1. FDA Monitoring and Enforcement
Forbidden Fruit | Illegal Pesticides in the US Food Supply: 1. FDA Monitoring and Enforcement
Criticism of FDA's pesticide residue monitoring program is not new. Since 1980 the GAO has published at least 22 reports and testified six times before the Congress about problems with FDA's pesticide monitoring program. Over the years, the GAO has repeatedly found that the FDA:
Fails to sample a statistically representative amount of the food supply (GAO 1994b, 1989c, 1989a, 1986c, 1986a, 1984);
Does not possess the knowledge and the analytical means to detect many pesticides commonly applied to food crops (GAO 1994b, 1993d, 1993b, 1989c, 1989a, 1986c, 1986a, 1984, 1980b);
Fails to exercise the analytical rigor needed to detect several heavily used pesticides, even when the equipment and knowledge are at hand (GAO 1994b, 1993d, 1993b, 1989c, 1989a, 1986c, 1986a, 1984, 1980b);
Lacks the authority, without a court order, to hold or detain U.S. produce while it is being tested, meaning that virtually all domestic food with illegal pesticide residues is eaten by the unsuspecting public (GAO 1994b, 1989a, 1986c, 1986b, 1984, 1980b);
Cannot require importers to post bonds of sufficient value to deter them from marketing shipments of produce containing illegal pesticides. As a result, two-thirds of imported produce with illegal pesticides is eaten by the public (GAO 1994b, 1993b, 1992d, 1989a, 1986c, 1986b, 1986a, 1984);
Has no authority to assess civil penalties against domestic producers whose food contains illegal pesticide residues. Because criminal charges must be brought against violators, enforcement actions are almost never initiated (GAO 1994b, 1989a, 1986c, 1984, 1981), and;
- Has no authority to fine importers who distribute food contaminated with illegal pesticides (GAO 1994b, 1993b, 1992d, 1989a, 1986c, 1986a, 1984, 1981).
The FDA has responded effectively to some of these criticisms. For example, since the mid-1980s sampling of imported food has dramatically increased and now represents about one-half of all fruits and vegetables tested. In addition, the agency has expanded its knowledge about what pesticides are being used in foreign countries that export food to the United States. Most problems identified by the GAO, however, have not been resolved because they require more money or new legal authority, neither of which have been made available to the agency.
The bottom line is that FDA continues to operate the routine pesticide monitoring program with poor sampling strategies, outdated equipment, inadequately trained and overburdened personnel, weak enforcement authority, and no ability to assess civil penalties. The combined effect of these flaws and disincentives is to perpetuate a situation that virtually guarantees underreporting of both legal and illegal pesticide residues in food.
Sampling will always be a problem for the FDA or for any agency that is saddled with enforcing pesticide tolerances. Without enhanced legal authority, the FDA must randomly sample for pesticides in food with no knowledge of what pesticides were applied to the crop being tested. A statistically-based random sample of pesticide residues in the food supply would require hundreds of additional samples of each crop tested, and the agency simply cannot afford it.
The FDA routinely claims that it samples about one percent of imported food shipments for pesticide residues (Young 1987, GAO 1994b, 1992d), a figure often extrapolated to incorporate the entire food supply. There are several problems with this claim. First, while it is possible that a one percent sample could provide reasonable insight into both illegal and overall residues in the food supply, such a sampling strategy would have to be extraordinarily well-targeted. The FDA surveillance program, which forms the basis for the FDA's one percent claim, is not targeted at all, but rather is designed as a random sample survey (FDA 1994b).
Second, although the FDA's one percent figure is apparently based on tested shipments, it is not clear what actually constitutes an average shipment or how the one percent figure was calculated. When pressed on the matter, FDA officials were unable to define exactly what a shipment is and how a shipment differs from crop to crop or among modes of transport. For example, a shipment of bananas would weigh considerably more than a shipment of blueberries. Likewise, a shipment from a truck would weigh less than a shipment from a cargo vessel.
