Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
c Comparing the Impact of the Scenarios . This chapter compares the four scenarios by their effects on herbicides, insecticides, and fungicides; individual pesticide active ingredients; and selected crop-pesticide combinations. The previous chapter describes the scenarios and their impacts. This chapter highlights the important convergent and divergent effects of these scenarios on oncogenic pesticides as measured by changes in estimated dietary oncogenic risk, acre treatments, and expenditures. THE IMPACTS OF THE SCENARIOS ON HERBICIDES INSECTICIDES, AND FUNGICIDES The four scenarios have markedly different impacts on the major classes of pesticides. In each scenario, fungicides suffer the greatest percentage of canceled crop uses and tolerance revocations; the revoca- tions account for the greatest percentage of total risk reduction. In each case, more than 50 percent of all fungicide risk is eliminated, and half or more of existing fungicide uses are affected. These scenarios have more diverse effects on insecticides and herbi- cides. The risk reduction for insecticides ranges from 19 percent in scenario 4 to 99 percent in scenario 3. For herbicides the range is from 11 percent in scenario 4 to 99 percent in scenario 3. Table 5-1 arrays the percentage of estimated risk reduction across the scenarios for each type of pesticide. A chief reason for the disparity among the scenarios' risk reduction is
COMPARING THE IMPACT OF THE SCENARIOS ~19 TABLE 5-1 Estimated Risk Reduction for Each Type of Pesticide by Scenario (in percent) Scenano Risk Standard 1 2 3 4 Tolerances Section 408 Zero Risk/benefit 10-6 risk trigger for Risk/benefit for raw risk each crop, processed foods with no and raw form com- processed fonn Lined; no consider- ation of benefits Section 409 Zero Zero risk for See above 10-6 risk trigger for risk processed processed foods foods tied (tied to parent raw to parent commodities); no raw com- consideration of modifies benefits Pesticides Fungicides 100 71 98 51 Herbicides 100 36 99 11 Insecticides 100 26 99 19 All pesticides 100 55 98 36 that much of the dietary risk from herbicides and insecticides examined in this study stems from residues in food derived from animals. The EPA does not currently recognize these foods as having processed forms. Any scenario that revokes or denies tolerances on the basis of oncogenic risk in processed foods will not touch tolerances for residues in beef, milk, poultry, or pork products. Thus, even though dietary risk from exposure to residues in animal products may exceed that associated with the human food forms of the major feed crops, tolerances for animal products will not be revoked under scenarios 2 or 4. For example, in a case where meat from animals fed pesticide-treated corn presents a greater risk than corn oil derived from the same treated corn, tolerances for meat would not be revoked. Of the compounds the committee examined, about 50 percent of all herbicide risk and 40 percent of all insecticide risk are derived from tolerances for animal products. This partially explains the difference in the treatment of herbicides and insecticides in scenarios 2 and 4 compared with scenario 3. On the other hand, less than 1 percent of all risk from fungicide residues is from tolerances for animal products. This helps explain the comparatively consistent treatment of fungicides across all scenarios. Although the EPA's criteria currently attribute very little oncogenic
120 REGULATING PESTICIDES IN FOOD risk to fungicide residues in animal products, the committee is suspicious of this low estimate. Indeed, much evidence suggests that many fungicide-treated crops whose by-products are fed to animals lack legally mandated feed-additive tolerances. Modern residue chemistry data are likely to demonstrate the need for such tolerances. If these tolerances are set, overall estimated risk would rise, as would the percentage of risk derived from meat, poultry, and dairy products, currently all defined as raw foods. The percentage reduction in fungicide risk under scenarios 2 and 4, on the other hand, would decline. This change in depiction of the baseline would not greatly change the overall performance of the scenar- ios, however. Further, the committee is aware that confirmation of residues in processed food or feed could result in the loss of tolerances for the major animal feed crops. This in turn could reduce the residue level in all food products derived from animals fed these crops, leading to a de facto reduction in risk without the loss of food tolerances for these crops. It is important to note, however, that total elimination of this risk is not certain in these cases. Crops with animal feed uses could retain both food and feed tolerances through the sensitivity-of-the-method procedure if it were demonstrated that residues of these oncogenic pesticides in human food posed very low risks. (See Appendix C for the thiodicarb case study.) Scenarios 2 and 4, which