The Helsinki Declaration [24
] indicates clearly that "... in medical research on human subjects, considerations related to the well-being of the human subject should take precedence over the interests of science and society." At face value, the Declaration prioritises the safety and rights of study participants ahead of the interests of other stakeholders. While different articles in the Declaration have to be interpreted in the context of all other articles in the declaration, including, for example recognition that "progress in the medical field is based on research which ultimately must rest on experimentation involving human research subjects", the protection of participants from exploitation in research lies at the heart of most ethical codes. Thus, in the risk-benefit estimation, the evidence of protection of third parties from risks from which they would otherwise not be protected, or of wider preventive benefits for society, is also critical to ethical assessment.
There are consequences of experimental human volunteer studies that can be considered to confer health benefits on parties other than the study participants. For example, having data that confirm a biomarker for a new pesticide might enable future monitoring of workers or the planning of field epidemiological studies with better exposure assessment, and better understanding of the ADME for a pesticide may indicate ways to improve control of exposure, or exclude certain pathways as critical for toxic effects, thereby informing risk assessment. It would appear, however, that the quantum of benefit declines relative to the risk when toxicity endpoints such as NOAELs are the object of the study. This is particularly problematic if there is a lack of study power, a common problem in NOAEL studies, as confirmed in our review. It is therefore less obvious how reducing uncertainty in risk assessment will, of itself, lead to reduced risks to humans As would be expected given the conservative assessment factors that are applied in extrapolation from animals to humans, none of the studies reviewed in this project that have used human volunteer data, have resulted in lower exposure limits to date. They have only confirmed existing values or, in a few cases, increased reference values.
The presumption of minimal risk is key to ethical consideration. To have any confidence in 'low-doses' being well below levels of toxicity, we rely on animal toxicity data, so experimental human volunteer studies would only be feasible for chemicals with extensive animal databases. But, given the fact that we may not have all information on all possible risks, some chronic outcomes may occur at concentrations lower than those at which other recognised effects occur. For example, by analogy, there is experimental evidence that Chlorpyrifos exerts neurodevelopmental impacts at lower concentrations than would be needed to impact on cholinesterase function [82
] and similar findings have been made in relation to other orgranophosphates [83
]. While it is possible that such effects may be judged irrelevant for single dose studies in adults, it is nevertheless important that the whole existing database is taken into account when making judgments on the risk to volunteers, as elaborated further below.
Resnik and Porter (2005) [10
] deal with this in relation to carcinogens by proposing that pesticides that are possible carcinogens should be excluded from any experimental human volunteer studies (although this begs the question of how a "possible carcinogen" should be defined). They also propose that the burden of proof of minimal risk be shifted onto the researcher to demonstrate that the effects sought are not serious and are reversible and suggest there should be long-term follow up of study participants to monitor for adverse outcomes. To some extent, their proposals mimic the general requirement for ethical review which must identify the appropriate balance between a credible or demonstrable benefit and the risks involved. Insistence on long-term follow-up of study participants could itself only be recommended where it was considered ethically justifiable - i.e. where there was sufficient prospect of benefit to the subject to justify the demands on their time and privacy from the follow-up process. Moreover, long-term follow up may not, of itself, be sufficient to ensure avoidance of any long-term unanticipated harms.
Furthermore, some of the debates on experimental human volunteer studies suggest that symptoms such as headache, sweating and nausea, are "not serious and are reversible" so would constitute minimal risk. But, given, for example, a view that neurobehavioural symptoms are part of a spectrum of neurotoxicity [86
], such symptoms may, however, also be markers of more profound damage to the CNS and their prognostic significance would need to be evaluated as part of the risk-benefit assessment.
Lastly, much unease with the idea of permitting human experimental volunteer studies rests with the concern that this will be a slippery slope, which will, on the one hand, encourage less than ethical studies in countries where there are not established oversight structures, and, on the other hand, justify arguments for very low dosing of volunteers with other well recognised toxins and toxicants such as lead if improvements of preventive action or health protection were envisaged. This remained a disputed issue within the Workgroup's discussions.
Based on the above considerations, the workgroup could not reach a uniform position on whether experimental human volunteer studies involving pesticides should be permitted or prohibited. The issues represent complex scientific and ethical challenges and their assessment entails personal value judgements.
Points of Agreement
However, we do propose that the following general principles may provide a starting point on which to base judgements as to the ethics of such studies.
• Uncertainty in risk assessment for pesticides is undesirable, as is harm or risk of harm to study participants. Harm or risk of harm to a study participant is less unacceptable where the participant derives a compensatory benefit in health or well-being.
