Jump to main content.

March, 1998 American Crop Protection Association Presentation

February 5, 1998

Dr. Lynn R. Goldman
Assistant Administrator for
Prevention, Pesticides and Toxic Substances
U.S. Environmental Protection Agency
Room E637
401 M Street, SW
Washington, DC 20460

Dear Lynn:

Thank you again for meeting with us on such short notice last week to discuss our concerns about the need for due process and transparency in the Agency's decision making process under the Food Quality Protection Act (FQPA), particularly with respect to application of the additional 10X uncertainty factor and other decisions which could result in changes in the reference dose (RfD) of pesticide chemicals. We especially appreciate your candor in acknowledging problems with the process in recent weeks and the importance of correcting those problems.

As you will recall, we requested an opportunity for our companies to participate early and in a constructive way in your efforts to develop criteria for application of the 10X uncertainty factor. You welcomed our suggestions, but pointed out that you were under serious time constraints due to your plan to send EPA's proposed criteria to the FIFRA Scientific Advisory Panel (SAP) for consideration in their upcoming March meeting. We committed to getting our recommendations to you this week. Those recommendations are enclosed with this letter.

We believe that the enclosed recommendations represent sound science as well as sound public policy and are true to both the spirit and letter of the FQPA. We request that you give them most serious consideration as you develop the Agency's proposed criteria for SAP review. Specifically, we request that a delegation representing our companies be given the opportunity to meet with your staff to discuss these recommendations before you finalize the 10X criteria document you will send to the SAP. We are prepared to meet at your earliest convenience.

At our January 28th meeting we also discussed the need for a process for individual companies to be notified-and have an opportunity to meet with the Agency'prior to EPA reaching a final decision on the application of additional uncertainty factors that will result in significant changes to individual pesticide chemical RfD's. Again, you welcomed our suggestions on the subject and we will be sending our recommendations to you shortly.

In closing, we want to reaffirm our support for the goals of the FQPA'including the protection of children's health-and for effective implementation of this landmark legislation. We believe that such implementation will be advanced if the pesticide industry can have the kind of early and meaningful participation in the Agency's decision making process that we discuss in this letter and its enclosure, and that you encouraged in our meeting on January 28th.

A representative of our companies will contact your staff shortly to schedule the meeting we have requested to discuss the enclosed recommendations.

Sincerely yours,

Emilio J. Bontempo
Novartis Crop Protection
A. Charles Fischer
Dow AgroSciences
Emil E. Lansu
Bayer Corporation
Agriculture Division
Alan Reade
Rhône-Poulenc Ag Company

c: Jay Vroom
American Crop Protection Association

Use Of Additional Uncertainty Factors to Account For Greater Sensitivity Of Infants And Children

 Executive Summary

The Environmental Protection Agency (EPA) must decide what criteria to use to carry out the Food Quality Protection Act (FQPA) provision on an additional safety factor of up to 10X to protect infants and children. If this provision is applied inappropriately, it could deprive agriculture of useful pesticides without any added protection of children's health.


Health and regulatory agencies around the world establish safe levels for chemical exposure (acceptable daily intake or reference dose) by applying safety or uncertainty factors (UFs) to NOAELs derived from appropriate animal studies. A 10X UF is typically used to deal with the uncertainty in extrapolating from animals to humans and another 10X factor is used to account for possible variability within the human population.

This paper recommends a balanced approach that EPA can use to implement the provision of the FQPA regarding the use of an additional uncertainty factor for the purpose of protecting infants and children. How EPA interprets and applies this provision can dramatically affect EPA decisions on whether to approve new pesticides or uses and whether already-approved uses should be continued. If it is applied too rigidly it can deprive agriculture of useful pesticides without adding any health protection.


The purpose of the FQPA provision is to provide a suitable degree of added protection to account for the potential of increased sensitivity or exposure of fetuses, infants or children to pesticides.

The FQPA states, in part:

"In the case of threshold effects .... an additional ten-fold margin of safety for the pesticide chemical residue and other sources of exposure shall be applied for infants and children to take into account pre- and post-natal toxicity and completeness of the data with respect to exposure and toxicity to infants and children. Notwithstanding such requirement for an additional margin of safety, the Administrator may use a different margin of safety for the pesticide chemical residue only, if, on the basis of reliable data, such margin will be safe for infants and children."

This provision (and other related provisions in the FQPA) indicate not only the need for prudence with regard to possible increased risk but also the importance of examining each case on its merits and a recognition that in some cases there will be no need to apply an added safety factor. Some cases will raise fewer concerns than others, and there is an allowable range of factors to use. Therefore, an additional 10X safety factor should not be applied if a lesser factor is adequate or if no extra factor is needed. A second principle should be that the goal is to lessen uncertainty by encouraging the development of relevant information. When a new criterion is adopted or when new data requirements are imposed to address uncertainty, affected parties should be allowed time to provide that information.

