[Federal Register Volume 77, Number 172 (Wednesday, September 5, 2012)]
[Rules and Regulations]
[Pages 54402-54421]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 2012-21844]
-----------------------------------------------------------------------
ENVIRONMENTAL PROTECTION AGENCY
40 CFR Part 180
[EPA-HQ-OPP-2002-0302; FRL-9359-9]
Dichlorvos (DDVP); Order Denying NRDC's Objections on Remand
AGENCY: Environmental Protection Agency (EPA)
ACTION: Final Order.
-----------------------------------------------------------------------
SUMMARY: In this order, EPA denies an objection to a prior order
denying a petition requesting that EPA revoke all pesticide tolerances
for dichlorvos under section 408(d) of the Federal Food, Drug, and
Cosmetic Act. The objection was filed on February 1, 2008, by the
Natural Resources Defense Council (NRDC). The original petition was
also filed by NRDC. Previously, in July 2008, EPA denied this same
objection but the United States Court of Appeals for the Second Circuit
vacated that decision, in part, and remanded the matter to EPA. This
order is being issued in response to the court's remand.
DATES: This order is effective September 5, 2012.
ADDRESSES: The docket for this action, identified by docket
identification (ID) number EPA-HQ-OPP-2002-0302, is available either
electronically through http://www.regulations.gov or in hard copy at
the OPP Docket in the Environmental Protection Agency Docket Center
(EPA/DC), located in EPA West, Rm. 3334, 1301 Constitution Ave. NW.,
Washington, DC 20460-0001. The Public Reading Room is open from 8:30
a.m. to 4:30 p.m., Monday through Friday, excluding legal holidays. The
telephone number for the Public Reading Room is (202) 566-1744, and the
telephone number for the OPP Docket is (703) 305-5805. Please review
the visitor instructions and additional information about the docket
available at http://www.epa.gov/dockets.
FOR FURTHER INFORMATION CONTACT: Melanie Biscoe, Pesticide Re-
evaluation Division (7508P), Office of Pesticide Programs,
Environmental Protection Agency, 1200 Pennsylvania Ave. NW.,
Washington, DC 20460-0001; telephone number: (703) 305-7106; email
address: [email protected].
SUPPLEMENTARY INFORMATION:
I. General Information
A. Does this action apply to me?
In this document EPA denies an objection by the Natural Resources
Defense Council (NRDC) concerning
[[Page 54403]]
EPA's denial of NRDC's petition to revoke pesticide tolerances. This
action may also be of interest to agricultural producers, food
manufacturers, or pesticide manufacturers. Potentially affected
entities may include, but are not limited to those engaged in the
following activities:
Crop production (North American Industrial Classification
System (NAICS) code 111), e.g., agricultural workers; greenhouse,
nursery, and floriculture workers; farmers.
Animal production (NAICS code 112), e.g., cattle ranchers
and farmers, dairy cattle farmers, livestock farmers.
Food manufacturing (NAICS code 311), e.g., agricultural
workers; farmers; greenhouse, nursery, and floriculture workers;
ranchers; pesticide applicators.
Pesticide manufacturing (NAICS code 32532), e.g.,
agricultural workers; commercial applicators; farmers; greenhouse,
nursery, and floriculture workers; residential users.
B. How can I get electronic access to other related information?
You may access a frequently updated electronic version of EPA's
tolerance regulations at 40 CFR part 180 through the Government
Printing Office's e-CFR site at http://ecfr.gpoaccess.gov/cgi/t/text/text-idx?&c=ecfr&tpl=/ecfrbrowse/Title40/40tab_02.tpl.
II. Introduction
A. What action is the agency taking?
In this order, EPA is issuing a revised denial of an objection to
an earlier EPA order, (72 FR 68662, December 5, 2007), denying a
petition to revoke all tolerances established for the pesticide
dichlorvos (DDVP) under the Federal Food, Drug, and Cosmetic Act
(FFDCA), 21 U.S.C. 346a. Both the objection as well as the petition was
filed with EPA by NRDC. (Refs. 1 and 2). EPA had previously denied this
objection, (73 FR 42683, July 23, 2008), but that order was vacated, in
part, by the United States Court of Appeals for the Second Circuit.
(NRDC v. US EPA, 658 F.3d 200 (2d Cir. 2011)).
NRDC's petition, filed on June 2, 2006, pursuant to FFDCA section
408(d)(1), asserted numerous grounds as to why the dichlorvos
tolerances allegedly fail to meet the FFDCA's safety standard. This
petition was filed as EPA was completing its reassessment of the safety
of the dichlorvos tolerances pursuant to FFDCA section 408(q). (Ref.
3). In response to the petition, EPA undertook an extensive review of
its dichlorvos safety evaluation in the tolerance reassessment
decision. Based on this extensive review, EPA concluded that dichlorvos
met the FFDCA safety standard and, therefore, denied the petition. (72
FR 68695). NRDC then filed objections with EPA to the petition denial
order and requested a hearing on its objections. The objections
narrowed NRDC's claims to two main assertions--that, in assessing the
risk to dichlorvos, EPA unlawfully reduced the statutory tenfold (10X)
additional safety factor for the protection of infants and children and
EPA unlawfully relied on a human toxicity study (the Gledhill study).
After carefully reviewing the objections and hearing requests, EPA
determined that NRDC's hearing requests did not satisfy the regulatory
requirements for such requests and that its substantive objections were
without merit. (73 FR 42709-42711). NRDC sought review of EPA's
decision in the United States Court of Appeal for the Second Circuit.
As noted, the Second Circuit court vacated a portion of EPA's order
finding that ``[b]ecause EPA failed to explain why it did not use a 10X
children's safety factor for dichlorvos risk assessments that relied on
the Gledhill study, EPA acted in an arbitrary and capricious manner.''
(658 F.3d at 218). Specifically, the court vacated ``those portions of
EPA's July 23, 2008 order assessing the risk of dichlorvos based on the
Gledhill study * * * '' (Id.). The court remanded the matter to EPA.
(Id. at 219).
On remand, EPA has carefully examined the court's opinion and has
reconsidered that portion of its prior decision that relied on the
Gledhill study in assessing dichlorvos risk. Because the court found
this portion of EPA's order to be arbitrary and capricious due to its
absence of an adequate explanation on the additional safety factor for
the protection of infants and children, EPA focused on a reexamination
of what additional safety factor for the protection of infants and
children should be applied for the assessments based on the Gledhill
study. EPA concludes, like it did in the July 23, 2008 order, that a
threefold (3X) additional safety factor will protect the safety of
infants and children. Accordingly, EPA again denies NRDC's objections
as to those portions of the July 23, 2008 order that were vacated.
Although EPA reaches the same conclusion on remand on the additional
safety factor for the protection of infants and children, EPA has
provided a revised, more extensive explanation for its position.
Because this revised explanation addresses the court's reason for
finding portions of the July 23, 2008 order to be arbitrary and
capricious, EPA has not otherwise reopened or reconsidered that prior
order.
B. What is the agency's authority for taking this action?
NRDC petitioned to revoke the dichlorvos tolerances pursuant to the
petition procedures in FFDCA section 408(d)(1). (21 U.S.C. 346a(d)(1)).
Under section 408(d), EPA may respond to such a petition by either
issuing a final or proposed rule modifying or revoking the tolerances
or issuing an order denying the petition. (21 U.S.C. 346a(d)(4)). Here,
EPA responded by issuing an order under section 408(d)(4)(iii) denying
the petition. (72 FR 68622, December 5, 2007).
Orders issued under section 408(d)(4)(iii) are subject to a
statutorily-created administrative review process. (21 U.S.C.
346a(g)(2)). Any person may file objections to a section 408(d)(4)(iii)
order with EPA and request a hearing on those objections. (Id.). EPA is
required by section 408(g)(2)(C) to issue a final order resolving the
objections to the section 408(d)(4)(iii) order. (21 U.S.C.
346a(g)(2)(C)). NRDC filed objections to EPA's denial of its dichlorvos
petition and EPA issued a section 408(g)(2)(C) order denying NRDC's
objections. (73 FR 42683, July 23, 2008). EPA's order denying NRDC's
objections was vacated, in part, and remanded to EPA. This revised
order on remand is also being issued under section 408(g)(2)(C).
III. Statutory and Regulatory Background
In this Unit, EPA provides background on the relevant statutes and
regulations governing the matter on remand as well as a much-
abbreviated discussion on pertinent Agency risk assessment policies. A
full discussion of EPA's approach to pesticide risk assessment is
included in EPA's prior order on NRDC's objections. (73 FR 42685-
42688). Because the court's decision focused on the explanation offered
by EPA for its use of safety factors, this Unit includes an expanded
discussion on use of safety or uncertainty factors, including the
additional safety factor required by the FQPA for the protection of
infants and children. Further, because Benchmark Dose Methods analysis
is discussed for the first time in this revised order, a short section
explaining that concept is included.
A. FFDCA/FIFRA and Applicable Regulations
1. In general. EPA establishes maximum residue limits, or
``tolerances,'' for pesticide residues in food and feed commodities
under
[[Page 54404]]
section 408 of the FFDCA. (21 U.S.C. 346a). Without such a tolerance or
an exemption from the requirement of a tolerance, a food containing a
pesticide residue is ``adulterated'' under section 402 of the FFDCA and
may not be legally moved in interstate commerce. (21 U.S.C. 331, 342).
Monitoring and enforcement of pesticide tolerances are carried out by
the U.S. Food and Drug Administration (FDA) and the U.S. Department of
Agriculture (USDA). Section 408 was substantially rewritten by the Food
Quality Protection Act of 1996 (FQPA), which added the provisions
discussed below establishing a detailed safety standard for pesticides,
additional protections for infants and children, and the endocrine
disrupting substances screening program. (Pub. L. 104-170, 110 Stat.
1489 (1996)).
