The Hormesis Challenge for Environmental Health Regulators
Michael Baram, Professor
Boston University School of Law
Center for Law and Technology
765 Commonwealth Ave., Room 866
Boston, MA 02215
Boston University School of Law
Center for Law and Technology
765 Commonwealth Ave., Room 866
Boston, MA 02215
The hormesis hypothesis provides that human, animal or plant exposure to a low external dose of a hazardous chemical substance or physical agent may have beneficial health outcomes1. For example, human exposure to a precisely defined, low external dose of a carcinogenic pollutant in the workplace or ambient environment may lead to an internal dose or concentration which stimulates a dormant defense mechanism that inhibits the growth of tumors.*
It is well established that a vaccine containing a weakened external dose of an infectious agent prompts human resistance to disease from subsequent exposure to the full strength agent. Now, laboratory research suggests that exposure of cells and test animals to low external doses of certain carcinogens similarly produces beneficial hormetic effects2. If further research conclusively establishes the clinical validity of the hormesis hypothesis for such toxins, regulators will be confronted with new scientific evidence which impugns their current analytic approaches to setting exposure-limiting standards for protecting worker and public health. They will also face the prospect that companies which contribute to such exposures will seek to invalidate existing standards in federal courts by arguing that new and conclusive evidence of hormetic effects, for the toxins involved, renders the standards scientifically unsound and arbitrary.
Thus, it seems advisable for regulators to anticipate growing evidence of hormesis, determine its relevance to their regulatory programs, and develop a principled approach for determining when, where and how to apply hormesis in making regulatory decisions.
Assuming that further research produces conclusive evidence of hormesis for specific toxins, the threshold issue for regulators to consider at EPA and OSHA is whether their statutory authority prevents consideration of hormesis in setting standards and taking other regulatory actions. Where authority is permissive, as is likely to be found, regulators will then need to determine the relevance and applicability of hormesis to the diverse types of regulatory actions they take to protect human health: e.g. technology-based discharge/emission standards for existing and new sources; health and safety margin-based ambient air pollution and drinking water quality standards; feasibility-based standards for workplace exposure, some of which also provide for medical and other interventions; cost-benefit-based registration and labeling of pesticides; rules requiring the reporting of new risk evidence on toxic substances; standards and rules requiring material safety data sheets and other risk information disclosures; criteria for designating certain wastes as hazardous; and variations on such regulations when "imminent" hazards, "emergency" conditions and other circumstances demand urgent action.
Having determined where and how hormesis seems to be relevant and applicable, regulators must then determine how it can be applied in a manner consistent with agency initiatives for incorporating other important scientific and policy developments in their regulatory programs: e.g. initiatives to mitigate and prevent disproportionate impacts on racial and ethnic minorities for environmental justice and to reinforce new policies to assure that environmental standards adequately protect the health of children; and efforts to regulate chemical mixtures and prevent harmful synergistic effects arising from exposure to multiple pollutants. Even more challenging will be determining how to simultaneously accommodate evidence of hormesis with evidence of variable genetic susceptibility. The Federal Environmental Genome Program is rapidly producing considerable evidence of differential genetic susceptibilities to disease from exposure to workplace and environmental pollutants, a scientific advance which will dramatically change current regulatory approaches for protecting human health from environmental and workplace contaminants3. Thus, any hormesis initiatives will need to be interwoven with other important regulatory developments.
At some early stage in developing a hormesis policy, regulators must consider their ability to fully implement hormesis-based standards and rules so they are efficacious as intended. This is critical because deep public concern about exposure to toxins and widespread mistrust of regulators are likely to fuel considerable opposition to regulatory adoption of hormesis. The notion that a small external dose of a toxin is to be imposed because it promotes human health is counter-intuitive, and feeds public suspicion about industrial influence over regulators4.
Thus, even if hormesis is
scientifically-defensible, other points of contention will arise, such as whether
the agency is capable of keeping human exposure
and internal dose within the requisite low levels in order
to avoid over-exposure and harm, an issue likely to
be raised in regulatory proceedings and the courts.
Although the same issue can arise in non-hormesis
regulatory scenarios, it is likely to be intensified here
because the regulatory action will be perceived as
authorizing and supervising human exposure to toxins known
to cause cancer and other irreversible harms.
Regulators should therefore prepare to deal with the many facets
of the implementation capability issue likely to be raised
by opponents of hormesis: e.g.
· can the agency continuously monitor human exposure to the toxin in the precise manner needed to prevent over-exposure and excessive internal dose, given the well-known limitations of using ambient monitoring to gauge human uptake (internal dose), especially in light of the very low levels involved?
