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 Although the information provided here was accurate and current when first created, it is now outdated.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 5/28-29/98

Staff Paper # 3.2



In evaluating food residue data, the Agency generally assigns a non-zero residue value to samples with no detectable residues. This practice may overestimate the risk associated with a particular use. In some cases, real risk associated with undetected residues may be inconsequential, but in other cases the risk may be significant, even at exposure levels which cannot be detected.

Measuring the residues. When EPA reassesses a tolerance, we try to refine the dietary exposure assessment to get a realistic picture of actual exposures. One step in the exposure assessment is estimating the level of residues in food. When possible, we use measurements of residues in commodity samples taken at grocery stores. Because we can sample only a tiny fraction of all the food that goes to market it is very important that the commodity samples be representative of all of that commodity in the U.S. food supply. The only sampling programs that will provide this assurance are nation-wide programs such as U.S. Department of Agriculture's Pesticide Data Program or Food and Drug Administration's residue monitoring programs. The handling and processing of the samples must be standardized so that the results of analyzing samples that had been collected at different times and places will be comparable. The next step -- analyzing the sample -- is key to the integrity of the data. Even with the best and most reliable methods, the level of detection varies depending on the instrument, the quality of the analytical chemicals, the skill of the chemist, and the composition of the food being analyzed. This variation in levels of detection can make it difficult to interpret and use the results.

What does "no residues detected" mean to the Agency? The fact that no residues are detected does not necessarily mean none are there. Residues may exist at levels that are too low to detect. Or the pesticide residue may have degraded into compounds that are not recognized by the method being used. "Degradates" may be of concern since it is possible that they may pose a risk.

Calculating the "anticipated residue." When conducting a risk assessment, EPA adjusts the residue measurements to reflect the percentage of the crop that is likely to be treated. If it is reported to EPA that 10% of the crop is treated, but only 5% of samples have detectable residues, it may be that either the percent crop treated data are wrong or residues are present in another 5% of the samples, but at levels too low to detect. Unless EPA has additional data to check out these possibilities, EPA generally assumes that the crop treated data are reliable for predicting pesticide use. We use half the level of detection for the 5% of samples that we believe were treated but do not have detectable residues. Because the percent crop treated data predicted that no residues would be found in 90% of the samples, we use a value of zero for 90% of the samples in the exposure assessment. The statistical basis of these procedures has been accepted by outside peer reviewers.


Further refinements of the anticipated residue value. Approaches for getting better estimates of actual residues in foods where no residues are detected include developing a more sensitive analytical method or conducting residue studies that use exaggerated application rates. If the exaggerated rate study shows detectable residue, the findings could be used to predict residue levels when normal application rates are used.

A way to demonstrate that the actual residue on the ready-to-eat food is lower than the tolerance level is to conduct studies that analyze the identity and levels of pesticide degradation products in commodity samples collected from the time of harvest until the food is ready to eat. In surveys intended to obtain residue estimates for acute dietary exposure assessments, single servings of the food may be sampled.

"No reasonable expectation of finite residues." Under 40CFR.180.6(a), EPA examines available data to see whether a tolerance must be established on milk, meat, poultry or eggs derived from animals fed pesticide-treated feed. If the Agency finds no reasonable expectation of finite residues in milk, meat, poultry or eggs, no tolerance is required. The Agency is considering evaluating data on other commodities to see if a finding of "no reasonable expectation of finite residues" can be made for these foods.


EPA has identified the commodities that constitute the bulk of the diets of infants and children. Now we are determining which OP pesticides are used on these commodities. Residue reduction studies, monitoring data, or other information to show actual levels of residue on the ready-to-eat food, may enable the Agency to further refine its assessment.


1. When EPA believes that a crop has been treated with a pesticide, but no residues are detected, how should EPA address this situation: assign a residue value of ½ the level of detection, assume some other value, require development of other data (e.g., improved analytical method, market basket survey)?

2. Are there circumstances in which the Agency should make a finding that a pesticide residue, for all practical purposes, can be ignored? What kind of criteria should be used to assure risks are inconsequential?

3. Should EPA have more than one approach depending on the reliability of its data?

EPA currently relies on a scientifically peer-reviewed justification for assigning a residue value of half the level of detection when no residues are detected in market survey or monitoring samples of a crop that EPA believes was treated with a pesticide. EPA is currently considering what conditions, such as potential risk and the quality of supporting data, may be used to support the use of "zero" residues in a risk assessment. The Agency believes a logical place to start is to examine food uses where residues are not detected using sensitive analytical methods; where the predicted levels are likely to be much lower than the level of detection; and where the possible risk from predicted levels of residues are acceptably low.

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