EPA also regulates pesticides under the Federal Insecticide,
Fungicide, and Rodenticide Act (FIFRA), (7 U.S.C. 136 et seq). While
the FFDCA authorizes the establishment of legal limits for pesticide
residues in food, FIFRA requires the approval of pesticides prior to
their sale and distribution, (7 U.S.C. 136a(a)), and establishes a
registration regime for regulating the use of pesticides. FIFRA
regulates pesticide use in conjunction with its registration scheme by
requiring EPA review and approval of pesticide labels and specifying
that use of a pesticide inconsistent with its label is a violation of
Federal law. (7 U.S.C. 136j(a)(2)(G)).
2. Safety standard for pesticide tolerances. A pesticide tolerance
may be promulgated or left in effect by EPA only if the tolerance is
``safe.'' (21 U.S.C. 346a(b)(2)(A)(i)). This standard applies when
responding both to petitions to establish and petitions to revoke
tolerances. ``Safe'' is defined by the statute to mean that ``there is
a reasonable certainty that no harm will result from aggregate exposure
to the pesticide chemical residue, including all anticipated dietary
exposures and all other exposures for which there is reliable
information.'' (21 U.S.C. 346a(b)(2)(A)(ii)).
Risks to infants and children are given special consideration.
Providing additional protection to infants and children was a
particular focus of the FQPA. Section 408(b)(2)(C) requires EPA to make
a specific determination regarding the safety of tolerances to infants
and children and to consider, among other things, information
``concerning the special susceptibility of infants and children to the
pesticide chemical residues * * *.'' (21 U.S.C. 346a(b)(2)(C)(i)(II)
and (ii)(II)). This provision also creates a presumptive additional
safety factor for the protection of infants and children. Specifically,
it directs that ``[i]n the case of threshold effects, * * * an
additional tenfold margin of safety for the pesticide chemical residue
and other sources of exposure shall be applied for infants and children
to take into account potential pre- and post-natal toxicity and
completeness of the data with respect to exposure and toxicity to
infants and children.'' (21 U.S.C. 346a(b)(2)(C)). EPA is permitted to
``use a different margin of safety for the pesticide chemical residue
only if, on the basis of reliable data, such margin will be safe for
infants and children.'' (Id.). For convenience's sake, the legal
requirements regarding the additional safety margin for infants and
children in section 408(b)(2)(C) are referred to throughout this Order
as the ``FQPA safety factor for the protection of infants and
children'' or simply the ``FQPA safety factor.''
3. Procedures for establishing, amending, or revoking tolerances.
Tolerances are established, amended, or revoked by rulemaking under the
unique procedural framework set forth in the FFDCA. Generally, a
tolerance rulemaking is initiated by the party seeking to establish,
amend, or revoke a tolerance by means of filing a petition with EPA.
(See 21 U.S.C. 346a(d)(1)). EPA publishes in the Federal Register a
notice of the petition filing and requests public comment. (21 U.S.C.
346a(d)(3)). After reviewing the petition, and any comments received on
it, EPA may issue a final rule establishing, amending, or revoking the
tolerance, issue a proposed rule to do the same, or deny the petition.
(21 U.S.C. 346a(d)(4)).
Once EPA takes final action on the petition by establishing,
amending, or revoking the tolerance or denying the petition, any party
may file objections with EPA to EPA's decision on the petition and seek
an evidentiary hearing on those objections. (21 U.S.C. 346a(g)(2)).
Objections and hearing requests must be filed within 60 days. (Id.).
The statute provides that EPA shall ``hold a public evidentiary hearing
if and to the extent the Administrator determines that such a public
hearing is necessary to receive factual evidence relevant to material
issues of fact raised by the objections.'' (21 U.S.C. 346a(g)(2)(B)).
EPA regulations make clear that hearings will only be granted where it
is shown that there is ``a genuine and substantial issue of fact,'' the
requestor has identified evidence that ``would, if established, resolve
one or more of such issues in favor of the requestor,'' and the issue
is ``determinative'' with regard to the relief requested. (40 CFR
178.32(b)). Further, a party may not raise issues in objections unless
they were part of the petition and an objecting party must state
objections to the EPA decision and not just repeat the allegations in
its petition. Corn Growers v. EPA, 613 F.2d 266 (D.C. Cir. 2010), cert.
denied, 131 S. Ct. 2931 (2011). EPA's final order on the objections is
subject to judicial review. (21 U.S.C. 346a(h)(1)).
B. EPA Risk Assessment for Tolerances--Policy and Practice
1. The safety determination--risk assessment. To assess risk of a
pesticide tolerance, EPA combines information on pesticide toxicity
with information regarding the route, magnitude, and duration of
exposure to the pesticide. The risk assessment process involves four
distinct steps: (1) Identification of the toxicological hazards posed
by a pesticide; (2) determination of the ``level of concern'' with
respect to human exposure to the pesticide; (3) estimation of human
exposure to the pesticide; and (4) characterization of risk posed to
humans by the pesticide based on comparison of human exposure to the
level of concern.
Toxicological hazards posed by a pesticide are identified through
use of testing in laboratory animals or humans. Generally, EPA will use
the lowest ``no observed adverse affect level'' (NOAEL) or ``lowest
observed adverse effect level'' (LOAEL) from the available studies or a
calculated value called a Benchmark Dose as a starting point (called
``the Point of Departure'') in estimating the ``level of concern'' for
human exposure to the pesticide. Points of Departure and levels of
concern will be identified for all exposure routes to the pesticide
(oral, dermal, and inhalation) and durations of exposure (acute, short-
term, intermediate-term, and chronic). Another critical aspect of the
``level of concern'' determination involves the use of safety or
uncertainty factors to compensate for the limitations of toxicology
testing. Safety and uncertainty factors are discussed in detail in Unit
III.B.2. below. Having identified a pesticide's hazards, the Point(s)
of Departure, and level(s) of concern, EPA then estimates exposure to
the pesticide taking into account the various routes of exposure, how
exposures vary over time, and the differences in exposure to different
subpopulations. Finally, EPA combines information on hazard, level of
concern, and exposure to produce a characterization of the risk posed
by the pesticide. Risks are calculated for all of the various routes
and durations of exposure scenarios associated with a pesticide. These
risk assessment
[[Page 54405]]
scenarios may be calculated separately for different age-based
population groups (e.g., non-nursing infants) or applied to all
population groups, including infants and children, depending on
information on the potential for exposure and data on differential
sensitivity. A more comprehensive discussion of this risk assessment
process is presented in EPA's previous order denying objections. (73 FR
42685-42689).
Before turning to a detailed discussion of safety and uncertainty
factors, EPA's risk characterization process is briefly summarized
because it is frequently referred to in this order. For pesticides that
pose a risk over a certain threshold of exposure, EPA's
characterization of risk is presented in one of two ways: Either using
the Reference Dose (RfD) approach or the Margin of Exposure (MOE)
approach. Importantly, these different approaches do not render
substantively different results. Both approaches use the same data--the
Point of Departure, the applicable safety/uncertainty factors, and
human exposure to the pesticide; they just express the characterization
of risk in a different metric. Under the RfD approach, EPA directly
extrapolates a dose from an animal or human study to an overall safe
dose for humans. An RfD is calculated by dividing all applicable
safety/uncertainty factors into the level of exposure from animal or
human studies determined appropriate for assessing risk (i.e., the
``Point of Departure''). Estimated human exposure to the pesticide is
then compared to the RfD to determine if it is excessive. Under the
Margin of Exposure (MOE) approach, EPA does not calculate a safe dose
in humans but rather focuses on the margin of exposure between a dose
from an animal or human study and human exposure to the pesticide. A
MOE is calculated by dividing human exposure to the pesticide into the
Point of Departure. To determine whether that MOE is considered
sufficiently protective of humans, EPA compares it to the product of
all applicable safety/uncertainty factors, referred to as the target
MOE. MOEs that are less than the target MOE indicate a risk of concern.
At bottom, both approaches extrapolate a safe measure of human exposure
from animal or human studies using a mixture of uncertainty/safety
factors.
2. Safety and uncertainty factors. i. History. It has long been a
standard risk assessment practice to use numerical factors in
conjunction with experimental toxicity data in assessing risk to humans
from exposure to chemical substances. (Ref. 4). These numerical factors
are designed to provide an additional margin of safety so that risks to
the populations covered by an assessment are not understated. The
practice was first developed by the Food and Drug Administration (FDA)
in the middle part of the last century. (Ref. 5). An influential 1954
paper by two FDA scientists called for a hundredfold margin of safety
when extrapolating from long-term animal experiments to calculate safe
doses in humans. (Ref. 6). The paper justified this safety factor on
the basis of, among other things, potential differences in sensitivity
between humans and laboratory animals as well as potential variations
in sensitivity within humans. Accordingly, the paper recognized that a
smaller factor would be appropriate where adequate human data are
available. An explicit recommendation for a factor ``as low as 10'' was
made by the Joint Food and Agricultural Organization/World Health
Organization (FAO/WHO) Meeting on Pesticide Residues in 1965 for
circumstances where human data was relied upon. (Ref. 7 at 12).
Eventually, it became common regulatory practice to treat the
hundredfold margin of safety as comprised of two tenfold factors: The
first addressing the potential difference in sensitivity between humans
and experimental animals (i.e., interspecies sensitivity) and the
second addressing variation within the human population (i.e.,
intraspecies sensitivity). The rationale for these two factors is
concisely summarized in a recent publication from the International
Programme on Chemical Safety:
The interspecies uncertainty factor can be considered to convert
the NOAEL/NOAEC [No observed adverse effect concentration] for
animals (derived from a small group of relatively homogeneous test
animals) into the NOAEL/NOAEC anticipated for an average
representative healthy human. The uncertainty factor for human
variability converts the NOAEL/NOAEC for the average human into a
NOAEL/NOAEC for susceptible humans. Although adverse effect data in
humans can be used directly without the need for an interspecies
factor, the paucity of such data means that the vast majority of
risk assessments are based on studies in experimental animals.
(Ref. 8 at 15).