do not directly affect tolerances for animal products, ensure less risk reduction than does scenario 3, which is not limited to finding residues in a processed food. THE IMPACTS OF THE SCENARIOS ON INDIVIDUAL ACTIVE INGREDIENT RISK Among individual pesticides, fungicides are the most heavily affected under all scenarios. This does not mean, however, that all scenarios have equal effects. In fact, the same scenario can have quite disparate impacts on different fungicides, herbicides, and insecticides. A single scenario can have very different impacts on two similar types of pesticides. For example, under scenario 3, tolerances accounting for more than 93 percent of all risk from benomyl and 98 percent from maneb are revoked. Scenario 4 treats the two fungicides quite differently: 80 percent of benomyl risk and 25 percent of maneb risk are eliminated. When one examines the effect of the different scenarios on all toler- ances for the same active ingredient, scenario 3 stands out sharply. It consistently revokes tolerances for pesticides presenting relatively high risks, but does not affect relatively low-risk compounds. Scenario 3 achieves only 2 percent less risk reduction than scenario 1 while revoking 1,500 fewer food tolerances for the 28 compounds that constitute the
COMPARING THE IMPACT OF THE SCENARIOS 121 TABLE 5-2 Impact of Scenarios on Different Pesticide Active Ingredients Reduction in Risk from Total Estimated Risk in Three Scenarios (%) Pesticide 2 3 4 Herbicides Alachlor (Lasso) 50 77 SO Linuron (Lorox) 36 99 11 Metolachlor (Dual) 30 0 0 Insecticides Chlordimeform (Galecon) 31 99 31 Cypermethrin (Cymbush, Ammo) S 77 0 Permethrin (Pounce, Ambush) 24 98 12 F. ., unglclaes Benomyl (Benlate) 93 93 80 Captan 71 98 48 Maneb 43 98 25 committee's risk estimate. Scenario 3 achieves greater risk reduction than either scenario 2 or 4. Scenario 1, of course, eliminates all dietary . . Oncogenlc rls. (. Sizable differences in risk reduction between scenarios 3 and 4 are apparent for many compounds, as shown in Table 5-2. For the insecticide permethrin, scenario 3 reduces dietary risk by 98 percent, whereas scenario 4 reduces risk by only 12 percent. For the herbicide linuron, scenario 3 reduces risk by 99 percent; scenario 4 reduces risk by 11 percent. The selectivity of scenario 3 is also highlighted at this level of analysis, particularly when contrasted with scenario 2. Scenario 2 (which allows no risk in processed foods) does not discrim- inate between active ingredients that pose more significant and relatively insignificant risks. Even though scenario 2 achieves a higher degree of risk reduction than scenario 4, it fails to do so efficiently particularly when compared with scenario 3. This is because the rigorous zero-risk standard in scenario 2 applies only to the processed portion of the food supply. As a result, tolerances for pesticides accounting for sizable estimated oncogenic risks, such as linuron and maneb, are relatively unaffected under scenario 2 because many foods on which they are found lack processed forms. At the same time, several pesticides presenting relatively low estimated dietary risks, such as glyphosate and metolachlor, lose tolerances under scenario 2 because they are presumed to be present in certain processed foods. This failure to discriminate
122 o ._ ._ rD o ._ v ._ q) v ao a~ s~ ~o C~ ~: o · _ v s~ ._ ao s~ Ct E~ s~ v ._ 4 - U) V , ~ ·_ _' c Cq ao , C C c, a~ Ck ~ X C, P: C ~L) .0 C ~ Ct C ~ Ct <~: ~ c o . .= 4) Z P: , C o ~o C C) ~, X ~ P: c . c o . C) . ~ ~, P: ~ t: .° .O E~ . Z ~ 4~ o C i~ C ~L) .0 Ct ~ o o ~ o o o o o ~ o 0 ax ~o ~ o ~ _ ~ o N ~ - ~ ~ ~ ~ oo o o oo ~ oN ~ ~ ~o ~ ~ ~ ~ ~ x o - ~ o oo - ~ x ~ ~ ~ ~o N ~ o ~4 ~ ~ ~ ~ x o o o o o o o o o o o o o o o - - - - - - - t'6 ~ oo oo ~ ~ ~ ~ N ~ o ~ - oo oN ~ ~ ~ oo c o - ~ ~ o o o o o o o o .- ~ o o o o o o o o ~ ~ - - - - - ~ - - : - a, o 4, r' ;^ s~ c~ a~ et c s: ¢ ~ ~ ~ ~ 1 c~ e . ~ o o : 1 o ~ - o
COMPARING THE IMPACT OF THE SCENARIOS 123 between high- and low-risk exposures to residues is the principal flaw in this scenario. Scenario 4 displays greater consistency than scenario 2 because it revokes relatively few tolerances for either high- or low-risk residues in food. Consequently scenario 4 results in less risk reduction than any other scenario. In sum, an analysis by individual active ingredients reveals that scenario 3 reduces more risk from herbicides, fungicides, and insecticides than do scenarios 2 and 4, while allowing continued use of several relatively low-risk compounds. Scenario 2 revokes more tolerances for these low-risk active ingredients by eliminating all tolerances for all crops with processed forms. Scenario 2 also revokes fewer tolerances for certain high-risk compounds applied to foods with no processed form, resulting in less overall risk reduction. Scenario 4 allows most tolerances for high- and low-risk compounds to continue. Scenario 1, of course, eliminates all tolerances for all oncogenic active ingredients. A CROP-LEVEL ANALYSIS: