Epigenetics in Health and Disease, in Epigenetics in Allergy and Autoimmunity (Christopher Chang & Qianjun Lu eds., 2020), https://doi.org/10.1007/978-981-15-3449-2_1.

smoke, polycyclic aromatic hydrocarbons, hormones, radioactivity, viruses and bacteria, to name a few.¹¹¹

As the role of epigenetic factors in cancer and in other diseases unfolds, these might also be grist for the mill of both plaintiff and defense lawyers.

Medical History

Is the Medical History of the Individual Consistent with the Toxicologist’s Expert Opinion Concerning the Injury?

One of the basic and most useful tools in diagnosis and treatment of disease is the patient’s medical history.¹¹² A thorough, standardized patient information questionnaire would be particularly useful for identifying the etiology, or causation, of illnesses related to toxic exposures; however, there is currently no validated or widely used questionnaire that gathers all pertinent information.¹¹³ Nevertheless, it is broadly recognized that a thorough medical history involves the questioning and examination of the patient as well as appropriate medical testing. The patient’s written medical records also should be examined.

The following information is relevant to a patient’s medical history: past and present occupational and environmental history and exposure to toxic agents; lifestyle characteristics (e.g., use of nicotine and alcohol); family medical history (i.e., medical conditions and diseases of relatives); and personal medical history (i.e., present symptoms and results of medical tests as well as past injuries, medical conditions, diseases, surgical procedures, and medical test results).

111. See Bob Weinhold, Epigenetics: The Science of Change, 114 Env’t Health Perspectives A160 (2006), https://doi.org/10.1289/ehp.114-a160; Giacomo Cavalli & Edith Heard, Advances in Epigenetics Link Genetics to the Environment and Disease, 571 Nature 489 (2019), https://doi.org/10.1038/s41586-019-1411-0.

112. For a thorough discussion of the methods of clinical diagnosis, see John B. Wong et al., Reference Guide on Medical Testimony, in this manual. A number of cases have considered the admissibility of the treating physician’s opinion based, in part, on medical history, symptomatology, and laboratory and pathology studies. See, e.g., Morrow v. Brenntag Mid-South, Inc., 505 F. Supp. 3d 1287, 1290 (M.D. Fla. 2020) (declining to find expert testimony reliable where the expert, a physician, had not reviewed any “prior medical history or treatment records”); cf. Galvez v. KLLM Transp. Servs., LLC, 575 F. Supp. 3d 748, 760–61 (N.D. Tex. 2021) (finding expert testimony reliable where causation opinion was supported by a review of the patient’s medical history, part of “the well-established ‘scientific method’ of diagnosis in the medical community”).

113. Office of Tech. Assessment, U.S. Congress, Reproductive Health Hazards in the Workplace 8 (1985), at 365–89.

114. See McManaway v. KBR, Inc., No. H-10-1044, 2012 WL 13059744, at *6 (S.D. Tex. Aug. 22, 2012) (expert review of medical records and in-person interviews, including review of acute symptoms, supports a finding of reliability); Coene v. 3M Co., No. 10-CV-6546-FPG, 2017 WL 1046749, at *3, *9 (W.D.N.Y. Mar. 20, 2017) (finding toxicologist’s opinion regarding causation of silicosis to be reliable when the opinion was based on a review of the plaintiff’s medical records as well as conversations with the plaintiff, the MSDS sheets for the chemicals the plaintiff worked with that allegedly caused the injury, and technical literature); Ledbetter v. Blair Corp., No. 3:09–CV–843–WKW, 2012 WL 2464000 (M.D. Ala. June 27, 2012) (finding medical toxicologist’s opinion as to impairment caused by certain prescription drugs to be reliable when it was based, in part, on a review of medical and prescription records). Contra Gilbert v. Lands’ End Inc., No. 19-cv-823-jdp, 2022 WL 2643514, at *12–13 (W.D. Wis. July 8, 2022) (in a case alleging injury from required work uniforms, finding expert’s opinion unreliable when it was based only on a review of partial medical records and when the expert did not talk to the plaintiffs, “attempt to link any particular plaintiffs’ symptom to any particular garment,” or “rule out other potential causes”).