· if periodic biological monitoring is to be used on a continuing basis instead, since it is usually superior to ambient monitoring for measuring human uptake of a toxin, does the agency have sufficient resources to carry out such monitoring for the entire population to be exposed? Will such monitoring, which involves periodic extraction and analysis of bodily fluids or tissue for toxin uptake, be imposed on the public involuntarily or by individual consent? Who will perform the intrusive tests and what protocols for consent, safety and privacy will apply?
· what interventions are possible and which will be undertaken by the agency if certain subsets of the population are found, through monitoring or clinical tests, to have developed excessive internal doses or concentrations of the toxin? OSHA's lead standard requires medical surveillance of exposed workers and medical removal from the workplace of those workers who develop excessive concentrations, and provides for their retention of salary and seniority status5. This type of intervention is obviously not feasible for protecting the public. Nor is it helpful if the toxin is a carcinogen or mutagen which has irreversible effects and no therapeutic remediation, unlike lead. Imagine public outrage if excessive concentrations indicating high cancer or reproductive risks are found and no suitable remediation is available.
· given that many industrial sources of pollution do not comply fully with emission, discharge and exposure-limiting standards, and that such violations, if detected, are punished by fines or negotiated settlements and corrected over a span of years, will more timely and effective enforcement be used to better assure that exposures and internal doses are kept within safe hormetic levels?
· since cumulative human exposure and consequent
internal dose of a toxin frequently arises from multiple
pathways (e.g. air, water, food, dermal) and multiple sources
(e.g. industrial discharges, workplace contaminants,
household products, etc.), as in cases involving lead, benzene,
mercury and formaldehyde, can the agency provide
sufficient oversight and control to assure that the cumulative
exposure and dose will not exceed hormetic bounds?
According to recent research on formaldehyde: "Outdoors,
major formaldehyde sources include power plants,
manufacturing facilities, incinerators and automobile
exhaust emissionsforest fires and other natural
combustion sourcesin indoor airit is released from plywood
and consumer productstobacco smoke [and] may also
be present in food" As a result, formaldehyde exposures
and consequent internal doses vary significantly: e.g.
"Formaldehyde levels in homes have been reported ranging
from between 100 and 3,680 ppb." In addition, "several
factors influence the emission of formaldehyde from
materialsAs the temperature and humidity increase, the emission
rate also increases."6 Is any agency capable of sufficient
oversight and total control of such circumstances?
Nevertheless assuming that high capability to implement can be established, the regulator must then determine how to incorporate the hormesis hypothesis and data on hormetic effects in agency fact-finding and decision-making. Standards and rules must be supported by adequate findings of fact, the inferences and conclusions that can be reasonably drawn from a sufficient compilation of data and expertise. To reach the ultimate findings, which are estimates of the risk of specific harms to human health, the regulator must do risk assessments: identify intrinsic hazards of a substance, consider exposure and internal dose circumstances, apply data and assumptions to estimate human biological responses to such exposures and then estimate harmful consequence probabilities7.
Hormesis will provide support for abandoning the long-held assumption that the dose-response relationship for carcinogens is linear at low levels of exposure, and for replacing it with a safety zone having a lower bound at zero internal dose, and an upper bound (threshold) at a designated maximum safe level of exposure or internal dose8. Thus, agency attempts to develop ever more stringent standards and ultimately eliminate exposure or internal dose regarding carcinogenic pollutants in the workplace and public environment would be thwarted by evidence of hormetic benefit-based safety zones.
Another consequence for agencies seeking to set protective standards is that the estimated incidence of harms to the exposed population will be considerably reduced. Thus, an agency such as OSHA or EPA will find it more difficult to meet judicial expectations that a standard which burdens industry will prevent a significant incidence of harms. For example, the U.S. Supreme Court has opined that an estimated harm (e.g. cancer) to one person in a thousand (1 x 10-3) would be "significant" and justify regulation, whereas harm to one person in a billion (1 x 10-7) would be de minimis and therefore provide insufficient justification9. Despite the casual home-spun quality of this observation by a plurality of the Supreme Court in a case involving an OSHA standard which would have reduced worker exposure to benzene, agencies have followed it as a rough guidance and it seems to have influenced lower courts as well in their review of agency standards10. As a result, hormesis, by reducing estimates of risk incidence, is also likely to deter agencies from enacting more stringent standards for limiting exposure or dose at levels greater than those in the safety zone.
