Introduction
to: Pesticides in Children's Foods Methods:
Updating the CU Scoring SchemeHighlights
of Results of the 1998 PDP AnalysisRisk
Drivers
Pesticides in Children's Foods1
Risk Drivers
Risk Drivers ||
Odds of Exceeding a Safe Dose
The Role of Chlorpyrifos in Dietary Risk || Organochlorine
Insecticides: A Persistent Problem
You can't control pesticide use by the agricultural industry...
...but you can remove the pesticides!
Food and Water Ozonator
|
1. Risk Drivers in the 1998 PDP Survey
We
apply the term “risk drivers” to residues of
pesticides that account for more than 10 percent of the total
Toxicity Index for a given food tested by the PDP. To be considered
a risk driver, a residue
must also account for at
least 10 TI points; i.e., we’re not concerned with cases in which a residue
accounts for a large share of a very low score. For this analysis, we have
focused primarily on residues that are risk drivers for more than one food
tested by the PDP.
Our 1999 report contains a detailed analysis
of risk drivers in the foods tested by the PDP in 1994-1997.
That analysis
is not repeated here; this section
focuses on foods tested in 1998. Table 5 shows the details of the contributions
of various pesticides to the total TI scores for each food.
The residues found in different foods differ, but most pesticides that
were risk drivers in 1998 were also risk drivers in previous
years. Tables 2
and 3 summarize
the acute and chronic toxicity data on the pesticides detected by the PDP,
and display the basis for our toxicity indices for each chemical. See those
tables for details on the nature of the risk concerns with
respect to each individual pesticide classed as a risk driver.
Twelve
chemicals stand out as risk drivers in this year’s
analysis; they are either major factors in the total TIs for
multiple foods, or the predominant
factor for a single food. A handful of other residues that each account for
more than 10 percent of the TI for a single food are not discussed, because
their contributions to total dietary risk are less prominent than those of
the chemicals that are discussed. See Table 5 for details on residues that
are factors in the TI for some foods but did not qualify as risk drivers this
year.
This year’s risk drivers include:
Dieldrin.
This long-banned chlorinated hydrocarbon insecticide persists
in soils and is absorbed through the roots by members
of the squash and melon
family. It is the top risk driver for fresh and frozen U.S. winter squash,
fresh Mexican winter squash, Mexican and U.S. cantaloupe, and U.S.
soybeans and sweet potatoes. Its TI contribution ranges from
46 percent for Mexican cantaloupe to 94 percent for U.S. frozen
winter squash.
Methamidophos, an organophosphate
insecticide, was the subject of a new EPA risk assessment in
1999. EPA decreased the Reference Dose,
which has increased methamidophos’s contribution to our
TI scores. It is the first- or second-ranked component of the
high TI scores received by Mexican and U.S. tomatoes, U.S.
canned green beans, and Mexican cantaloupe, and of the lower
score for Chinese apple juice. Its share of the total TI score ranges from
23 percent for Mexican cantaloupe to 75 percent for U.S. tomatoes.
Chlorpyrifos.
This organophosphate insecticide was also the subject of
a new risk assessment by the EPA in the past year,
which decreased its RfD and
increased its relative importance in our TI scores. In the 1998 PDP tests,
chlorpyrifos plays a role in the high TI scores for Chilean
pears and Mexican and U.S. tomatoes, accounting for from 11
to 19 percent of the TIs
for those foods. (See page 19 for additional details.)
Azinphos-methyl. Another
organophosphate insecticide, azinphos–methyl
was found in pears from Argentina, Chile and the U.S., where it contributed
42, 52 and 40 percent of the score, respectively, for composite
samples, and slightly less for pears analyzed as single servings.
Methyl
Parathion. This organophosphate insecticide, which was the
top risk driver in our 1999 report because it was found in
several foods tested in 1994-1996, is a top factor in the TI
for U.S. grown pears in the 1998 PDP survey, accounting for
61 percent of the score for single-serving samples and 39 percent
of the score for composite samples. Methyl parathion accounts
for 29 percent of the TI for canned green beans.