• Disparities in the apportionment of harms, risks and benefits (e.g. a substantial risk and little benefit to study participants with a substantial benefit and little or no risk to other people) are undesirable.
• It must be mandatory that participants in research involving experimental exposure to pesticides give free and properly informed consent. This implies that subjects whose capacity to give informed consent is compromised should not be involved.
• Independent ethical review of experimental studies in humans must be mandatory.
We therefore propose that the following criteria must be considered in deciding whether a study involving experimental exposure of humans to pesticides can be scientifically and ethically justified:
1) Experimental studies in humans should be considered if (a) they will provide data that cannot be obtained by other methods - i.e. there is cogent evidence that there is there no other practical way to get the information needed that does not involve deliberate human exposure to pesticides; and (b) the information to be obtained is needed to address the optimal protection of health or prevention of illness; and (c) the study has been designed in a manner which ensures that its conduct, analysis and reporting will be adequate scientifically to answer the question(s) at hand.
2) Pesticides are a group of chemicals with a very large toxicological database in animals, larger than that for most other chemicals, and experimental studies in human volunteers should only be performed after consideration of all relevant toxicological and epidemiological data. Experimental exposure of humans to pesticides should not be permitted in the absence of an adequate prior toxicological database.
Further, the onus should be placed on the researcher to provide evidence based on extensive animal and, where appropriate, in vitro testing, that no adverse effects are likely in study participants.
In addition, data from experimental studies of potential adverse effects in humans should only be used to set reference values if it is clear from the toxicological database that a) the chosen endpoint is both appropriate and sensitive, and b) there are no other toxicological effects seen in animal studies for which there would be inadequate margins of exposure if the human data were used.
3) Consideration of risks relative to benefits establishes that there is minimal risk to participants relative to potential benefits from the study, and that the societal benefits of the study should substantially outweigh any expected risks to participants.
4) All of the recognized ethical standards and procedures for protecting the interests of study participants are observed, implying equitable standards of benefit-risk assessment and rigorous informed consent procedures. This should include full explanation of the current knowledge and uncertainty around the hazards of the agent/s and the risks of exposure. Payment of volunteers for their participation beyond the immediate costs that they incur could constitute an undue inducement in the context of no evident therapeutic benefit and needs careful examination.
5) While informed consent is a critical and necessary component of ethical requirements for a study, it does not alone comprise sufficient grounds to render a study ethical. Rather, in a stepwise approach to ethical assessment, the first action is to determine whether the study meets the risk-benefit threshold before considering whether the requirement for proper informed consent is also met. It would be incorrect to focus alone on questions of informed consent in these studies without a comprehensive consideration of all aspects of their ethics.
6) No studies should be conducted without prior assessment and approval by an ethics committee that is independent of both the researcher and the sponsors of the study. The review system should place the onus on the researchers to provide an ethical analysis and ethical justification for his/her study to the reviewing ethics committee as part of ethical oversight. As a check and balance on ethical conduct, this measure is critically dependent on the capacity of such ethics committees, which may vary widely between different countries and contexts, as well as the institutional environment that ensures their independence. Without properly constituted, adequately resourced and consciously independent Ethics Committees, with the requisite skills-mix for scientific and ethical assessment, this proviso may be ineffective in ensuring adequate ethical protections for studies of human exposure to pesticides. It is also particularly important that this provision apply regardless of country or region in the world, given the potential for serious global inequalities in research oversight.
7) The ethics of using findings from past studies involving experimental human exposure to pesticides should be considered on a case-by-case basis.
These above seven points of agreement concur with ethical guidelines for epidemiological studies published by CIOMS in 2009 [80
Points of Disagreement
Whilst agreement within the authors was possible for the above recommendations, areas of dispute that could not be resolved included the following:
1) Many criticisms of studies entailing experimental exposure of human volunteers to pesticides have focused on their use to reduce the need for uncertainty factors in regulatory risk assessments. The workgroup were divided over whether this was relevant to ethical analysis. One view held that use of data from experimental volunteer studies to set reference values when the studies were not appropriately designed for this purpose (and were perhaps conducted for other reasons), was a scientific problem rather than an ethical issue. It should also be taken into account that the setting of reference values is most often decided by others than the researchers involved. An alternative view was that the use of study data should always be integral to consideration of the ethics of a study at the outset, and so was not separable from ethical considerations.