EPA, as well as other Federal and international regulatory bodies,
also will, where appropriate, apply additional numerical factors to
take into account chemical-specific considerations affecting the risk
assessment. (Ref. 9) Use of these additional factors is further
explained in Unit III.B.2.v., vi, and vii.
ii. Terminology. Different terminology has been used to label
numerical factors used in calculating safe doses of chemical
substances. As noted, they were first referred to as ``safety''
factors. The terminology has evolved over the decades, however, such
that what was once generally called a safety factor has come to be
generally referred to as an uncertainty factor. (Ref. 10 at A-3). The
rationale for the change was that, although the use of such factors
does promote safety, there was a concern that the use of the term
``safety'' implied that these factors provided absolute safety. (Ref.
11). The FQPA reintroduced the term ``safety'' factors with its
reference to a ``margin of safety.'' 21 U.S.C. 346a(b)(2)(C).
Subsequent to the passage of FQPA, EPA's Office of Pesticide Programs
(OPP) has used the terms safety factor and uncertainty factor
interchangeably. Both terms have been criticized by the National
Academy of Sciences (NAS). The NAS explained that the terms safety and
uncertainty imply that factors ``are simply added on for safety or
because of a lack of knowledge or confidence in the process.'' (Ref. 12
at 132). To the contrary, according to the NAS, these factors are
scientifically-based and used ``to adjust for differences in individual
human sensitivities, for humans' generally greater sensitivity than
test animals' on a milligram-per-kilogram basis, for the fact that
chemicals typically induce harm at lower doses with longer exposures,
and so on.'' (Id.).
iii. Scientific basis for inter- and intraspecies factors. Only
limited scientific data, involving differing sensitivity of humans and
animals, are cited in the 1954 article in justification of the
recommendation for a hundredfold safety factor. Subsequent
investigations of both animal and human toxicity data, however, have
provided general support for the protectiveness of the tenfold factors
for interspecies and intraspecies sensitivity differences if an
adequate toxicity database is available. (Refs. 9, 13, 14, and 15). The
interspecies factor has been investigated through comparisons of
toxicity testing in laboratory animals and humans. (Refs. 15 and 16).
The protectiveness of the human intraspecies factor has been assessed
through examining sub-population differences both among various human
age groups (the young, adults, and elderly) as revealed in
pharmaceutical trials and between juvenile and adult laboratory animals
identified in toxicity testing. (Ref. 13 at 211 (``For substances other
than pharmaceuticals, age-related differences in toxicity have been
primarily investigated in rodent studies.''); Ref. 17 at 462-463
[[Page 54406]]
(describing pharmaceutical trials involving humans and comparative
studies in juvenile and adult laboratory animals)). For example, the
NAS, in its report ``Pesticides in the Diets of Infants and Children,''
looked to both human data and animal data in evaluating the potential
for increased sensitivity in infants and children to pesticides. (Ref.
18 at 344-345).
iv. Adjustment of inter- and intraspecies factors. In addition to
evaluating the protectiveness of the intra- and interspecies
uncertainty factors, scientists have also examined both generic
biological as well as chemical-specific factors that may affect intra-
and interspecies variability with the aim of deriving more accurate
uncertainty factor values than the default tenfold values.
One reason humans are considered to be potentially more sensitive
to toxic agents than laboratory animals is that otherwise equivalent
external doses of such agents for humans and animals on a milligram-
per-kilogram of body weight basis may result in a greater internal dose
for humans. This is due to species differences in general metabolic
processes--commonly referred to as toxicokinetics--and ``is thought to
be related to species differences in exchange surfaces and distribution
networks that constrain concentration and flux of metabolic
reactants.'' (Ref. 19 at 4-35; see Ref. 15 at 228).
In addition to toxicokinetic effects on internal dose, differences
between humans and laboratory animals are also driven by toxicodynamic
factors. Toxicodynamics refers to the manner in which the target tissue
and body respond to the toxic agent. Thus, interspecies differences are
a factor of both differences in the internal dose received by humans
and animals and differences in how humans and animals react to the
internal dose received. Similarly, sensitivity differences between
juveniles and adults, whether humans or animals, are also considered to
be tied to toxicokinetic and toxicodynamic factors. Accordingly, both
the inter- and intraspecies uncertainty factors are considered to have
toxicokinetic and toxicodynamic components. EPA typically has
considered both the tenfold (10X) inter- and intraspecies factors to be
roughly equally divided on a logarithmic basis (i.e., 10\0.5\ or
roughly a 3X factor) between toxicokinetics and toxicodynamics. (Ref.
19 at 4-29; see also Ref. 19 at 4-40 (explaining why two 3X factors
[technically, 3.16X] would be equivalent to a 10X factor)). Other
organizations have recommended that, while toxicokinetics and
toxicodynamics play an equal role in intra-human variability,
toxicokinetics has a greater effect on interspecies differences and
thus recommend that the tenfold interspecies factor be divided into a
fourfold factor for toxicokinetics and 2.5-fold factor for
toxicodynamics. (Ref. 8 at 17; see Ref. 14).
Of the toxicokinetic and toxicodynamic differences between humans
and animals and among various human subgroups, the most is known about
the toxicokinetic differences between humans and animals. For
inhalation exposures, EPA has used toxicokinetic information on humans
and animals to create generic dosimetric adjustment factors that
replace that portion of the interspecies factor tied to toxicokinetic
differences. (Refs. 19 at 4-29; 20). Where such dosimetric adjustment
factor is used, the interspecies factor is reduced to 3X.
EPA guidance entitled ``A Review of the Reference Dose and
Reference Concentration Processes'' (``RfD Guidance'') also urges that
data be developed to support substitution of chemical-specific
adjustment factors (sometimes referred to as data-derived factors) for
the default 10X uncertainty factors for inter- and intraspecies
variability. (Ref. 19 at xviii -xix, 4-47). This guidance recognizes
that chemical-specific data from both humans and animals has been
relied upon by EPA to adjust the human intraspecies uncertainty factor
citing an article by Dourson et al. That article collects instances in
which EPA has adjusted uncertainty factors on a chemical-specific
basis. (Ref. 9). For example, Dourson et al. point to a 1996 EPA
assessment of Aroclor that reduced the human intraspecies factor to 3X
given that the Point of Departure came from a sensitive animal
population--there, infant rhesus monkeys. In discussing the Dourson et
al. article, the RfD Guidance notes that:
In those cases where developmental effects were the most
sensitive endpoint (0 RfCs, 6 RfDs), reduction of the intraspecies
[uncertainty factor] from 10 to 3 was based on data derived either
from human data showing which age groups or time periods were most
susceptible (e.g., methyl mercury exposure to the developing fetus)
or from an animal study with support from strong human or other data
(e.g., Aroclor 1016 in utero exposure in monkeys, strontium-induced
rachitic bones in young rats).
(Ref. 19 at 4-43). The RfD Guidance endorsed a view similar to that
expressed in an agency-wide paper prepared in development of EPA's
Children's Safety Factor Policy. That paper also noted that there were
circumstances where data from human studies or from animal studies
might support reduction of the human intraspecies uncertainty factor:
``The Toxicology Working Group recommends that reduction of the
intraspecies uncertainty factor from a default of 10 be considered only
if data are complete and the age group or window of vulnerability
during development has been clearly delineated, preferably based on
human data or on animal data with supporting human data.'' (Ref. 21 at
28). On the other hand, the RfD guidance also recognized that a 10X
intraspecies factor ``may sometimes be too small because of factors
that can influence large differences in susceptibility, such as genetic
polymorphisms.'' (Ref. 19 at 4-44).
In sum, the 10X inter- and intraspecies factors are default values.
Although there is substantial scientific support for these default
values, chemical-specific human and animal data may be relied upon in
reducing, confirming, or increasing these default values.
v. Additional Safety/Uncertainty Factors. In addition to the inter-
and intraspecies factors, risk assessors from EPA as well as other
Federal and international regulatory agencies also apply ``additional''
or ``modifying'' safety/uncertainty factors based on specific
circumstances related to the toxicity data, particularly with regard to
deficiencies in that data. Like the inter- and intra-species factors,
these additional factors help to ensure that risks to populations
covered by an assessment are not understated. Additional factors are
applied to address: (1) An absence of critical toxicity data; (2) the
failure of a study to identify a NOAEL; (3) the necessity of using sub-
chronic data to choose a Point of Departure for estimating chronic
risk; and (4) results in a study that suggest the inter- or
intraspecies factors may not be sufficient (sometimes referred to as a
``modifying factor''). (Ref. 10 at 9). Generally, a safety factor value
of 10X or 3X (which is considered to be one-half of 10X on the
logarithmic scale) is used to address these concerns.
The protectiveness of these default values has also been the
subject of scientific examination. Studies have been done on the
variations in the levels of NOAELs in the databases for various
pesticides. They confirm the need for an additional factor when core
data are lacking. (Ref. 22). Examination of the completeness of the
animal database remains important even when human data are used as the
Point of Departure for calculating the RfD. The latest EPA guidance on
RfDs emphasizes that in
[[Page 54407]]
these circumstances ``[i]nformation on life stages and organ systems
may come from either animal or human studies.'' (Ref. 19 at 4-45). The
guidance notes that ``the lack of a two-generation animal reproduction
study might be considered a deficiency even if the reference value is
based on human data.'' (Id.). Similarly, research has been conducted on
existing databases to determine the adequacy of uncertainty factors
used to address reliance on a LOAEL instead of a NOAEL, or subchronic
data to estimate chronic risk. (Refs. 9 and 15).
Selection of particular values for these additional uncertainty
values depends on what is known from the full body of information about
the chemical, including both data from testing with animals and humans,
about the chemical. For example, as EPA's RfD Guidance advises: ``the
size of the database factor to be applied will depend on other
information in the database and on how much impact the missing data may
have on determining the toxicity of a chemical and, consequently, the
POD [Point of Departure].'' (Ref. 19 at 4-45). With regard to an
additional factor for extrapolation of a NOAEL from a LOAEL, Dourson et
al. report that ``[a]nalysis of several data bases suggest that a
factor of 10 or lower is adequate and that use of data does support a
lower factor with certain chemicals.'' (Ref. 9 at 112). The critical
consideration, according to Dourson et al., is the severity of the
effect at the LOAEL: ``The data indicate that when faced with a LOAEL
and not a NOAEL, the choice of uncertainty factor should generally
depend on the severity of the effect at the LOAEL.'' (Id.).