Third, the FDA is conspicuously silent on the issue of domestic sampling. Although imported foods make up more than half of the samples each year, domestic fruits and vegetables respectively constitute 63 percent and 92 percent of all fruits and vegetables consumed annually in the United States (USDA 1993b). In percentage terms, this means that far less than one percent of the domestic fruit and vegetable supply is sampled and tested for pesticides.
The Environmental Working Group attempted to determine what percentage of the food supply the FDA tests by comparing the agency's routine pesticide monitoring data with information from the Agriculture Department on fruit and vegetable production and consumption. Our analysis revealed that the agency samples only about .00002 (two one-hundred-thousandths-of-one) percent of the pounds that are shipped each year of the 42 different crops analyzed in this report. In order for the FDA's one percent figure to be accurate, the average imported fruit or vegetable shipment would have to weigh 18,500 pounds, or about nine tons per shipment.
If we assume that the FDA actually does sample one percent of all imported fruits and vegetables, then an equivalent sample size applied to the domestic fruit and vegetable supply means that the FDA samples about 0.2 percent of all fruits and vegetables eaten by the American public. (Compared to imports, Americans eat 12 times more domestically grown vegetables, and nearly two times more domestically grown fruit.) Under this scenario, the average American would have to eat about 100 pounds of fruits and vegetables in order to eat from a shipment tested for pesticides by the FDA.
In recent years, however, sample sizes have been slashed. In FY 1993, the FDA took 3,550 fewer samples of fruits and vegetables than it did in FY 1992, a whopping 30 percent cut in one year. Although exact numbers for 1994 are not yet available, the overall sample size declined again and is expected to continue to shrink over the next several years (Jones 1994).
In response to criticisms about its sampling, the FDA instituted the Pesticide Incidence Level Monitoring Program in FY 1993, which provided statistically-representative sampling (800 to 1,000 samples per crop) for two commonly consumed crops each year. In the third year of the project, however, Congress eliminated the program citing the existence of a similar project at USDA which analyzed more commodities and produced superior data.
Samples of less than 800 to 1,000 per crop could provide a reasonably accurate picture of pesticides in the food supply, but only if the sampling strategy is redesigned. For example, full disclosure of pesticides used, supported by third-party sampling and residue testing, could dramatically improve the quality and volume of information on pesticide residues in food. An efficient public-private partnership such as this, however, is not possible under current law.
Data and Analytical Constraints
In its search for pesticide residues, the FDA is flying blind into an avalanche of food, looking for hundreds of tough-to-find residues, using outdated technology and haphazardly trained personnel (GAO 1994b, 1993d, 1993b, 1989a, 1986c).
The equipment available in the average FDA lab can detect only about one-half of the pesticides used on food grown around the world. This, however, is a best case scenario, which assumes that each sample will be tested thoroughly for all the pesticides likely to be present on that crop. Testimony presented by the Environmental Working Group in January 1994, before the Subcommittee on Trade and the Environment of the House Foreign Affairs Committee, illustrates how rarely FDA labs thoroughly utilize the available equipment and methods when testing for pesticides. The results presented in this report confirm that the findings presented to the Congress apply with equal force to FDA's monitoring of food produced in the United States.
Because reporting of pesticide use is not required, when the FDA gathers a sample for pesticide residue analysis, inspectors and chemists have no idea what pesticides were applied to that crop. To compensate for this deficiency, the agency relies on multi-residue pesticide detection methods (MRMs), which under optimal conditions can detect several hundred pesticides in a single sample.
Our review of three years of FDA routine residue monitoring for 22 imported fruits and vegetables shows that nearly 90 percent of FDA samples analyzed relied exclusively on multi-residue detection methods (Wiles, Elderkin et al. 1994). This means that under the best of circumstances, 90 percent of FDA's analyses of imported food could detect only half the pesticides that might have been used on a given crop. The remaining ten percent of samples analyzed by the FDA were tested with so-called single residue detection methods (SRMs), which are needed to detect the remaining 50 percent of pesticides used on food.