When considering the enactment of the FQPA, Congress mistakenly thought that EPA already used a 10X uncertainty factor whenever there was evidence of developmental toxicity. The House committee report (1) states that the section is based upon the recommendation of the 1993 National Research Council (NRC) "Pesticides in the Diets of Infants and Children" report (2), which includes the statement that:

"This third 10-fold factor has been applied by the EPA and FDA whenever toxicity studies and metabolic/disposition studies have shown fetal developmental effects."

Insofar as EPA is concerned, that statement was incorrect both in 1993 and in 1996. The NRC report does not furnish a basis for the statement. EPA's own documents on how to calculate reference doses do not mention any extra factor for developmental effects. (EPA's rules do allow-but do not require-the use of an added uncertainty factor when particularly severe effects of any kind are noted.) In fact, prior to the 1996 enactment of the FQPA, the EPA routinely and consistently used a 100-fold safety factor, not a 1000-fold safety factor, when developing RfDs from animal-study NOAELs on products whose complete data bases showed developmental toxicity, as stated in a 1996 EPA presentation on this subject to the FIFRA Scientific Advisory Panel (SAP) (3). The FQPA requires the reevaluation of hundreds of tolerances using a presumption that Congress thought would be little different from that used in the original evaluation insofar as developmental effects are concerned, but that in fact is potentially much more stringent. Accordingly, EPA should be open to allowing a transitional period when applying the additional 10X uncertainty factor as regards potential developmental effects.

This is especially important with regard to previously approved products. While the FQPA requires EPA to reevaluate existing tolerances according to a schedule, Congress also included various provisions in FFDCA §408(f) and (g) to provide an orderly way to obtain newly important data and to provide a transition that allows studies to be conducted and evaluated before tolerance reevaluations have to be made effective.

Clearly, an incomplete database is a justifiable cause for concern. In a regulatory context, however, fairness dictates that the criterion for completeness be announced, and the regulated party be given an opportunity to satisfy the criterion, before adverse action is taken on the basis of incompleteness. When deciding whether a data set is complete, the Agency should take into account not only what study it lacks but also how recently the study was deemed to be required, when the registrant or other sponsor of the compound was informed of the need for the study, and whether the registrant has had a reasonable time to conduct the study.

If the Agency asserts that a particular kind of study is required to provide data pertinent to risks to infants or children, the same rules should apply as for all other types of data requests:

  1. The data requirement should be made known to the registrant/sponsor. If the data requirement will apply to all compounds, or all within a particular category, then, before it is imposed, the Agency should provide an opportunity for public comment on the reasonableness of the requirement.
  2. Sufficient time for the registrant to generate and submit the data should be allowed before any adverse action is taken, or adverse inference made, because of the absence of the data. If the registrant elects not to perform the study, fails to make progress, or fails to submit the study on time, the registration and associated tolerances could be suspended (if the data call-in procedures were used) and/or an additional safety factor could be imposed. Otherwise, no sanction, additional safety factor, or other adverse inference is warranted.

EPA has imposed many data requirements on pesticides by publishing rules (40 CFR Part 158) and a series of implementing guidelines, or by informing registrants that particular studies are required through actions such as re-registration data call-in notices. A multi-generation reproductive study and developmental studies in two species have been requirements for a long time.

On the other hand, until very recently, the developmental neurotoxicity study has been required only rarely, and no general announcement of the kinds of pesticides to which it applied had ever been made. Indeed, the Agency informed the SAP in September 1996 (3) that the study was required if the pesticide had been shown to cause "central nervous system malformations" in developmental studies, but that otherwise it was only being required on a case-by-case basis. Some registrants were asked to provide the study and were given registrations conditioned on their agreement to supply the study in the future. In late November 1997, the Agency announced for the first time (in a letter to three environmental organizations) that a developmental neurotoxicity study is ordinarily required for all organophosphorus pesticides. The industry has not yet been given any opportunity to comment on the propriety or reasonableness of this pronouncement as to when the study would be required.

If EPA determines that a developmental toxicity or reproductive effects study is needed to replace or supplement a previously accepted study (e.g., to address recent revisions in the guidelines or a reinterpretation of the earlier study), or if the EPA determines that it needs a particular kind of study not previously required (e.g., a developmental neurotoxicity study), of course the new study may be requested. No additional UF should be applied based on completeness in this kind of situation until an opportunity to conduct and submit the study is provided.

