Comments on Legal Implications of Hormesis

Susan R. Poulter*

Professor of Law

University of Utah College of Law

332 S. 1400 East Front

Salt Lake City, UT 84112

Phone: 801-581-6545

Fax: 801-581-6897


Renewed interest in hormesis comes at a time when the bases for risk assessment and risk-based standards are being reconsidered. EPA recently proposed to revise its carcinogen risk assessment guidelines1; the methodology of noncancer risk assessment is undergoing revision;2,3 and at least some courts seem to be holding EPA and other regulatory agencies to heightened requirements in setting environmental and health standards.3,4

Professor Cross's proposal5 sets forth a thoughtful framework for incorporating information on hormesis into risk assessment and standard setting. Incorporating evidence of hormesis in these processes could produce a number of beneficial results, including the demonstration that threshold doses exist, and even the possibility of optimizing exposures at levels where net beneficial effects are greatest.6 Professor Calabrese has even suggested that hormesis might become a default assumption, either generally or for certain classes of chemicals.7

At the present time, however, the theory and evidence for hormesis need further development before hormesis is incorporated into health and safety regulation. It appears that hormesis has not gained widespread acceptance among toxicologists and risk assessors, and further scientific investigation is needed before a mechanistic understanding can be gained.8 There are also scientific questions about the significance of hormesis in an environment where most of the population already experiences background exposure to many toxic substances. Additionally, there are policy issues about the meaning of conservative assumptions in assessing beneficial effects and allocation of burdens of proof. The purpose of this commentary is to raise those issues for discussion.

Should Resources be Invested in Hormesis Research?

There is a considerable body of opinion that investments in pollution control and safety measures should be prioritized,9,10 or even balanced against the expected benefits by utilizing cost-benefit analysis.11 Although these proposals have engendered considerable criticism, they are being utilized in various regulatory contexts.12 Information costs, in this case, the costs of assessing and quantifying hormesis effects, would have to be included in determining priorities and in balancing potential costs and benefits. These costs are likely to be considerable.

Assessing hormesis effects requires more toxicology testing than is currently conducted or available for most substances; more studies at low doses are required, and these studies require larger numbers of test subjects, in order to achieve statistically significant results.6 Not to put too fine a point on it, with limited resources, the issue is whether it is better to conduct the current panoply of bioassays on several previously uninvestigated substances, or to expend the same resources evaluating possible hormetic effects of only one such substance.

Of course, the cost of regulating toxicants at very low levels also increases steeply as permitted levels are decreased10 and identification of hormetic effects may justify less stringent standards, although that will not always be the case.5 Thus, it may be worthwhile for regulated industries to produce the necessary research, and perhaps they could be permitted to submit such information, without requiring that the regulatory agency undertake such research as a precondition to setting a standard.

How does Hormesis Work Where There are Multiple Exposures?

Incorporating an assessment of hormesis into standard setting also raises some knotty problems in determining how standards relate to actual human exposures. Professors Cross and Calabrese have both noted the problem of translating ambient standards, which are usually risk-based and therefore are the appropriate locus for consideration of hormesis, into emission standards for sources, standards that are generally technology- or feasibility-based.5,6 There is the further complication, however, that ambient standards usually represent maximum permitted levels, rather than actual exposures. Mean or median exposures in an area in compliance with the ambient standard should generally be less than the ambient standard, and usually would vary over time and geographic area.13

Further, exposure levels tend to vary among exposed populations for other reasons, such as differences in residence times, respiration or consumption rates, activity levels, and other factors.13 While these variations can be modeled,14 the complexity of superimposing an exposure distribution on a toxicity curve that includes hormesis is not going to make standard setting easier. Moreover, it will likely raise questions about whether some groups (those whose risk levels are optimized) are benefitted at the expense of others (the "tail" of the risk distribution whose exposures would be at a level where harm is more likely than benefit).15

Even where hormesis is established for individual substances, there is likely to be a lag before scientists have a clear understanding of how dose-response curves are shaped in the real world of multiple exposures. Would low-dose exposure to one substance saturate the hormesis region for another related substance, or would they act independently? If hormetic effects occur at background levels of exposures, the effects of additional exposures may occur primarily in the "above-the-line" region of harmful effects.16 In such a situation, including hormesis should not result in less stringent standards. If savings in control costs are unlikely to be realized, the perceived benefits of investigating hormetic effects may be lessened.

What Risk Assessment Assumptions Are "Conservative" for Hormesis?

Hormesis also presents challenges to the prevailing use of "conservative" assumptions in low-dose risk assessment, particularly the use of the lower 95% confidence level of the dose-response curve.1 What kind of assumption about the hormesis region represents a "conservative" (i.e., more protective) standard? Should risk assessors use the maximum likelihood estimate of hormesis, or one that represents the upper or lower 95% confidence level of the maximum effect? Or should the precautionary principle embedded in the use of "conservative" assumptions be abandoned?17 These questions need to be addressed,18 and incorporation of hormetic effects into risk assessment and standard setting requires that a rationale be developed for these science policy choices.19

How to Get There from Here

As the foregoing suggests, there are a number of obstacles to be overcome before hormesis can be fully incorporated into risk assessment and environmental standards. The first is establishing, to the satisfaction of the scientific community, that hormesis is a frequent if not universal phenomenon. Calabrese and Baldwin have surveyed the history of the theory of hormesis and identified a number of reasons why hormesis has received less attention and perhaps acceptance than it deserves.8 Although they make a good case that hormesis is probably much more common than has been assumed, it appears that the scientific case is not yet convincing to toxicologists and regulators, perhaps in part because the mechanistic explanations are still largely unverified.

