Hormesis: Possible Legal Implications of Hormesis

Wendy E. Wagner

Professor, University of Texas School of Law

727 Dean Keeton St.

Austin, TX 78705-3299

Email: wagner9@attglobat.net

Professor Cross provides an exceedingly clear and accessible account of the largely neglected intersection of U.S. risk regulation and hormesis. He raises a number of difficult and important questions about the future of environmental regulation in a world where hormesis dominates risk assessments. I must begin my commentary, then, with a hearty round of applause for his valuable contribution.

In the remainder of the commentary, I accept Professor Cross's invitation to explore how or whether hormesis will change the regulatory landscape. After highlighting several key assumptions regarding hormesis that I refrain from challenging, but that dramatically affect the analysis, I then proceed to posit a different view of how hormesis might affect current regulatory practices. My general conclusion is that while as a scientific matter hormesis may matter a great deal, as a regulatory matter it likely will add only more complexity to an already overwhelmed system. Because it will produce more questions than answers and is only one factor in a much longer list of considerations about how best to conduct environmental regulation, I suspect that hormesis will change little about how we regulate toxic substances in the foreseeable future.

Critical Assumptions about Hormesis

Purely for the sake of argument, I make several assumptions about hormesis in this commentary that appear to be consistent with Professor Cross's assumptions. Because they are fundamental to the legal implications of hormesis, however, I want to be clear about what these assumptions are and underscore the possibility that they may not yet be resolved. First, I assume that hormesis is a validated theory that is generally accepted within the scientific community. If there is a significant debate within the scientific community about whether the hormesis hypothesis constitutes "the best available science,"1 however, then hormesis will likely not be integrated into environmental regulatory policy, or it will only be integrated to the extent that it suggests that more protective regulations are in order.

Second, I assume for the sake of argument that the very low concentrations at which beneficial effects are maximized (Et on Cross's Figure 4) tend to be well above background concentrations, even though it is not clear from the literature whether this is the case.2 For example, it is not clear how substances interact with each other, particularly with respect to the beneficial effects portion of the hormetic curve. (The literature does reveal that the general "very low" exposure level at which beneficial responses to a single contaminant occur is approximately 4 to 5 times below the no adverse effects level (NOAEL)).3 Yet if substances within a family react similarly (or in an additive fashion), then for some families of substances it seems possible that these "very low" additive exposure levels that maximize beneficial response (Et) are close to background levels already found in the environment.1 Indeed, to the extent that Eo does approach background levels, hormesis is likely to suggest standards that are uniformly more stringent because of the steeper curve of harms that occur after this low, background concentration is exceeded.

Hormesis and Regulatory Policy in the U.S.

Having made my working assumptions explicit, I can turn to the crux of Professor Cross's argument that hormesis will significantly change the regulatory landscape and might actually "dejustify some regulatory actions." Both in the short and long term, I agree that hormesis could and perhaps should change the way toxicology studies are conducted and the nature of the data that is collected.

As a matter of regulatory policy, however, I doubt that hormesis will make much difference to environmental regulations in the U.S., even in an ideal world of pure motives and complete competency. This is because it seems unlikely that hormesis will provide a distinctly different picture of risk than the exceedingly fuzzy picture that already exists. Even if hormesis does ultimately justify making the range of plausible health or environmental standards less stringent, however, it still may not change regulatory policy in most settings because the result of a risk assessment is but one of a number of different considerations that are used to determine the best regulatory solution to a particular environmental problem.

Hormesis will make risk assessments more, rather than less, uncertain and variable

First, hormesis appears more likely to exacerbate, rather than alleviate, the current uncertainties that plague risk assessment. This feature will not only make hormesis unpopular to the regulatory world, but it might also serve to undercut the confidence one can place in the different exposure levels it recommends. This greater uncertainty and variability in a hormetic risk assessment occur for a number of reasons. Perhaps the largest source of uncertainty is the paucity of toxicological information on most substances, particularly with respect to information on beneficial responses at low doses.4,5 At least over the next few decades, then, there will be a dearth of data for determining points Et, Em, and Eo (see Professor Cross's Figure 4) for any given substance or family of substances. Without data, however, these exposure levels must be based on unvalidated guesses. The possibility of variations in the shape of the "U" (or "J") curves among substances or classes of substances6 adds further complications when toxicity data is scarce, made all the more precarious due to the necessity of extrapolating the scant data on laboratory organisms to humans.2 The possibility of synergistic (or additive) reactions between toxic substances raises still other critical, but unresolved questions in risk assessors' efforts to integrate the low dose beneficial effects into a risk assessment, 1 while variability among the population with respect to the Et, Em, and Eo exposures6 substantially complicates the ability of risk assessors to narrow the exposure level that will maximize beneficial responses among the population.

