October 15, 2007

Debra Edwards, PhD

Director, Office of Pesticide Programs

US Environmental Protection Agency

1200 Pennsylvania Ave NW

Ariel Rios Building

Washington, DC 20460

Dear Dr.  Edwards,

We appreciate your July 2, 2007, response to our letter, however, we
continue to have serious concerns about EPA’s occupational risk
assessment methods.

The Agricultural Re-entry Task Force (ARTF) data are flawed and should
not be used.

First, methods for assessing hand or neck exposure using wipes or rinses
tend to systematically underestimate the extent of exposure because they
fail to capture pesticide residues that are absorbed through the skin
during the time between exposure and data collection.  In a
methodological study of techniques for assessing exposure through the
hands, researchers found that nearly twice as much of the residue was
removed from the hands if wiped immediately after exposure as opposed to
waiting an hour (based on a standard residue loading).  The difference
represents the amount absorbed through the hands during the delay before
washing.  The study concluded that estimates of exposure based on hand
wipes were likely to underestimate actual exposure by two to five times.
 In a workplace study conducted with apple workers, hand wipes produced
a 10-fold underestimate of exposure.  Without a systematic correction
for this underestimation, these data cannot be deemed valid.  Such a
correction is particularly critical here since hand exposure accounts
for a significant proportion of a post-application worker’s exposure.

Second, as the Agency itself acknowledges, the crop groupings used by
ARTF have never been evaluated by farmworker representatives,
independent experts or the FIFRA Scientific Advisory Panel (SAP). 
Rather, they have been cobbled together by the growers alone and are
therefore suspect and tainted by obvious bias.

Third, the ARTF data should have been subjected to notice and public
comment as well as SAP review, and should not be used until it has
passed muster under such scrutiny.

The EPA’s existing algorithm is not sufficiently conservative to
protect agricultural applicators and post-application workers.

Despite its claims, EPA’s occupational risk assessment algorithm is
not composed of unrealistic high-end exposures, and thus does not
produce a protective estimate of worker exposure:

The 8-hour workday assumption is insufficient:  Basing risk assessments
on an 8-hour day does not accurately reflect the length of exposure
frequently experienced by a significant proportion of the agricultural
workforce.  According to the National Agricultural Worker Survey (NAWS),
conducted by the US Department of Labor, the majority of farmworkers
(56%) worked on average between 30 and 50 hours per week (in 1997-98);
15% worked an average of more than 50 hours per week.

Limited access to showers and laundry facilities leads to prolonged
exposure:  Delayed access to bathing and laundry facilities is a
widespread problem.  One study of settled farmworkers in the Yakima
Valley found that only 50% of agricultural workers bathed and changed
clothes as soon as they returned home from work.  This suggests that
among migrant workers who may live in shacks, their cars or even the
fields, and therefore lack ready access to shower facilities, the number
who wear their contaminated work clothes all evening before changing
would be far higher.  In addition, some farmworkers sleep in their work
clothes and/or wear the same pesticide-contaminated clothing all week
long.  The unfortunate reality is that farmworkers are often too poor to
have multiple work outfits, and their living conditions do not readily
allow them to wash their clothes during the workweek.

EPA assessments are not always based on maximum application rates: The
Agency claims that it relies on DFR data from studies involving repeated
applications of a pesticide at the maximum labeled application rate.  We
have reviewed multiple EPA risk assessment using studies involving
exposure at less than maximum legal rate or assumptions of typical use
patterns below the maximum label rates.

For example, in the 1999 Chlorpyrifos: HED Preliminary Risk Assessment
for Reregistration Eligibility Decision, Docket No.  OPP-34203, the EPA
termed its agricultural handler analysis a “reasonable high end”
estimate, even while the Agency acknowledged that the studies submitted
by Dow AgroScience had too few replicates, inadequate quality control
and included “typical” application rates (chosen by Dow) rather than
maximum label rates.

Use of a maximum label rate would not overestimate exposure: Because the
label rate is the legal limit, it can and will be used on many
occasions, and workers will be exposed at that rate.  Therefore, the
maximum label rate should be viewed as an exposure norm rather than an
“upper bound assumption.”

Strict compliance with a restricted entry interval (REI) is not a
conservative or protective assumption:  Putting workers in the field
immediately upon expiration of the REI is legal and does occur.  As
such, it does not represent an unduly “high end” exposure.

Use of transfer coefficients that are arithmetic means based on the
highest exposure day that occurred in the study does not capture
high-end exposures.  Using the arithmetic mean is not protective of the
most productive workers, who often have greater foliage contact per time
exposed.  The 95th or 99th percentile should be used instead, as we have
recommended previously.