Dicofol is a chlorinated
hydrocarbon insecticide, one of the few members of this family
still registered for food uses.
In the 1998 PDP data, it contributes 20 percent of the TI
scores for Chilean pears and 16 and 21 percent for fresh and
frozen strawberries, respectively,
from the U.S.
Methomyl, a carbamate insecticide,
is the top risk driver on strawberries from the U.S., accounting
for
47 percent of
the score for fresh samples and 33 percent of the score for
frozen samples.
Anilazine, a fungicide, is the
top risk driver for Mexican strawberries, and accounts for
65 percent of
the score for
that food.
Iprodione, a fungicide, was
found on fresh and frozen strawberries. It accounted for 10
percent of the score
for fresh U.S. samples,
and 13 percent of the score for both fresh Mexican and frozen
U.S. strawberries.
Permethrin, a synthetic pyrethroid
insecticide, is the top contributor to the score for canned
spinach from
the U.S., accounting for 83 percent of the
TI.
Demeton-S sulfone, a metabolite
of an organophosphate insecticide, accounts for 41 percent
of the comparatively high
TI score
for Mexican tomatoes.
Heptachlor epoxide, a breakdown product
of heptachlor, a long-banned chlorinated hydrocarbon insecticide,
contributes to the very
high TI score for frozen
winter squash from the U.S. Since the dieldrin component of this TI
is huge, heptachlor accounts for just 5 percent of the total. But the
absolute score for heptachlor epoxide is 181 points, greater than the
total TI for
all residues combined in most other foods, qualifying it as a risk driver.
2.
Odds of Exceeding a Safe Dose
A report we published last year explained how
legal limits on exposure to pesticide residues, defined by
EPA tolerances, and safe exposures, defined by Reference Doses,
are not the same. Our analysis showed many cases in which the
legal residue level (tolerance) exceeded the “safe” level
(chronic RfD) by from 10- to 200-fold. (See our report “Legal
Does Not Equal Safe” at http://www.ecologic-ipm.com/legal_safe.pdf.)
As the EPA has reviewed and reassessed the risks of exposure to pesticides
under the FQPA, the agency has lowered its definition of safe exposure to
many active ingredients, widening the gap between its definitions of “legal” and “safe” exposure
in many cases.
“Safe” exposure, expressed as the EPA Reference
Dose (RfD), is defined in the Food Quality Protection Act as
a level of exposure that has a “reasonable
certainty of no harm” to public health. The RfD is based on tests that
found no adverse effects in animals, and includes “uncertainty” or “safety” factors
intended to ensure a margin of safety against adverse effects in humans. The
chronic
RfD describes a dose that should be
safe if ingested day after day
for a lifetime. The acute RfD describes a dose that should be safe for any
single exposure event.
EPA RfDs can be compared with residues in foods, to
see whether the actual amounts of pesticides detected in foods approach
or exceed levels the EPA has defined as safe. In last year’s
report, we compared residues from PDP Tests with EPA chronic
RfDs, and found
many instances in which a large
fraction of the samples tested had residues that would give a child a dose
of a pesticide that exceeded the RfD. (Table 6 of our 1999 report.) Since
the RfDs incorporate a safety margin, a dose above the RfD is not necessarily
harmful, per se. But it means that exposure is higher than the level that
EPA deems “reasonably
certain” to be free of harm, and that the safety margin between actual
exposure and that known to be harmful is narrower than the agency, acting
for society, has judged acceptable. Exposure above the RfD does not indicate
an immediate health hazard, but it does indicate a need to take steps to
reduce exposure and restore an acceptable safety margin.
We received some
critical comments on our report last year, particularly from Carl Winter
of U.C, Davis, to the effect
that we erred in comparing residues
in single food servings with the EPA’s chronic RfDs. Since the chronic
RfD defines safe lifetime average exposure, Winter argued, it is not hazardous
to health for an occasional residue in a food to exceed that
RfD,
even by a wide margin, as long as the long-term average remains below the
chronic RfD. According to Winter, we should have compared single-serving
intakes of residues with EPA’s acute RfDs for the pesticides present.