THE IMPACTS OF THE SCENARIOS ON BENEFITS AND RISKS This section examines the immediate impact of the four scenarios on risks and benefits associated with eight crop-pesticide combinations. Longer-term impacts on crop production and the effect of the Delaney Clause on new product development are discussed in the next chapter. All calculations of risk in the following crop-level analyses reflect estimated total risk for a crop adjusted (multiplied) by the percentage of planted acres actually treated with the pesticides in question. Risk reduction estimates are based on this adjusted risk estimate. Only the crop-level analyses in this chapter and Chapter 4 incorporate the percent- age of planted acres treated. Benefits are measured by several rough indicators of pesticide use and expenditures. The impact of the scenarios is measured by the acre treatments and expenditures associated with pesticides that lose toler- ances as a percentage of all herbicide, insecticide, or fungicide use on that crop. The committee believes that a better measure of a pesticide benefit is the difference between the total benefits received from its use, minus the total benefits of using the next best pest control method. It lacked the time and resources to perform such estimates, however. The committee decided to study crops having processed-food forms. Because of this choice, scenarios 1 and 2 produce nearly identical results. Scenarios 3 and 4, on the other hand, differ markedly. Table 5-3 displays the effects of the scenarios on these eight crops in terms of risk reduction and acre treatments lost.
124 REGULATING PESTICIDES IN FOOD Corn and Soybean Herbicides Because of their zero-risk standards, scenarios 1 and 2 would revoke tolerances for eight active ingredients, which account for 39 percent of all corn herbicide acre treatments and 40 percent of expenditures. Under scenario 3, however, corn tolerances for only one of the four active ingredients for which risk estimates were possible would be revoked. Eliminating use of this single pesticide would reduce dietary risk from corn herbicides by 99 percent while affecting 30 percent of all corn herbicide acre treatments and 27 percent of expenditures. Scenario 4 leaves all corn herbicide tolerances untouched because oncogenic risk from herbicide residues in processed corn products in no case exceeds lo-6. Soybean producers would be harder hit than corn producers under scenarios 1 and 2. Both scenarios would revoke tolerances for 11 herbicides, which currently account for 67 percent of all acre treatments and 58 percent of all expenditures. Scenario 3 would eliminate two herbicides, resulting in 99 percent risk reduction. Scenario 4 would eliminate tolerances for only one herbicide, linuron, but this would eliminate 94 percent of the estimated dietary risk from soybean herbi- cides. This substantial risk reduction is striking considering linuron's share of total acre treatments (9 percent) and expenditures (7 percent). Because of the availability of a range of new, effective, non-oncogenic herbicides, the impact of tolerance revocations on corn and soybean producers would probably be modest even under scenarios 1 and 2, which would repeal tolerances for all oncogenic compounds. It is also true, however, that this result would eliminate less than 1 percent of total estimated dietary risk. Cotton Insecticides Scenarios 1 and 2 would end the use of eight insecticides accounting for about 80 percent of all cotton insecticide acre treatments and 61 percent of all expenditures. The loss of cypermethrin, which accounts for about 45 percent of all acre treatments, would produce most of this impact. The repeal of all cotton tolerances for oncogenic insecticides would reduce total estimated dietary risk by only about 0.2 percent. The loss of eight active ingredients accounting for nearly 80 percent of all acre treatments would require a sizable adjustment in insect control for cotton. Although state agricultural experiment stations and extension entomologists recommend many of the 35 remaining registered com- pounds for control of the Heliothis complex and other cotton insect pests, a number of these compounds are not as economical as the agents that
COMPARING THE IMPACT OF THE SCENARIOS ~ 25 would lose tolerances particularly the synthetic pyrethroid cyper- methrin. It is also important to note that virtually all estimated dietary risk from insecticides used on cotton is from chlordimeform, which accounts for only 9 percent of all cotton acre treatments and 7 percent of cotton insecticide expenditures. Eliminating the most widely used oncogenic cotton insecticides, cypermethrin and parathion, under scenarios 1 and 2 would reduce estimated dietary risk from cotton insecticides by less than 1 percent and total risk by only 0.0002 percent. Scenarios 3 and 4, which apply to the six cotton insecticides for which risk estimates were possible, would revoke tolerances for only one insecticide, chlordimeform. As in the case of the herbicide linuron on