Information provided for informational purposes only

Note: This information is provided for reference purposes only. Although the information provided here was accurate and current when first created, it is now outdated.

TRAC 5/27/98

Staff Background Paper # 4



The Office of Pesticide Programs (OPP) evaluates the safety of pesticides through a process that is known as risk assessment. This process involves assessing the toxicity or hazard potential of a chemical and determining how much exposure is likely to occur to ensure that when a pesticide is used both humans and the environment are adequately protected.

Three provisions of the Food Quality Protection Act (FQPA) are having a major effect on the way OPP assesses risk:

Each of these three provisions is thoroughly described in the 1996 Food Quality Protection Act Implementation Plan (March 1997).


To perform a risk assessment, we need data. Under the provisions of FIFRA, the pesticide manufacturers (i.e., registrants), are required to submit a full and comprehensive battery of toxicity and residue chemistry data. Residue chemistry is the term OPP uses for the data required to determine the amounts of pesticide residues in and on all foods and food products, including milk and meats. All the submitted data are reviewed by Agency scientists for conformity with standard practices within the discipline and conformity with Agency Testing Guidelines.

Toxicity Data

On the toxicity side, the tests are conducted on animals dosed by different routes of exposure: oral, dermal, and inhalation. Finally, the toxicity tests are designed such that they explore all types of effects that may occur (e.g., birth defects, cancer, changes in fertility or ability to reproduce, neurotoxicity, harmful effects to the kidney or liver, etc.).

Other sources of toxicity data include the open literature, epidemiological information, and voluntary submissions by the registrants. A good example of a voluntary toxicity data submission would be the test for dermal absorption, which measures how much of the pesticide is actually absorbed through the skin and thus gets taken into the body where it could potentially cause an effect. OPP does not routinely require dermal absorption data. Rather, in assessing risk, we assume that all the pesticide contacting the skin gets absorbed. That is, we conservatively assume that dermal absorption is 100 percent in the absence of data to support a different assumption. Registrants will often voluntarily provide dermal absorption data so that OPP can more realistically assess the actual risk. Where appropriate, EPA has also exempted some pesticides from certain test requirements, e.g., pheromones and certain biologicals, since these tests are not necessary to assess the safety of these types of pesticides.

Residue Chemistry Data

As with toxicity, EPA also requires a full battery of testing so that it can estimate the amount of pesticide residue that may be in the foods we eat and drink. The actual pesticide residue measurements are done using raw agricultural commodities (i.e., fruits and vegetables that are grown in the fields). Raw agricultural commodities are the most basic food forms. To estimate the amount of pesticide residue that would be found in other food forms such as apple juice and raisins, OPP does a calculation, using its knowledge of how much pesticide concentrates in processing, etc.

It is important to note how the actual crop field trials are run. In conducting these types of studies, the person running it will have the pesticide applied at the highest rate possible, according to the label instructions. When the crop is harvested, sampling is done at the 'farm gate' which means that sampling occurs before the crop has gone through any sort of processing such as washing or has entered the channels-of-trade. The sampling is done here so that the residue measurement will represent the highest level of pesticide that might ever be on that fruit or vegetable. It is this level of pesticide residue on food that is used to set the tolerances, which are the legal maximum amount of pesticide that may remain in or on food.

PDP is a joint USDA/EPA effort that
was started in the early 1990s to
generate pesticide residue data
targeted toward foods that infants
and chidren typically eat in larger
proportions than adults, including
milk. For more information, see the
USDA Background Paper on PDP

The NPRD is a newly created
centralized database of all pesticide
monitoring resduedata on food
(provided in Appendix 3 is a fill

In reality, consumers are not exposed to pesticide residues in food at the tolerance levels. So, in assessing dietary risk for purposes other than setting tolerances, OPP often uses (as appropriate) pesticide residue measurements that were taken from foods sampled under more 'real-life' situations, such as at the grocery store or through FDA or USDA monitoring efforts. Sometimes this type of pesticide residue data is referred to as 'dinner plate' residue data. Important sources of 'dinner plate' residues are: FDA and state monitoring; USDA's Pesticide Data Program; and, the Agency's newly created National Pesticide Residue Database (NPRD).