We
believe that both chronic and acute exposure to pesticide residues are
public health concerns. Children who eat a variety
of fresh fruits and vegetables
will be exposed to several pesticide residues a day, and the odds are relatively
high that, on any given day, they may exceed an RfD for one or more of
those residues. Even if no residue exceeds the RfD for a single
chemical on a given day, multiple residues with the same mechanism
of toxic action may add up to an unsafe exposure. We therefore
disagree with Winter’s
dismissal of chronic exposure concerns. We believe this is a valid public
health issue, and have compared residues to chronic RfDs again
this
year as a useful
way to identify individual pesticide uses on specific foods that contribute
significantly to chronic overexposure.
We do agree, however,
with Winter’s suggestion that
we should compare residues in single food servings with acute
RfDs, where EPA has established
them. In this year’s analysis, therefore, we have done the comparisons
both ways, using chronic and acute RfDs. As we did last year, we have assumed
a “standard” scenario
in which a 20-kg child consumes a 100-gram serving of a PDP-tested food. Using
the EPA RfDs, we have calculated reference concentrations (RfCs, acute and
chronic) for each pesticide. The RfC is the residue level that, if present
in a 100-g serving of food, would give a 20-kg child the RfD of that pesticide.
Using PDP data on residues in individual samples,
we counted the number of samples with residues above the chronic or acute RfC,
as applicable. That number, expressed as a percent of the total number of samples,
yields the “odds” that a child who eats that
food would exceed that RfD for that pesticide.
Table 7 presents our calculations
of the odds of exceeding a safe dose, for selected food/pesticide combinations
in the 1998 PDP tests. Here are the
highlights:
As we reported last year, this year’s PDP data again
show that the risk of exceeding a chronic RfD is greatest for
dieldrin residues in frozen winter
squash. Our standard 20-kg child eating 100 grams of this food would exceed
the EPA chronic RfD for dieldrin 66 percent of the time. (EPA
has not set an acute RfD for dieldrin.)
Several other food/pesticide
combinations had significant probabilities of exceeding a
safe chronic dose in our standard scenario. Methamidophos residues
on U.S. canned and frozen green beans and on Mexican and U.S. tomatoes
would exceed the RfD about 20, 19 and 13 percent of the time,
respectively. Mexican
tomatoes also exceed the chlorpyrifos chronic RfD 11 percent of the time.
Methomyl residues on U.S. strawberries exceeded the safe chronic
dose 2 percent of the time.
In its 1998 tests, the PDP did a
dual analysis of fresh pears. Their normal sampling looked
at residues in composite (5 pound)
samples, and they also
tested for residues in single servings (individual pears), at the request
of the EPA, to get better data on possible acute exposures. Composite samples
may average out the residues on single pieces of fruit, and
underestimate maximum levels in single servings, an acute exposure
concern. Additional
tests were
done to address this analytical issue.
The PDP analysis of single-serving
pears found potential acute and chronic exposure problems
with residues of two pesticides.
Methyl parathion was present
at greater than the chronic RfC on 14 percent of single pear servings and
exceeded the acute RfC on nearly 3 percent of the samples.
Azinphos
methyl exceeded the chronic RfC in 2.5 percent of the samples,
and was over the acute RfC in 0.31 percent of the servings.
The findings for single servings and composite samples of pears
were similar for azinphos methyl with respect to the chronic
RfC, but the composite samples showed a higher risk of exceeding
the acute RfC (0.82 percent). The composite samples of
pears also did not have as frequent or as high methyl parathion residues
as the single-serving sampling showed. The implications of the single-serving
data in terms of the accuracy of composite samples for identifying potential
acute exposure problems remain to be fully explored.
The PDP data show concerns
about potential acute exposure to residues of a few other
especially toxic insecticides. Residues of methamidophos in
U.S.
canned/frozen green beans and Mexican tomatoes, and methomyl in U.S. fresh
strawberries, exceeded the acute RfCs in 0.30, 1.25 and 0.36
percent of the samples, respectively.