soybeans, actions against one pesticide could dramatically reduce esti- mated dietary risk from cotton, while affecting a relatively small share of total expenditures on, and acre treatments with, cotton insecticides. Apple Fungicides Ten oncogenic active ingredients currently account for more than 50 percent of all expenditures on apple fungicides. Scenarios 1 and 2 would eliminate all the benefits associated with this use. Chemicals accounting for 59 percent of all acre treatments and 53 percent of all expenditures on apple fungicides would lose tolerances. Total baseline risk would be reduced by more than 5 percent. Applied to these same 10 fungicides, scenarios 3 and 4 would affect fewer active ingredients and achieve slightly less risk reduction by revoking tolerances representing 54 and 50 percent of acre treatments, respectively (see Table 5-31. In each scenario, a significant percentage of all apple fungicides would be lost, creating the possible need for replacement compounds or other control methods. A substantial number of presumably non-oncogenic apple fungicides are in development, currently registered, or both. Copper, sulfur, and the fungicide triadimefon are the primary currently registered non-oncogenic alternatives. The committee's survey of compounds in development identified a relatively high rate of product discovery for new apple fungicides. At least nine compounds are now being field tested for control of the major apple fungal diseases. Nearly all appear to be better than the best commercially available standard for eradicating apple scab and powdery mildew. However, more than half of these are poor in protecting against apple scab. None of the new compounds was found to be as good as the commercial standard for treating the seven summer diseases: bitter rot, black rot, white rot, sooty blotch, fly speck, brooks spot, and black pox. The revocation of tolerances for all oncogenic apple fungicides in
126 REGULATING PESTICIDES IN FOOD scenarios 1 and 2 would be felt hardest in the Southeast, where adequate replacement fungicides currently do not exist for all diseases. In the northeastern and north central growing regions, apple production also currently relies on oncogenic fungicides; however, non-oncogenic re- placements for the dominant fungal diseases, powdery mildew and apple scab, are more readily available. The 10-6 risk standard of scenarios 3 and 4 would initially preserve tolerances for certain oncogenic fungicides on apples. It is likely, how- ever, that over time, few oncogenic apple fungicides would retain tolerances. This would occur when the use of agents that do not now trigger the 10-6 risk criterion increases after tolerances for other fungi- cides are revoked. Even under scenario 4, which would initially revoke tolerances for only two active ingredients, a substitution pattern could emerge in which risk from residues of each remaining oncogenic apple fungicide would even- tually exceed 10-6 and trigger tolerance revocation. In fact, revocation of tolerances for the apple fungicides that currently present the greatest risk could increase the total dietary risk if little-used, more potent fungicides come into broader use. Several scenarios that could increase estimated risk are presented below in the section on tolerance reduction. This finding highlights the importance of ensuring that regulatory actions at the crop level actually reduce risk, taking into account the probable actions of growers to find and apply substitute chemicals. Potato Fungicides Nine oncogenic fungicides currently account for around 81 percent of all expenditures and nearly 91 percent of all acre treatments for potato diseases. Scenarios 1 and 2 would terminate uses of these agents, eliminating all potato fungicide risk and reducing overall estimated risk by 2 percent. The complete loss of all oncogenic fungicides currently used in potato production would again have different regional impacts. Northeastern potato growers currently apply around 55 percent of all oncogenic fungicides on potatoes; midwestern growers apply around 33 percent. Western potato growers apply less than 15 percent of the total, although they plant more than 50 percent of all potato acres. Partly because blight and other fungal diseases are not generally a problem in the West, potato production has been moving there over the past 20 years.) Most of the oncogenic fungicides used in potato production are applied as preventatives on a routine basis. There is little field monitoring and forecasting to make more accurate determinations of when fungicide applications are necessary. It is increasingly possible, however, to use