A final piece of information that can be used in assessing dietary exposure and risk is the percent of a given crop that is actually treated with the pesticide. Percent of crop treated corrections are applied for foods with national or broad regional distribution and for chronic consumption patterns (e.g., the food is eaten frequently). As with dermal absorption described under toxicity data, in the absence of percent crop treated data, OPP will conservatively assume that 100 percent of the crop gets treated, whether this is true or not. Such an assumption can lead to an overestimate of the actual exposure level.


In risk assessment terminology,
the identified effects are
referred to as toxicological
endpoints. Short-lived effects
are known as acute and longer
term effects are called chronic.
Risk assessment follows a four-step process:

(1) Hazard Identification; (2) Dose-Response Assessment; (3) Exposure Assessment; and (4) Risk Characterization.

Hazard Identification

An important component of hazard identification is toxicological testing of animals with an assumption that, unless proven otherwise, test results in animals are relevant to the identification of hazards in humans. During hazard identification, all available toxicology data are reviewed to see what harm the pesticide might cause. Some of these harms or effects may be short-lived and will completely disappear after a little while (e.g., headache and nausea). Other effects will appear only after years of exposure (e.g, cancer). Knowing whether the effects are acute, chronic, or both is important in dietary exposure assessment. This is further discussed under Exposure Assessment.

Dose-Response Assessment

Dose-response is the relationship of a dose given to the effect seen in a test. In running a toxicological test, the study investigator is trying to find out at what dose level the effects will occur. To do this, different levels of pesticide (i.e., doses) are given to different groups of animals and they are then studied to see what happens. In some cases, there will be no response in the animals until a certain dose level is reached; in other cases, there will be some response at every dose level. An effect that operates under the first case - no response until a certain dose level is reached - is called a threshold effect. Sometimes people refer to these type of effects as 'all or nothing' - either you see something or you don't. Examples of threshold effects include: general systemic toxicity (weight loss, kidney damage); developmental toxicity (cleft palate); and reproductive toxicity (decreased sperm count). An effect that operates so that there is some response (however small) at every dose level is called a nonthreshold effect. This is sometimes referred to as 'one-hit'- one exposure in a lifetime gives you some chance of experiencing the effect. The classic example of a nonthreshold effect is cancer. This is a model that assumes a linear relationship at low doses between dose and effect. OPP uses this linear model for cancer risk assessment unless it is proven by the registrant that a non-linear model or threshold approach is actually more appropriate, with the support of the broader science community. The distinction between threshold and nonthreshold effects is important when considering the extra 10-fold safety provision of FQPA because, according to the statute, this provision only applies to threshold effects. The 10-fold issue is discussed under risk characterization.

In practice, a threshold effect is evaluated by looking at all the doses given to the animals and identifying the highest one where no effect was seen. This level is called the No-Observed-Effect-Level (NOEL). Nonthreshold effects are evaluated differently. All the doses and their corresponding effects are fed into a computer model which calculates something called a q1*. In technical terms, q1* represents a the upper 95th percentile estimate of the dose-response. Use of the upper 95th percentile is believed to account for possible interspecies or intraspecies differences in sensitivities to the carcinogen. In practice, a q1* can be thought of as the slope of the doses plotted against their corresponding effects. Provided in Appendix 4 is a graphical presentation of how threshold and nonthreshold effects are analyzed.

In assessing risk, one of the goals is to come up with a number that represents a person's margin of safety (e.g., a margin of 100 would mean people would be exposed at a level 100 times lesser than the level believed to cause a toxic effect) or in some cases their probability of experiencing the toxic effect (e.g., a 1x10-6 cancer risk means that the person has a one in a million chance of developing cancer from exposure to the pesticide). As mentioned in the Introduction, there are two prongs to risk assessment: the toxicity and the exposure. Simply put, RISK = toxicity × exposure. The process of putting a number (i.e., quantifying) on the toxicity portion of RISK is called dose-response assessment.