Specifically, Dourson et al. note that ``[l]ess severe effects would
not require a large factor, because, presumably, the LOAEL is closer to
the unknown NOAEL.'' (Id.).
vi. FQPA safety factor--integration with traditional uncertainty
factors. EPA's safety/uncertainty factor practice with regard to
pesticides was altered to a degree by the Food Quality Protection Act
(FQPA). (Ref. 10). That Act established a presumptive additional
``safety'' factor of 10X to protect infants and children. The
additional factor was designed to account for the completeness of the
toxicity and exposure databases and the potential for pre- and post-
natal toxicity. EPA has interpreted this legislation as both a
``codification and expansion'' of prior EPA practice with regard to
additional safety/uncertainty factors. (Ref. 10 at A-3--A-5). It
codified EPA's prior practice by requiring the additional presumptive
factor to address toxicity data completeness issues (i.e., absence of a
particular study, lack of a NOAEL in a completed study, or absence of
chronic data). These traditional additional uncertainty factors became
FQPA safety factors for the protection of infants and children. This
accords greater protection to infants and children because for FQPA
safety factors, unlike pre-FQPA additional factors, there is a
presumption, which can only be overcome by reliable data, that they
will be applied. At the same time, EPA concluded that Congress had not
intended EPA to double-up on safety factors by, for example, applying
an additional uncertainty factor due to missing data, and applying an
FQPA additional safety factor as well to address the same missing data.
(Ref. 10 at A-4). Congress expanded EPA's prior practice by providing
that the additional FQPA safety factor for the protection of infants
and children was designed to address not just toxicity data
deficiencies but exposure data deficiencies as well and by its emphasis
on protecting against potential pre- and post-natal toxicity. In
theory, EPA could have, prior to the enactment of the FQPA, used an
``additional'' or ``modifying'' factor to address health risks to
children not otherwise protected by the interspecies, intraspecies, or
data deficiency safety factors, but use of such a factor was not
common. The FQPA also modified the status quo by making the additional
safety factor for infants and children presumptive in nature.
The narrowly-focused and highly-prescriptive nature of the FQPA
safety factor provision has required careful integration with pesticide
risk assessment approaches under other statutes and, more generally,
with Agency risk assessment practices. As noted above, the FQPA, with
regard to the assessment of risks to infants and children, essentially
codified EPA's prior risk assessment practice as to additional
uncertainty factors and it expanded the use of additional uncertainty
factors into new areas. The FQPA, however, did not speak to use of
traditional (non-additional) uncertainty factors (i.e., the inter- and
intraspecies factors). Thus, the end result was that some uncertainty
factors for FFDCA pesticides remained unaffected by the new statutory
requirements (the inter- and intraspecies factors), some uncertainty
factors became FQPA safety factors (additional uncertainty factors that
addressed toxicity data deficiencies), and some safety factors that
either had previously never existed or were at least extremely rare
were created as a statutory phenomenon (a factor to address exposure
data base deficiencies and a factor to address potential pre- and post-
natal toxicity). This selective inter-weaving of statutory requirements
with Agency science policy made FFDCA risk assessments for pesticides
unique compared to general Agency risk assessment practice.
Pesticide risk, however, is not regulated under a single statute.
Risks to workers or the environment from pesticide use are regulated by
EPA under FIFRA, not the FFDCA. Further, EPA may address risks posed by
pesticide contamination of the environment under several other
statutes, including the Safe Drinking Water Act, 42 U.S.C. 300f et
seq., the Resource Conservation and Recovery Act, 42 U.S.C. 6901 et
seq., and the Comprehensive Environmental Response, Compensation, and
Liability Act, 42 U.S.C. 9601 et seq. Prior to enactment of the FQPA's
specific provisions on pesticide risk assessment, a pesticide risk
assessment performed by EPA's Office of Pesticide Programs under the
aegis of FFDCA section 408 could generally be easily translated for use
by the Office of Pesticide Programs under FIFRA, or by the other media
offices within EPA for use under other statutes. However, once
pesticide risk assessment under the FQPA became not simply a matter of
good scientific practice but was channeled by explicit statutory
requirements, it became incumbent upon the Office of Pesticide Programs
to prepare its FFDCA pesticide risk assessments in a manner that
clearly delineated what aspects of the assessment were driven solely by
science and what aspects primarily by FQPA statutory requirements.
Specifically, the Office of Pesticide Programs had to be transparent
with regard to whether it was relying on FQPA safety factors based on
unique FQPA requirements (exposure database deficiencies and potential
pre- and post-natal toxicity) or FQPA safety factors that are
essentially a codification of prior general EPA ``additional'' safety/
uncertainty factor practice.
EPA addressed these transparency issues at length in its 2002
policy statement on the FQPA safety factor. To clarify how the FQPA
safety factor provision left a portion of prior safety/uncertainty
practice unchanged, codified another portion, and also expanded the use
of safety factors, EPA explained the overlap between the FQPA safety
factor and additional safety factors in depth and included the
following figure to graphically illustrate the issue:
[[Page 54408]]
[GRAPHIC] [TIFF OMITTED] TR05SE12.000
With regard to providing transparency on the FQPA safety factor
decisions, EPA took two steps. First, it adopted a new term, the
``special'' FQPA safety factor, for children safety factors that were
based solely on the new FQPA requirements. Second, it adopted the
approach of calculating two different safe doses for a pesticide: one
that excluded any ``special'' FQPA safety factors and one that included
them. The former was referred to, in line with standard EPA policy, as
a Reference Dose (RfD), and the latter as a Population Adjusted Dose
(PAD). Introducing the new terminology on FQPA safety factors into
long-established safety factor practice has proved challenging. EPA
staff on occasion drafted documents that (1) claimed no FQPA safety
factor was needed but applied an additional uncertainty factor to
address the completeness of the toxicity data base or reliance on a
LOAEL; or (2) treated the ``special'' FQPA safety factor as the only
type of FQPA safety factor. However, as EPA's policy made clear, EPA
interpreted FFDCA section 408(b)(2)(C) as codifying prior practice as
to additional uncertainty factors such that these factors became FQPA
factors. The mislabeling of uncertainty factors did not substantively
change risk assessment outcomes but it did raise the confusion level on
an already complex topic. Eventually, EPA determined that the term
``special'' FQPA safety factor caused more problems than it solved and
abandoned it. However, EPA has retained the approach of continuing to
calculate both a safe dose with, and without, what was once referred to
as ``special'' FQPA safety factors.
vii. FQPA safety factor--decision-making guidance. In 2002, EPA
issued detailed policy guidance for Agency risk assessors on decision-
making under the FQPA safety factor provision. The purpose of this
guidance was concisely set forth by EPA: ``[T]his guidance explains how
OPP intends to `take intoaccount * * * potential pre- and post-natal
toxicity and completeness of the data with respect to exposure and
toxicity to infants and children'' as directed by FFDCA section
408(b)(2)(C)(i).' '' (Ref. 10 at ii). Although the guidance is
structured around these statutory considerations, EPA also emphasizes
throughout that the FQPA safety factor decision is a weight-of-the-
evidence decision that must consider all available data. Thus, the
policy specifies that ``[b]efore any decisions are made on the
appropriate FQPA safety factor applied to ensure the safety of infants
and children from the use of a particular pesticide, all of the
relevant submitted data for the pesticide should be assembled and
reviewed by Agency scientists.'' (Id. at 8).
This emphasis on the broadness of the inquiry is repeated in the
discussion of the statutory consideration related to the completeness
of the toxicity database. According to EPA, this consideration should
not be narrowly focused on EPA's existing database requirements.
Rather, ``the `completeness' inquiry should be a broad one that takes
into account all data deficiencies.'' (Ref. 10 at 23). At the same
time, the guidance stresses that ``a determination of the possible need
for and size of the database uncertainty factor will necessarily
involve an assessment that
[[Page 54409]]
considers the overall weight-of-evidence to evaluate the significance
of the data deficiency.'' (Id. at 26).
With regard to potential pre- and post-natal toxicity, the policy
emphasizes that evaluation of this consideration cannot be divorced
from the existing process for choosing levels of concern (i.e., RfDs,
PADs, and target MOEs). Thus, EPA instructs risk assessors to evaluate
the concern with data showing pre- and post-natal toxicity by
considering, among other things, ``the degree to which protection for
infants and children is provided by the standard approach for deriving
RfDs through the application of traditional uncertainty factors.'' (Id.
at 29). The guidance stresses that ``[i]n particular, the risk assessor
should consider the protection accorded infants and children by the
intraspecies uncertainty factor.'' (Id.). EPA notes that the scientific
literature as well as the National Academy of Sciences has concluded
that the intraspecies factor is generally adequate to protect infants
and children; however, the policy points out that certain chemicals may
display greater than 10X age-related variability. For this reason, EPA
reiterates that ``[t]he adequacy of the standard intraspecies factor to
address the potential for greater sensitivity or susceptibility of
children should be considered in the context of evidence on potential
pre- and post-natal toxicity as discussed below.'' (Id.; see also Id.
at 51-52). The policy paper went on to provide numerous examples of
weight-of-the-evidence considerations relevant to evaluation of human
and animal data on pre- and post-natal toxicity. (Id. at 30-33).
The discussion on the completeness of the exposure database focuses
on whether the various approaches EPA uses to assess exposure are
likely to understate it. Risk assessors are to evaluate whether their
assessments ``have addressed all significant exposure routes'' and
whether ``there may be uncertainty about whether OPP's approach to
estimating exposure for a particular use pattern, pathway, or aggregate
exposure is sufficiently health protective.'' (Id. at 48).