COMPARING THE IMPACT OF THE SCENARIOS 127 information about weather and crop conditions to prescribe the use of fungicides when most needed. This practice can reduce the use of fungicides up to 30 percent in some areas in some years. It has not become common practice, though, perhaps because growers lack confi- dence in these forecasts. Currently registered non-oncogenic potato fungicides are little used. Generally these agents (triphenyltin hydroxide, metalaxyl, and sulfur) are far less effective than currently used fungicides. Further, some alterna- tives (triphenyltin hydroxide) pose other toxicological problems. Ridomil (metalaxyl) shows some control of potato blight but is far more expensive than the current fungicides of choice (EBDCs, chlorothalonil), is not widely used, and is known to have led to pathogen resistance in other crops. Scenarios 3 and 4, which would apply to seven potato fungicides for which risk estimates were possible, would have much different results than scenarios 1 and 2. Scenario 3 would initially revoke tolerances for only one fungicide, mancozeb, which accounts for about 68 percent of potato fungicide risk and about 31 percent of acre treatments and expenditures. Over time, however, scenario 3 would probably have the same effect as scenarios 1 and 2. Because of the shortage of non- oncogenic substitutes, the elimination of one oncogenic fungicide would very likely increase the percentage of all acres that are treated with other registered oncogenic compounds. The risks posed by these alternatives would eventually exceed 10-6, leading to the loss of all tolerances for all oncogenic potato fungicides. Scenario 4, which revokes tolerances for a pesticide on a crop when the risk derived from the processed form of the crop is greater than 10-6, would revoke no tolerances for fungicides used on potatoes. The EPA's Tolerance Assessment System (TAS) currently assumes that most pota- toes are consumed in the nonprocessed form. They are usually cooked fresh and consumed as baked, boiled, or fried potatoes. In the average U.S. diet, the TAS calculates that less than 1 percent of all potatoes are consumed in processed forms, such as chips or dried instant potatoes. Accordingly, the risk from consumption of processed potatoes calculated on the percentage of acres treated is less than 10-6 for all fungicides. Even though the risk from at least one fungicide on whole potatoes exceeds 10-6, scenario 4 would revoke no tolerances for potato fungicides. Tomato Fungicides Scenarios 1 and 2 would revoke tolerances for 11 oncogenic active ingredients accounting for approximately 50 percent of all acre treatments and 51 percent of all tomato fungicide expenditures. All dietary oncogenic
128 REGULATING PESTICIDES IN FOOD risk from tomatoes would be eliminated, and 11 of the 20 fungicides registered for use on tomatoes would lose tolerances. Scenarios 3 and 4, which would apply to 10 of these 11 oncogenic fungicides for which risk estimates were possible, would revoke tolerances for only five active ingredients but the impact would be nearly the same. The estimated dietary oncogenic risk from fungicides on tomatoes would be reduced by 99 percent, and tolerances would be lost for active ingredients accounting for 49 percent of all acre treatments and 51 percent of expenditures. As with all the fungicide-crop combinations examined, the impact of scenar- ios 3 and 4 would most likely continue past these initial tolerance revocations. Tolerance revocations for the five compounds would result in the increased use of other oncogenic tomato fungicides. As this occurred, the risk from these replacement compounds would probably exceed 10-6 and trigger tolerance revocations for them as well. The midwestern and southeastern growing regions would feel the loss of these oncogenic tomato fungicides the most. Growers in the Midwest, East, and Southeast apply 85 percent of all oncogenic fungicides to tomatoes, even though less than one-third of all tomatoes are grown in these regions.2 Approximately 80 percent of these applications are made in southeastern states. Nearly two-thirds of all tomatoes consumed in the United States are grown in California, yet less than 10 percent of the total pounds of oncogenic fungicides used on tomatoes are applied there.3 The committee's survey of non-oncogenic tomato fungicides in devel- opment or in field testing indicates a moderate degree of activity in this area. Nonetheless, the committee is unable to judge the relative efficacy of these compounds. It appears that as many as eight new compounds are currently being tested for control of the eight major tomato diseases, however. This finding suggests that additional non-oncogenic alternative fungicides for control of tomato diseases may be available within several years. Peanut Fungicides Because TAS assumes no processed peanut food forms for the 53 oncogenic pesticides examined, the results of the scenarios on peanuts differ from the committee's other crop-level analyses. Neither scenario 2 nor 4 would result in any lost tolerances. Because no single fungicide risk exceeds 10-6, no tolerances would be revoked under scenario 3. Only scenario 1 would revoke peanut fungicide tolerances. Oncogenic compounds account for about 86 percent of all acre treat- ments and 83 percent of all expenditures for fungicides on peanuts. Under scenario 1, all benefits associated with these fungicides would be lost. Because peanuts are grown entirely in the Southeast, and federal mar-