If the Agency asserts that there is sufficient uncertainty about a particular toxicity issue to raise serious concerns about possible risks to infants or children but has not determined what kind of study or data are needed to address the uncertainty, it may wish to use an additional UF. In such a case, to be fair to registrants and users and to encourage the development of information to lessen uncertainty, a process is needed to explore how to lessen the uncertainty, such as the following:



Pesticide exposure is typically estimated from a combination of measured consumption and residue values (food and water) or from measures of exposure resulting from a variety of non-dietary uses (residential indoor/outdoor) and activity patterns. The amount and quality of exposure data available vary considerably for different pesticides and specific population subgroups such as infants and children. Uncertainty is handled by the use of appropriate assumptions and default values on a case-by-case basis or by the selection of reasonable upper-bound values from probability-based distribution analysis. Caution should be taken to avoid using unreasonable worst-case scenarios or unreasonably high percentiles. EPA is already using high-percentile food consumption data, worst-case food and water residue data or assumptions, and highly conservative residential exposure assumptions. Therefore, additional uncertainty factors based on exposure considerations are not needed.


1. Use of Uncertainty Factors

Health and regulatory organizations throughout the world have long recognized the existence of dose thresholds for most adverse health effects and typically establish safe human levels for regulatory purposes [RfD, acceptable daily intake, etc.] by applying a variety of "safety" or "uncertainty factors" (UFs) to NOAELs from appropriate animal or human studies (4-6). The UFs typically applied are a 10X factor for uncertainty in extrapolating from animals to humans and a 10X factor for possible variability within the human population (7). Additional UFs may be added to account for data gaps, for particular severity of effects, etc., and if good human data are available, the 10X interspecies UF may be reduced or omitted (8). While the 100X UF originally was not based on extensive data (9), the use of the UF approach has come to be accepted widely. UF values should not be applied routinely as across-the-board defaults but should reflect case-by-case judgment decisions made by experts who have carefully evaluated the data available.

Until very recently EPA ordinarily assumed that the 10X UF for variability within the human population accounted satisfactorily for any age-related differences and was protective for infants and children (3). The FQPA now requires that EPA consider whether an additional UF of up to 10X should be applied to protect children with regard to potential effects that may result from prenatal or early postnatal exposure.

2. Relative Sensitivity Of Infants, Children, and Adults To The Effects Of Pesticides

There is no a priori reason to think that any particular pesticide (or other chemical) is more likely to cause adverse effects in infants and children than in adults. The NRC report Pesticides in the Diets of Infants and Children (2) clearly states (pp. 106-107):

 "Studies in laboratory animals have demonstrated an age-related difference in acute toxicity; however, the direction of the difference is dependent on the chemical, and the magnitude of the effect is usually no more than 1 order of magnitude and often is considerably less. A developing animal may be more, less, or equally sensitive to a given chemical than is an adult. . . . Data from studies in humans (e.g., in children and adults treated with cytotoxic chemotherapeutic agents) show that toxic effects are similar qualitatively but may differ quantitatively. . . . Indeed, for some drugs, the maximum tolerated dose is greater in developing than in human adults (i.e., such drugs are less toxic to infants and children)."

The sensitivity of developing organisms to the same dose of a chemical may be less than, similar to, or greater than that of an adult depending on differences in adsorption, distribution, metabolism, elimination, or target organ sensitivity. Because of legal and ethical constraints, it is highly unlikely that studies will ever be conducted to measure directly possible adult vs. neonate differences in humans. But toxicity studies on developing animals are routinely conducted to investigate this possibility (i.e., developmental toxicity studies in rat and rabbit and multi-generation rodent reproduction studies). If such studies demonstrate that in utero or neonatal exposure results in severe adverse effects while the same dose has no effect in the adult animal, then in the absence of human data, an added safety factor to account for the possible increased sensitivity of human infants and children may be warranted. It should be noted, however, that in the case of developmental and reproductive studies, the developing fetus or neonate (not the adult) is being studied directly; for endpoints from these studies there is no need to extrapolate from effects observed in adults to those that might occur in infants. This should clearly serve to decrease the degree of uncertainty involved.

