Ethics of Human Dosing Trials

Clinical trials involving the intentional dosing of humans with pesticides or other toxic substances for the purpose of establishing regulatory limits have given rise to intense reactions on all sides. Some say such studies are inherently repugnant and should never be used. Others contend that dosing trials provide valuable information that contributes to science-based decision-making and cannot be obtained in any other way. Nearly all experts argue, however, that human studies should be conducted under very strict controls. And some say data from such trials should not be used to establish a safe threshold limit for human exposure--a level above which adverse effects occur.

In late June, an Environmental Protection Agency draft proposal on how to conduct human dosing studies was leaked to congressional staff members and later to the press. Immediately, it led to a storm of criticism because it did not include key requirements recommended in the 2004 National Research Council (NRC) report on the same topic. Nor did the draft proposal include some of the important recommendations of a report on human studies prepared by a joint committee of the EPA Science Advisory Board and science advisory panel.

For example, the EPA proposal does not require that the agency set up a human studies review board to approve study protocols before trials are begun and after they are completed. Nor does it stipulate an almost total ban on trials involving children, newborns, pregnant women, or prisoners. And it does not demand that subjects be given medical treatment free of charge if they are injured during trials.

The NRC report contains all of these provisions. It puts a great deal of emphasis on establishing a human studies review board, which would provide an additional layer of protection to subjects over that provided by institutional review boards (IRBs) at the laboratories carrying out the trials. "It was not clear to the committee that local IRBs can be expected to conduct a thorough assessment of this kind of research," the NRC report says.

Instead, the EPA proposal states, "The agency has decided not to include any proposed requirements relating to a human studies review board." The agency's rationale is that acting on NRC's recommendation would "unnecessarily confine EPA's discretion."

UNDER SIEGE
Animal Welfare Groups Sue EPA Over Animal Test Data"Why doesn't the agency acknowledge openly that it is not following current trends [of human studies review boards] and explain its position?"

While a number of experts and public interest groups are trying to prevent EPA from using data from human pesticide trials for standard-setting purposes, there are at least two advocacy groups that are attempting to prevent EPA from using animal test data.

On July 11, People for the Ethical Treatment of Animals (PETA) and the Physicians Committee for Responsible Medicine filed suit in federal court to prevent EPA from using data from animal tests to set pesticide exposure guidelines for children. They asked EPA to repeal its developmental neurotoxicity test guidelines, which entail the use of animal data.

The groups allege that the proscribed neurotoxicity tests have not been validated to verify that they are reliable and relevant predictors of pesticidal effects in humans. They also claim that using animal test results leads to pesticide exposure standards that endanger children. They argue that the protocols should have been adopted only after formal rule-making required under the Administrative Procedures Act.

In October 2004, PETA submitted a similar petition to EPA, and the agency denied it. In its written response, EPA said the test guidelines were developed over many years with broad public participation and significant involvement of the scientific community. It claimed the test guidelines were issued through legally correct procedures and are "reliable, relevant, and validated" for their intended purposes. Pesticide law gives EPA broad authority to specify the information it needs to evaluate the potential risks of pesticides, the agency wrote.

It remains to be seen whether the federal court will decide the current suit in EPA's favor.

THE AGENCY declines to comment on the draft proposal because the proposal is still incomplete and going through interagency review and revision, says EPA spokeswoman Eryn Witcher. "We don't really know what the final piece is going to look like," she says.

EPA's controversial proposal prompted the passage of some conflicting legislation. On June 29, the Senate addressed pesticide testing in two amendments to the appropriations bill for the Department of Interior and related agencies. One measure, sponsored by Sen. Barbara Boxer (D-Calif.), prohibits EPA from spending any funds to consider third-party pesticide studies--those conducted by industry or outside organizations--in which humans are intentionally dosed. The second measure, introduced by Sen. Conrad Burns (R-Mont.), instructs EPA to thoroughly review the human pesticide dosing studies that have already been submitted to the agency and issue a final rule on the conduct of future studies within 180 days.

The interest groups behind the amendments apparently understand their implications. The American Farm Bureau and CropLife America, which advocate testing pesticides on humans, support the Burns amendment, whereas the Environmental Working Group (EWG), which strongly opposes human pesticide trials, favors the Boxer proposal. But it is not clear that the senators who voted for both measures were aware they were approving diametrically opposed bills. In June, the House approved an amendment similar to the Boxer proposal. The conference committee for the appropriations legislation is expected to retain one of the amendments in the final bill.

Before 1996, when the Food Quality Protection Act (FQPA) was passed, human trials with pesticides were not an issue. Pesticide makers submitted very few human studies to EPA to support registration or reregistration of their products.

