The Risks of Residue: How Much Is Too Much?

The Environmental Protection Agency is charged with being sure that the public comes to no harm from the use of 600 active pesticide ingredients and 1,200 “inert” ones sprayed on 376 agricultural commodities by 2 million farmers. That charge has always been and still is impossible.

In October, President Reagan signed into law a set of revisions to the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA). The revised law has been nicknamed “FIFRA Lite” by environmentalists, not because it weakens pesticide regulations--in fact it strengthens them--but because it still falls far short of what is needed to protect the ground water, the environment and the population--especially children.

The new FIFRA authorizes $110 million in federal funds over the next nine years and assesses chemical companies $150 million to review the hundreds of pesticides now in use that have never been properly tested for health effects. To put that corporate $150 million in perspective, it is about half a cent out of every dollar of pesticide sales the companies will make, assuming prices and quantities stay constant.

FIFRA Lite also removes a crippling requirement that the EPA compensate companies for inventories of banned pesticides (for some chemicals the amount of compensation would exceed the EPA’s entire pesticide budget).

These are important and hard-won gains. But the new FIFRA does not protect ground-water supplies or farm workers, does not prevent U.S. companies from marketing banned pesticides abroad and does not address the rising concern about pesticide exposure among children.

I first worried about children and pesticides four years ago when Prof. John Wargo moved to an office across the hall at Dartmouth College. He and two graduate students, Ross Brennan and Stewart Dary, were working on a project for the National Academy of Sciences, converting EPA data tapes on pesticides to a microcomputer.

The buzz of conservation across the hall about alachlor and benomyl and tolerances interested me, and I kept asking questions. Slowly I began to understand what the EPA goes through to decide how much atrazine we can eat with our corn oil, how much Alar with our apples.

According to FIFRA, new pesticides (and, eventually, all old pesticides) must go through a battery of tests to determine short-term toxicity and long-term carcinogenicity. Toxicity is measured by feeding the chemical to laboratory animals in increasing doses until it has killed half of them. Tests of internal and external exposure over periods up to two years determine whether the pesticide causes tumors or birth defects.

Test data is submitted to the EPA. They are the base of scientific observation upon which the whole regulatory process is built. Already, however, some questionable assumptions have been made.

--Assumption 1. The results have been interpreted correctly. (There was a big fight, for instance, about whether the lesions caused by dicofol were cancerous--it was decided they weren’t.)

--Assumption 2. The test lab is honest. (One of the big ones, Industrial Biotest Laboratories, was discovered not to be.)

The EPA must convert animal tests to permissible human exposures. To do so, it must string together a long chain of even more uncertain assumptions.

First the EPA determines a No Observable Effects Level (NOEL)--the daily dose that produces no visible ill effect on test animals.

--Assumption 3. No observable effect means no actual effect.

--Assumption 4. There is some safe level of exposure.

No assumption has been more fought over than that one. It may be that carcinogens cause tumors even at very low concentration. Federal law has been inconsistent, permitting no carcinogens in processed foods, but “safe” levels in raw foods. In fact the EPA has permitted carcinogenic pesticide residues if they pose a “negligible risk” of inducing cancer. On Oct. 12, they made that practice a publicly declared rule.

Next the EPA has to convert the NOEL from mice to men.

-- Assumption 5. Humans are 10 to 10,000 times more sensitive to pesticides than are test animals. The EPA takes the NOEL, divides it by a number between 10 and 10,000 (exactly what number is a matter for regulatory judgement), and calls that the Acceptable Daily Intake (ADI). For carcinogens it sets the ADI at what it assumes will cause no more than one cancer in 1 million people over a lifetime of exposure.

I can find no one who can describe exactly how the ADI from toxicity tests and the ADI from carcinogenicity tests are combined to determine an overall acceptable public exposure. There is no visible rationality in the final amounts of pesticide residues permitted by the EPA. There is no correlation between permitted amounts and either the toxicities or the carcinogenicities of the chemicals.

