Environmental Backgrounder: Pesticides and Food Safety
Environmental Backgrounder: Pesticides and Food Safety
The Environmental Protection Agency (EPA) is responsible for regulating the
amount of pesticide residue that may remain in or on food and animal feed. EPA's
regulatory authority derives from the Federal Food, Drug, and Cosmetic Act
(FFDCA) and the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA).
Overview of the Pesticide Tolerance Process
Pesticide tolerances established by EPA are enforced by the Food and Drug
Administration (FDA), the U. S. Department of Agriculture (USDA), and state
enforcement agencies. The U.S. food supply is monitored not only for compliance
with pesticide tolerances, but also for potential contamination by aflatoxin,
salmonella, and other microbiological contaminants. Since surveillance resources
are limited, they must be allocated on the basis of priority concerns; pesticide
residues are generally considered to be of a lesser concern than microbiological
contamination.
Federal and state inspectors commonly sample food and feed produce for the
purpose of tolerance enforcement soon after a farmer markets the treated
commodity, so that any tolerance violations may be traced to their source. Thus,
tolerances are intended to apply to treated raw agricultural commodities as soon
as they enter into commerce, starting when the produce leaves "the farm gate."
Pesticide residues tend to dissipate or break down as time passes after harvest,
and in the majority (but not all) cases, residues are further reduced by
washing, peeling, cooking, and processing of foodstuffs. Almost by definition,
tolerances represent levels of pesticide residue that are not expected to occur
as actual residues in food commodities that reach the consumer.
In general, the data required for a tolerance are of two kinds -- residue
chemistry data and toxicology study results. Such data are not generated by EPA
laboratories; rather, EPA uses its data-collection authorities under the law to
require the pesticide manufacturer to sponsor testing to produce these data.
All of EPA's tolerance data requirements are designed to answer three key
questions. First, what is the chemical residue? Second, how much residue is
there? The "what" and "how much" information, derived from residue chemistry
data, is then used by EPA toxicologists to answer the third question Does the
residue represent an acceptable dietary level of exposure? In other words, is
there a reasonable assurance that under the prescribed conditions of use of the
pesticide, no unreasonable adverse effect will result in humans even after a
lifetime of exposure?
Residue Chemistry Data
Residue chemistry data required for tolerance purposes include several kinds of
data ranging from "product chemistry" information to food processing study
results.
Product Chemistry Data:
To characterize the pesticide substance, EPA requires composition data for
pesticides, called product chemistry data. In particular these data include (1)
information on the manufacturing process, (2) chemical analysis of the active
ingredient, (3) "certified limits"for unavoidable variations in the amounts of
the ingredients in a product, and (4) analytical methods used to determine the
composition of the pesticide.
EPA evaluates this product chemistry data set to determine whether impurities
could constitute a significant component of the residue in food or animal feed.
This is an important consideration because impurities created in the manufacture
of a pesticide may become a residue problem, if they are not identified before
tolerances are established.
Metabolism in plants and animals:
Plant metabolism data are required so that EPA may characterize the nature of
the residue that occurs on crops intended for consumption as food or animal
feed. To obtain this kind of data, the pesticide is labelled with a radioactive
atom, usually Carbon fourteen, and applied to the crop plant in accordance with
proposed use directions. Since the pesticide molecule is radiolabelled, any
metabolite or degradation product remaining in the plant at maturity will be
radioactive. The Carbon fourteen activity is separated into various fractions,
and the chemicals associated with the activity are identified. It is very
important to identify most of the radioactivity before tolerances are
established. If this is not done, previously unidentified residues may become
problematic in the future as more highly sensitive analytical methods are
developed which could possibly detect residues of concern.
Plant metabolism studies are required for a minimum of three diverse crops,
usually a root crop, an oilseed, and a leafy vegetable. If the metabolism in
each of these crops is similar, then the metabolism in other crops is assumed to
be similar. At the end of this process, EPA has enough information to answer the
question, what is the residue in plants? Whenever a proposed use of a pesticide
may result in residues in animal feed, or when a pesticide is intended for
treatment of livestock, animal metabolism studies are required in addition to
plant metabolism data. Animal metabolism studies are generally carried out on
ruminants (cows or goats) and poultry (chickens).
Like plant metabolism tests, animal metabolism studies use radiolabelled
pesticides. The animals are dosed, and the level of radioactivity resulting in
potential meat or poultry products (muscle, liver, kidney, milk, and eggs) is
analyzed. If significant activity is found, then the chemical identity
associated with the activity is determined. This process answers the question,
what is the residue in animals?
