Will Fiddling With Pesticide Tolerances
Fine-tune Dietary Exposure?

Dr. Allan S. Felsot, Environmental Toxicologist, WSU

After selecting one or two chemicals from a screening of 20,000 and conducting more than 100 chemistry and toxicology tests on them, an agrichemical manufacturer has spent millions of dollars and many years without profit. All this investment leads to the big hurdle before registration-obtaining the tolerance. A tolerance is the amount of pesticide residue legally allowable on a fresh or processed food commodity. Tolerances are not safety standards. Yet, for any one pesticide, when the tolerances for all registered food uses are multiplied by the amount of food eaten in a day, the sum total should not exceed the reference dose (RfD). The RfD, expressed as micrograms of pesticide per kilogram of body weight per day (g/kg/d), reflects the amount of pesticide that can be consumed each day throughout a lifetime with reasonable certainty that no harm will occur. Thus, when pesticide registrants suggest a tolerance level to the EPA, they consider all current registrations and possible future registrations to avoid exceeding the RfD.

Pesticide Tolerances and the Risk Cup

With passage of the Food Quality Protection Act (FQPA), the RfD has been analogized to the brim of a "risk" cup. Potential exposures that are added to the cup were formerly only from food but now include water and residential use. When the cup is full, or in other words, when the RfD is reached for all possible exposures, no more exposures will be allowed.

EPA has estimated that food contributes 80% of all exposures to pesticide residues over a lifetime, and water and residential use each contribute 10%. When a new pesticide is first proposed for registration, and extensive residue data are not yet available, the EPA will assume that the residues will occur at the level of the registrant's proposed tolerance. Furthermore, EPA will assume that all the crops to be registered have the residues. This estimation process is known as the Theoretical Maximum Residue Contribution (TMRC). Thus, for compounds of comparatively high toxicity and correspondingly low RfD, high tolerances will tend to fill the risk cup very rapidly when subjected to the TMRC. More space can be made in the risk cup by lowering proposed tolerances.

Real Data Bails Out the Risk Cup

The FQPA encourages EPA to continue its standing policy of bailing out the risk cup by making exposure assessments based on actually measured residue data. Such assessments rely on an anticipated residue concentration (ARC), which can be estimated from registrants' crop residue studies. Such studies are used to determine the level of residues under the worst of conditions, for example, a doubling of the recommended application rate or harvest of the commodity very close to the time of application. Such residues could approach the tolerance level, but proposed tolerances will always be set at a level greater than the worst case residues, to guard against inadvertent violations when the pesticide is finally registered and widely used.

Another alternative for bailing out the risk cup applies to pesticides with longstanding registrations. During reassessments of the toxicology of these compounds, sometimes the EPA decides that newly discovered hazards demand lower exposures. In other words, the risk cup becomes smaller. However, exposure assessment can be accomplished by examining residues in the food as purchased and consumed. For example, when EPA proposed canceling the registration of the EBDC fungicides (zineb, maneb, mancozeb) during the late 1980s, the pesticide registrants conducted a Market Basket Study. An extensive collection and analysis of foods in the market place showed that EBDC residues in food were too low to fill the risk cup. In this case, adjustment of tolerances was not as important to managing risk as collecting data to more accurately estimate exposure.

Controversy about the use of organophosphate (OP) insecticides has been stirring in the agricultural community, especially among growers of minor crops.The EPA, recently lowered the RfDs for the OPs largely because of the perceived lack of information about effects in children (see A&EN issue 143). Furthermore, the FQPA is forcing EPA to consider cumulating the exposure of pesticides with identical mechanisms of toxicity. As a result, possible simultaneous exposure to all OP insecticide residues will be considered when reassessing the tolerance of any one of the active ingredients. Thus, in a short period of time, the risk cup for the OPs became very small. If the risk cup for OPs overflows, can it be bailed out? More specifically, will lowering the tolerances for OP residues create more space in the risk cup?

What Do We Know About OP Insecticide Residues in Food?

To determine the potential for reducing OP exposure in food first requires an accurate assessment of residues in food. Fortunately, we taxpayers have supported residue monitoring through the longstanding Pesticide Program run by the Food & Drug Administration (FDA) and since 1991, the Pesticide Data Program (PDP) run by the U.S. Department of Agriculture (USDA). Both programs serve as independent sources of pesticide residue data that can be used to accurately gauge dietary exposure. With the advent of the World Wide Web, anyone can download recent annual reports as well as residue databases from these government programs (FDA at http://vm.cfsan.fda.gov/~lrd/pestadd.html; USDA at http://www.ams.usda.gov/science/pdp/index.htm).

The FDA regulatory monitoring program is designed to enforce tolerances. The program is not designed to represent statistically the entire United States, but it does include commodities that are consumed in high quantities as well as commodities that receive a lot of pesticide use. Furthermore, FDA regulatory monitoring has been a recurring program since the 1960s and involves a much wider array of commodities than occurs in the USDA PDP. The PDP focuses solely on fresh and processed fruits and vegetables, wheat, and milk. These commodities and their products account for a large part of the American diet.

The USDA PDP involves the cooperation of 10 state department of agriculture laboratories. Food is collected as close to the point of consumption as possible, based on statistically reliable sampling protocols. For example, commodity samples are collected randomly, but the number is apportioned according to state population. The sampling protocol also accounts for different volumes of produce distributed annually from each site. For example, 1,123 samples of any one of 12 commodities were collected from California in 1996, but only 319 were collected from Washington.