3. Benchmark dose approach. As indicated above, EPA has
traditionally used a NOAEL or LOAEL as a Point of Departure in
estimating an exposure level of concern for a pesticide or other
substance. Increasingly, however, EPA uses a more sophisticated
modeling tool known as the Benchmark Dose approach as an alternative to
using NOAELs or LOAELs for Point of Departure selection. (Refs. 23). A
benchmark dose, or BMD, is a point estimate along a dose-response curve
that corresponds to a specific response level. For example, a BMD\10\
represents a 10% change from the background level (the background level
is typically derived from the control group). In addition to a BMD, a
confidence limit may also be calculated. Confidence limits express the
uncertainty in a BMD that may be due to sampling and/or experimental
error. The lower confidence limit on the BMD is termed the benchmark
dose limit (BMDL). Use of a BMD or BMDL for deriving the Point of
Departure allows more precise estimates of the Point of Departure,
resulting in tighter confidence intervals. Use of the BMDL also helps
ensure with high confidence (e.g., 95% confidence) that the selected
percentage of change from background is not exceeded. Numerous
scientific peer review panels over the last decade have supported the
Agency's application of the BMD approach as a scientifically
supportable method for deriving Point of Departures in human health
risk assessment, and as an improvement over the historically applied
approach of using NOAELs or LOAELs. (Refs. 24, 25, and 26). The NOAEL/
LOAEL approach can look at the dose response at only the few doses used
in a study, and is therefore limited by the characteristics of the
study design, such as dose selection, dose spacing, and sample size.
(Ref. 23 at 3-5). With the BMD approach, all the dose response data are
used to derive a dose response curve. For all of these reasons, BMD
analysis is preferred by EPA to the NOAEL/LOAEL approach of selecting a
Point of Departure from studies when the available data are amenable to
BMD modeling consistent with the biological processes relevant to the
study in question.
IV. Dichlorvos
Dichlorvos is a chlorinated organophosphate pesticide that inhibits
plasma, red blood cell (RBC), and brain cholinesterase in a variety of
species. (Ref. 3 at 122-123). Cholinesterase inhibition is a disruption
of the normal process in the body by which the nervous system
chemically communicates with muscles and glands. Although
cholinesterase inhibition in the nervous system is not itself regarded
as a direct adverse effect, it is ``generally accepted as a key
component of the mechanism of toxicity leading to adverse cholinergic
effects.'' (Ref. 27 at 25; see 73 FR 42688-42689). Inhibition of blood
cholinesterase ``is not an adverse effect, but may indicate a potential
for adverse effects on the nervous system'' and thus serves as a
``surrogate'' for cholinesterase inhibition in the nervous system (Ref.
27 at 28). Subchronic and chronic oral dichlorvos exposures to rats and
dogs as well as chronic inhalation dichlorvos exposure to rats resulted
in significant decreases in plasma, RBC, and/or brain cholinesterase
activity. Repeated, oral subchronic dichlorvos exposures in male humans
were associated with statistically and biologically significant
decreases in RBC cholinesterase inhibition. These cholinesterase
effects occurred at dose levels below levels at which any other adverse
effect was seen. Generally, there was no evidence of increased
sensitivity to young animals following exposure to dichlorvos. No
evidence of increased sensitivity to young animals was seen following
in utero dichlorvos exposure to rat and rabbit fetuses as well as pre/
post natal dichlorvos exposure to rats in developmental, reproduction,
and comparative cholinesterase studies. The only evidence of
sensitivity in the young was seen in one parameter, auditory startle
amplitude, in a developmental neurotoxicity study; however, the effects
in the rat pups in that study were at levels well above levels that
result in RBC cholinesterase inhibition.
Because inhibition of cholinesterase activity was identified as the
most sensitive effect, it was selected as the toxicity endpoint for
assessment of risks for all acute and chronic dietary exposures, as
well as short-, intermediate-, and long-term (chronic) dermal,
inhalation, and incidental oral residential exposures. For each risk
assessment scenario, EPA selected a Point of Departure based on either
an animal or human study taking into account the duration of the study
and the route of exposure used in the study. (Ref. 3 at 130-135). These
Points of Departure were used in calculating RfD/PADs and acceptable
MOEs. Due to the lack of sensitivity differences between adults and
juveniles, the resulting RfD/PADs and acceptable MOEs were designated
as applicable to all population subgroups, including infants and
children. Animal studies were used in choosing levels of concern for
evaluating risk from acute and chronic dietary exposure; acute dermal
exposure; and acute and chronic inhalation exposure. A human study (the
Gledhill study) was used in evaluating risk from short-term incidental
oral exposure; short-, intermediate-, and long-term dermal exposure;
and short- and intermediate-term inhalation exposure. All of the
studies from which a Point of Departure was selected were conducted in
adults
[[Page 54410]]
(adult humans or adult animals). (See Table 1).
Safety factor determinations used in determining the level of
concern for each risk assessment scenario differed based on whether EPA
relied on one of several different animal studies or a human study for
the Point of Departure for that scenario. For levels of concerns
derived from a Point of Departure from an animal study, EPA generally
applied a 100X safety factor (10X for interspecies variability and 10X
for intraspecies variability). Based on a weight-of-the-evidence
evaluation, EPA removed the 10X FQPA safety factor for risk assessments
based on an animal study. (See Table 1). EPA's weight-of-the-evidence
evaluation concluded that (1) the toxicity database was complete; (2)
most of the data indicated no increased sensitivity in the young and
the only evidence of increased sensitivity occurred at levels well
above the Points of Departure used for establishing the levels of
concern; and (3) its estimate of human exposure to dichlorvos was not
understated.
For levels of concerns derived from a Point of Departure from the
human study, EPA applied a 10X safety factor for intraspecies
variability and a 3X FQPA safety factor. (72 FR 68694-68695). No
interspecies factor was applied because EPA was not extrapolating a
level of concern in humans from a dose in an animal study. The weight-
of-the-evidence balance for the FQPA safety factor was slightly
different for risk assessments relying on the Gledhill human study for
the Point of Departure. In addition to all of the considerations
pertaining to the assessments with an animal-derived Point of
Departure, the Gledhill-based risk assessments introduced another
factor to consider--namely, that the Gledhill study raised a data
completeness issue due to the fact that it only identified a LOAEL.
This latter factor convinced EPA to retain a portion of the FQPA safety
factor when relying on the human study for the Point of Departure. EPA
concluded, however, that reliable data supported reduction of the 10X
factor to 3X because the effect seen at the LOAEL in that study was so
marginal (16 percent RBC cholinesterase inhibition) that a lower dose
would have been unlikely to detect any adverse effect. (72 FR 68694-
68695; see Table 1).
Table 1--Summary of Risk Assessment Scenarios, Population Groups, and Uncertainty/Safety Factors for Dichlorvos
----------------------------------------------------------------------------------------------------------------
Study from which Population groups
Scenario point of departure Age and species of covered by risk Uncertainty/safety
taken study subjects assessment factors
----------------------------------------------------------------------------------------------------------------
Acute Dietary................... Rat acute oral Adult rats........ All population Interspecies--10X;
cholinesterase groups, including Intraspecies--10X
study. infants and ; FQPA--1X.
children.
Chronic Dietary................. 1-year dog study.. Adult dogs........ All population Interspecies--10X;
groups, including Intraspecies--10X
infants and ; FQPA--1X.
children.
Short-term Incidental Oral...... Human 21-day oral Adult humans...... All population Interspecies--1X;
study. groups, including Intraspecies--10X
infants and ; FQPA--3X.
children.
Acute Dermal and Acute Rat acute oral Adult rats........ All population Interspecies--10X;
Incidental Oral. cholinesterase groups, including Intraspecies--10X
study. infants and ; FQPA--1X.
children.
Short-, Intermediate- and Long- Human 21-day oral Adult humans...... All population Interspecies--1X;
term Dermal. study. groups, including Intraspecies--10X
infants and ; FQPA--3X.
children.
Acute Inhalation................ Rat acute oral Adult rats........ All population Interspecies--10X;
cholinesterase groups, including Intraspecies--10X
study. infants and ; FQPA--1X.
children.
Short- and Intermediate-term Human 21-day oral Adult humans...... All population Interspecies--1X;
Inhalation. study. groups, including Intraspecies--10X
infants and ; FQPA--3X.
children.
Long-term Inhalation............ 2-year rat Adult rats........ All population Interspecies--10X;
inhalation study. groups, including Intraspecies--3X;
infants and FQPA--1X.
children.
----------------------------------------------------------------------------------------------------------------
V. NRDC's Petition to Revoke Dichlorvos Tolerances and the
Administrative Proceedings on the Petition
A. NRDC's Petition and EPA's Denial of the Petition
On June 2, 2006, the NRDC filed a petition with EPA which, among
other things, requested that EPA conclude the dichlorvos tolerance
reassessment process by August 3, 2006, with a finding that the
dichlorvos tolerances do not meet the FFDCA safety standard and issue a
final rule by August 3, 2006, revoking all dichlorvos tolerances.
NRDC's petition contained dozens of claims as to why dichlorvos' FFDCA
tolerances should be revoked. After carefully considering all of NRDC's
claims, the public comment received on the petition, and a revised risk
assessment EPA conducted in response to the petition, EPA issued an
order pursuant to FFDCA section 408(d)(4)(iii) denying the request to
revoke dichlorvos' FFDCA tolerances. (72 FR 68662, December 5, 2007).
B. NRDC's Objections and EPA's Denial of the Objections
On February 1, 2008, NRDC filed, pursuant to FFDCA section
408(g)(2), objections to EPA's denial of its tolerance revocation
petition and requested a hearing on those objections. NRDC's objections
and requests for hearing included two main claims: (1) That EPA has
unlawfully failed to retain the full 10X safety factor for the
protection of infants and children; and (2) that it was unlawful for
EPA to rely on a toxicity study for dichlorvos (the Gledhill study)
that was conducted with humans. Because NRDC did not seek judicial
review on EPA's substantive conclusions on the latter issue but only
challenged EPA's denial of a hearing on the issue, and because the
Second Circuit court on review did not reach the hearing issue, the
Gledhill study is
[[Page 54411]]
further discussed only to the extent it bears on the FQPA safety factor
decision.
NRDC cited several grounds for its assertion that EPA unlawfully
lowered the 10X children's safety factor. However, only two of its
arguments were later raised in NRDC's judicial challenge to EPA's
decision. First, NRDC claimed that EPA lacked adequate data on
dichlorvos' potential effects on the endocrine system because EPA had
not received data on endocrine effects through the Endocrine Disruptor
Screening Program. Second, NRDC argued that EPA's choice of a 3X
additional safety factor was based on generic data and ``not [ ] on any
data specific to DDVP.'' (Ref. 1 at 5).