2) One view regarded research intended to help relieve industry of a regulatory burden as not representing a societal but rather a private benefit, consideration of which is integral to the assessment of the relative burdens and benefits of the research. Further, research presented as having a different endpoint (e.g. an ADME study) but actually intended for the purpose of facilitating a waiver of an uncertainty factor, or discounting of a toxic effect as not relevant to humans, should not be allowed to alter reference values upward so as to avoid use of experimental human volunteer studies as a strategy to subvert stricter regulatory standards. A counter argument was that unnecessarily strict controls on one pesticide may lead to use of alternative products that in fact carry a higher risk. In this argument, provided the risks from participation in an ADME study are sufficiently small, their use for the purpose of facilitating a waiver of an uncertainty factor to allow ongoing use of a pesticide could be acceptable.
3) Related to the above, is the question of how to identify and quantify potential societal benefits from such studies. There was disagreement on the extent to which potential benefits of improved risk assessment for pesticides could be linked to potential benefits such as reducing unsafe use of pesticides, avoiding substitution by more hazardous methods of pest control, reducing hazards from vector-borne disease and improving food availability and security. The more distal the causal attribution of benefits, the more contested claims to benefits are likely to be. For example, arguing for better food availability on the basis of human testing presupposed a very long set of assumptions, which stray very far from the science and ethics of human testing studies.
4) The group also differed with regard to the nature of benefits related to data to inform more cost-effective regulation, or reducing the costs of regulation. One view held that, unless they demonstrated a need for protective measures which had earlier not been known, studies aimed at providing more cost-effective regulation, or simply reducing the cost of regulation, would not meet the criterion for benefit and therefore could not be justified in relation to the risks of harm from which ethical codes seek to protect participants, as has been argued elsewhere [12
An alternative view in the workgroup was that such studies should be examined and reviewed following the principles of balancing benefits and risks, the ascertainment of an informed consent procedure and an independent assessment by an independent ethical review board or its equivalent. One formulation of such a weighing up of risks and benefits might be to say that intentional human dosing studies that are to be used only to improve the accuracy of a reference dose (RfD), and that otherwise provide no health or environmental benefit, can be justified only when there is reasonable certainty that participants will experience no adverse effects.
5) There was also disagreement related to estimations of minimal risk. One perspective took the view that if an adequate animal toxicology database on the chemical existed, the risks of long-term effects from appropriately low experimental exposures could be trusted to be negligible. Within the group, however, there was concern about how easy it was to assume such low risk, particularly with repeat exposures in some study designs.
6) Additionally, the group considered but was divided over a further criterion - whether there should be an absolute proscription on human studies involving pesticides that are genotoxic or that have effects that are not easily predictable with simple dose-response relationship, such as might be the case for some endocrine disrupting chemicals. One view argued that no studies should be permitted with such agents, or at the very least, they should be exceptional based on cogent motivation. Another view was more permissive and did not automatically exclude studies with such chemicals, but rather proposed that they should be evaluated on a case-by-case basis, again with careful consideration of the risk-benefit analysis.
7) There was also considerable discussion within the group over the question of scientific independence in relation to the source of the funding and the affiliations of the researchers, and its relevance to ethical considerations. One view was that the concept of 'full independence' was meaningless, given that no researchers are independent of their own preconceptions or the need to have their research funded. As a result, it was mistaken to distinguish between industry researchers and academics who erroneously considered themselves to be independent. A contrary view argued that epidemiologists and occupational health practitioners have well recognised ethical obligations to strive for full professional independence in research [21
]. The fact that there are, in practice, threats to researcher independence should not be mistaken for a normative view of the absence of independence or one that renders all threats to independence as equivalent. For example, the presence of a direct commercial interest in the outcome of a study, through funding or employment relationships, was viewed as of a qualitatively different nature from the kinds of interests that a publicly funded researcher might hold. To say that is not to say that all research by industry or industry-funded researchers is problematic or unethical, only to recognize that there are additional ethical challenges.
8) Concerns were expressed that, in accepting data from experimental human volunteer studies, regulators were generating a market for such studies, which will create a climate which encourages cutting of ethical corners. In discussing this scenario, the group was unable to achieve a consensus in its analysis. While there was agreement that the role of all scientists involved was to be vigilant about opportunistic pressures applied to researchers, and to consistently uphold ethical requirements in research, there was disagreement as to how likely this scenario was, at least in the developed world. Moreover, within the group, there was a concern that pressure to secure ethical approval from private sector ethical review organisations would be likely to escalate. This position was not arguing that all human studies are a product of this cutting of corners, but for greater concern about this possibility.
9) Lastly, a perspective was advanced that any experimental studies performed with human volunteers for purposes of generating information to inform a reference value should be assumed to be unethical as the default position, until the researchers can provide a coherent case as to why it would be ethically justified in taking into account foreseen health benefits and risks/hazards involved. An alternative view was that such studies were not inherently different from other human experimental studies with pesticides and that the same criteria should be applied to judge their acceptability.