Finally, since hormesis predicts that at least some of the benefits that occur at low levels are different in type from the harms, the regulatory agency will also face a number of complicated decisions about how to define, measure, and compare these benefits against the harms. The agency will first have to identify a uniform and measurable endpoint or series of endpoints for the beneficial effects of a substance, an endeavor that the literature suggests is far from straightforward.1,7 Second, the agency must ensure that the beneficial effects (such as elevated defense mechanisms or longevity) do not compromise other survival features, such as long-term defenses or reproduction. Yet this critical issue also appears unresolved, at least at the ecological level.1,8 Finally, to the extent that there is some variability among the population with respect to Eo, the agency will be required to balance one or more types of benefits against different types of harms. Yet trading one person's longevity against another person's increased risk of incurring cancer is an exercise that is fraught with policy judgments that may be difficult, if not impossible, to resolve.9

Replacing the linear default model with the hormetic model requires value choices that diverge from those currently in place in our laws and regulations

Professor Cross's apparent response to these messy hormetic risk assessments is to sidestep the rampant uncertainties and variabilities entailed in determining hormetic curves on a substance-by-substance basis10 by substituting a general hormetic default curve for the linear default dose-response curve currently in use. Professor Cross's advocacy for the hormetic curve as the default dose-response model not only seems to ignore the many unanswered questions that hormesis generates, but it misframes the selection of default curves as predominately a scientific exercise, despite the National Academy of Sciences' repeated admonitions against doing so when discussing competing default models.11,12 In truth, the decision to replace the linear default curve with the hormetic default curve, at least at this point in the science of hormesis, is at least as much a policy judgment as a scientifically based conclusion.12 Although biological plausibility is certainly one consideration in selecting among alternative dose-response curves, because the underlying science is very incomplete and incapable of being validated, the biological plausibility of alternative models competes with other considerations, such as a preference to err on the side of the public health.12 While it is important to revisit our environmental policies when presented with new information such as hormesis, the terms of the discussion (i.e., scientific uncertainties and variabilities associated with each alternative theory; potential consequences for current protective levels) should be made explicit and accessible and not buried beneath expert dialogue.

Of still greater import in predicting the regulatory implications of hormesis, however, is the fact that substituting a hormetic default curve for the linear model conflicts with the precautionary preferences currently codified in most of our current laws and regulations.12,13 This does not mean that we are locked into the values of our past. However, it is important to acknowledge that because the selection of the hormetic curve as a default model for dose-response seems to diverge from past policies, advocates of this new approach will need to muster strong arguments in its favor before hormesis achieves the rank of an official "default option." They will thus not only need to present evidence that the approach is clearly superior from a scientific perspective, but also that conservative precautions codified in current regulatory policies are unwise or will be unaffected by substitution of the hormetic model. Yet the many unresolved features of both the science and policy implications of hormesis make it difficult to advance these more focussed arguments, much less to support them.

The choice of environmental regulatory tools is based on a number of considerations, only some of which relate to science

Even if concerns about the uncertainty and variability associated with hormetic risk assessments are misplaced, it seems doubtful that hormesis will bring about the dramatic regulatory changes that Professor Cross envisions. Regulatory policy is not only based on science and value choices, but on practical considerations, such as whether a regulatory tool will provide fair and predictable requirements for regulated sources.14 In my view, Professor Cross neglects these various other factors that influence the selection of regulatory instruments when he predicts an almost revolutionary shift from feasibility-based standards to market or ambient approaches. For example, Professor Cross suggests that hormesis should cause feasibility standards (such as technology-based standards, which require sources to reduce pollution to the level achieved by installing the "best available pollution control technology") to be replaced with a "risk cup" approach that ties discharges and emissions to the assimilative capacity of the environment and public health. Yet technology-based standards are used in large measure because they can be employed in an expeditious fashion with minimal administrative costs for multiple types of sources,15 not because the linear dose-response model for a substance produces standards with damaging economic consequences. Similarly, markets in pollution prove most effective under circumstances that changes in the underlying risk assessment for a pollutant are unlikely to affect.14,16 Indeed, the most successful pollution market occurred for lead, where the ultimate goal was reducing lead additives in gasoline to zero,17 a level that Professor Cross implies might not be suitable for market trading.


In summary, although I agree that advances in science should influence environmental law and policymaking, it is not clear that the relationship between science and law is as direct, immediate, or simple as Professor Cross implies. Hormesis may be an important new theory for science, but its impact on environmental regulation will likely be much more subtle and, at least in the foreseeable future, much more modest that Professor Cross suggests.


1 Elliott KC, A Case for Caution: An Evaluation of Calabrese and Baldwin's Studies of Chemical Hormesis. Risk: Health, Safety & Environment 2000; 11:177-196.

2 Davis JM, Farland WH, Biological Effects of Low-level Exposures: A Perspective from U.S. EPA Scientists. Environmental Health Perspectives 1998; 106 (Suppl 1):379-381.

3 Calabrese EJ, Baldwin LA, Holland CD, Hormesis: A Highly Generalizable and Reproducible Phenomenon With Important Implications for Risk Assessment. Risk Analysis 1999; 19:261-281.