115. EPA, Recognition and Management of Pesticide Poisonings 45 (6th ed. 2013), https://perma.cc/Z4UL-9NCQ.

116. Complex issues of exposure to oil and dispersants, chemicals, fumes, and odors and reported adverse health effects are addressed in a medical settlement establishing a Periodic Medical Consultation Program and a Specified Physical Conditions Matrix as discussed in In re Oil Spill by Oil Rig Deepwater Horizon, 295 F.R.D. 112 (E.D. La. 2013) (medical class certified for settlement purposes only). An example of a disease that may be caused by exposure to toxins but may present only with nonspecific symptoms is Legionnaires’ Disease, explored at length in Mueller v. Chugach Federal Solutions, Inc., No. 12-S-00624-NE, 2014 WL 2891030 (N.D. Ala. June 25, 2014). There, the decedent’s wife alleged that her husband had been exposed to Legionella pneumophila at his workplace. Id. at *1. The court noted that the “disease may present early in the illness with nonspecific symptoms, so it can be difficult to diagnose.” Id. at *20. Indeed, the decedent in Mueller had been hospitalized for pneumonia and treated for the same prior to finally being diagnosed with Legionnaires’ Disease after a urine antigen test. Id. at *4. In opining that that decedent’s illness was

caused by exposure in the workplace, the expert was permitted to base his opinions on various materials produced during litigation, his own experience, and the fact that other employees also exhibited symptoms consistent with Legionnaires’ Disease. Id. at *6.

117. Failure to rule out other potential causes of symptoms may lead to a ruling that the expert’s report is inadmissible. See, e.g., Hardeman v. Monsanto Co., 997 F.3d 941, 953, 965 (9th Cir. 2021) (expert’s specific causation opinion was admissible where hepatitis C was ruled out as a cause of plaintiff’s non-Hodgkins lymphoma based on research studies and a temporal analysis of exposures and active disease); Pluck v. BP Oil Pipeline Co., 640 F.3d 671, 678–81 (6th Cir. 2011) (affirming district court’s conclusion that expert’s opinion was unreliable when the expert failed to rule out alternative causes of the plaintiff’s non-Hodgkins lymphoma); Scott v. Dyno Nobel, Inc., No. 4:16-CV-1440 HEA, 2021 WL 1750238, at *10–11 (E.D. Mo. May 4, 2021) (finding expert opinion reliable when it ruled out other possible causes of symptoms and noting that another expert’s opinion was found inadmissible for failure to do the same).

118. See, e.g., Myers v. United States, No. 02cv1349–BEN, 2014 WL 6611398, at *35–38 (S.D. Cal. Nov. 20, 2014) (explaining at length the necessity of reliable laboratory testing and reasons why tests may be found unreliable).

119. However, if a laboratory is certified for the testing of blood lead levels in workers, for which the OSHA action level is 40 micrograms per deciliter (µg/dl), it does not necessarily mean that it will give reliable data on blood lead levels at the much lower CDC Reference Level of 3.5

µg/dl (defined as the blood level used “to identify children with blood lead levels that are higher than most children’s levels,” CDC, Blood Lead Levels in Children, https://perma.cc/XY53-3A7K).