COMPARING THE IMPACT OF THE SCENARIOS 129 keting orders bar expansion or movement of peanut cultivation to other regions, the impact of scenario 1 would be concentrated in these south- eastern states. The committee's survey of plant pathologists and examination of the most recent test results in Fungicide and Nematicide Tests indicate that there are three or four new fungicides that control major peanut diseases as well as the currently used oncogenic fungicides.4 Although none of these have tolerances for use on peanuts yet, it appears that non- oncogenic fungicides could become available in the near future. ALTERNATIVES TO THE SCENARIOS On the basis of an analysis of the eight crops discussed, the potential agricultural impact of all scenarios seems severe when measured by the percentage of acre treatments and expenditures associated with revoked tolerances. Efforts to eliminate oncogenic residues in fungicide-treated crops could yield the greatest public health benefits, but could also force significant adjustments in agricultural practices. In light of this finding, the committee decided to explore the impact of cropwide tolerance reduction as a way to reduce risk from fungicide residues in food. Fungicides: A Special Case Fungicide sales in 1985 totaled $269 million, or slightly more than 7 percent of all agricultural pesticides sales.5 In contrast to their small market share, fungicides account for 60 percent of all estimated oncogenic risk. They also provide significant benefits per acre for producers of high-value fruit and vegetable crops. Many growers rely heavily on fungicides, particularly in humid regions. Implementation of any of the scenarios could present problems because relatively few new fungicides have gained registration or been granted tolerances in recent years. Although 14 percent of all R&D expeditures by major U.S. pesticide manufacturers were spent on new fungicide research and development- a commitment roughly twice the percentage of fungi- cide sales- the fungicide market remains an elusive target for most major agrichemical firms.6 Only four products registered since 1972 have gained market shares greater than 5 percent for any food crop.7 In contrast, several herbicides and insecticides introduced during the same time have gained significant market shares, especially the pyrethroid insecticides for use on cotton and vegetables and several new corn and soybean herbi- cides. Almost all newer fungicides are systemic in action; that is, the material translocates to another part of the plant from where it was applied and
1 3 0 REG ULA TING PES TI CIDES IN FOOD residues are generally found inside the plant rather than on its surface. Because of this, systemic fungicides often encourage pathogen resistance. On the other hand, fewer of the new fungicides are oncogenic, and those that are oncogenic tend to be less potent. Older oncogenic fungicides tend to be nonsystemic in action, and most have yet to develop any serious resistance problems. The problems of resistance and the lack of product diversity and depth are complicated by the following circumstances peculiar to fungicides as a class: · Approximately 90 percent of all fungicides used in agriculture are animal oncogens. Many of these compounds are substitutes for each other. Regulatory action taken to reduce oncogenic risks from use of one fungicide will often result in wider use of another oncogen. Indeed, unless the sequence and timing of regulatory actions are carefully planned, total dietary cancer risk from fungicide residues could rise. · In general, raw commodity tolerances for most oncogenic fungicides (except benomyl and some EBDCs on certain crops) were established in the absence of modern residue chemistry data. These tolerances are generally well above actual residue levels and tend to overstate risks from residues of these fungicides on crops. · The dietary risk from residues of these compounds in certain processed foods is probably understated, however, because of the scar- city of processing studies and consequent lack of processed-food toler- ances. Complete data on residue concentration in processed foods exist only for benomyl. Except for tolerances for captan on raisins and mancozeb on raisins, rye, oats, and wheat, no oncogenic fungicide other than benomyl has any section 409 tolerances. Yet it is certain that fungicide residues concentrate in processed foods made from several crops. Fungicide tolerances for residues in animal products are also incomplete. · The use of fungicides in agriculture is concentrated in humid regions of the country, principally the East and particularly the Southeast. Important regional implications need to be considered in evaluating alternative regulatory policies. The combination of these factors makes regulation of oncogenic risk from fungicides a complex and delicate problem. The committee believes that the Delaney Clause, as traditionally interpreted, is not responsive to these considerations. Literal implementation could complicate EPA at- tempts to reduce dietary cancer risk from fungicides. The following analyses are the results of the committee's effort to better understand the challenge confronting the EPA. The committee empha- sizes the need for an approach to reduce fungicide risk that takes all the