EPA denied both of NRDC's reasons for its objection to the choice
of a 3X FQPA factor. EPA rejected NRDC's endocrine data argument on
both legal and factual grounds. EPA concluded that the statute gave it
broad discretion to determine what data are needed in making a
determination on the FQPA safety factor and that nothing in section
408(p), creating the Endocrine Disruptor Screening Program, overrode
that broad discretion. As a factual matter, EPA found that it had
adequate data on endocrine effects from the existing dichlorvos
database. (73 FR 42697-42698).
EPA also rejected NRDC's claim that it relied on wholly generic
data, rather than dichlorvos-specific data, in choosing a 3X FQPA
factor. NRDC's argument here was that EPA chose 3X because EPA
considers 3X to be a half-value of a 10X factor rather than on data
pertaining to dichlorvos. In response, EPA noted that its petition
denial order had comprehensively restated its basis for its FQPA safety
factor decision, and that restatement focused in great detail on the
toxicology data for dichlorvos, particularly, the data on the
sensitivity of the young. (73 FR 42695). EPA further pointed out that
although the statutory considerations underlying the FQPA safety factor
generally supported removal of the 10X additional factor, the reason
EPA chose to retain a 3X FQPA safety factor for some assessments was
directly tied to a deficiency in a dichlorvos study (the Gledhill
study) that is critical to those assessments. (Id.). Thus, there was no
basis for NRDC's claim that EPA had not relied on dichlorvos-specific
data in making its FQPA safety factor decision.
VI. Judicial Review of EPA's Denial Order
A. NRDC's Petition for Judicial Review and the Matters Presented on
Review
NRDC petitioned the Second Circuit court for review of EPA's denial
of certain of its objections and hearing requests. As to its hearing
requests, NRDC argued that EPA improperly denied its request for a
hearing on statistical and informed consent issues presented by the
Gledhill study. As to its objections, NRDC asserted (1) that, as a
legal matter, EPA was required to retain the 10X FQPA factor if it did
not have data from the Endocrine Disruptor Screening Program; and (2)
that EPA's choice of a 3X FQPA factor was arbitrary and capricious
because EPA had relied upon ``generic assertions that unlawfully fail
to take into account any dichlorvos-specific information for infants
and children.'' (Ref. 28 at 37). NRDC supported the latter argument in
the following fashion. First, it argued that EPA chose 3X solely
because it was half of 10X. Second, NRDC asserted that EPA's
consideration of the Gledhill study did not constitute ``dichlorvos-
specific information for infants and children'' because the Gledhill
study was conducted with adults. Third, NRDC dismissed EPA's reliance
on dichlorvos developmental studies in animals on the ground that a
prior case had held that EPA had not, in that particular case, offered
an adequate explanation of how the data on developing animals supported
the FQPA factor chosen.
In response, EPA explained that NRDC's focus on EPA's discussion of
why 3X is considered half of 10X ignored the central part of EPA's
analysis: An assessment of whether the dichlorvos data showed 3X would
be safe. EPA responded to the claim of a failure to consider
``dichlorvos-specific information for infants and children'' by noting
that the Gledhill study had not been considered in isolation in the
decision on the FQPA safety factor but in the context of ``the animal
data showing no difference in adult-young sensitivity'' because it was
``that very data that shows why the Gledhill study is appropriate for
the entire population * * *'' (Ref. 29 at 63). Further, EPA noted that
NRDC's argument that EPA reliance on animal sensitivity data does not
justify a choice of 3X contradicted the core of NRDC's claim--that EPA
had not considered ``dichlorvos-specific information for infants and
children.'' (Id. at 62).
B. The Second Circuit Court's Decision on Review
On review, the Second Circuit court addressed three issues: (1) Was
EPA legally compelled to retain the 10X FQPA safety factor in the
absence of obtaining data from the Endocrine Disruptor Screening
Program; (2) did EPA adequately explain its decision on the FQPA safety
factor; and (3) was NRDC entitled to an evidentiary hearing with regard
to its claims regarding the alleged statistical and informed consent
deficiencies in the Gledhill study.
1. Endocrine data. The court held that EPA was not statutorily
required to retain the 10X FQPA factor in circumstances where it has
not obtained the data required under the Endocrine Disruptor Screening
Program. (658 F.3d at 219). The court found ``no indication in the
statute or legislative history that Congress * * * intended the
children's safety factor to be mandatory in assessing the risks of all
pesticides until EPA completed the estrogen disruptor screening program
* * *'' (Id.). According to the court, ``Congress allowed EPA to
determine, based on all available data, whether there was `reliable
data' supporting a reduced or waived children's safety factor * * *''
(Id.).
2. FQPA safety factor. Contrary to the narrow FQPA safety factor
issue presented to EPA in NRDC's objections--did EPA's decision on the
FQPA safety factor rely on ``a generic assertion [instead of being]
based on any data specific to DDVP''?--the court framed the issue on
the FQPA factor more broadly: ``NRDC now seeks review of that EPA
order, arguing in part that EPA failed to explain why, when assessing
the safety of dichlorvos for certain exposure scenarios, EPA did not
apply an additional tenfold children's safety factor, to account for
potential pre- and post-natal toxicity and completeness of data with
respect to exposure and toxicity to infants and children.'' (Id. at
201).
The court found that, for risk assessments relying on the Gledhill
study in deriving the Point of Departure, EPA had provided essentially
no explanation with regard to the FQPA safety factor. The court noted
that EPA had retained an additional 3X safety factor for these risk
assessments but the court concluded that it was EPA's express position
that this factor was not based on any evaluation of the risks to
infants and children but rather was intended to address the lack of
NOAEL in the Gledhill study only. According to the court, ``[i]n EPA's
IRED and two published orders, EPA consistently
[[Page 54412]]
reiterated this position and declined to claim that the 3X factor was
based on any evaluation of the risk to infants and children.'' (Id. at
216). Further, the court concluded that, unlike the risk assessments
that were not based on the Gledhill study, EPA did not rely on the
developmental animal studies showing no differential sensitivity
between adult and juvenile animals. According to the court, ``EPA
explicitly stated that it did not rely on any animal studies.'' (Id. at
217). The court thought this abnegation of reliance of animal studies
was confirmed by EPA's decision not to apply an interspecies factor to
the Gledhill-based assessments. (Id.). Although the court noted that
EPA called the 3X factor a FQPA factor, the court found that label to
be insufficient absent an explanation ``[i]n []either its IRED []or its
two orders [of] how the 3X factor was designed `to take into account
potential pre- and post-natal toxicity and completeness of the data
with respect to infants and children.' '' (Id.). The court held that
EPA's reasoning concerning the marginal effects seen at the LOAEL in
the Gledhill study did not constitute a sufficient explanation because
EPA did not relate that reasoning ``to `potential pre- and post-natal
toxicity and completeness of the data with respect to infants and
children.' '' (Id.). Finally, the court questioned EPA's analysis that
the effects at the LOAEL were marginal suggesting that EPA had not done
a proper statistical analysis. (Id. at 218).
Accordingly, the court concluded that, as to risk assessments that
used the Gledhill study to derive the Point of Departure, EPA's order
was arbitrary and capricious due to EPA's failure to provide an
adequate explanation with regard to its decision on the FQPA safety
factor. (Id.). Given this conclusion, the court vacated the aspect of
EPA's order pertaining to risk assessments based on the Gledhill study
and remanded the matter to EPA. (Id. at 220).
3. Evidentiary hearing. With regard to NRDC's request for an
evidentiary hearing on issues it raised concerning the Gledhill study,
the court determined that it did not need to resolve this question
given its disposition of the FQPA safety factor issue. As the court
pointed out, ``EPA may decide, on remand, not to rely on the Gledhill
study or to rely on the study in a different manner or for different
reasons.'' (Id. at 219).
VII. FQPA Safety Factor Determination for Gledhill-based Assessments
A. Introduction
FFDCA section 408(b)(2)(C) expressly requires EPA to apply a
default additional 10X safety factor for the protection of infants and
children unless EPA determines, based on reliable data, that a
different factor would be safe. Under the terms of the statute, this
additional safety factor is imposed ``to take into account potential
pre- and post-natal toxicity and completeness of the data with respect
to exposure and toxicity to infants and children.'' (21 U.S.C.
346a(b)(2)(C)). To implement these statutory commands, EPA has released
detailed guidance that advises EPA risk assessors in making decisions
on the FQPA safety factor to focus on potential pre- and post-natal
toxicity and the completeness of the toxicity and exposure databases.
In the dichlorvos IRED and the two orders responding to NRDC'S
dichlorvos petition, EPA devoted several pages to explaining how its
decision to apply a 3X FQPA safety factor complied with the statutory
directives on the FQPA safety factor and was consistent with its policy
guidance document. (See Ref. 3 at 128-132; 72 FR 68694-68695; 73 FR
42695-42696). From start to finish this discussion centered on the
issues of completeness of the toxicity and exposure databases for
dichlorvos and the potential increased sensitivity of infants and
children to dichlorvos from pre- and post-natal toxicity.
Nevertheless, in vacating, in part, EPA's dichlorvos order, the
Second Circuit court held that there was a complete absence of an
explanation from EPA as to how EPA's choice of a safety factor
protected infants and children. As the court repeatedly stated, ``EPA
did not explain why a children's safety factor less than 10X would
`take into account potential pre- and post-natal toxicity and
completeness of the data with respect to infants and children.' '' (658
F.3d at 217). In fact, the court rejected EPA's claim to have applied
any FQPA safety factor at all. According to the court, the additional
safety factor applied by EPA could not be considered a FQPA safety
factor given what the court viewed as EPA's denial that the additional
safety factor had anything to do with infants and children. (Id. at
211, 216).