Another hurdle to be overcome is that regulators are likely to resist changes that make their work more complicated, divert scarce resources from their current priorities, or increase the risk of protracted and highly contentious rulemaking, exemplified in a number of recent cases.20,21 Thus, changing risk assessment and standard setting to routinely include consideration of hormesis will not only require convincing evidence of scientific merit, but will also require that regulators be convinced that obtaining and utilizing hormesis information will produce better standards and will not make their decisions more vulnerable to challenge. Consideration needs to be given to the questions of who is obligated or permitted to produce evidence of hormesis, as well as how complete the evidence must be in order to modify the default assumptions of the current risk assessment paradigm.

The foregoing is not intended to suggest that hormesis should be dismissed and ignored - where hormetic effects exist, they would be significant in a number of ways. At the least, evidence that a substance exhibits hormesis would demonstrate that net adverse effects have a threshold, an issue that has been particulary controversial where carcinogens are concerned,1,21 and one that may dramatically affect concepts of "safe" levels and the corresponding control costs. Further, hormetic effects could be important, perhaps essential, in cost benefit analysis of regulatory controls, especially where a range of exposure scenarios is under consideration.13 More generally, hormesis has implications for the design of epidemiologic studies and animal bioassays, since the hormesis effects of low dose exposures may cancel out the adverse effects at higher doses and make it less likely that a study will detect those adverse effects.22,23

At the very least, more research should be done on hormesis - to determine how commonly hormetic effects occur, their magnitude, and the mechanisms through which they operate. Perhaps a small number of demonstration projects could be funded to develop data on hormesis for risk assessment. That kind of research will provide a sound basis for working out the significance of hormesis and the policy choices that are necessary for incorporating hormetic effects in risk assessment and standard setting. Professor Cross's framework provides an excellent starting point for that discussion.


1. Environmental Protection Agency, Proposed Guidelines for Carcinogen Risk Assessment, 61 Fed. Reg. 17960 (1996). The guidelines have not been made final as of this writing.

2. Environmental Protection Agency, Notices, Benchmark Dose Peer Consultation Workshop, 61 Fed. Reg. 44308 (1996).

3. AFL-CIO v. OSHA, 965 F.2d 962 (11th Cir. 1992)(OSHA's use of safety factor analysis to establish permissible exposure levels was held to have been inadequately explained)

4. American Trucking Ass'ns, Inc. v. U.S. E.P.A., 175 F.3d 1027 (D.C.Cir.1999), cert. granted, 120 S.Ct. 2003 (2000) (statutory delegation of authority invalid where EPA failed to articulate an intelligible standard for setting ambient air quality standard for ozone).

5. Cross FB, Legal Implications of Hormesis (this issue).

6. Calabrese EJ, Baldwin LA, Holland CD, Hormesis: A Highly Generalizable and Reproducible Phenomenon with Important Implications for Risk Assessment, Risk Anal.19, no. 2: 261, 273(1999).

7. Calabrese EJ and Baldwin LA, Hormesis as a Default Parameter in RfD Derivation, Human and Exptl. Toxicology (1998) 17, 444-447.1.

8. Calabrese EJ, Baldwin LA, The Marginalization of Hormesis, Toxicologic Pathology (1999) 27, no. 2:187-194. The author takes no position on the scientific validity of hormesis. New scientific theories that do not fit the prevailing paradigm (in toxicology, the prevailing theory is generally that the lower the dose, the lesser the effect) often encounter resistance for a period of time.

9. U.S. EPA, Unfinished Business: A Comparative Assessment of Environmental Problems (1987).

10. Breyer S, Breaking the Vicious Circle: Toward Effective Risk Regulation (1993).

11. Risk vs. Risk: Trade-Offs in Protecting Health and the Environment (Graham JD, Weiner JB, eds., 1995).

12. Applegate JS, Comparative Risk Assessment and Environmental Priorities Projects: A Forum, Not a Formula, Northern Kentucky Law Review1997, 25:71.

13. National Research Council, Science and Judgment in Risk Assessment, 1994.

14. Poulter SR, Monte Carlo Simulation in Environmental Risk Assessment: Science, Policy and Law, Risk: Health, Science and Law Winter, 1998, 9:7.

15. Lave LB, Hormesis: Policy Implications, Journal of Applied Toxicology 2000, 20(2):141-145.

16. Paperiello CJ, Risk Assessment and Risk Management Implications of Hormesis, Journal of Applied Toxicology 2000, 20(2):147-148.

17. Cross FB, Paradoxical Perils of the Precautionary Principle, Wash. & Lee L. Rev. 1996, 53:851.

18. Sielken RL Jr., Stevenson DE, Some Implications for Quantitative Risk Assessment if Hormesis Exists, Human & Exptl. Toxicol. 1998: 17(5):259-262.

19. Wagner WE, The Science Charade in Toxic Risk Regulation, Colum L. Rev., 1995 95:1613.1.

20. Corrosion Proof Fittings v. EPA, 947 F.2d 1201 (5th Cir. 1991)(invalidating EPA ban on asbestos after 10 years of rulemaking).

21. Chlorine Chemistry Council v. E.P.A., 206 F.3d 1286 (D.C.Cir. 2000)(vacating drinking water standard for chloroform based on assumption of low-dose linearity).

22. Cook R, personal communication.

23. Bailer AJ, Oris, JT, Incorporating Hormesis in the Routine Testing of Hazards, Human & Experimental Toxicology 1998, 17(5): 247-50.

* This work was supported by the College of Law Research Fund. The author thanks Professors Calabrese and Cross for the opportunity to submit these comments and appreciates the comments of Ralph Cook, RC Consulting, on the significance of hormesis in risk assessment.