Although there is no broad-based study comparing NOAELs in adult and young animals, there is useful information in comparisons of other endpoints. Studies by EPA investigators (10-12) on 20 anti-cholinesterase compounds showed that the ratio of LD50's for adults to the LD50s of weanling rats ranged from 0.18 to 3.64. The mean and median ratios for these 20 chemicals were 1.38 and 1.31, respectively. In these studies, adults were more sensitive than the weanlings in 7 of 20 cases whereas weanlings were more sensitive in 12 of 20 cases, but only in one case were weanlings more than twice as sensitive. Clearly, in this class of similar chemicals with a similar effect (cholinesterase inhibition), there is no clear pattern of greater toxicity in young animals than mature animals and that the differences in young and adult animals' susceptibility to any particular chemical are not large. Similarly, Sheehan and Gaylor (13) compared the LD50 ratios of adult to newborn mammals for 238 chemicals and found a median ratio of 2.6 (adults to newborn). In more than 85 percent of these evaluations the ratios were less than 10. It is clear that from these acute comparisons that a 10-fold intraspecies uncertainty factor would most likely cover any difference in sensitivity between adults and children without the introduction of an additional uncertainty factor.

3. Relevant Studies

The routinely-required toxicological studies of most relevance in evaluating potential pre- and post-natal toxicity are a developmental toxicity study in two species (rats and rabbits) and a multi-generation reproductive toxicity study in rats. These studies constitute a stringent test because the developmental toxicity studies typically involve maximally tolerated doses administered throughout the sensitive period of organogenesis and the multi-generation study involves long-term exposure of a large number of animals (parents and pups) and observations of many postnatal and reproductive endpoints. The adequacy of these studies to detect the effects of chemical exposure on fetal and neonatal development is clearly of central importance in determining whether a pesticide is likely to present a special potential risk to human infants and children.


EPA's usual inter- and intra-species uncertainty factors allow for the possibility that human are 10 times more sensitive than test animals and that some humans may be 10 times more sensitive than other humans (14). Experimental evidence has supported that this approach is reasonable and workable (15). An additional uncertainty factor, however, could be applied on a case-by-case basis (7). These cases would be where there is inadequate toxicological information or for those pesticide residues producing severe and irreversible developmental or reproductive effects (7,16).

If it is determined that an additional UF is needed to protect infants and children, the resulting RfD must only be used in relation to scenarios relevant to the exposure of fetuses, infants, and children. The safety factor used for developmental endpoints should not be applied indiscriminately to NOAELs derived from other toxicity endpoints obtained from studies in adult animals.

Table 1 includes a number of different example scenarios where based on a weight-of-evidence approach, an expert might conclude that an additional uncertainty factor might be warranted.

1. Completeness and Quality of the Database

An incomplete database may be a justifiable cause for the issuance of a data call-in (DCI) or the assignment of an additional UF pending completion of the study, especially when there is little information available on the toxicity of the chemical class.

2. Evidence Of Developmental And/Or Reproductive Toxicity

The availability of a complete set of required developmental and reproductive toxicity studies conducted according to acceptable guidelines ordinarily should provide a reliable, high quality database on which to reach conclusions regarding potential adverse effects on infants and children. There is, of course, a wide range of effects that may be observed in developmental and/or multi-generation toxicity studies, and not all of these should be considered to be equally serious or relevant to humans. The database for each pesticide should be subjected to the type of careful evaluation described in the Agency's September 30, 1996, submission to the FIFRA SAP on this subject. The following factors should be considered in determining the level of concern regarding potential effects on infants or children:

3. Additional Information

Other kinds information that should be taken into consideration as appropriate in a total weight-of-evidence evaluation of whether the additional FQPA safety factor is needed are:


Because of the many factors involved in determining the completeness and reliability of the database and evaluating the severity and relevance of the effects vis-à-vis possible effects on infants and children, it should be clear that an across-the-board application of an additional UF of 10X is neither necessary nor desirable. Indeed, in the case of a pesticide with a complete and reliable database that shows no indication of any adverse developmental or reproductive effects, there is no justification for adding any additional UF. In this case, the RfD should be evaluated by applying the usual 100-fold UF to the NOAEL or a 10-fold uncertainty factor when the NOAEL is derived from a human study.

If it is concluded that the results of the studies do raise concerns about possible developmental or reproductive effects and that these might be relevant to humans, an additional UF may be considered appropriate. It is important, however, that the magnitude of this UF should accurately reflect the data available and the level of concern. Sound scientific judgment is needed and, when necessary, an appropriate UF developed for a specific chemical.