The obvious limitation of SRMs is that each SRM can detect only one or two pesticides. The less obvious limitation of an SRM is that on average, each SRM analysis is as expensive as an MRM that potentially can detect over 200 pesticides when fully engaged. Because the FDA has no idea what pesticides are applied to a specific shipment of food, the agency is reluctant to use single methods, even though without them, the best the agency can do is find half of all pesticides in the food supply.
A 1993 review by the FDA's Center for Food Safety and Applied Nutrition (CFSAN) reports that single residue tests are required to find approximately 300 out of 630 pesticides used worldwide (FDA 1994a). Fourteen of these pesticides are canceled in the United States and more than 100 are listed as "foreign use" pesticides, meaning that they are not registered for any use in the United States (FDA 1994a). Federal regulators often have no idea of the health or environmental effects of pesticides used only in foreign countries. CFSAN also reports 363 metabolites, byproducts, and impurities of these pesticides that are required as part of the tolerance. Single residue tests are required to find 246, or two-thirds of these pesticide byproducts.
Many pesticides that require single residue analyses are of significant health concern. These include the carcinogens benomyl, Alar, and the EBDC fungicides, as well as many compounds canceled in the U.S. for various reasons, such as the arsenical compounds, carbon tetrachloride, EDB, DBCP, and dinoseb (Wiles, Elderkin et al. 1994).
Discrepancies Between Eastern and Western FDA Labs. Our January 1994 testimony also identified major discrepancies between FDA's eastern and western pesticide labs, and provided compelling evidence that the FDA's eastern labs are missing a significant percentage of legal and illegal pesticide residues.
To find the 200 pesticides that can be identified using the most common multi-residue detection method (the Luke extraction method), technicians must put each sample through six different detection screens. Our analysis of three years of test results (1990-1992) on 8,000 samples of 22 imported fruits and vegetables found that FDA labs rarely use all six screens. In particular, eastern labs typically use only half of the detection screens necessary to find all the pesticides that can be identified using the Luke method. In essence, eastern labs are operating at half capacity. This pattern held true for all 22 fruits and vegetables--each was tested equally poorly--and produced some serious and consistent disparities in the pesticides reported in food by the eastern, as compared to western, FDA pesticide labs (Wiles, Elderkin et al. 1994).
For 20 out of 22 crops evaluated, FDA's eastern labs detected significantly fewer pesticides than FDA's western labs. For the other two crops, the rates were nearly identical. The clear link between the failure to use detection methods and the failure to find residues, and the near universal detection of fewer pesticides in eastern labs, strongly suggests that eastern FDA labs are underreporting the levels of both legal and illegal pesticides present in the food supply.
Lack of Enforcement Authority
This report documents the fact that the FDA underreports illegal pesticides in fruits and vegetables. Several previous GAO reports, however, have shown that even when the FDA reports a violation, the public almost always eats the illegal pesticide.
Although the FDA is supposed to enforce pesticide tolerances in food, it has no authority to detain products during testing or to assess civil penalties on violators. As a result, even when food contains illegal pesticides, the FDA cannot keep it from reaching the consumer (GAO 1994b, 1992d).
This is especially true for domestically grown produce. Unless the FDA has a court order, the Federal Food Drug and Cosmetic Act (FFDCA) does not give FDA the authority to hold a shipment while it is being tested. The court order cannot be obtained, however, until after the product is found to contain illegal residues.
To make matters worse, the FDA has no authority to assess civil penalties against domestic producers whose food contains illegal pesticides. The only recourse is to pursue criminal charges through the U.S. Justice Department, which is costly, time-consuming, and therefore rarely feasible.
This system was sorely tested in 1994, when the FDA detected an illegal pesticide, chlorpyrifos-ethyl, in Cheerios