Following a close review of EPA's prior explanations and the
court's opinion, EPA now recognizes that the discussion of the FQPA
safety factor in its dichlorvos IRED and orders was less than
transparent. EPA's explanation for its position on the FQPA safety
factor used, at times, a form of short-hand that hid rather than
elucidated its reasoning. In particular, EPA's short-hand appears to
have led the court to the following two misunderstandings: (1) That
EPA's use of a 3X safety factor to address the lack of a NOAEL in the
Gledhill study had nothing to do with the safety of infants and
children; and (2) that EPA did not consider the animal developmental
data in making a determination on the FQPA safety factor for
assessments relying on the Gledhill study. Clarification of EPA's
position on these two issues is critical to an understanding of EPA's
FQPA safety factor decision. Accordingly, on remand, EPA has first
addressed how the Gledhill-based assessments relate to protection of
infants and children and how EPA used animal developmental data in
these assessments. Only then does EPA offer its explanation as to how,
in light of the court's opinion, its choice of a FQPA safety factor for
the Gledhill-based risk assessment is protective of the safety of
infants and children, as required by FFDCA section 408(b)(2)(C).
B. Clarifications
1. Applying a FQPA safety factor to address the lack of a NOAEL in
the Gledhill Study. Numerous times in the IRED as well as its
dichlorvos orders, EPA stated that an additional 3X safety factor was
applied in risk assessments using the LOAEL in the Gledhill study as
the Point of Departure due to a ``lack of a NOAEL'' in the study. (Ref.
3 at 133; 658 F.3d at 217 (collecting cites)). EPA explained that the
safety factor was used to project a NOAEL for the study. The court
interpreted these statements as meaning the 3X factor had nothing to do
with the protection to infants and children. According to the court,
``EPA explained that the 3X factor [used in conjunction with the
Gledhill study] was not based on any risk to children or infants, but
accounted for EPA's `failure to identify a NOAEL in the [Gledhill]
study.' '' (Id. at 214). Certainly, the narrow issue addressed by the
use of the 3X factor was the lack of a NOAEL in the Gledhill study.
However, extrapolating a NOAEL through use of a safety factor is not an
end in itself. Rather, the safety factor was used to ensure that
dichlorvos risk assessments relying on the LOAEL in the Gledhill study
adequately protect the population groups covered by those assessments.
Importantly, the population groups covered by the Gledhill-based
assessments include infants and children. Thus, the 3X factor to
account for the lack of a NOAEL in the Gledhill study was critical to
[[Page 54413]]
protecting infants and children. However, EPA's orders and IRED failed
to make this linkage between the 3X factor and the safety of infants
and children clear. That linkage is fleshed out in detail below.
As discussed in Unit III.B.2.v., prior to the passage of FQPA, EPA
had applied an additional uncertainty factor to address a data
deficiency such as when adverse effects were seen in the lowest dose of
a toxicological study (i.e., when the study did not provide a NOAEL).
Such a factor is used to essentially extrapolate a NOAEL for the study.
Without an additional safety factor, there is uncertainty as to whether
reliance on the LOAEL as a Point of Departure in calculating a RfD/PAD
or MOE is adequately protective of the populations covered by the risk
assessment scenario relying on that RfD/PAD or MOE.
EPA has interpreted the FQPA as codifying this LOAEL-to-NOAEL
uncertainty factor as a FQPA safety factor when the factor is used in a
portion of a risk assessment (i.e., in a particular exposure scenario)
that assesses, at least in part, the risk to infants and children.
(Ref. 10 at 11-16, A-3--A-4). The logic here is straightforward. A
study that fails to produce a NOAEL is considered to be a data
deficiency that affects the completeness of the toxicity database. The
statute specifically references completeness of the toxicity database
as a reason for requiring an additional safety factor for the
protection of infants and children. Thus, when the LOAEL from a study
that lacks a NOAEL is chosen for the Point of Departure for a risk
assessment applying to infants, children, or women of child-bearing age
(for the purpose of protecting fetuses), the safety factor used to
address this data deficiency is a FQPA safety factor for the protection
of infants and children. This is the case whether or not the Point of
Departure is used for infants, children, or women of child-bearing age
only or for both adults and all other population groups, including
infants and children. Many risk assessments for particular exposure
scenarios use the same Point of Departure for both adults and infants
and children because frequently the relevant toxicity data show a lack
of differential sensitivity between adults and the young. However, use
in a risk assessment of the same Point of Departure for both adults and
the young does not make the FQPA safety factor provision inapposite.
EPA's position is that any assessment of risk for a particular exposure
scenario that includes, at least in part, an assessment of risks to
infants and children triggers the FQPA safety factor provision. Nothing
in section 408(b)(2)(C) limits the safety factor provision only to
situations where infants or children are more sensitive than adults.
For similar reasons, it is also irrelevant to application of the FQPA
safety factor provision whether the Point of Departure is from a study
involving juveniles or adults. Points of Departure for assessing risks
to infants and children are based on the studies showing the most
sensitive effects, whether the studies are conducted in adults or
juveniles. (See Ref. 17 at 452 (``[C]hronic and subchronic tests in
[adult animals] have value in assessing potential risks to children by,
for example, identifying target sites for toxicity and providing dose-
response information that may be useful for human safety assessment,
irrespective of life stage.''). The critical factor for the FQPA safety
factor provision is whether the study is being used for a Point of
Departure for assessing risk to infants and children.
With this background, the connection between the use of a 3X safety
factor to address the Gledhill study LOAEL and the protection of the
infants and children can now be explicated. Because the Gledhill study
produced cholinesterase effects at the lowest level in the subchronic
studies in the dichlorvos database and the database showed no age-
related sensitivity, (see discussion in Unit VII.C.), EPA chose the
Gledhill LOAEL as the Point of Departure for assessing risks for short-
and intermediate-term exposure scenarios to all population groups,
including infants and children. In other words, the Gledhill LOAEL was
selected as the Point of Departure for all population groups for these
exposure scenarios because the dichlorvos database demonstrated that
the Gledhill study not only provided the best measure of cholinesterase
inhibition for protecting adults but that it was the best measure for
protecting infants and children. Nonetheless, EPA also recognized that
the data deficiency in the Gledhill study--the failure of the Gledhill
study to identify a NOAEL--raises uncertainty as to what that study
indicates regarding the threshold below which exposure to dichlorvos
will not result in cholinesterase inhibition. To address this
uncertainty and thus protect the safety of all population groups
covered by the risk assessments, including infants and children, EPA
chose to apply an additional safety factor of 3X. This choice of a
safety factor was made under the rubric of the FQPA safety factor
provision because the uncertainty raised by reliance on a LOAEL both
(1) affected the assessment of the risk to infants and children; and
(2) was driven by a data deficiency affecting the completeness of the
toxicity database. (73 FR 42695; 72 FR 68694-68695; Ref. 3 at 133,
134). Thus, the additional 3X safety factor used in assessments relying
on the Gledhill study was not simply to address the lack of a NOAEL in
that study but rather to ensure the protection of infants and children
(among others) given that a LOAEL was used as the Point of Departure
for assessing risk to infants and children for several exposure
scenarios. Regrettably, the connection between a safety factor used to
address the lack of a NOAEL in a study in adults and the protection of
infants and children was not transparent in EPA's IRED or its denial of
NRDC's petition and objections. That linkage should now be clear.
2. Reliance on animal developmental data. EPA's FQPA safety factor
policy emphasizes the importance of considering the ``weight-of-
evidence analyses for the completeness of the toxicity database, the
degree of concern for pre- and postnatal toxicity, and results of the
exposure assessments'' in making a safety factor determination. (Ref.
10 at 50). In particular, the policy stresses ``taking into account all
pertinent information in evaluating potential pre- and postnatal
toxicity.'' (Id. at 29). The policy recognizes that human data on pre-
and postnatal toxicity is ``difficult to obtain'' and for that reason
discusses, in detail, how animal developmental data should be
considered in evaluating the potential for pre- and post-natal toxicity
in humans. (Id. at 28-31). Although EPA did discuss the animal data on
juvenile sensitivity in its FQPA safety factor determination, (72 FR
68694-68695), the court concluded that EPA had not considered that data
in making a determination on the FQPA safety factor for assessments
relying on the Gledhill study for the Point of Departure.
To support this conclusion, the court opined that EPA's orders
specifically referenced the animal developmental studies in conjunction
with the safety factor determination for the non-Gledhill-based
assessments but had not done so as to the Gledhill-based assessments.
The court is correct that EPA did not clearly explain that its
discussion of the animal developmental data related both to the
assessments based on a Point of Departure from animal data as well as
the assessments relying on the Gledhill study for the Point of
Departure. EPA's discussion of
[[Page 54414]]
the Gledhill study, and the data deficiency therein, followed the
analysis of the animal developmental data but did not directly
reference that data or the statutory considerations bearing on the FQPA
safety factor decision. (Id.). To avoid this error in its revised
safety factor finding below, EPA has included a discussion of the data
deficiency in the Gledhill study under the topic of ``completeness of
the data with respect to * * * toxicity'' and also explicitly discussed
how the statutory consideration pertaining to the potential for pre- or
post-natal toxicity, and the animal data bearing on this issue, was
considered in the context of the Gledhill-based assessments.
The court also concluded that ``EPA explicitly stated that it did
not rely on any animal studies'' in connection with the Gledhill-based
assessments, (658 F.3d at 217), citing to language in the IRED that
specified that where the Point of Departure was chosen from the
Gledhill study ``there was no need to account for interspecies
extrapolation * * * [s]ince the study was conducted in human
subjects.'' (Ref. 3 at 133, 134). According to the court, ``[w]hen EPA
did rely on the animal studies * * * [it] properly applied a safety
factor of `10X for interspecies differences.' '' (658 F.23d at 217).
The court appears to have drawn the conclusion that the interspecies
factor should be applied whenever EPA considers animal studies in any
aspect of the risk assessment. Thus, the court reasoned that because
EPA did not apply an interspecies factor for the Gledhill-based
assessments, it could not have considered the animal developmental data
in the FQPA safety factor determination for dichlorvos.
The court has misapprehended the reason EPA uses an interspecies
factor in risk assessments. The factor is not automatically applied
whenever animal data are considered in any aspect of a risk assessment.