Dose- Response Assessment:

For threshold effects, an RfD

For nonthreshold effects, a q1*

For threshold effects, dose-response is quantified by a reference dose (RfD). A reference dose is an estimate of the level of daily exposure to a pesticide residue, which, over a 70-year life span, is believed to have no significant deleterious effects. The Food and Drug Administration refers to this as an acceptable daily intake or ADI. The pesticide program routinely calculates an RfD by dividing the no-observed-effect level from an animal study by two uncertainty factors - a 10-fold factor to account for uncertainty in extrapolating from animals to humans (i.e., interspecies) and a 10-fold factor to account for the variation within the human population (i.e., intraspecies). In addition to these two 10-fold uncertainty factors, there is also the FQPA 10-fold factor. When more or fewer data are available, OPP has sometimes used different uncertainty factors. Until recently, this FQPA factor was also applied to the RfD. However, it now gets handled at the risk characterization stage. This topic is more thoroughly discussed under Risk Characterization. For nonthreshold effects, the dose-response quantification is the q1* which was discussed in a previous paragraph.

Exposure Assessment

Pesticide exposure occurs through the three routes of exposure - oral, dermal, and inhalation, depending on what the person is doing and where they are. For instance, someone who is eating may be exposed orally through their food and drinking water. Homeowners applying pesticides around the house may be exposed dermally and through the inhalation route; their children may be exposed dermally by touching treated surfaces such as a lawn and they may be exposed orally by inadvertent ingestion (i.e., sucking on their fingers that have contacted the lawn). Workers in an agricultural setting may be exposed dermally and through the inhalation route from mixing, loading, and applying pesticides. Finally, all pesticide exposures may be of a short-term (acute) or long-term (chronic) duration. Provided below is a discussion of how OPP estimates dietary exposure.

Aggregate Exposure

As mentioned earlier, FQPA is having a major impact on how OPP assesses risks from exposure to pesticides. The provision that affects how we assess dietary exposure is 'aggregate exposure;' it states: "In establishing...a tolerance...for a pesticide chemical residue, the Administrator shall consider...available information concerning the aggregate exposure levels of consumers (and major identifiable subgroups of consumers) to the pesticide chemical residue and to other related substances, including dietary exposure under the tolerance and all other tolerances in effect for the pesticide chemical residue, and exposure from other non-occupational sources"

§ 408(b)(2)(D)(vi).

EPA has interpreted this provision to mean that in addition to the pesticide exposure that occurs through the diet, OPP must also include exposure that occurs from non-occupational sources, which include drinking water and residential exposure.

Exposure through the Diet

The Residue Chemistry Data section of this paper describes how OPP obtains the information for assessing the amount of pesticide exposure that occurs through the diet and briefly how the exposure is calculated. The one piece of information to add is the acute vs. chronic exposure. In acute dietary exposure assessments, OPP generally has used the worst-case (i.e., those at tolerance level) exposure estimates while in chronic exposure assessments more real-world estimates are used. OPP in now beginning to use other techniques such as Monte Carlo analyses to provide more realistic acute dietary assessments (see staff paper #3.1).

Exposure through Drinking Water

Data on how a pesticide behaves in the environment are routinely required under the 40 CFR Part 158 regulations but drinking water residue data are not. Monitoring data for pesticide residues in ground and surface water are available from various sources (U.S. Geological Survey, states, and academia) and provide information on ambient water quality only and not drinking water quality, per se. The Agency currently uses simulation models, which utilize conservative assumptions (health-protective), as screening tools to estimate pesticide residues in ground and surface water. When the pesticide passes the screen there is no need for further evaluation. If the screening numbers are high, EPA will evaluate data from the U.S. Geological Survey, states, and other sources to attempt to provide a more realistic estimate.

To address the problem of estimating pesticide residues in drinking water, EPA has been working with ILSI, the International Life Sciences Institute, which is a nonprofit, worldwide foundation established in 1978 to advance the understanding of scientific issues related to nutrition, food safety, toxicology, and the environment.

For a more thorough discussion on this topic and how OPP has evolved from the process description in the interim decision logic to today, see the staff paper #4.1 on: "Issues Associated with Drinking Water Assessments."