While it may be tempting
to conclude that odds of less than 1 percent of exceeding most
RfCs indicate “no problem,” this analysis in fact
shows a significant acute exposure risk. About 20 million U.S.
children are six
years old or younger. Of a million of those children who eat a serving of
canned green beans, 3,000 of them (0.3% times 1,000,000) would get more than
the acute RfD for methamidophos. If a million children ate a fresh pear,
27,800 would get more than the acute RfD of methyl parathion, and 3,100 would
get too much azinphos methyl. (The odds say that 83 children
would get too much of both pesticides in their pear.) These
estimates reflect a simplifying assumption that the RfC for
a given residue is uniform for the entire child
population; in fact, an individual child’s RfC depends on both body weight
and food serving size, so these figures are approximate, and may understate
actual
risks. (Our 20-kg child represents a five-year-old; smaller children would
have greater odds of getting an excessive dose.) While the odds that any individual
child will be overexposed are small, the odds that significant
numbers of children who eat certain foods with high residues will exceed an
acute safe intake of pesticides are substantial. This analysis also examined
only single food/pesticide combinations. Likely additive effects of multiple
residues with the same mechanism of toxicity in children’s overall daily
diets would increase the magnitude of the problem.
The Environmental Working
Group has done an analysis, using Monte Carlo simulation techniques, of the
odds that the combination
of all residues in all foods
typically consumed by children would add up to exposure above their estimated
safe acute intake for the organophosphate insecticides as a class. The
EWG found that on a typical day, about 600,000 U.S. children
age
5 and younger get a dose of OPs that exceeds the acute “safe” dose.”
Chronic
overexposure to pesticide residues is also a clear public health
risk. As Table 7 shows, children who consume tomatoes, strawberries,
pears, green beans or winter squash have odds ranging from
2 to 66 percent that they will get more than a safe chronic
dose of at least one pesticide. These foods and residues by
no means exhaust the
list of potentially problematic
exposures; they are simply selected examples from foods tested by the PDP in
1998. When the presence of residues on multiple foods that
children eat and the additive nature of exposures to pesticides
with the same mechanism of toxicity are considered, it is clear
that many children are being repeatedly exposed to doses of
pesticides that exceed “safe” chronic intake.
When the Environmental Working Group modeled multi-residue dietary exposure
in a 1998 report (http://www.ewg.org/pub/home/reports/ops) they estimated that
one million U.S. children—five percent of the population
less than six years old—exceed safe chronic cumulative
intake of organophosphate insecticides on any given day. Since
EWG made that estimate, the EPA has substantially lowered chronic
RfDs for several of the OPs, which would tend to increase the
magnitude of the overexposure problem.
3. The Role of Chlorpyrifos
in Dietary Risk
The EPA has recently published a revised risk assessment for
chlorpyrifos, and is expected to announce soon (in June 2000)
its regulatory strategy for this widely used insecticide. Chlorpyrifos
is an ingredient in a vast array
of pesticide products designed for professional and consumer use around the
home, lawn and garden, and EPA’s risk assessment shows
that most if not all of those uses pose unacceptable exposure
risks. Chlorpyrifos is also an agricultural insecticide, used
against many different pests on a wide variety of crops, in
the U.S. and around the world. Thus, chlorpyrifos residues
are commonly detected on many foods tested by the PDP.
Table
8 presents data on chlorpyrifos residues detected by the
PDP in all foods tested from 1994 through 1998. We have extracted
the chlorpyrifos data from Table 5, and arranged it to show
the foods in which chlorpyrifos is found, its frequency of
detection and mean residue levels, and the share of the toxicity
index
for each food/country combination that this insecticide
accounts for.
As the table shows, chlorpyrifos was detected in 22 different
foods the PDP tested from 1994 through 1998, and is commonly
detected in both imported
and domestic samples. Chlorpyrifos is a major risk driver (a TI of over
100 points from chlorpyrifos residues alone) in seven cases,
with imported foods, apples from New Zealand and Mexican tomatoes,
showing the highest TIs and highest or most prevalent residues.