Basing an algorithm on the assumption that workers work between 3 and 13
weeks doing the same tasks on the same crop is not a conservative
assumption.  For many crops, such as apples, strawberries, lettuce, and
grapes, workers do consistently perform the same task on the same crop
over an extended period of time.

Finally, EPA’s risk assessment method leaves out much of the real
world experience of farmworkers.  Many of the required protections are
not provided, i.e., workers are sent back into fields too soon, early
entry workers do not receive the protective equipment to which they are
entitled, and workers are often exposed to multiple chemicals at a time,
which may have additive or synergistic effects.  As such, the EPA
algorithm produces exposure estimates that are closer to the best-case
rather than worst-case scenario.

The use of rat dermal absorption studies does not necessarily provide a
conservative absorption rate estimate

EPA states that it is widely recognized that rat skin is more permeable
than human skin and thus is considered a conservative variable.  The
comparison assumes intact skin, however, farmworkers are likely to have
increased skin permeability due to the fact that their skin is often
hot, moist, irritated or abraded, yet EPA uses no adjustment factor to
account for these circumstances.  Between 1992 and 2001, higher rates of
dermatitis were consistently reported in agriculture when compared to
all other industries (NIOSH Worker Health Chartbook 2004). 
Consequently, many workers have abraded skin.  Vigorous work in hot and
humid conditions also promotes dermal absorption of chemicals.

Furthermore, dermal absorption in humans is increased by a number of
variables that are not applicable to rats, including pre-existing
medical conditions, poor nutrition, potentially conflicting medications,
and the generally poorer health of farmworkers as compared with
laboratory animals.

It should be noted that the Agency is increasingly using data from human
dermal absorption studies.  In addition to raising ethical concerns,
these studies similarly underestimate absorption because subjects do not
have abraded skin and are not exposed while conducting vigorous work
under hot and humid conditions.

Use of a 10-fold intraspecies uncertainty factor is inadequate.

EPA’s risk assessment requires inclusion of an uncertainty factor to
account for the variation between individual responses or susceptibility
to pesticide exposure.  This is due to many factors, including
differences in nutritional status, health status, pre-existing medical
conditions, activity level, lifestyle, and exposure to other chemicals
or agents.  In addition, inherent genetic differences exist between
individuals in the activity of the enzymes that break down toxic
chemicals in the body.  Generally, the Agency uses a standard
intra-species factor of 10X, presuming no more than a 10-fold difference
in susceptibility across a diverse human population.

The enzyme paraoxonase (coded by the gene PON1) is involved in the
detoxification of many organophosphates (OPs), particularly
chlorpyrifos.  Paraoxonase works by catalyzing the hydrolysis of its
toxic oxon metabolite, breaking down the by-products of chlorpyrifos
metabolism rather than allowing them to build up in the body.  A
slow-acting genotype of PON1 is less efficient at detoxifying the oxon
and is therefore associated with increased pesticide toxicity.

Furlong et al. (2006) reported a 14-fold difference between adults in
PON1 activity.  The Agency explicitly acknowledged this study, then
subsequently disregarded it, instead choosing to rely on a 2002 study
that used a physiologically-based pharmacokinetic (PBPK) model for
chlorpyrifos that showed that the “response was relatively insensitive
to changes in oxonase activity at low doses.”  Consequently, the
Agency applied only a 10X inter-species uncertainty factor to all the
organophosphates.  Despite the availability of significant informative
data derived from unintentionally exposed people, the Agency instead
relied on a model to support its assessment.  PBPK models are only as
reliable as the data used to design them; they are therefore meant to
help bridge data gaps, rather than override evidence-based data.

EPA’s treatment of the PON1 studies with respect to the calculation of
the intra-species uncertainty factor provides a stunning example of the
Agency turning a blind eye to existing relevant, robust data.  Not only
did the EPA fail to use the 14-fold uncertainty factor, as would be
indicated by the Furlong study, it failed to take in to account factors
beyond genetics, such as poor nutrition and complicating medical
conditions that introduce additional variability in exposure response
among humans.

EPA’s risk assessments fail to adequately account for post-application
workers’ inhalation exposure.

