Statement on Human Testing Before NAS Committee

Statement on Human Testing Before NAS Committee

Thursday, October 11, 2007

Statement by Richard Wiles, Environmental Working Group, Before the Committee on Use of Third Party Toxicity Research with Human Participants

January 8, 2003

Thank you for the opportunity to appear before you today.

In 1998, Environmental Working Group first uncovered the practice of dosing human subjects with pesticides, and exposed the use of these human experiments in pesticide standard setting in our report, The English Patients. The report and the ensuing public outcry resulted in a moratorium on the use of human study results at the EPA. This committee was formed chiefly in response to renewed concern about an attempt by the Bush EPA to quietly repeal the Clinton moratorium on human studies.

As a result of The English Patients and subsequent research, EWG is now firmly opposed to direct dosing of human subjects with pesticides, industrial chemicals, or pollutants on both scientific and ethical grounds.

We hope that the committee will see that human subject experiments with pesticides, industrial chemicals and pollutants are so ethically compromised, so biased in conduct and intent, so scientifically questionable, and such an ethical burden for the regulators that you will recommend that they never be done.

The scope of the committee’s work now extends beyond pesticides, which should inform the debate considerably, because when one looks beyond pesticides, the chemical industry’s record is quite illustrative of its true intent.

Pesticides are covered by the thorough animal testing requirements of the Federal Food Drug and Cosmetic Act (FDCA) and the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA). Some 120 animal studies and other tests are required to fully register a pesticide. These study requirements are enforceable by the EPA administrator’s comprehensive and unencumbered authority to request these tests under FIFRA. For pesticides, companies must submit a thorough battery of toxicity studies to gain market access, and the EPA has clear authority to suspend the registration of a pesticide if the data are not forthcoming.

For commercial chemicals there are no similar requirements. Industrial chemicals, which are typically made by the same companies that make pesticides, are regulated under a very different law, the Toxic Substances Control Act (TSCA). Commercial and industrial chemicals include all chemicals not directly added to food, and under TSCA, no health and safety testing is required prior to manufacturing and marketing a chemical. None. To quote figures from a 1998 EPA review of the subject:

Fewer than half of all applications to EPA for the manufacture of new chemicals are accompanied by basic toxicity screens (none have real toxicity tests – two year feeding studies and the like). Eighty (80) percent of these are approved with no restrictions and no request for additional data – and 80 percent of these are approved in less than three weeks, at an average rate of 7 approvals per day.

If the committee and the EPA allow human experiments for commercial chemicals regulated under TSCA, most will proceed with a substantially incomplete battery of animal toxicity tests to support them. In these cases two basic ethical principals will have been violated. First, without a complete animal toxicity testing battery, scientists will not know whether the information gathered in human experiments could have been gained another way, and second, scientists will not have fully investigated the hazards of a chemical in animals prior to experimenting with it on humans.

In addition to blocking even minimum toxicity testing requirements, the industry has effectively thwarted biomonitoring for its chemicals in the environment and the human population.

There is no requirement that industry monitor for the presence of its chemicals in the environment or in people and in general they do not do so. If monitoring is performed, the results are not published or made public until the industry is forced to reveal the data by regulators or the courts. Industry is similarly not required to provide government researchers or anyone else with methods to detect their chemicals in the human population, and in general they are not volunteering these methods either.

In our view, comprehensive biomonitoring for pesticides and chemicals in the human population combined with comprehensive animal studies would largely rebut any rationale for human subject experiments.

Eroding Safety Margins

When we look behind the curtain what we have is an industry that opposes mandatory animal testing on its chemicals (excepting pesticides), complains constantly about the costs of required studies for pesticides, and systematically stymies research into the extent to which it is contaminating the human race with its products, but seemingly, and sometimes literally, is willing to go to the ends of the earth to directly dose human subjects in experiments with pesticides and industrial chemicals.

The question the committee should be asking is why?

Human experiments with pesticides and pollutants are not biomedical research and they should never be confused as such. This is not research designed to find a cure for a disease, or to generate a new scientific advance.

This is research designed to further the goals of those who pay for it. Those goals are to relax pollution limits, relax drinking water contaminant limits, or to allow more pesticides in or on food.

Excellent evidence of this fact comes from the debate surrounding perchlorate, a component of rocket fuel and drinking water contaminant, where according to internal company documents reported just three weeks ago in the Wall Street Journal, direct dosing of human volunteers with perchlorate is part of an overt strategy on the part of defense contractors “to provide EPA with a scientific-based argument to justify a higher reference dose and thus a more reasonable remediation standard." (Waldman, Peter, WSJ, 12/17 2002)

With pesticides this goal was first explicitly articulated in the “Road Map” report of the industry “Implementation Working Group” published in 1998:

“Registrants will find it increasingly undesirable to rely on endpoints derived from animal data, since this customarily requires the application of a 10-fold uncertainty factor (UF) to account for interspecies variation in addition to the other UF’s used. For this reason, there probably will be an increased reliance by registrants on data from human studies on acute or short term toxicity of the OP’s that could avoid the need for that 10-fold UF for interspecies extrapolation.”

We believe that the industry has made its motives for conducting these studies quite clear.

With the motives established it is helpful to consider the issues before the committee in three distinct groups.

  1. The quality of the science. If a study involving human subjects is bad science, then it is inherently unethical. The committee must clearly articulate the rigid boundaries of what is and is not sound and acceptable science when human subjects are involved. The most obvious problem here is that there are no standard protocols for the conduct of human experiments with pesticides and chemicals.