120. See, e.g., McManaway v. KBR, Inc., No. H-10-1044, 2012 WL 13059744 (S.D. Tex. Aug. 12, 2012) (expert conducted a differential diagnosis for each plaintiff, in which he ruled out other causes for their injuries, and also conducted blood testing to assist with differential diagnoses regarding sodium dichromate exposure); In re E.I. du Pont de Nemours & Co. C-8 Personal Inj. Litig., 348 F. Supp. 3d 680, 696 (S.D. Ohio 2016) (differential diagnosis is an “appropriate method for making a determination of causation for an individual instance of disease”) (citing Hardyman v. Norfolk & W. Ry. Co., 243 F.3d 255, 260 (6th Cir. 2001)); see also Milward v. Rust-Oleum Corp., 820 F.3d 469, 476 (1st Cir. 2016) (affirming exclusion of expert testimony as unreliable for failure to rule out an idiopathic diagnosis); Kovach v. Wheeling & Lake Erie Ry. Co., 556 F. Supp. 3d 762, 771 (N.D. Ohio 2021) (quoting Best v. Lowe’s Home Ctrs., Inc., 563 F.3d 171, 181 (6th Cir. 2009)) (noting that “doctors need not rule out every conceivable cause for their differential-diagnosis-based opinions to be admissible[,]” but they must at least consider and rule out alternative causes).

121. See generally Angelo Moretto et al., A Framework for Cumulative Risk Assessment in the 21st Century, 47 Critical Revs. Toxicology 85 (2017), https://doi.org/10.1080/10408444.2016.1211618; Devon C. Payne-Sturges et al., Methods for Evaluating the Combined Effects of Chemical and Nonchemical Exposures for Cumulative Environmental Health Risk Assessment, 15 Int’l J. Env’t Rsch. & Pub. Health. 2797 (2018), https://doi.org/10.3390/ijerph15122797.

enhancement of the effect of another agent, the response is termed potentiation.¹²² These interactions can often be explained if the metabolic pathways of the agent or the toxicological mechanism by which the individual agents cause their effects are known.

Three types of toxicological approaches are pertinent to understanding the effects of mixtures of agents. One is based on the standard toxicological evaluation of common commercial mixtures, such as gasoline. The second approach is from studies in which the known toxicological effect of one agent is used to explore the mechanism of action of another agent, such as using a known specific inhibitor of a metabolic pathway to determine whether the toxicity of a second agent depends on this pathway. The third approach is based on an understanding of the basic mechanism of action of the individual components of the mixture, thereby allowing prediction of the combined effect, which can then be tested in an animal model.¹²³

The toxicologist may also need to consider whether the mixture contains agents whose effects would allow estimation of exposure to a putative causative agent that was also part of the mixture. For example, the upper boundary for exposure to benzene present at a level of 0.1% in commercial grade toluene can be based on the level of toluene that causes central nervous system effects, including coma. A worker’s daily exposure of 1 part per million (ppm) benzene to this mixture would have imposed a concomitant exposure of approximately 1,000 ppm toluene—incompatible with the National Institute for Occupational Safety and Health (NIOSH) rating of Immediately Dangerous to Life and Health level for toluene of 500 ppm for fifteen minutes. Conversely, in a situation in which the worker did become woozy or comatose following exposure to a toluene-containing mixture, the toxicologist could use information about the level of toluene that caused such symptoms as a means to estimate the extent to which there was exposure to benzene or another contaminant of concern.

122. Courts have been called on to consider the issue of synergy. See Russell v. U.S. Dep’t of Labor, No. 3:15-CV-320-DCP, 2018 WL 2054566, at *8–9 (E.D. Tenn. May 2, 2018) (finding that the Department of Labor properly considered synergistic effects of four substances in arriving at its conclusion that plaintiff’s chronic obstructive pulmonary disease was not caused by exposure to the substances “individually or in combination”); but see In re E.I. Du Pont De Nemours & Co. C-8 Personal Inj. Litig., 337 F. Supp. 3d 728, 747 (S.D. Ohio 2015) (noting that the expert’s opinion on synergistic action between C-8 and obesity was “mere speculation” and, in this context, “synergism as a theory is unreliable and inadmissible under Daubert and Rule 702”).

123. See generally Moretto et al., supra note 121; Payne-Sturges et al., supra note 121. The EPA has been addressing the issue of multiple exposures to different agents within a community under the heading of cumulative risk assessment. This approach is particularly of importance in dealing with environmental justice concerns. See Michael A. Callahan & Ken Sexton, If Cumulative Risk Assessment Is the Answer, What Is the Question?, 115 Env’t Health Persps. 799 (2007), https://doi.org/10.1289/ehp.9330.