There are eleven additional cases in which the TI score for
chlorpyrifos contributes from 25 to 100 points to the
overall score for a particular food from a particular country. Most of these
cases involve U.S. apples and Chilean grapes, pears and peaches. A third
tier of 12 additional cases had lower or less frequent chlorpyrifos
residues, which contributed from 10 to 21 points to the overall
TI score. In 33 more
cases, chlorpyrifos residues were detected, but contributed less than 10
TI points to the overall score for that food/country of origin.
Chlorpyrifos
residues on single foods in some cases exceed safe doses
for children (see previous section), and chlorpyrifos makes
major contributions to the cumulative risk of exposure to the
organophosphate insecticides as a class, in the diet as a whole.
To reduce
these exposures and risks, tighter
restrictions on chlorpyrifos applications on crops that are major sources
of dietary intake are needed, both here and abroad.
One of
the most striking findings evident from Table 8 is that 15
of the top 20 TI scores for chlorpyrifos residues (and 7
of the top 8) are for imported foods. As this report is being
written,
the EPA is poised to announce its regulatory
strategy for reducing the risks of chlorpyrifos exposure. The U.S. apple
industry has said it expects tighter restrictions on chlorpyrifos
use on apples, probably in the form of an extended pre-harvest
interval (PHI), i.e., limiting chlorpyrifos use to early-season
applications that should leave no detectable residues on the
fruit.
Extending the PHI could achieve lower residues on U.S.-grown
fruit, while allowing EPA to avoid the political costs of
banning use of this insecticide on a major crop like apples.
But EPA also must also lower its tolerances for chlorpyrifos,
to a level consistent with the “reasonable
certainty of no harm” standard of the FQPA. Extending
the PHIs applicable to U.S. growers while leaving current (very
high) tolerances on the books will reduce
residues on domestic produce, while allowing imported produce to contain much
higher residues. U.S.-grown apples dominate the market here, and residues in
U.S. fruit account for most population exposure to chlorpyrifos from apples,
even though U.S. apples contain residues less often than New Zealand apples
do. Unless the EPA substantially lowers the chlorpyrifos tolerance for apples,
however, imported apples could become the primary source of chlorpyrifos exposure
from this key children’s food. Mexican tomatoes, Chilean pears, peaches
and grapes are already the largest sources of chlorpyrifos exposure from those
foods, and EPA must take steps to reduce these imported foods’ contributions
to overall dietary risk by lowering the tolerances, not merely adjusting application
limits for U.S. growers.
4. Organochlorine Insecticides: A Persistent Problem
The prominence of dieldrin residues as risk drivers for winter
squash and cantaloupe in the 1998 PDP data (and in winter squash
in 1997 PDP tests, (see Table 5) suggests a larger general
problem. Dieldrin uses on foods were banned in the early 1970s,
as were
uses of several other chlorinated hydrocarbon
insecticides, including DDT, aldrin, endrin, heptachlor and chlordane.
These
chemicals were banned both because of suspected of risks to
public health (primarily, cancer risk), and because of ecological
hazards; they are toxic
to a wide range of organisms and very persistent in the environment. Organochlorine
pesticide residues last for decades in soils, accumulate in
food chains,
adversely affect the reproduction of raptorial birds and other wildlife,
and accumulate in human body fat. Bans on the organochlorine insecticides
were among the EPA’s most prominent early decisions on pesticide risks.
A
few members of this chemical family are still registered today for use
on U.S. crops. Residues of dicofol, endosulfan and methoxychlor
may reflect current applications, instead of or in addition
to persistent residues
from past uses.
But residues of dieldrin, endrin, heptachlor, chlordane, and DDT and its
breakdown products DDE and DDD, are detected in foods tested
by the PDP because of persistent environmental contamination,
not current use.