The federal pesticide monitoring programs sample thousands of individual food items annually. Insecticides primarily and then fungicides comprise the types of pesticide residues detected in the FDA and USDA programs. Herbicides are infrequently detected. Rates of pesticide detections fluctuate annually. For example, in 1987 about 57% of all domestic foods collected by the FDA had no detectable pesticide residues. In 1996, 49% had no residues. This difference may be explained partly by the improvement in efficiency and sensitivity of analytical technology in the intervening years.

The 1996 PDP results showed no residues detected in 32% of all food samples. This percentage seems much lower than the proportion found by the FDA regulatory program, but PDP commodities are those generally perceived to have the greatest residues. For example, 35% more fruit and vegetable samples are analyzed in the PDP than in the FDA program. The PDP analyzed 340 wheat grain samples, but the FDA analyzed a combination of 185 wheat grain and wheat products. The PDP found only 17% of fresh fruits and vegetable samples and 8% of wheat samples to be residue free. Rates of residue-free produce in the 1996 FDA program were 53% for fruits, 35% for vegetables, and 40% for wheat and wheat products.

The FDA regulatory program does not, like the PDP, cover processed commodities. The PDP reported 61% of the processed fruit and vegetables to be residue free, indicating that processing effectively eliminates some residues.

Comparison of Reported Residues and Tolerances

While the data for incidence of residues in food are revealing in regards to the likelihood of our eating a food with pesticide residues, it is not informative of whether a hazard exists. Nor is it helpful to assessing the effect of lowering tolerances on residue distribution and exposure. The actual residues discovered are needed to assess compliance with the established tolerances and to assess dietary exposure.

A tolerance may be violated in two ways--exceeding the magnitude of concentration or containing a residue of a pesticide not registered for the specific commodity. During 1996 the FDA reported an overall violation rate of 1.3% for fruits and 1.1% for both vegetables and grain products. The PDP reported 4% of fruits and vegetables in violation of tolerances, while only 0.3% of wheat samples were violative. Violation rates for imported foods were similar to rates for domestic foods.

In both the FDA and PDP studies, the overwhelming majority of violations occurred solely because no tolerance was established for a particular commodity and pesticide combination. Such data indicate a need for improved management of pesticide use decisions. Design and content of labels may also need improvement, to help growers quickly recognize specifically defined uses of a pesticide formulation. On the other hand, the very infrequent occurrence of residues greater than the tolerance suggests that growers are using the legal application rates.

Reducing Tolerances 10-fold Will Not Affect Dietary Exposure

Some regulatory officials and environmental groups have asserted that lowering tolerances will reduce pesticide exposure and meet the stricter requirements of the FQPA. Close examination of the actual residue data reported in the PDP reports for 1992-1996, however, suggests that lowering tolerances will not accomplish much; the vast majority of detected residues are already far below tolerances. Residues of two of the most frequently used OP insecticides, azinphosmethyl (Guthion) and chlorpyrifos (Lorsban), on apples and peaches, important children's foods, are highlighted to illustrate their magnitude with respect to their tolerances (Tables 1 and 2). The frequency of residue detection on both commodities seems to have increased over the five reporting years. Nevertheless, the majority of apples and peaches contain no detectable insecticide residues.


As mentioned previously, however, incidence of detection does not describe hazard. More important is the distribution of actual residue concentrations. Distributions can be described by looking at the proportion of residue concentrations below a certain level. These distributions are categorized as percentiles. For example, 90% of all residue concentrations (i.e., the 90th percentile) of azinphos-methyl on apples were less than 6% of the listed tolerance (2 ppm) during all five years of sampling (Table 1 and 2). No samples exceeded the tolerance level.

The 90th percentile of chlorpyrifos residues on apples did not exceed 2% of the tolerance. Because the tolerance for chlorpyrifos on peaches is so low (0.05 ppm), chlorpyrifos residues approached 20% of the tolerance. One sample of peaches had a chlorpyrifos residue exceeding the tolerance. Fortunately, few peaches had detectable residues (Table 2).

Decreasing the apple tolerances for azinphos-methyl and chlorpyrifos by 10-fold would not change dietary exposure, because 9 out of 10 detectable residues are already at least 20 times less. On peaches, the tolerance for chlorpyrifos is already close to the reported detection limit (about 0.01 ppm).

One could argue that tolerances should be decreased by a factor of 100. Resulting tolerance concentrations would require chlorpyrifos residues to be detectable at levels less than currently feasible detection limits and azinophosmethyl residues at levels barely greater than their detection limit. Such demands on analytical capability would decrease confidence in accuracy and precision of the data, confounding the need for actual exposure data to estimate hazard. From the perspective of an analytical chemist, demands to lower tolerances by two or more orders of magnitude seem to be a move to eliminate registrations altogether.

One effect of lowering the tolerances more than 10-fold would be an increased probability that more apples would be categorized as violative. Commodities that are violative cannot be shipped. The only way to achieve residue targets that meet the requirements of a greater than 10-fold lowered tolerance would be to substantially reduce application rates. Such actions, however, greatly increase the risk of ineffective pest control. Growers use OP insecticides on orchards because the few registered alternatives are not as effective. To trade the ability to produce a safe food supply economically for no apparent increase in public health benefits hardly seems a rational trade-off.


Return to Table of Contents for the June 1998 issue

Return to Agrichemical and Environmental News index