124. See generally Moretto et al., supra note 121; Payne-Sturges et al., supra note 121; see also Marissa B. Kosnik & David M. Reif, Determination of Chemical-Disease Risk Values to Prioritize Connections Between Environmental Factors, Genetic Variants, and Human Diseases, 379 Toxicology & Applied Pharmacology 114674 (2019), https://doi.org/10.1016/j.taap.2019.114674.

125. See, e.g., Supratim Choudhuri et al., From Classical Toxicology to Tox21: Some Critical Conceptual and Technological Advances in the Molecular Understanding of the Toxic Response Beginning from the Last Quarter of the 20th Century, 161 Toxicological Scis. 5 (2018), https://doi.org/10.1093/toxsci/kfx186.

126. See, e.g., Sarkees v. E.I. DuPont de Nemours & Co., 15 F.4th 584, 592–93 (2d Cir. 2021) (expert identified numerous factors that can elevate risk of bladder cancer and then sufficiently ruled them out, thereby establishing differential etiology); see also Cooper v. Smith & Nephew, Inc., 259 F.3d 194, 202 (4th Cir. 2001).

127. See, e.g., Mohamed A. Abdelmegeed et al., Role of CYP2E1 in Mitochondrial Dysfunction and Hepatic Injury by Alcohol and Non-Alcoholic Substances, 10 Current Molecular Pharmacology 207 (2017), https://doi.org/10.2174/1874467208666150817111114.

128. See, e.g., Waite v. AII Acquisition Corp., 194 F. Supp. 3d 1298, 1313, 1315 (S.D. Fla., 2016) (holding that the weight-of-evidence approach used by the expert was “thoroughly reasoned and based on sound methodology” and explaining that the approach “requires consideration of all available scientific evidence, including epidemiology, toxicology (animal studies), cellular studies (in vitro), and molecular biology”); In re Flint Water Cases, No. 17-10164, 2021 WL 5631706, at *7–8 (E.D. Mich. Dec. 1, 2021) (finding expert’s opinion on general causation reliable where the opinion was based on “reasonable scientific inferences” from “the great weight of the evidence” and where the expert “cite[d] to peer-reviewed, epidemiological studies which account for confounding variables”); In re Abilify (Aripiprazole) Prods. Liab. Litig., 299 F. Supp. 3d 1291, 1311 (N.D. Fla. 2018) (explaining that the “‘weight of the evidence’ approach to analyzing causation can be considered reliable, provided the expert considers all available evidence carefully and explains how the relative weight of the various pieces of evidence led to his conclusion . . . , it is crucial that the expert describe each step in the process by which he gathered and assessed the scientific evidence”). But see Daniels-Feasel v. Forest Pharms., Inc., No. 17 CV 4188-LTS-JLC, 2021 WL 4037820, at *5 (S.D.N.Y. Sept. 3, 2021), aff’d, No. 22–146, 2023 WL 4837521 (2d Cir. July 28, 2023) (quoting In re Abilify, 299 F. Supp. 3d at 1311) (expert’s opinion that relies on cherry-picked data is not reliable). See generally Liesa L. Richter & Daniel J. Capra, The Admissibility of Expert Testimony, in this manual.

toxicology. Toxicology is a heterogeneous field. A number of indicia of expertise can be explored, however, that are relevant to both the admissibility and weight of the proffered expert opinion.

Does the Proposed Expert Have an Advanced Degree in Toxicology, Pharmacology, or a Related Field? If the Expert Is a Physician, Are They Board Certified in Toxicology or in a Field Such as Occupational Medicine?

A graduate degree in toxicology demonstrates that the proposed expert has a substantial background in the basic issues and tenets of toxicology. Many universities have established graduate programs in toxicology. These programs are administered by the faculties of medicine, pharmacology, pharmacy, or public health.