    EPA spends years, sometimes more than a decade, developing animal study methods and study designs for pesticide testing through an open and deliberate peer-review process. Yet when it comes to experiments on people, where the stakes are so much higher, there are no peer-reviewed protocols at all. This has contributed significantly to the shoddy science that has been submitted to the agency to date, most, if not all of which has NOT been subjected to independent peer review. In the absence of both peer review and standard protocols, we have seen selective reporting of results, biased interpretation of symptoms, severely underpowered studies, and studies that examine effects in adults when it is established that the effect measured provides information that at best is irrelevant, and at worst is intended to divert attention from the most serious health effects, which occur in the fetus or infant. At a minimum:
    • The committee must recommend extending the moratorium on human studies until protocols for their conduct are developed in an open and public process equivalent to that used for the design of animal study protocols.
  2. The ethics of the study itself. This includes what many perceive as the core issue in this debate: informed consent of study participants, competent and deliberate reviews by Independent Review Boards (IRBs), the involvement of physicians in each and every study, dose regimes that avoid over-toxic effects in test subjects and the like. There is a consensus emerging within the testing community on what these ethical requirements should be. While we understand the importance of these issues, they are secondary to the fundamental ethical question of whether or not the studies should be done in the first place, and whether one could get to the end result, sufficiently protective health standards, another way.

    With pesticides, the overarching health standard must drive the ethics of human testing: specifically, the statutory requirement that regulations protect the fetus, infant, and child. Before any additional human experiments are conducted there must be a specific articulation of which study designs best inform this regulatory requirement. In that process, four questions (among many) need to be answered:

    • Does a human experiment on a relatively small population of adults provide more meaningful information on the health risks to the fetus and infant than a properly designed, multi-generation, in utero feeding study in animals? If so, when?

    • Does a complete animal test battery combined with thorough biomonitoring of the human population obviate the need for human tests? If not, why not, and under what specific conditions?

    • Should human studies ever be done in the absence of a full animal test battery on the chemical?

    • Should human studies – which must be performed on adults – ever be done if animal studies have established that the fetus or infant is the most sensitive population? A vivid example of this problem is ongoing with the rocket fuel component and drinking water contaminant, perchlorate. Perchlorate disrupts the thyroid, and the fetus is well established as the most sensitive population. This did not stop polluters from conducting several studies with perchlorate on adults. Adults, not surprisingly, show a far greater tolerance to perchlorate than would the fetus or the neonate, and these results have been used by polluters to aggressively push for contaminant limits for perchlorate in tap water that are clearly unsafe for the fetus. This type of study, even if ethically conducted, is being used in highly unethical, scientifically inappropriate ways. In our view, the only way to eliminate this unethical use of study results is to severely constrain the types of studies that can be considered by the EPA in the first place.

  3. The ethics of what type of studies are done, and how they are used. It is easy enough for scientists and ethicists to define, in the abstract, the ethics and scientific parameters for a direct dosing experiment. But issues like informed consent, attending physicians, prohibiting studies on children, prohibiting studies designed to find a toxic threshold, are but a small part of the larger issue. Limiting the committee’s recommendations to the ethical mechanics of the science will simply shift the toughest ethical decisions to the scientists, bureaucrats and the political appointees at the EPA who are under enormous pressure from industry to use human study results in ethically marginal ways.

    Some of the tough questions that must be addressed by the committee are:
    • Should any human studies be considered by the EPA before protocols for their conduct are devised and formally adopted?
    • In the absence of protocols should a human study be considered if it has not been published in a peer-reviewed, refereed journal? In our view, studies that are unpublished, not peer reviewed and do not conform with established EPA protocols do not meet the standards of the Data Quality Act of 2002, and should be thrown out.
    • Should studies on adults ever be considered when standard setting is focused on infants and children, or when infants and children are the most sensitive population?
    • Can studies on adults ever be used to erode safety margins for children?
    • Should underpowered studies be considered in any way by the EPA? Right now, most studies submitted to the EPA are severely underpowered. These studies are particularly dangerous because they almost uniformly give a false impression of safety due to their inherent inability to detect effects that could occur in a small but significant percentage of a large population.


    EWG recommends that the committee reject all human subject experiments with pesticides, and industrial or commercial chemicals on both ethical and scientific grounds.

    If the committee does not take that view, then the committee must set rigid standards for the design, conduct, and most importantly the use of these studies.

    We urge the committee to recommend the following:

    Before a single experiment with pesticides or commercial chemicals is conducted on human subjects,

    The chemical industry must:

    1. Monitor the human population for the presence of the chemical in question, focusing on populations likely to be most vulnerable or highly exposed (not including workers).
    2. Provide a reliable analytical detection method for that chemical to the Centers for Disease Control and Prevention.
    3. Perform a full battery of animal toxicity, environmental fate and metabolism studies on the chemical that it is proposing to experiment with in humans. This test battery should be equivalent to the tests required to register a food-use pesticide under the Food Quality Protection Act (FQPA). Screening tests under the High Production Volume (HPV) program do not satisfy this requirement.

    The EPA must:

    1. Establish peer-reviewed standard protocols for human studies, as it does with animal studies. These protocols must include specific requirements for sample size, endpoint measurements, and all other aspects of the conduct of the study. Only studies that conform with these protocols should be considered for review by the EPA.
    2. Clearly articulate what types of studies it will not accept. At a minimum, the EPA must reject all human studies:
      • That are designed to determine a no adverse effect level, or a threshold for any adverse effect.
      • That exposes children or pregnant women.
      • That involve a persistent or bioaccumulative chemical or metabolite.
      • When animal studies establish that the fetus or infant is the most sensitive group within the population.
      • Where a complete animal test battery has not been performed.
      • Where biomonitoring data are not available.

    Thank you for your time and this opportunity.