COMPARING THE IMPACT OF THE SCENARIOS 131 above factors into account. The approach considered here cropwide tolerance reduction differs significantly from the automatic tolerance revocations by the scenarios analyzed above and from current EPA practice. Cropwide Tolerance Reduction The most compelling argument for addressing the risk posed by all oncogenic pesticides of a given class, such as fungicides, on a given crop is the difficulty of ensuring that regulatory actions against a single compound will reduce dietary risks. Addressing the risk posed by all oncogenic agents in a single class used on a crop requires consideration of different risk reduction measures than those typically used for one compound. Strategies of tolerance reduction or even establishment of zero tolerances have particular appeal here for several reasons: · Zero and level-of-detection tolerances, which are tolerances set when the only residue allowed is that undetectable at the limit of detection, as well as tolerance reductions, have been applied in regulating some of the oncogenic fungicides on tobacco and certain vegetable crops exported to Canada and elsewhere. · The EPA has used level-of-detection tolerances to bypass the Delaney Clause (see the permethrin case study in Appendix C) and to allow the use of an economically valuable oncogenic compound. · Tolerance revocations for individual fungicides, implemented one compound at a time, could actually cause estimated dietary risk to rise. A cropwide tolerance reduction strategy is useful for fungicides be- cause the market is dominated by oncogenic compounds that are ready substitutes for each other; few viable non-oncogenic alternative fungi- cides are in development; and tolerances for many of the widely used older compounds are well above the levels that properly treated crops at the time of harvest should have. In contrast, cropwide tolerance reduc- tion is probably not the optimal strategy for reducing dietary oncogenic risk from corn and soybean herbicides and cotton insecticides because there are numerous non-oncogenic substitutes; risk is attributable to one or two compounds; and tolerances, because they are generally newer, more accurately reflect actual residue levels in food. Table 5-4 presents the order in which the EPA will face regulatory decisions on oncogenic fungicides. As described in Chapter 3, these dates represent the time by which the EPA expects to have completed a special review and registration standard for each of these oncogenic fungicides. At the conclusion of these processes, the EPA will probably know with relative certainty whether a pesticide is an animal oncogen and whether
132 REGULATING PESTICIDES IN FOOD TABLE 5-4 Potential Short-Term Impact of the Delaney Clause on Selected Fungicides Fungicide Date of Market Possible Estimated Risk Share Active Tolerance (% pounds Ingredient Action Action Raw Processed Total applied) Benomyl 1986 Rsa 3.42 x 10-5 7.91 x 10-5 1.13 x 10-4 10 Captafol 1987 sRb 4.34 x 10-4 1.59 x 10-4 5.94 x 10-4 5 EBDCs 35 Mancozeb 1987 RS 2.43 x 10-4 9.44 X 10-5 3.38 x 10-4 Maneb 1987 RS 3.90 x 10-4 5.22 x 10-5 4.42 x 10-4 Metiram 1987 RS 7.65 x 10-5 3.91 x 10-5 1.15 x 10-4 Zineb 1987 RS 4.71 x 10-4 2.45 x 10-4 7.17 x 10-4 Folpet 1987 RS 1.81 x 10-4 1.43 x 10-4 3.24 x 10-4 5 Captan 1988 SR 2.80 x 10-4 1.93 x 10-4 4.74 x 10-4 15 Chlorothalonil 1988 SR 1.89 x 10-4 4.82 x 10-5 2.37 x 10-4 10 NOTE: These risk estimates are derived using EPA data and methods described on pages 5(K6 and in Appendix B. aRS is registration standard. bSR is special review. residues of that pesticide concentrate in processed foods. With such information available, the EPA will have to decide whether or not to revoke or modify tolerances under the Delaney Clause. To determine whether tolerance revocations might increase the risk presented by oncogenic fungicide residues in major fruit and vegetable crops, the committee performed a simple analysis for benomyl and the EBDCS. For selected crops, the percentage of acres treated with a fungicide is assumed to be zero to simulate the effect of tolerance revocation. Then, the shares of total acres treated with the likely replacements are raised to compensate for the loss of the compound eliminated. A new estimate of dietary risk is then computed. All risk estimates analyzed below assume residues at the tolerance level, incorporate TAS residue estimates for processed foods for which no sec- tion 409 tolerances have been established, and are adjusted to reflect the percentage of planted acres assumed to be treated with each fungicide. BENOMYL The fungicide benomyl is the first fungicide active ingredient registered before 1978 for which complete residue and oncogenicity data are available. Benomyl residues concentrate in several processed foods, and