In addition to toxicologists with a specific Ph.D. in toxicology, many highly qualified toxicologists are physicians or hold doctoral degrees in related disciplines (e.g., veterinary medicine, pharmacology, biochemistry, environmental health, or industrial hygiene). For a person with this type of background, a single course in toxicology is unlikely to provide sufficient background for developing expertise in the field. Additional specific training and board certification in toxicology is highly relevant to evaluating the credentials of a toxicologist.

A proposed expert should be able to demonstrate an understanding of the discipline of toxicology, including statistics, toxicological research methods, and disease processes. A physician without particular training or experience in toxicology is unlikely to have sufficient background to evaluate the strengths and weaknesses of toxicological research. Most practicing physicians have little knowledge of environmental and occupational medicine.¹²⁹ Subspecialty physicians may have particular knowledge of a cause-and-effect relationship (e.g., pulmonary physicians have knowledge of the relationship between asbestos exposure and asbestosis); anesthesiology is largely focused on the effects of certain chemicals; and certified addiction medicine physicians focus on chemical agents that are not pharmaceuticals or are illegally used pharmaceuticals, such as

129. For recent documentation of how rarely an occupational history is obtained, see Barry J. Politi et al., Occupational Medical History Taking: How Are Today’s Physicians Doing? A Cross-Sectional Investigation of the Frequency of Occupational History Taking by Physicians in a Major U.S. Teaching Center, 46 J. Occupational Env’t Med. 550 (2004), https://doi.org/10.1097/01.jom.0000128153.79025.e4.

fentanyl.¹³⁰ Physicians who receive additional training in classic toxicology include those who are certified by the American Board of Preventive Medicine in medical toxicology, addiction medicine, or occupational and environmental medicine.¹³¹ Knowledge of toxicology is particularly strong among occupational health specialists who work in the chemical, petrochemical, and pharmaceutical industries, in which the surveillance of workers exposed to chemicals is a major responsibility.¹³²

130. Courts look at specific circumstances in determining whether a treating physician may testify as a fact witness or an expert witness. See, e.g., Mueller v. Chugach Fed. Solutions, Inc., No. 12-S-00624-NE, 2014 WL 2891030, at *3–4 (N.D. Ala. June 25, 2014) (permitting treating physician to testify as fact witness and offer his “opinion regarding the cause of death . . . based upon his personal experience of treating” the decedent); but see N.K. by Bruestle-Kumra v. Abbott Lab’ys, 731 F. App’x 24, 26 (2d Cir. 2018) (noting that a treating physician may not testify as a fact witness and must “be admitted as a Rule 702 expert witness to provide expert testimony on specific causation”). Some courts have allowed treating physicians to testify regarding causation. See, e.g., Lassalle v. McNeilus Truck & Mfg., Inc., No. 16-cv-00766-WHO, 2017 WL 3115141, at *2 (N.D. Cal. July 21, 2017) (permitting the treating physician to offer his opinion regarding causation because it could have been formed within the scope of treatment); Burton v. Am. Cyanamid, 362 F. Supp. 3d 588, 599 (E.D. Wis. 2019) (allowing treating physician to testify regarding lead exposure and causation because “his professional training and his experience treating . . . patients with elevated levels are sufficient to qualify him to give testimony”). Contra Zellers v. NexTech Ne., LLC, 533 F. App’x 192, 198 (4th Cir. 2013) (excluding testimony of treating physician as to causation because, inter alia, the physician was a neurologist who did not have “specialized training in the field of toxicology”); Higgins v. Koch Dev. Corp., 794 F.3d 697, 704–05 (7th Cir. 2015) (excluding treating physician opinion regarding causation because the treating physician had no “training in toxicology”).
Treating physicians also become involved in considering cause-and-effect relationships when they are asked whether a patient can return to a situation in which an exposure has occurred. The answer is obvious if the cause-and-effect relationship is clearly known. However, this relationship is often uncertain, and the physician must consider the appropriate advice. In such situations, the physician will tend to give advice as though the causality was established, both because it is appropriate caution and because of fears concerning medicolegal issues.