Prior to FQPA's passage, the agency used two safety factors in setting health standards for human exposures to pesticides. By conducting animal tests on a pesticide, EPA established a "no-observed-effect level (NOEL)"--the highest level of pesticide that caused no adverse effect in lab animals. To set a standard for humans, EPA then divided the NOEL by two safety factors--a 10-fold factor to account for the possibility that humans are more sensitive to the chemical than lab animals and an additional 10-fold factor to account for variations in human sensitivity.

FQPA was designed to increase human protection from pesticide exposure. It required a third 10-fold safety factor to account for the increased sensitivity of infants and children unless reliable data show a different factor is safe. This new precaution increased the total safety factor to 1,000.

But if human dosing studies with a pesticide showed that humans were no more sensitive than lab animals--or only two times more sensitive, for example--the 10-fold safety factor for extrapolation from animal to human exposures could be eliminated or reduced. After 1996, pesticide manufacturers began asking EPA to use safety factors of less than 10 for certain chemicals when extrapolating from safe animal to safe human exposures. To justify this, they began submitting human experiments to the agency.

In a 1998 report, "The English Patients," EWG revealed the details of experiments that pesticide makers had conducted in England and Scotland. In these trials, participants were given oral doses of pesticides, and in some studies, the pesticide was described as a drug and no risks were revealed. EPA does not require companies that run human trials to follow any specific protocol for protection of human subjects. In its report, EWG asked EPA to stop accepting human studies until it adopted procedures for such experiments.

In 1998, EWG reported that researchers at the Medeval Laboratories in Manchester, England, dissolved the neurotoxic insecticide dichlorvos in corn oil and paid six men to eat it. The purpose of the experiment was to test the chemical's acute effects. Dichlorvos is used to kill fleas, caterpillars, and other bugs on fruits and vegetables.

In 1992 at the Inveresk clinical laboratory in Edinburgh, Scotland, aldicarb--an extremely toxic insecticide--was dissolved in orange juice and given to 38 men and nine women at breakfast. Some subjects experienced a 70% drop in cholinesterase levels, which is considered dangerous. A 50% decline is supposed to trigger the use of the antidote, atropine.

After a public outcry following publication of "The English Patients," EPA announced in 2000 that it would no longer accept privately funded human experiments until the ethical and regulatory issues were resolved. It asked a joint committee of its Science Advisory Board (SAB) and the Federal Insecticide, Fungicide & Rodenticide Act Scientific Advisory Panel (SAP) to formulate ethical guidelines for pesticide testing in humans.

The SAB/SAP report, issued in 2000, concluded that human trials posed serious risks to subjects and should be carried out with stringent oversight. It also stated that human trials should be conducted only if they provide a public health or environmental benefit. "Trials with sample sizes inadequate to support reasonable inferences about the matter in question are unjustifiable" and inherently unethical, the report said. It concluded that all of the available studies relied on sample sizes far too small to yield statistically meaningful results. It also said that human trials should not be conducted if their primary purpose is to establish a NOEL. Two members of the committee wrote a strong dissenting opinion, saying all intentional pesticide dosings of humans are unethical.

In 2002, EPA asked NRC to produce a report on the issue. Many of NRC's recommendations were similar to those in the SAB/SAP study, including emphasizing that EPA should not accept data from previous experiments, which, according to NRC, did not meet ethical and scientific standards.

IN THE MEANTIME, CropLife America, the pesticide manufacturers' trade group, sued EPA, alleging that the agency's refusal to accept and consider human dosing trials violated the law. EPA lost the suit in a decision issued in 2003 and was forced to begin accepting the results of human trials.

Currently, EPA is reviewing 24 studies in which humans were intentionally dosed with pesticides. According to EPA information and a variety of published sources, organophosphate pesticides were used in 11 of the experiments, and carbamates in five. Six of the trials were conducted to determine a NOEL in humans.

In one trial, subjects wore modified lab goggles and different concentrations of methyl isothiocyanate (a breakdown product of the soil fumigant metam sodium) were pumped inside the goggles for up to eight hours at a time. The purpose of the trial was to determine the NOEL for eye irritation. At higher levels of exposure, some participants experienced an extreme degree of irritation. The trial report concluded that human eyes are more sensitive to the chemical than lab animals' eyes.

In another trial conducted in 1998, the subjects were given doses of the organophosphate insecticide azinphos-methyl that were as high as 1 mg/kg of body weight, even though the trial design states that doses above 0.75 mg/kg of body weight "would require particular caution."

Also in 1998, dozens of college-age Nebraskans were each paid up to $460 to swallow a pill containing chlorpyrifos--the active ingredient in Raid roach spray. Students were recruited with school newspaper ads urging them to call 402-474-PAYS.