Significant Metabolites and Tolerance Expression:
Using the results of plant and animal metabolism studies, EPA determines which
metabolites are of concern and need to be included in the tolerance. In each
case, this decision is based on (l) the toxicity of the metabolite, (2) the
percent and magnitude of its residue, and (3) whether a practical analytical
methodology is available or can be developed to detect and measure the
metabolite. For metabolites that are toxicologically significant and occur at
significant levels, a suitable analytical methodology is mandatory. Considered
together, the pesticide active ingredient and any significant metabolites are
called the "total toxic residue."
Residue Field Trial Data:
Once the metabolism data have indicated what to look for, and methods have been
developed to measure the total toxic residue, field experiments are carried out
to answer the question, how much residue is there? These are studies in which
the pesticide is applied to crops at known application rates, in a manner
similar to the use directions which will eventually appear on the pesticide
label if the tolerance and registration are approved. The field trial must
reflect the use conditions that could lead to the highest possible residues.
Generally this means the highest permissible application rate, the maximum
number of applications allowed, and the shortest pre-harvest interval permitted
by the use directions. Data are normally required for each crop or crop group
for which a tolerance and registration are requested. In addition, data are also
required for each raw agricultural commodity derived from the crop. As an
example, for corn, residue analyses would be required for both grain and feed
items -- forage, silage, and fodder.
Analytical Methods:
Based on plant and animal metabolism study results, EPA requires tolerance
petitioners to develop analytical methods to determine all components of the
total toxic residue. In some cases, it is not possible to develop a single
method that can determine all components of the residue, and several methods are
required. Pesticide analytical methods are used for two purposes: (l) to obtain
residue data on which dietary exposure assessments and tolerances are based, and
(2) to enforce the tolerance after it is established. EPA validates each new
analytical method using a method trial, to ensure that the procedures can
actually be used for tolerance enforcement purposes by FDA, USDA, and state
agencies.
Determining the Tolerance Level:
A petitioner for a tolerance proposes a tolerance level, based on residue field
trial data, which reflects the maximum residue that may occur under "worst-case"
conditions as a result of the proposed use of the pesticide. The tolerance must
include significant metabolites and must be high enough to cover all components
of the total toxic residue. If one component of the residue is significantly
more toxic than other components, two levels may be included in the tolerance.
Processing data:
While residue field trials provide data for estimating tolerance levels in raw
agricultural commodities, processing studies are required to determine whether
residues in raw commodities may be expected to degrade or concentrate during
food processing. If residues do concentrate during processing, one or more food
or feed additive tolerances must be established. However, if residues do not
concentrate in processed commodities, the tolerance for the parent raw commodity
applies to all processed food or feed derived from it.
Feeding Studies:
Whenever pesticide residues result in feed items, data on the transfer of
residues to meat, milk, poultry, and eggs are required. These studies are also
required if a pesticide is to be applied directly to animals. Data from these
studies tell EPA how much and what kind of secondary residues may result in
meat, milk, poultry, and eggs, in cases where this question arises.
Toxicology Studies
For tolerance purposes, the toxicology data obtained from studies in which test
animals are exposed to the pesticide chemical, mainly through oral exposure, are
used to determine a lifetime "no observed effect level" (NOEL) for non-oncogenic
effects, or a cancer risk estimate if the pesticide has oncogenic potential. The
studies begin with young (post-weanling) animals, and exposure continues through
their adulthood (thereby mimicking human exposure beginning in childhood and
continuing over a lifetime).
The toxicological effects of concern here are not the severe and immediate
poisoning symptoms that could result from accidental massive ingestion of a
pesticide, or skin and eye irritation characteristics. Rather, these long-term
feeding tests are designed to reveal potential adverse effects which may result
from continuous low-level ingestion of a pesticide -- such as bone marrow
damage, cancer, blood effects, and other chronic effects. Special test
procedures also determine the potential of the chemical to cause birth defects,
adverse developmental effects, reproductive effects, neurotoxicity, and gene
damage. In addition, the multi-generation reproduction study looks at the
reproductive effects on animals exposed to pesticides in the womb and during
nursing.
The Reference Dose or Acceptable Daily Intake (ADI) Level
Using the NOEL, an acceptable daily intake (ADI) level -- which scientists now
call a reference dose -- can be proposed for humans by applying a suitable
safety factor. The safety factor is intended to allow an extra margin of safety
to compensate for (l) the scientific uncertainty inherent in the process of
extrapolating human risk projections from animal data, and (2) the possibility
of differing sensitivities to the pesticide in individuals or subgroups (such as
children) among the general population. The magnitude of this factor may vary,
depending on the toxicological effects observed in laboratory animals, and the
amount of toxicity data available, but a 100-fold safety factor is used in most
instances.
In general, the ADI or reference dose can be defined as an estimate of a daily
exposure to the human population (including sensitive subgroups) that is likely
to be without an appreciable risk of adverse effects.