131. Before 1990, the American Board of Medical Toxicology certified physicians. But beginning in 1990, medical toxicology became a subspecialty board under the American Board of Emergency Medicine, the American Board of Pediatrics, and the American Board of Preventive Medicine, as recognized by the American Board of Medical Specialties.

132. Clinical ecologists, another group of physicians, have offered opinions regarding multiple chemical hypersensitivity and immune system responses to chemical exposures. These physicians generally have a background in the field of allergy, not toxicology, and their theoretical approach is derived in part from classic concepts of allergic responses and immunology. This theoretical approach has often led clinical ecologists to find cause-and-effect relationships or low-dose effects that are not generally accepted by toxicologists. Clinical ecologists often belong to the American Academy of Environmental Medicine.
In 1991, the Council on Scientific Affairs of the American Medical Association concluded that until “accurate, reproducible, and well-controlled studies are available . . . multiple chemical sensitivity should not be considered a recognized clinical syndrome.” Council on Sci. Affairs, Am. Med. Ass’n, Council Report on Clinical Ecology 6 (1991). Multiple chemical sensitivity has been held by courts to be “a controversial diagnosis that has been excluded under Daubert as unsupported

by sound scientific reasoning or methodology.” Madej v. Maiden, 951 F.3d 364, 374–75 (6th Cir. 2020) (quoting Summers v. Mo. Pac. R.R. Sys., 132 F.3d 599, 603 (10th Cir. 1997)) (collecting cases). The Madej court recognized that the “mountain of precedent” regarding multiple chemical sensitivity consists mostly of cases “over a decade old[,]” but reiterated that it has not been “shown that more recent scientific advancements have led the scientific community to come to accept a multiple-chemical-sensitivity diagnosis.” 951 F.3d at 375. The “diagnosis remains unrecognized by the American Medical Association and unlisted in the World Health Organization’s International Classification of Diseases.” Id.

133. As of 2024, there are twenty-nine specialty sections of the SOT that represent the different specialty areas involved in understanding the wide range of toxic effects associated with exposure to chemical and physical agents. These sections include Mechanisms, Molecular and Systems Biology, Inhalation and Respiratory, Metals, Neurotoxicology, Carcinogenesis, Risk Assessment, Biological Modeling, Immunotoxicology, and Sustainable Chemicals through Contemporary Toxicology, to name a few.

134. Examples of reputable, peer-reviewed journals are the Journal of Toxicology and Environmental Health; Toxicological Sciences; Toxicology and Applied Pharmacology; Science; British Journal of Industrial Medicine; Clinical Toxicology; Archives of Environmental Health; Journal of Occupational and Environmental Medicine; Annual Review of Pharmacology and Toxicology; Teratogenesis, Carcinogenesis and Mutagenesis; Fundamental and Applied Toxicology; Inhalation Toxicology; Biochemical Pharmacology; Toxicology Letters; Environmental Research; Environmental Health Perspectives; International Journal of Toxicology, Human and Experimental Toxicology; Regulatory Toxicology and Pharmacology; and American Journal of Industrial Medicine.

university has a graduate education program in that area. However, some universities offer adjunct or affiliate faculty positions to local professionals (consultants, government employees, etc.), often to assist in teaching entry-level toxicology courses. Although many well-qualified toxicologists working in the private sector or government agencies often have adjunct academic appointments, such an appointment does not necessarily confer the same level of expertise or toxicology credentials as a regular faculty appointment.

Acknowledgments

The authors gratefully appreciate the excellent research assistance provided by Stephanie N. Patton, associate, Buchanan Ingersoll & Rooney, P.C.; Lyndsey Arneson Field, Wake Forest Law School class of 2024; and Marvin Astrada, senior research associate, Federal Judicial Center, in the preparation of this fourth edition of the reference guide.