Finally, the regulator must reach a determination on "how safe is safe enough," i.e. set the "stopping point" on the dose-response curve, and designate the duties to be met in order to monitor and maintain exposure or internal dose accordingly. In addition to supporting this determination with findings of fact and evidence of significant risk, as discussed, the regulator must demonstrate adherence to the "intelligible principles" for decision-making that are set forth in the governing statute11. Each statute presumably contains such principles. For example, laws governing EPA regulation of pesticides and toxic substances call for use of cost-benefit analysis to determine "unreasonable risk," the level of risk to be prevented by EPA action, whereas the OSH Act requires a feasibility analysis for OSHA to make such determination. Other statutes contain other prescriptions for decision-making: e.g. "best available technology." Thus, setting a hormesis-based standard, or any other for that matter, requires that the relevant findings of fact are assembled and used in accordance with Congressionally-determined decision principles.
To sum up, it seems advisable for regulators to identify and assess the issues involved in
accommodating hormesis theory in their disparate regulatory
programs for protecting worker and public health and to
develop and articulate a principled approach for
determining when, where and how hormesis will be applied in
rules and standards. But the issues discussed indicate that
such policy must have more: it must convincingly
evince agency capability for implementing the rules
and standards in a manner which assures that
hormetic benefits will be achieved. Pragmatic issues of
monitoring, intervention and enforcement, control of
multiple sources and pathways, and related matters must be
fully dealt with for agency use of hormesis to be credible,
for the public to be confident that despite its suspicions,
the promise of hormetic benefit will be kept.
1 E. Calabrese, "Hormesis Revisited: New Insights Concerning the Biological Effects of Low-Dose Exposures to Toxins," Environmental Law Reporter, v. 27 (10/1997) pp.10526-10532. E. Calabrese, L. Baldwin, "The Dose Determines the Stimulation (and Poison): Development of a Chemical Hormesis Database," Int'l. Jnl. Toxicology, v. 16 (1997).
2 For a brief summary of such research see F. Cross, n. 7 infra.
3 R. Sharp, J. Barret, "The Environmental Genome Project and Bioethics," Kennedy Institute of Ethics Journal, v. 9, n. 2 (1999) at 177; R. Preston, "Genetic Susceptibility and Sensitivity to Cancer," Chemical Industry Institute of Toxicology Activities, v. 16, n. 8 (8/1996); and M. Baram, "Genetic Testing for Susceptibility to Disease from Exposure to Toxic Chemicals: Implications for Public and Worker Health Policies," paper presented at Conference on Predictive Genetic Testing, Arizona State College of Law (April 7, 8, 2000). (Jurimetrics Jnl. in press).
4 O. Renn, "Implications of the Hormesis Hypothesis for Risk Perception and Communication," BELLE Newsletter (Biological Effects of Low Level Exposures), v. 7, n. 1 (May 1998) pp. 2-9; and P. Slovic, "If Hormesis Exists: Implications for Risk Perception and Communication," BELLE Newsletter, v. 7, n. 1 (May 1998) pp. 9-10.
5 29 C.F.R. § 1910.1025 for "general industry," 29 C.F.R. 1926.62 for the construction industry. Also see "Medical Guidelines" and "Model Contract for a Lead Medical Program," Occupational Health Surveillance Program, Massachusetts Dept. Public Health (June 1997) for an example of state action.
7 See discussion in F. Cross, "Incorporating Hormesis in Risk Regulation," Environmental Law Reporter, v. 30 (9/2000) pp. 10778-10785. Also see Chapter III, "The Scientific Predicate for Environmental Regulation," in R. Revesz, Foundations of Environmental Law and Policy, Foundation Press (1997) for useful background.
8 Id. Also see BELLE Newsletter, v. 8, n. 1 (7/1999) on Low Level Exposures: Implications for Regulatory Agencies.
9 Industrial Union Dept., AFL-CIO v. American Petroleum Institute, 448 U. S. 607 (1980) (plurality opinion).
10 See, for example, EPA's benzene regulation; 54 Fed. Reg. 38044 (1989); Natural Resources Defense Council, Inc. v. EPA, 824 F. 2d 1146 (D.C. Cir. 1987); and F. Cross, et al, "Discernible Risk A Proposed Standard foe Significant Risk in Carcinogen Regulation," Administrative Law Review, v. 43 (1991) 61, 69.
11 American Trucking Assoc. v.
EPA, 175 F. 3d 1027 (D.C. Cir. 1999), cert. granted sub
nom. Browner v. American Trucking Assoc., No. 99-1257, 2000 U. S. LEXIS 3577 (May 22, 2000).
*This paper employs a simplified model of disease etiology and terminology derived from D. Bennett and M. Waters, "Applying Biomarker Research," Environmental Health Perspectives, v. 108, n. 9 (9/2000) pp. 907-910.
The model, as adapted and the terms as used herein are