4 Calabrese EJ, Baldwin, LA. The Dose Determines the Stimulation (and Poison): Development of a Chemical Hormesis Database. International Journal of Toxicology 1997; 16:545-559 (summarizing that "the reason for the uncertainty surrounding the existence of hormesis as a 'real phenomenon' is believed to be the result of its relatively infrequent observation in the literature due to experimental design considerations, especially with respect to the number of doses, range of doses, and endpoint selection").

5 National Research Council, Toxicity Testing: Strategies to Determine Needs and Priorities. National Academy Press: Washington, DC, 1984 (detailing the general paucity of toxicity testing on chemicals in commerce). For more recent studies on the dearth of toxicity testing on high volume chemicals in commerce, see Environmental Defense Fund, Toxic Ignorance. Graphic Illusions: Dennisport, Mass, 1997, and Environmental Protection Agency, Office of Pollution Prevention and Toxics. What do we really know about the safety of high production volume chemicals? Chemical Regulation Reporter 1998; 22: 261-78.

6 Calabrese DJ, Baldwin LA. A general classification of U-shaped dose-response relationships in toxicology and their mechanistic foundations. Human & Experimental Toxicology 1998; 17:353-364 (endeavoring to provide a general classification of U-shaped dose-response relationships that emerge from the diverse and distinctive hormetic responses reported in the literature). At page 362 the authors note that "[t]he adaptive response is likely to vary based on age, sex, diet and health status and this will contribute to observations of variability in susceptibility in the population."

7 Calabrese EJ, Baldwin LA, Holland CD. Hormesis: A Highly Generalizable and Reproducible Phenomenon With Important Implications for Risk Assessment. Risk Analysis 1999; 19:261-281. At page 264, based on their comprehensive review of the literature on hormetic reactions, the authors observe that the "range of biological effects observed with respect to hormesis is also widespread and includes growth, longevity, reproduction, disease incidence, and behavioral aspects."

8 Calabrese EJ. Expanding the Reference Dose Concept to Incorporate and Optimize Beneficial Effects While Preventing Toxic Responses from Nonessential Toxicants. Regulatory Toxicology and Pharmacology 1996; 24:S68-S75. At page S74, the author cites work by Caslow suggesting that "even though a 'beneficial' response (e.g., survival) may be observed at low doses it does not come without some metabolic costs which my affect growth and/or reproductive production," and as a result advocates "the need for a more detailed, ecologically based critique of the present proposal."

9 Calabresi G., Bobbitt P. Tragic Choices: The conflicts society confronts in the allocation of tragically scarce resources. W.W. Norton & Company: New York, NY, 1978.

10 Calabrese EJ, Baldwin, LA. Hormesis as a default parameter in RfD derivation. Human & Experimental Toxicology 1998; 17:444-447. On page 446 the authors state that "If hormesis were required to be 'proven' for each agent, . . . this would place such an extensive demand on the sponsoring group that the specific integration of hormesis in any RfD derivation would be relegated to very rare exceptions."

11 Committee on the Institutional Means for Assessment of Risks to Public Health, National Research Council, Risk Assessment in the Federal Government: Managing the Process. National Academy Press: Washington, DC, 1983.

12 Committee on Risk Assessment of Hazardous Air Pollutants, National Research Council, Science and Judgment in Risk Assessment. National Academy Press: Washington, DC, 1994. At page 7 in the executive summary, the committee observes that because default options "are used in the absence of convincing scientific knowledge on which of several competing models and theories is correct" they "inevitably involve[] policy choices" for their resolution. The committee also lists at least some of the criteria that should be taken into account when selecting among default models: "protecting the public health, ensuring scientific validity, minimizing serious errors in estimating risks, maximizing incentives for research, creating an orderly and predictable process, and fostering openness and trustworthiness."

13 Applegate JS. The Precautionary Preference: An American Perspective on the Precautionary Principle. Human and Ecological Risk Assessment 2000; 6:413-443.

14 Office of Technology Assessment, Congress of the United States, Environmental Policy Tools: A User's Guide pp. 145-148. Government Printing Office: Washington, DC, 1995 (listing as the seven factors used to evaluate environmental regulatory tools ­ "assurance of meeting goals," "environmental equity and justice," "fairness and cost-effectiveness," "demands on government," "adaptability," "pollution prevention," and "technology innovation and diffusion.").

15 Wagner WE. The Triumph of Technology-Based Standards. University of Illinois Law Review 2000; 2000:83-113 (citing considerable literature on when technology-based standards are effective).

16 Foster V, Hahn RW, Designing More Efficient Markets: Lessons from Los Angeles Smog Control, Journal of Law and Economics 1995; 38:19-44.

17 Ackerman BA, Stewart, RB. Reforming Environmental Law. Stanford Law Review 1985; 37:1333-65 (see pages 1348-49 for a discussion of the lead phase-out market).