EPA risk assessments showed a much higher acute toxicity for pesticides
such as chlorothalonil when exposed via inhalation.  Nevertheless, only
one acute experiment was cited on the effects of inhaled chlorothalonil,
in which no NOAEL was detected.  In the risk assessment, a NOEL value
was used from an oral study on rats and converted to what was determined
to be an equivalent inhalation margin of exposure (MOEI).  This was
considered sufficient, although no further chronic or subchronic
inhalation tests were cited. The California Department of Pesticide
Regulation conducted acute inhalation studies in which adverse health
effects were found at every dose level. Draper et al. (2003) reported
cases of increased sensitivity and occupational asthma due to repeated
exposure to powdered pesticides such as chlorothalonil, indicating that
it “can induce specific immunological reactions in the airways” of
workers in close proximity to the chemical.

Hazardous pesticides contaminate the air.

When conducting occupational risk assessments, EPA states that its
“current approach to estimating workers’ post-application
exposures” does not “[underestimate] high end exposure levels.”
(DE Letter, 7/7/2007)  However, post-application worker risk is
calculated using measurements and thresholds for oral and dermal routes
of exposure only.  The EPA fails to adequately consider inhalation
exposure.

California, with arguably the best pesticide illness reporting system in
the US, has shown that over half of all reported acute poisoning cases
among farmworkers result from exposure to pesticide drift (and in about
half these cases no regulatory violation is found). While some of that
exposure may be dermal, a substantial portion is likely to be via
inhalation.  As part of the implementation of the California Toxic Air
Contaminant Act, a series of application site monitoring studies were
conducted by the California Air Resources Board (ARB) to monitor
chemicals of concern in the air, both application-specific and ambient
air concentrations.  Recent air monitoring studies, conducted by both
state agencies and community groups, report the presence of substantial
levels of pesticides in the air of agricultural communities, especially
following spray applications.

When compared with published EPA data, these results show that pesticide
concentrations in air near application sites commonly exceed reference
concentrations for volatile pesticides such as chlorpyrifos and
diazinon.  For example, the peak concentration of diazinon in air near a
legal application exceeded EPA’s inhalation reference concentration
(determined without using a 10-fold FQPA factor) by 17 times for
short-term, intermediate and long-term exposure for an adult 72 feet
from the field boundary. Over a three-day monitoring period,
concentrations exceeded the acute reference concentration in 82% of all
samples taken and 100% of samples collected downwind from the field.

A study by the Pesticide Action Network (PAN) linking community air
monitoring and biomonitoring was conducted in Lindsay, CA, in June and
July of 2004, 2005, and 2006.  Lindsay is a community dominated by
orange groves and with high levels of chlorpyrifos use.  The study
results demonstrated that daily exposure to pesticides in agricultural
communities can lead to substantial absorption into the body.  Even
though measured levels in air were generally well below the adult
reference exposure levels (RELs), corresponding levels found in the
urine of individuals living near the locations of air samplers suggested
that acceptable air concentration levels (derived from the RELs) do not
prevent the occurrence of unacceptable levels of a chlorpyrifos
breakdown product (TCPy) in the body.  Of the 12 participants of this
small pilot project, 11 had TCPy in their urine above the CDC average
for US adults (1.5 (g/L) and seven out of eight women had levels above
the “acceptable” level for pregnant and nursing women (1.5 (g/L).

In another study conducted by California ARB measuring chlorpyrifos
concentrations near orange groves in Tulare County, concentrations
exceeded RELs in 95% of samples, with three-day, time-weighted averages
ranging from 5,312 to 8,112 ng/m3 (1.4 to 2.1 times the 24-hour adult
REL).  At the end of the monitoring period, concentrations were still
30% above the adult REL at 4,900 ng/ m3 57 ft from the field boundary. 
Also, the breakdown product chlorpyrifos oxon was observed in 100% of
the samples, but the toxicity of this substance was not taken into
account because no RELs are available.  However, because the oxon is
more acutely toxic than the parent compound, neurotoxic effects
associated with breathing air contaminated with both chlorpyrifos and
its oxon at the measured levels will be greater than chlorpyrifos
concentrations alone would suggest.  In sum, inhalation exposures must
be part of the EPA’s risk assessment algorithm for workers.

Sincerely,

Shelley Davis

Pamela Rao

Farmworker Justice

Anne Katten

California Rural Legal Assistance Foundation

Margaret Reeves

Pesticide Action Network

Jennifer Sass

National Resources Defense Council

1126 16th Street, NW, Suite 270• Washington, DC 20036

(202) 293-5420 • (202) 293-5427 fax  • email: fj@nclr.org   • 
www.farmworkerjustice.org

  PAGE  2 

 Fenske R and Lu C (1994).  Determination of Handwash Removal
Efficiency: Incomplete Removal of the Pesticide Chlorpyrifos from Skin
by Standard Handwash Techniques.  American Industrial Hygiene
Association Journal 55:425-32.