In a recent experiment, performed in December 2004, 127 college students and members of minority groups were put in chambers and exposed to chloropicrin to assess irritation and inflammation. Chloropicrin, a suspected neurotoxicant, was employed as a chemical warfare agent in World War I and is now used as a soil fumigant to kill insects and plant root fungi.

In four of the 24 studies EPA is now reviewing, informed consent of the subjects is not even mentioned, and in five, sample consent forms were not given to EPA. In most of the remaining trials, the consent forms are deficient in some way. Several do not explain the risks involved or fail to reveal the real purposes of the trial. These are violations of the Common Rule--rules governing the ethics and scientific quality of human studies.

The Common Rule was established in 1991, when 16 federal agencies adopted a single set of regulatory provisions governing medical research on humans that is conducted, funded, or overseen by the agencies. Under the rule, persons injured during a trial must be compensated, and informed consent explaining the real risks involved in a trial is required. Trial procedures are reviewed by local IRBs and by the human studies review boards at the agencies. So far, EPA has not adopted specific regulations that codify the Common Rule into agency programs. It has no mechanism in place to ensure that pesticide companies follow the rule.

Although some of the arguments over human pesticide trials are arcane, the results of the studies have real-world consequences. Many experts say that if EPA applies a 1,000-fold safety factor when calculating a NOEL for humans, it will be impossible to reregister many of the older, more toxic organophosphate and carbamate pesticides. The calculated safe dose would be so low that it would not be feasible to use the chemicals effectively and still keep residues low enough.

On the other hand, if human dosing studies with some of these pesticides show that a 10-fold safety factor is not needed when extrapolating from animal to human safe exposures, the additional safety factor would be eliminated or reduced and these pesticides could most likely stay on the market.

As it turns out, the human experiment on aldicarb resulted in EPA raising the safe level of exposure by a factor of five over what it would have been without the human trial. The dichlorvos study in Manchester resulted in EPA eliminating the 10-fold interspecies safety factor for dietary exposure.

IN CONTRAST, the methyl isothiocyanate trial convinced EPA to lower the maximum allowable level by a factor of four, says Patrick J. Donnelly, executive vice president of CropLife America.

EPA's draft proposal is out of the mainstream in that it includes few of the important safeguards recommended by NRC and EPA's own advisory committees, says Karen J. Maschke, associate for ethics and science policy at the Hastings Center. Nearly all other government agencies that accept, commission, or conduct human studies have set up the equivalent of human studies review boards, she explains, "yet EPA refuses to take this step. The question is why is EPA reluctant to follow current trends? Another question is why doesn't the agency acknowledge openly that it is not following current trends and explain its position?"

One reason EPA's proposal on human testing may be out of the mainstream, Maschke explains, is that health care professionals who have spent a great deal of time thinking about the issue have not yet come to a consensus. Even among those who consider human testing necessary and potentially valuable, there are a wide variety of opinions, she says.

For example, some experts believe it is never ethical to deliberately test pesticides on children, Maschke says. But some contend that in certain very controlled circumstances, it should be allowed. Concerning human testing in general, she explains, nearly all experts in the area say that EPA should establish a human studies review board to examine test protocols before and after the trials are conducted.

CropLife America's Donnelly, on the other hand, says that setting up a human studies review board would be of questionable value in terms of utilization of resources. IRBs can perform those functions, he explains. He points out, however, that EPA's proposal asks for comments on whether it should establish a human studies review board.

In many cases, Donnelly observes, human data result in more restrictive safety standards. There have been published studies, he says, comparing outcomes if only lab animal data are considered to outcomes if both animal and human data are considered. "In a full 36% of cases, having the human data available resulted in a more restrictive regulatory standard," he explains.

For this reason, Donnelly opposes the Boxer amendment. "It's a prohibition on data that have already been submitted to EPA," he says. "When is having less data more protective of public health?

"Our members are very supportive of science-based regulation and the safety of our products," he says. "The policies that Sen. Boxer is trying to institute are certainly not in the public interest."

A July commentary in Environmental Health Perspectives by David B. Resnick and Christopher J. Portier of the National Institute of Environmental Health Sciences sums up the primary opposing positions over human trials with pesticides [2005, 113, 813]. One is that they should never be allowed; the other is that they should be allowed, but only under "stringent scientific and ethical standards," the article says. Resnick and Portier argue that human studies should be allowed if the knowledge gained is expected to promote human health and cannot be obtained by any other means, and if the study is not expected to cause serious or irreversible harm to the subjects. It may be that a compromise involving all of the elements will eventually be reached regarding human trials.