Can Pesticide Tests on Humans Ever Meet Standards for Ethics?
Sharon Begley, Science Journal

The World Health Organization lists aldicarb, a pesticide, as "extremely hazardous." It calls the pesticide dichlorvos "highly hazardous"; the U.S. Environmental Protection Agency classifies it as a "possible human carcinogen."

Yet, if you are healthy and up for an adventure, pesticide makers will pay you $500, $800, even $1,500 to imbibe one or the other, either once or daily for 18 days with your morning OJ.

The controversy over the scientific value and ethics of testing pesticides on people is approaching full boil, spurred by intense pressure from the pesticide industry to loosen federal policies and by a lawsuit that manufacturers filed against the EPA to make it accept data from such experiments. At last count, 14 studies of 11 pesticides have been submitted since 1996.

The push for human testing was triggered by the 1996 Food Quality Protection Act. In exchange for allowing traces of carcinogenic pesticides to remain on food (previously illegal), the act requires the EPA to use an additional "safety factor" in setting allowed residue levels, to protect children and fetuses.

Pesticide makers have that safety factor in their cross hairs, hoping to persuade the EPA that if healthy adults suffer no adverse effects after consuming a bit of pesticide, then restrictions should be relaxed.
For more health coverage, visit the Online Journal's Health Industry Edition at wsj.com/health and receive daily Health e-mails.

The recent experiments dosing people with pesticides haven't exactly covered themselves in glory, ethically speaking. In one testing dichlorvos, in Scotland, many of the volunteers were cash-strapped college students.

The possibility of financial coercion always is an ethical no-no. And in several places, the consent form referred to dichlorvos as a "drug" and said it was used as a medicine, misleading volunteers, as The Wall Street Journal reported in 1998.

But leave aside ethics concerns for a moment and focus on the science, for unless an experiment on humans yields valuable data it is by definition unethical. The argument in favor is simple: "There is no substitute for the knowledge gained from human volunteer studies," as Monty Eberhart of Bayer CropScience, a major pesticide maker, puts it.

True -- but only in the abstract. The current round of studies have used small numbers (10 to 50, typically) of healthy adults, raising doubts about the studies' statistical power. "If you see an effect, you can be reasonably confident in that finding even if you used only 10 or 15 people," says toxicologist Michael Gallo of the Environmental and Occupational Health Sciences Institute in Piscataway, N.J. "But to be confident that you have no effect is much more difficult. You may just happen to have studied 10 people who are not sensitive to the compound. To prove a negative you need many more people," probably hundreds.

The existing studies border on junk science for another reason: You can't volunteer for them unless you're a healthy adult. Given this "voluntarism bias," as epidemiologist Lynn Goldman of Johns Hopkins University, Baltimore, calls it, there are real doubts that a level found to produce no adverse effects in such a group also is safe for children, the ill and the elderly.

When you run a toxicology study, you don't ask, "So, what does this chemical do, anyway?" You define an endpoint, such as, "How does this chemical affect acetyl cholinesterase activity?"

"This is the basic fallacy of relying on these studies," says Dr. Goldman, a former EPA pesticide official. The critical effects are things like developmental neurotoxicity in children and fetuses, and those might reflect chemical pathways unrelated to whatever is causing toxicity in adults. Organophosphate pesticides have been found to alter gene expression in the brains of fetal rats, for instance, yet are tested for how they inhibit enzyme activity in adult red blood cells.

But let's say the endpoints are chosen properly, and the tests have enough subjects to offer statistical power. Let's say, too, that the studies adhere to ethical standards. Are they OK now?

You still have an inherent ethical problem. When people volunteer to test the safety of drugs, "those Phase 1 trials are for things intended to make people better," says bioethicist Jeffrey Kahn of the University of Minnesota, Minneapolis. "If the purpose is to allow industry to get higher levels of pesticides into the environment, then it's very questionable."

When an EPA Science Advisory Panel studied the acceptability of experimenting with pesticides on people, it concluded in 2000 that if the study is conducted with rigorous ethical controls, if there is no other way to fill data gaps and if the goal is to protect public health, then the EPA should consider it. "We felt that only under the most extraordinary circumstances should human testing be done," says toxicologist Ronald Kendall of Texas Tech University in Lubbock, who chaired the panel.

Interestingly, very few experiments would likely meet those guidelines. As Prof. Gallo, who is viewed as sympathetic to industry, asks, "For existing pesticides, where we have a great deal of human data from epidemiology and 'biomonitoring' of farmworkers exposed to pesticides, why do we need these studies?"

Hoping for scientific and political cover, the EPA has asked a panel of the National Academy of Sciences to answer that. Its report is expected late this year.