Tolerance Decisions
Traditionally, in making tolerance decisions, EPA has compared the reference
dose with the Theoretical Maximum Residue Contribution (TMRC) of the pesticide
to the daily diet. The TMRC is a mathematical construct obtained when the
proposed tolerance(s) is summed together with any other tolerances that have
already been established for the pesticide, and multiplied by average food
consumption estimates based on USDA data from its Nationwide Survey of food
consumption (Currently, EPA is using data from USDA's 1977-78 survey since these
are the most recent comprehensive survey results available; USDA is in the
process of analyzing results from its 1987-88 Nationwide Survey, and EPA will
use these more recent survey results as soon as they become available). As a
rule of thumb, where basic data requirements are satisfied, EPA has routinely
ruled in favor of a proposed tolerance if the TMRC is less than the ADI.
However, as a routine practice, before making tolerance decisions on a
pesticide, EPA uses the TAS to calculate TMRC and risk estimates for the two
most sensitive subgroups identified by the system as well as the population as
an average. In some cases, peak exposures to certain subgroups, such as infants
and children, may exceed the ADI by some percentage even though this is not the
case for the composite average of the population. Since the ADI is not a precise
indicator of risk, exceeding the ADI by a small factor is not necessarily a
cause for concern. However, if the TAS analysis indicates that exposure and thus
risk to such a subgroup is so high that adverse effects are likely to occur, the
Agency will not approve a tolerance even if the risks to the average population
are acceptable.
Examples of cases where dietary concerns for infants and children in particular
have been the overriding consideration include:
> A decision in 1985 not to approve tolerances for the pesticide pyrdrin for
proposed new uses on alfalfa and sorghum, due to risks to children from
secondary residues in milk.
> An action level for the occurrence of heptachlor epoxide in milk from cows fed
pineapple greens that had been treated with heptachlor. Based on short- term
liver toxicity in non-nursing infants and small children, EPA issued a lower
action level than would have been required to protect adults from adverse
effects.
> EPA's recent announcement to expedite a notice of cancellation for all uses of
Alar (daminozide), meanwhile extending a temporary tolerance for Alar residues
in apples and urging voluntary restraint on its use. In the case of Alar, EPA is
particularly concerned about the upper-limit cancer risk for Alar exposure to
infants and toddlers for the first year and one-half of childhood.
It is important to note that the TMRC is a very rough-hewn tool of the tolerance
process. Taken out of context, and applied literally, the TMRC can be used to
make calculations that misrepresent risks from pesticides as artificially high.
It is necessary to go beyond the TMRC approach, using real data on pesticide
usage rates and-anticipated residue levels, in order to evaluate actual dietary
exposure and risk to consumers from pesticide residues in the food supply.
Certain standard TMRC assumptions tend to greatly exaggerate the dietary risks
attributable to pesticides; for example: (l) that 100% of each crop that may
legally be treated with a pesticide is in fact treated with the pesticide, and
(2) that pesticide residues in each commodity that may be treated with the
pesticide are always present at tolerance levels at the time of consumption.
Thus, for example, since 92 pesticides are registered for use on tomatoes, the
TMRC calculations would lead one to assume that all tomatoes contain tolerance
level residues of all 92 pesticides. This is never the case in reality. To
illustrate, Chart 1[*] gives an overview of the pesticides registered for use on
tomatoes, by type: 54 insecticides, 20 fungicides, 7 herbicides, 4 soil
fumigants, and 3 growth regulators; plus 4 additional miscellaneous pesticides
(not shown). By comparison, Chart 2 shows recent pesticide usage data on
tomatoes grown for processing in the State of California, which produces 86% of
tomatoes grown for processing in the U.S. California processing tomatoes grown
in 1986 were actually treated with very few of these 92 pesticides. Only 3% of
California farms reported use of as many as 3 fungicides; 4% used as many as 4
herbicides; 4% used as many as 5 insecticides, 2% used 2 soil fumigants, and
none used more than 1 growth regulator.
[*] NOTE: Charts and figures are not included in this version. They are
available from the Chemicals-Pesticides Program upon request.
To illustrate the difference between a tolerance limit and actual residues
present in consumer products, Chart 3 shows recent data on actual residues of
the pesticide chlorothalonil on tomatoes. The established tolerance is 5 parts
per million (ppm), indicated by the leftmost vertical bar. All post-harvest
"farm gate" residues in tomatoes in these studies were well below 5 ppm.
Moreover, there was little or no detectable residue of chlorothalonil in
tomatoes by the time the commodity reached the consumer. In fact, 95% of
residues were removed during washing at the packing house.
There is an important exception to the reference dose approach described above.
EPA has not used the reference dose concept, which implies a threshold level of
risk, in considering tolerances for pesticides that induce cancer in test
animals.