The apparent irrationality may be due to regulatory lag. The test data doesn’t come in all at the same time (carcinogenicity tests take longer). Many older chemicals have not been tested. New findings keep turning up. The ADIs are constantly being revised, even as the chemicals go on being used in the fields and the foods of the nation.

Out of the chaos, the EPA somehow determines, let’s say, an ADI of .05 milligrams of captan per kilogram of body weight per day. Captan is a fungicide used on apples, tomatoes, peas, sweet corn, onions, beans, squash, carrots and oranges. Now to be sure you don’t get more than that acceptable .05 milligrams in your total diet, the EPA has to know how much of each of those foods you eat.

-- Assumption 6 . We all eat the average American diet. The EPA has survey data on what a sampling of Americans eat. It averages us all together, you and me and your finicky two-year-old and my vegetarian grandfather. This, as we shall see in a moment, is the assumption that most threatens children.

Given the average diet, the EPA sets tolerances for each chemical and crop. Altogether the EPA has established 8,444 tolerances. For each one it sets forth field procedures specifying the maximum amount of each chemical a farmer can use on each crop and what period must elapse between application and harvest. The Department of Agriculture has the responsibility to publicize the field procedures and monitor compliance on the farms--another impossible task.

More assumptions have crept in here.

-- Assumption 7. The public’s only exposure to a pesticide comes through food ingestion.

--Assumption 8. When several pesticides are present in the diet together, they do no more damage than they would separately.

-- Assumption 9. Farmers can and will obey field guidelines.

-- Assumption 10. Obeying field guidelines results in tolerances being met.

Fewer than 1% of commodity shipments are sampled by the Food and Drug Administration to test for compliance. They are not tested for all pesticides, or even for some of the most common or most carcinogenic ones. From 5% to 8% of imported foods are found to be over tolerance, and 2% to 4% of domestic ones.

My colleagues across the hall put the data for toxicities, NOELs, ADIs, average dietary exposures and tolerances on a computer spreadsheet. Then they asked “what if” questions. What would be the risk if metolachlor is substituted for alachlor? What if we define “negligible risk” as one cancer in 10 million instead of 1 million? What if some subsets of the population don’t eat the average American diet?

That last question led to the children. Growing children eat twice as much food per unit of body weight as adults. They eat simpler foods with less variety, much more milk, basic grains and fruits. Average American apple juice consumption is .22 grams per kilogram of body weight per day, but the average non-nursing infant ingests 3.46 grams per kilogram per day--more than 10 times as much.

Wargo and aides started checking tolerances for 18 commonly used fungicides against the average diets of children. They found that children 1-6 years old could be exposed to more than the acceptable daily intake of 10 out of those 18 fungicides. Some of the calculated overexposure levels were enormous--from 5 to 200 times the ADI.

That was in 1985; the numbers have been expanded and changed somewhat now. The basic point has not changed at all. If all foods contain pesticides at the tolerance level (which they don’t), and if the EPA’s definition of safe exposure is accurate (which is questionable), a child eating an average child’s diet is exposed to pesticides at well above levels defined as safe.

Children might be safer than Wargo’s calculations suggest, because most crops probably contain residues well below tolerance levels--but the data is too scattered to be sure of that. Some children might be more endangered, because they eat much more of some foods than the average child. Children are also more vulnerable than adults because they have poorly developed immune systems, lower abilities to detoxify foreign substances and more permeable skins, lungs and digestive tracts.

Wargo updated his numbers, triple-checked them, expanded them, made his program more comprehensive and easier to use. In 1987 he presented his results to the National Academy and EPA. Thanks to such efforts, Congress has authorized the academy to conduct a $600,000 study of childhood exposure to pesticides. That work has just gotten under way, overseen by a committee of the nation’s best pediatricians and toxicologists.

John Wargo is also a parent. His personal reaction to his findings has been to scour southern Connecticut, where he now lives, to find organic apple juice for his two-year-old son.