COMPARING THE IMPACT OF THE SCENARIOS 133 TABLE 5-5 Estimated Change in Dietary Oncogenic Risk in Some Crops from Revoking Benomyl Tolerances Benomyl Acres Treated with Dietary Oncogenic Risk (5~) in Replacement Fungicide (Jo) Apples Peanuts Potatoes Tomatoes No replacement -4.8 -1.1 NA -1.5 Captan (50) +26.4 NA NA +3.2 Mancozeb (50) Chlorothalonil (100) NA +14.8 NA +1.5 NOTE: NA means the pesticide is not used or is not considered a likely substitute pesticide on that crop. benomyl is a confirmed animal oncogen. If benomyl tolerances were revoked pursuant to the Delaney Clause (residues are assumed to concentrate in all crops examined), little or no reduction in risk would be achieved (see Table 5-51. This is because, on the basis of current data, benomyl generally poses a lower dietary risk than the most likely substitute fungicides. For each of the crops analyzed, substitution of other compounds for the revoked benomyl tolerances raised the estimated dietary oncogenic risk. Significant increases are evident in apples where benomyl-treated acres were evenly divided between captan and the EBDC fungicide mancozeb. Following this substitution, risk from apple fungicides rose more than 26 percent. In peanuts, dietary oncogenic risk would rise nearly 15 percent if all acres now treated with benomyl were subsequently treated with chlorothalonil. EBDCs Revocation of tolerances for the EBDC fungicides would yield mixed results. For some crops the dietary risk from fungicides would rise, for others it would stay about the same, and for some it would be reduced significantly. Captafol and chlorothalonil are considered the most likely replace- ments for EBDC use on tomatoes. If the acres previously treated with EBDCs were evenly divided between captafol and chlorothalonil, the risk from fungicide residues in or on tomato products would rise almost 50 percent (see Table 5-61. There are more replacement fungicides for use on apples than other crops; thus, many more substitution scenarios could unfold. In all cases examined by the committee, the risk from fungicide residues in apples
~34 REGULATING PESTICIDES IN FOOD TABLE 5-6 Estimated Change in Dietary Oncogenic Risk in Some Crops from Revoking EBDC Tolerances EBDC Acres Treated with Dietary Oncogenic Risk (%) in Replacement Fungicide (56) Apples Potatoes Tomatoes Chlorothalonil (50) NA NA +46 Captan (100) - 18 NA NA Captan (40) -6.7 NA NA Captan (60) -33 NA Benomy} (40) NA Chlorothalonil (100) NA -62.1 NA NOTE: NA means the pesticide is not used or is not considered a likely substitute pesticide on that crop. was reduced. This reduction was not always significant, however. When captan was applied to 40 percent of EBDC acres and folpet to 60 percent, the result was a 6.7 percent risk reduction. The greatest risk reduction, 33 percent, would occur if 60 percent of the former EBDC acres were treated with captan and 40 percent with benomyl. Revocation of EBDC tolerances achieved the greatest reduction in dietary oncogenic risk for potatoes. The committee assumed that all acres treated with EBDCs would be treated with chlorothalonil. This assump- tion reduced estimated risk by 62 percent. The point of these projections is that any regulatory strategy, whether based on the Delaney Clause or any other standard, that attempts to reduce dietary oncogenic risk from fungicides by addressing compounds one at a time will not produce significantly lower risks. The one-pesticide- at-a-time approach may actually increase risk for many widely consumed crops that currently present significant dietary oncogenic risks. The Delaney Clause could worsen this phenomenon if the EPA revoked tolerances for oncogenic fungicides in the order in which data are available to make such decisions. The committee concludes that reducing tolerance levels for all oncogenic fungicides crop by crop will yield greater risk reductions than sequential actions to control individual fungicides across all of their uses. Furthermore, tolerance reductions and even zero tolerances are viable options that could be applied to many pesticides on many crops.
COMPARING THE IMPACT OF THE SCENARIOS 135 NOTES 1. U.S. Department of Agriculture. 1985. Agricultural Statistics 1985. Washington, D.C.: U.S. Government Printing Office. 2. U.S. Department of Agriculture. 1984. Vegetables. Washington, D.C.: U.S. Government Printing Office. 3. U.S. Environmental Protection Agency. 1986. Unpublished data. Washington, D.C. 4. Ritchie, D., ed. 1985. Fungicide and Nematicide Tests. Vol. 40. St. Paul, Minn.: American Phytopathological Society. 5. National Agricultural Chemicals Association. 1985. P. 3 in Impact of Current Law on Agricultural Pesticide Research Productivity. Washington, D.C. Photocopy. 6. Ibid., p. 13. 7. U.S. Environmental Protection Agency. 1986. Unpublished data. Washington, D.C.