Instead, the Agency takes a conservative approach, based on widely used
quantitative risk assessment models, which projects upper-bound (worst-case)
estimates of additional risk above the background cancer risk in the general
population of 1 in 4 or 0.25 (2.5 x 10 to the minus 1). Basically, the approach
involves determining a quantitative estimate of a pesticide's oncogenic potency,
called a "Q star," and comparing the Q star to dietary exposure estimates.
Dietary exposure estimates are based on the tolerance level of residue unless
verifiable data on actual residues of the pesticide in agricultural commodities
and consumer products are available.
In regulating pesticides that induce cancer in laboratory animals, EPA applies
the "negligible risk" concept recommended by the National Academy of Sciences
for making registration and tolerance decisions under FIFRA and FFDCA.
In sum, the tolerance process is highly protective in that it is based on the
most sensitive animal tests results available and a combination of highly
conservative assumptions and risk assessment practices. Tolerances are set at
the lowest level necessary to accommodate the maximum application rate and
frequency being proposed, even when higher levels would be safe for human
consumption.
Issues Related to Infants and Children
Several points should be highlighted regarding the tolerance process that relate
to the protection of children:
> As mentioned above, in animal studies used for human risk assessment purposes,
chemicals are administered to test animals beginning with young animals
(post-weanling) and continuing through adulthood (mimicking human exposure that
begins in childhood and continues over a lifetime). The body dose received by
the young animals may be double that of the adult animals, due to changes in
their consumption patterns; however, the lower (adult) body dose is typically
used in reference dose (ADI) estimates. This results in a lower (and more
protective) reference dose than would otherwise be the case.
> In setting ADIs or reference doses, EPA generally uses a lO-fold safety factor
to compensate for the uncertainty inherent in the process of extrapolating human
dietary risk projections from animal data and, in addition, another 10-fold
factor to compensate for the possibility of differing sensitivities in
individuals or subgroups -- such as children -- among the general population.
> In cases where carcinogenicity is the potential effect of concern, EPA
considers the upper limits of risk that may result from a lifetime exposure.
Where these upper limit risks are not unreasonable from lifetime exposure, EPA
concludes that the risk from a portion of lifetime exposure (e.g., between ages
1 and 12 or between 1 and 21) is likewise not unreasonable -- regardless of
changes in eating habits expected to occur between infancy and adulthood. This
practice is consistent with EPA's published cancer risk assessment guidelines
and with current, mainstream thinking in the scientific community. EPA does not
at this time have data to support specific modifications to this approach with
respect to infants and children. However, the NAS is currently examining issues
surrounding pesticides in the diets of infants and children. When the NAS
concludes its report, EPA will be considering whether its present approach
should be adjusted in light of the Academy's findings.
> Data on the comparative toxicity of pesticides in adult versus young weanling
mammals are very limited and pertain to acute toxic effects only. These studies
show mixed results. In studies of 37 chemicals administered to adult and
weanling rats, weanlings demonstrated greater sensitivity in 8 cases. Adults
showed greater sensitivity than weanlings in 23 cases. In 6 cases the
sensitivity of adult and weanling rats was roughly the same.
> When estimating dietary exposure to infants and children, EPA's Tolerance
Assessment System uses milligram-per-kilogram body weight (rather than body
surface) comparisons with dietary intake levels for risk assessment purposes. As
the FIFRA Scientific Advisory Panel has pointed out, this practice is likely to
overstate dietary risk to infants.
Other Areas of Concern
Neurotoxicity testing
EPA is in the process of upgrading its guidelines for neurotoxicity testing
requirements. On this initiative, EPA's Office of Pesticide Programs (OPP) is
working in concert with the Office of Toxic Substances (OTS), which has
published and received public comment on its neurotoxicity testing guidelines.
Once these guidelines are revised in light of public comments, they will be
submitted to the FIFRA Scientific Advisory Panel for review in summer 1989. EPA
will then incorporate the resulting guidelines into its data requirement
regulations under 40 CFR Part 158 and will require such data on a case-by-case
basis until the revisions become final and effective.
Inert Inqredients
EPA published a Federal Register notice on April 22, 1987, describing a policy
of increased scrutiny of inert ingredients. New inerts are subject to data
requirements and registrants are being urged to replace existing "inerts of
toxicological concern" with less toxic chemicals. Registrants who fail to
substitute must label their products and will be required to submit the
extensive data necessary for a product-by-product risk/benefit evaluation.
Data Gaps
Although a large number of active ingredients are required for pesticide use on
food, fewer than half of the active ingredients account for approximately 90
percent of the poundage of agricultural pesticides used. Many of these products
were first registered after 1984 and EPA has very extensive information on these
pesticides. Other pesticides have been registered for some years and are
undergoing reregistration. Since EPA has concentrated its reregistration efforts
on reviewing these large agricultural use pesticides, EPA either has now, or
will have in a very few years, a very complete data base on the pesticides
contributing most heavily to dietary exposure.
2/23/89