  1. House Committee on Commerce, Food Quality Protection Act of 1996. 104cCongress, 2nd session. Report 104-669, part 2, pp. 1-89. Washington, D.C.: Government Printing Office.
  2. National Research Council. 1993. Pesticides in the Diets of Infants and Children. Washington, DC, National Academy Press.
  3. EPA, 1996. Is an additional uncertainty factor necessary and appropriate to assess pre- and post-natal developmental and reproductive effects in infants and children exposed to pesticides thru chronic dietary exposure? (Submission to FIFRA Scientific Advisory Panel), 9/30/96
  4. Lehman, A.J. and Fitzhugh, O.C. 1954. 100-fold margin of safety. Assoc. Food Drug. Off. U.S.Q. Bull. 18:33-35.
  5. IPCS. 1987. Principles for the Safety Assessment of Food Additive sand Contaminants in Food. Environmental Health Criteria 70, World Health Organization, Geneva.
  6. Barnes, D.G. and Dourson, M.L. 1988. Reference Dose (RfD): Description and use in health risk assessment. Reg. Toxicol. Pharmacol. 8:471-486.
  7. Dourson, M.L., Felter, S.P., and Robinson, D. 1996. Evolution of Science-Based Uncertainty Factors in Noncancer Risk Assessment. Reg. Toxicol. Pharmacol. 24: 108-120.
  8. IPCS. 1994. Assessing Human Health Risks of Chemicals: Derivation of Guidance Values for Health-based Exposure Limits. Environmental Health Criteria 170, World Health Organization, Geneva.
  9. Dourson, M.L. 1996. Editorial: Uncertainty Factors in Noncancer Risk Assessment. Regul. Toxicol. Pharmacol. 24:107.
  10. Gaines, T.B. and Linder, R.E. 1986. Acute toxicity of pesticides in adult and Weanling rats. Fund. Appl. Toxicol. 7: 299-308.
  11. Padilla, S., McPhail, R., and Reiter, L. 1985. Neurotoxic potential of pesticides: Age-related effects of pesticides relevant to youth in agriculture. Division Report, U.S. Environmental Protection Agency, Health Effects Research Laboratory, Research Triangle Park, N.C.
  12. Durham, W.F. and Mitchell, C.T. 1987. Final Report: Health Effects for Youth in Agriculture Program., U.S. Environmental Protection Agency, Health Effects Research Laboratory, Research Triangle Park, N.C.
  13. Sheehan, D.M. and Gaylor, D.W. 1990. Analysis of the adequacy of Safety Factors. Teratology 41. 590-591.
  14. Renwick, A.G. 1993. Data-derived safety factors for the evaluation of food additives and environmental contaminants. Food. Addit. Contamin. 10: 275-305.
  15. Renwick, A.G. 1991. Safety factors and establishment of acceptable daily intakes. Food. Addit. Contamin. 8: 135-149.
  16. Renwick, A.G. 1995. The use of an additional safety or uncertainty factor for nature of toxicity in the estimation of acceptable daily intake and tolerable daily intake values. Regul. Toxicol. Pharmacol. 22: 525-545.




Completeness of Database
1. No developmental or multi-generation study available. 1. Data Call-In. 10X UF applied.
2. One developmental toxicity study missing but no evidence of adverse effects from other studies. 2. Data Call-In. Up to a 3X UF applied.
3. Studies available but do not meet latest guideline requirements; no evidence of adverse effects. 3. Data Call-In. No additional UF applied.
4. Agency request for a new extra-guideline study (e.g., developmental neurotoxicity). 4. Data Call-In. No additional UF applied.
Reliability of Database
1. Poor quality of required study due to non-compliance with GLP or current EPA or OECD guidelines. 1. Study considered incomplete. Data Call-In. 3X-10X UF applied.
Developmental Effects in the
Absence of Maternal Toxicity
1. Malformations, possible hormone-mediated effects or neuropathological effects in pups. 1. Application of up to a 10X UF depending on total weight of evidence (e.g., steepness of dose-response curve).
2. Non-specific, reversible effects such as decreased pup weight, skeletal variation, delayed ossification. 2. Application of up to a 3X UF depending on total weight of evidence (e.g., steepness of dose-response curve).
Developmental Effects Observed Only in the Presence of Maternal Toxicity No additional UF applied.
Reproductive Effects in the Absence of Parental Toxicity
1. Irreversible histopathological effects on reproductive organsor fertility/reproductive effects in F1 or F2 animals. 1. Application of up to a 10X UF depending on total weight of evidence (e.g., steepness of dose-response curve).
2. Less serious, non-specific, or reversible effects in F1 or F2 animals. 2. Application of up to a 3X UF depending on total weight of evidence (e.g., steepness of dose-response curve).

1 As discussed in the text, the need to apply any uncertainty factor should reflect the result of a total-weight-of-evidence evaluation of many factors that include the severity of the effect, the steepness of the dose-response curve, mechanistic information, and relevance of the effect to humans (biological plausibility or relevance of dosing regimen to likely human exposure).

Scientific Advisory Panel (SAP) March 1998 Meeting
American Crop Protection Association Presentation

Local Navigation

Jump to main content.