 Fenske R, et al. (1999).  Comparison of Three Methods for Assessment of
Hand Exposure to Azinphos-methyl (Guthion) During Apple Thinning. 
Applied Occupational & Environmental Hygiene 14:618-33.

 Findings from the National Agricultural Workers Survey 1997-98, US
Department of Labor, Office of Assistant Secretary for Policy, Research
Paper No.  8.  March 2000 at p.  32.

 The Dow studies are also of limited value because: 1) Dow did not
monitor all crops, tasks or application rates; 2) they include
application rates that Dow termed “typical” (p.  38) even though
farmers can and do routinely apply pesticides at maximum label rates;
and 3) they ignore the fact that Dow’s test methodology excludes
“inadvertent exposures” (as is explicitly noted in reference to the
PCO/LCO testing), and thus significantly underestimates actual exposure.

 Furlong CE, Holland N, Richter RJ, Bradman A, Ho A, Eskenazi B. 
(2006).  PON1 status of farmworker mothers and children as a predictor
of organophosphate sensitivity.  Pharmacogenetics and Genomics
16(3):183-90.

 Organophosphorus Cumulative Risk Assessment (CRA) – 2006 Update.  US
Environmental Protection Agency.  Available from   HYPERLINK
"http://www.epa.gov/pesticides/cumulative/2006-op/index.htm" 
http://www.epa.gov/pesticides/cumulative/2006-op/index.htm , pg 55.

 CRA at Section I.B, page 55.

 Reregistration Eligibility Decision: Chlorothalonil.  US Environmental
Protection Agency.  April 1999.  Available from   HYPERLINK
"http://www.epa.gov/oppsrrd1/REDs/0097red.pdf" 
http://www.epa.gov/oppsrrd1/REDs/0097red.pdf .

 Lim L, et al. (2005).  Chlorothalonil Risk Characterization Document
for Dietary Exposure.  Dept.  of Pesticide Regulation, California
Environmental Protection Agency.  Available from   HYPERLINK
"http://www.cdpr.ca.gov/docs/risk/rcd/chlorothalonil.pdf" 
http://www.cdpr.ca.gov/docs/risk/rcd/chlorothalonil.pdf .

 Draper A, et al. (2003).  Occupational asthma from fungicides fluazinam
and chlorothalonil.   HYPERLINK
"javascript:AL_get(this,%20'jour',%20'Occup%20Environ%20Med.');"
Occupational  and Environmental Medicine 60(1):76-7.

 Reeves M, Katten A, Guzmán M.  (2002).  Fields of Poison 2002. 
Californians for Pesticide Reform.

 Report for the Application (Kings County) and Ambient (Fresno County)
Air Monitoring of Diazinon during Winter, 1998.  California Air
Resources Board, Project No.  C97-070.  Nov 1998.  Available at  
HYPERLINK
"http://www.cdpr.ca.gov/docs/emon/pubs/tac/tacpdfs/diamapl.pdf" 
http://www.cdpr.ca.gov/docs/emon/pubs/tac/tacpdfs/diamapl.pdf , Analyzed
by S.  Kegley, Pesticide Action Network, in Docket number
EPA-HQ-OPP-2006-0618.

 Airborne Poisons: Pesticides in our air and in our bodies. 
Californians for Pesticide Reform.  2007.  Available at   HYPERLINK
"http://www.pesticidereform.org/article.php?id=297" 
http://www.pesticidereform.org/article.php?id=297 .

 Report for the Application and Ambient Air Monitoring of Chlorpyrifos
(and the Oxon Analogue) in Tulare County during Spring/Summer 1996. 
California Air Resources Board.  Test Report #C96-040 and # C96-041. 
April 7, 1998.  Available from:   HYPERLINK
"http://www.cdpr.ca.gov/docs/empm/pubs/tac/chlrpfs.htm" 
http://www.cdpr.ca.gov/docs/empm/pubs/tac/chlrpfs.htm .  Analyzed by S. 
Kegley and K.  Mills, Pesticide Action Network, in Air Monitoring for
Chlorpyrifos in Lindsay, California June-July 2004 and July-August 2005,
July 2006.  Available from   HYPERLINK
"http://www.panna.org/campaigns/docsDrift/Lindsay-CP_7_18_06.pdf" 
http://www.panna.org/campaigns/docsDrift/Lindsay-CP_7_18_06.pdf .