The problem of persistent residues in soils
is aggravated by the tendency of certain food crops, members
of the family
that includes squash and melons, to extract the organochlorine
residues from soil, absorbing them through their roots, translocating
them within the plant,
and accumulating the residues in the edible portions. The
most feasible solution for this problem is to avoid growing
crops like squash on soils contaminated with these pesticides.
As the PDP data show clearly, U.S. squash growers, in particular, have yet
to adopt that strategy.
Table 9 summarizes the data on organochlorine
insecticide residues in PDP- tested foods.
Dieldrin residues
in squash stand out as the largest single factor in any food’s
overall toxicity index, and dieldrin is also a risk driver for U.S. and Mexican
cantaloupe and a significant factor in scores for soybeans, sweet potatoes
and U.S. spinach. In 1994, but not in subsequent test years,
dieldrin
residues substantially increased scores for U.S. potatoes and U.S. carrots.
Such sporadic impacts on scores reflect geographic variation
in
dieldrin levels in soils.
Dicofol, used for mite control on
fruit crops, shows up as a problem residue on apples, grapes,
peaches, pears, strawberries, and tomatoes. It is either the
top or second-ranked risk driver on U.S.
grapes and Chilean pears.
Heptachlor, in the form of its
epoxide breakdown product, was detected in both fresh and frozen
U.S. winter squash
in both 1997 and 1998 tests. The TI’s
for these foods are overwhelmed by dieldrin, but the heptachlor score would
warrant concern even if no other problem residues were present.
DDT, and its
breakdown products, DDE and DDD, are ubiquitous in the
global environment, and are detected in 21 of the 28
foods listed in Table 9. In most cases, these residues contribute
only a small amount to overall TI scores, but carrots and spinach
are exceptions. DDT and its byproducts are primary risk drivers
for carrots, and account
for 10 to 17 percent of the total
score for spinach.
Endosulfan and its breakdown products are also found in
many foods, and generally contribute in minor ways to overall
TI scores. Mexican spinach and green beans are exceptions,
with endosulfan accounting for from 17 to 43 percent of the
TIs for those foods.
Other chlorinated organics, including aldrin,
methoxychlor, chlordane, lindane and benzene hexachloride
all show up occasionally
in PDP-tested foods, but
generally not at frequencies or levels comparable to those for the residues
discussed above.
The high TI scores for dieldrin and heptachlor epoxide,
in particular, are consequences of the comparatively great
toxicity
of these residues. Each
has a very low chronic Reference Dose, and they are also the two most potent
carcinogens among the carcinogenic pesticides detected by the PDP. These
attributes of
toxicity account for the high TI scores of winter squash, even though the
mean residues are quite low (0.03 ppm for dieldrin, 0.004 ppm for heptachlor).
The
EPA can manage the risks of currently-used pesticides by setting
strict tolerance limits, designed to keep residues within the “reasonable
certainty of no harm” range required by the FQPA. The dietary risk
contributions of dicofol and\ endosulfan, for example, could be managed this
way. But for
a banned chemical, such as dieldrin or heptachlor, EPA tolerances are already
set at zero. Unavoidable residues caused by environmental contamination are
legal, and are governed by “action levels,” set by the Food
and Drug Administration. An action level defines a level of contamination that
may render a food “injurious” and warrants keeping it off the
market.
Current action levels for the banned organochlorine insecticides
are relatively high; the action level for dieldrin is 0.1
ppm. High action levels sanction serious residue problems,
such as those observed in winter squash. As long as it remains
legal, squash growers
will continue to sell product
that contains significant dieldrin and heptachlor residues. If these action
levels were lowered, say to 0.01 ppm, growers would have an incentive to
seek out uncontaminated cropland, for food crops that take
up organochlorines as effectively as squash does. FDA depends
on EPA for risk assessments on pesticides. To provide a basis
for setting more health- protective action levels for the banned
organochlorine insecticides, the
two agencies need to work together. Under the FQPA, ensuring a wider safety
margin
for these residues should be a high priority.1
Goto the Next Section: Recommendations |
Search Our Site