Exposure through Residential Activities

As with drinking water, reliable and fairly refined residential and other non-occupational exposure estimates are needed to aggregate exposure. Also, as with drinking water, these data are not routinely required under the 40 CFR Part 158 data requirements. Before FQPA, HED was able to rely on the Pesticide Handlers Exposure Database (which was developed for worker exposure) and very conservative assumptions regarding exposure in a residential setting. However, with our current requirement to aggregate residential exposure with dietary, we need better, more refined residential exposure data. To obtain these data, EPA is working with industry in the development of indoor and outdoor residential exposure data. Currently, the Outdoor Residential Exposure Task Force is developing exposure data for professional and nonprofessionals (i.e., homeowners) who use pesticides on turf or lawns and for exposures incurred while coming into contact with a lawn that has been treated. Additionally, industry is now initiating a similar effort to generate indoor residential exposure data.

In the meantime, as with drinking water, OPP is using currently available information. For a more thorough discussion on this topic, see the staff p`aper #4.2 on: "Exposure Data for Residential Risk Assessments."

Risk Characterization

The final step in risk assessment is characterization, which is the process of combining the dose-response and exposure information to describe the overall magnitude of the public-health impact. For threshold effects, risk can be expressed via a margin-of-exposure (MOE) or as a percent of the reference dose (% RfD). For nonthreshold effects, risk is expressed as a probability (e.g., 1x10-6). The formulas for these are:

MOE = NOEL ÷ Total Dietary Exposure

% RfD = Total Dietary Exposure÷Reference Dose

Probability (of Developing Cancer) = q1* × Total Dietary Exposure

Total Dietary Exposure is the dietary exposure from the food residues aggregated with the nonoccupational residues combined with what we know about what people in the United States eat and in what proportions. This information is known as food consumption and is supplied by the U.S. Department of Agriculture. Food consumption data allows EPA to estimate dietary risks for the U.S. population as a whole along with 26 different population subgroups, including eight that are specific to infants and children.

The FQPA Factor

EPA's approach to implementing the FQPA safety factor has continued to evolve on the basis of new data, along with experience and a better understanding of potential toxicity to young animals from pre-natal or post-natal exposure to pesticides and continuing peer review of decisions and policies. Before the passage of the FQPA, EPA analyzed all the available toxicology studies to identify the most appropriate endpoints (e.g., cholinesterase inhibition, effects on central nervous system, reduced birth weight in young test animals) for risk assessments.

Following passage of FQPA, while still analyzing all of the available data, EPA began to look at developmental and reproduction studies in a more focused manner. After the entire toxicity database was studied and endpoints identified in the traditional manner, the developmental/ reproduction studies were analyzed further to determine whether different endpoints should be used in the assessment of hazard to children and women 13 years of age and older. In general, decisions on the FQPA safety factor have been based on weight of the evidence and professional judgment with few formal decision criteria.

In January 1998, upon evaluation and review of the initial approach, EPA submitted a new formal approach to the FQPA safety factor to the Scientific Advisory Panel for review. This approach gives more clarity to how the Agency now considers the completeness of the toxicity database, the type and severity of the effect observed, and the nature and quality of the available exposure data. The application of the FQPA safety factor is not a matter simply of uncertainty, but is also a way of assuring an extra measure of protection for infants and children in cases where special sensitivity or exposure for these subgroups is identified. The retention, reduction, or removal of the FQPA safety factor is still based upon a weight-of-evidence evaluation of all applicable data and reflects sound scientific judgment and principles. An internal FQPA Safety Factor Committee consisting of toxicologists, exposure scientists, and risk managers recommends whether to retain, reduce, or remove the FQPA safety factor


Appendix 1 - Overview of Human Health Toxicity Data Required For the Registration of Pesticide Chemicals.

Appendix 2 - Overview of Residue Chemistry Data Required for the Registration of Pesticide Chemicals

Appendix 3 - The National Pesticide Residue Database

Appendix 4 - Threshold and Nonthreshold Effects

TRAC Home | OPP Home | EPA Home | Comments
Site Map | Search OPP | Search EPA
updated May 22, 1998