Rather, as explained in Unit III.B.2., the interspecies factor is used
when extrapolating from a dose in an animal study (generally a NOAEL or
LOAEL) on a milligram-per-kilogram of body weight basis to a dose in
humans. (See Ref. 10 at 10 (an interspecies factor is used ``if animal
data have been used as the basis for deriving the hazard values''). The
interspecies factor is designed to account for possible toxicokinetic
and toxicodynamic differences in humans and laboratory animals that may
result in differences in internal dose and organ sensitivity between
humans and animals. Thus, in the dichlorvos animal assessments in which
EPA relied on animal data for the Point of Departure, EPA did apply an
interspecies factor. For those assessments, EPA was either
extrapolating a RfD for humans from animal data or comparing the margin
between human exposure and the dose in animals that was judged to be a
NOAEL. No interspecies factor was necessary in assessments based on the
LOAEL from the Gledhill study because EPA was not extrapolating from a
NOAEL or LOAEL in laboratory animals to humans or comparing human
exposure to a dose from an animal study. Rather, EPA had data in
humans--the Gledhill study--and was relying on that data for the Point
of Departure. There was no need to account for the toxicokinetic and
toxicodynamics differences between humans and animals when deriving a
safe dose for humans from a study conducted with humans.
EPA, however, did rely on the animal developmental data in the FQPA
safety factor determination for the Gledhill-based assessments. But
that reliance was for a purpose distinct and separate from use of the
data for extrapolating a dose from animals to humans. In accordance
with Agency FQPA safety factor policy, EPA considered the dichlorvos
animal developmental data with regard to the important information it
provides on whether the 10X intraspecies factor for dichlorvos is
protective of infants and children. (Ref. 10 at 29). A primary focus of
the animal developmental data (the rat and rabbit developmental
studies, the rat reproduction study, the rat developmental
neurotoxicity study, and comparative cholinesterase studies) is on the
relative sensitivity of adult and juvenile animals. Because EPA would
rarely have data on the relative sensitivity among different age groups
of humans to a pesticide, these animal data help inform, as EPA policy
makes clear, whether the 10X intraspecies factor is sufficiently
protective of infants and children. (Id.).
Considering animal developmental data in evaluating the
intraspecies factor is a standard part of EPA's risk assessment
process. As discussed in Unit III.B.2 and above, animal developmental
data are central both to establishing the justification for the 10X
default value for the intraspecies factor and for evaluating the
protectiveness of this default value for specific chemicals. Although
broad-based surveys of data on adult/juvenile sensitivity in both
humans and animals generally support the use of a 10X default value for
the intraspecies factor, there is wide recognition that the possibility
of heightened sensitivity in infants and children warrants obtaining
particularized data on juvenile/adult animal sensitivity for individual
chemical risk assessments. When these data are available, they may
indicate that there is no heightened concern warranting an additional
safety factor or that an additional factor is necessary above and
beyond the default 10X value for the intraspecies factor. In a few
cases, EPA has even relied, at least in part, on animal data as
supporting a reduction in the default 10X intraspecies factor.
Yet, despite the centrality of animal data to the justification for
and selection of the intraspecies factor, EPA is not aware of any
instance where an interspecies factor has been applied solely for
reliance on animal data on adult-juvenile sensitivity to evaluate the
protectiveness of the human intraspecies factor. For example, EPA's
long-established and consistent practice is not to apply an
interspecies factor when relying on a human study for the Point of
Departure even though a decision on the intraspecies factor is still an
essential part of such assessments. Dourson et al. collected a summary
of all EPA's RfDs on EPA's Integrated Risk Information System (IRIS) as
of May 2000 that used human data for the Point of Departure. (Ref. 17).
All 24 such assessments identified used an interspecies factor of 1X
(i.e., no factor). EPA has identified 9 additional such risk
assessments on IRIS post-dating May 2000, and each one of those also
does not apply an interspecies factor. (Ref. 30). Even more on point
are EPA pesticide risk assessments relying on human data. Since the
promulgation of the 2006 Human Research Rule, EPA has accepted 10 human
studies for use in pesticide risk assessments other than the Gledhill
study. (Id.). A Point of Departure was selected from 9 of those 10
studies.\1\ Yet, in none of those assessments did EPA apply an
interspecies factor in conjunction with a Point of Departure from a
human study even though the assessments do not focus on the human data
exclusively. Animal developmental data play a critical part in these
assessments, particularly where a FQPA safety factor analysis is
required.
---------------------------------------------------------------------------
\1\ The one human study that was not used for selection of a
Point of Departure was conducted with the pesticide oxamyl. The
oxamyl human study was submitted for the purpose of justifying a
reduction of the 10X interspecies factor despite use of an animal
study for the Point of Departure. The Human Studies Review Board
concluded that the ``intentional human dosing study of oxamyl was
sufficiently robust to be used for reducing the 10x inter-species
(i.e. animal to human) uncertainty factor in the cumulative risk
assessment for the N-methyl carbamates.'' (Ref. 36 at 28). Thus, it
is not even a given that a full interspecies factor will be applied
when an animal study is relied upon to extrapolate a dose in humans.
---------------------------------------------------------------------------
[[Page 54415]]
The FQPA safety factor analysis in the tolerance reassessment
document for the pesticide ethephon provides a good example of this.
With ethephon, ``[t]he conventional UF of 10X for interspecies
extrapolation was not applied because the endpoint selected for the
risk assessment was from a human study.'' (Ref. 31 at 6). At the same
---------------------------------------------------------------------------
time, EPA noted that:
The Agency concluded that no FQPA Safety Factor is necessary to
protect the safety of infants and children in assessing ethephon
exposure and risks because the toxicology database for ethephon
contains acceptable guideline developmental and reproductive studies
as well as acute and subchronic neurotoxicity studies. [Guideline
studies are conducted in animals. (40 CFR 158.500)]. The Agency also
concluded that there is no quantitative or qualitative evidence of
increased susceptibility following in utero or postnatal exposure in
any of the developmental or reproductive studies. The RfDs and
toxicity endpoints established are protective of pre/postnatal
toxicity following acute and chronic exposures.
(Id.). A variation on the approach in ethephon is the safety/
uncertainty factors chosen in assessing the risk of the pesticide
methomyl. (Ref. 32 at 5). For the methomyl risk assessments that relied
on a human study for the Point of Departure, the Agency applied a 10X
intraspecies, a 1X interspecies factor (no extrapolation from a dose in
animals to humans), and a 2X (data-derived) FQPA safety factor. The 2X
FQPA factor was chosen because, unlike dichlorvos, the adult/juvenile
comparative cholinesterase data in rats showed that juveniles were
approximately twice as sensitive to methomyl as adults. Thus, a 2X FQPA
safety factor was applied to ensure that the 10X intraspecies factor
was sufficiently protective. However, just as with dichlorvos and
ethephon, no interspecies factor (1X) was used because the Point of
Departure was derived from a human, not animal, study. A final example
illustrating that consideration of animal data in conjunction with
choice of a Point of Departure from a human study does not result in
use of a 10X interspecies factor is the assessment of the pesticide
chloropicrin. With chloropicrin, EPA relied upon a human study for the
Point of Departure and thus no interspecies factor (1X) was applied.
However, EPA's consideration of the data from humans and animals also
led EPA to conclude that no intraspecies factor (1X) was needed either.
(Ref. 33). No interspecies factor was applied as a result of
consideration of animal data in evaluating the need for an intraspecies
factor.
Use of a 10X interspecies factor for reliance on animal
developmental data to evaluate the protectiveness of the intraspecies
factor would also lead to illogical results. For example, animal
developmental data are now considered so critical to evaluating pre-
and post-natal toxicity that the FQPA imposes a presumptive 10X safety
factor in their absence. Yet, once the data are submitted, it does not
make sense to replace the 10X safety factor that addressed their
absence with a safety factor of equivalent value to address their mere
use for evaluation of pre- and post-natal toxicity. Leaving aside what
the animal developmental data show, there cannot be equal need for
safety factors both in the absence and presence of adequate animal
developmental data.
In sum, it would not only be unprecedented, but inconsistent with
well-established safety factor practice, to suggest that the mere
consideration of animal data in evaluating the protectiveness of the
intraspecies factor triggers application of an interspecies factor.
Importantly, under the FFDCA section 408, EPA is only authorized to
consider ``safety factors which in the opinion of experts qualified by
scientific training and experience to evaluate the safety of food
additives are generally recognized as appropriate for the use of animal
experimentation data.'' 21 U.S.C. 346a(b)(2)(D)(ix).
Unfortunately, EPA's short-hand description of its FQPA
determination misled the court regarding EPA's consideration of the
animal developmental data. Further, EPA's brief explanation for why it
did not apply an interspecies factor did not clarify the situation.
This, in turn, resulted in confusion regarding the role of the
interspecies factor. EPA's revised FQPA safety factor explanation
attempts to avoid such pitfalls.
C. Revised FQPA Safety Factor Decision
1. Introduction and background. The Second Circuit court has
vacated that portion of EPA's order on NRDC's objections ``assessing
the risk of dichlorvos based on the Gledhill study * * * .'' (658 F.3d
at 220). The court found that EPA had ``failed to explain why it did
not use a 10X children's safety factor'' for those assessments. (Id.).
In the IRED, EPA relied on the Gledhill human study for selection
of the Point of Departure for assessing dermal (short-, intermediate-,
and long-term), incidental oral (short-term), and inhalation (short-
and intermediate-term) risk for all population subgroups, including
infants and children. Agency-wide guidance on Reference Dose selection
emphasizes that human data provides the best source for assessing human
risk: ``Adequate human data are the most relevant for assessing risks
to humans. When sufficient human data are available to describe the
exposure-response relationship for an adverse outcome(s) that is judged
to be the most sensitive effect(s), reference values should be based on
human data.'' (Ref. 19 at 4-12; see Ref. 10 at 33 (``human data are the
most relevant data for assessing health risks'')). EPA chose the
Gledhill study, in particular, for determination of the Point of
Departure because it evaluated cholinesterase inhibition, the most
sensitive effect for dichlorvos as shown by animals studies, and
because the Gledhill study has ``the lowest LOAEL established for RBC
cholinesterase inhibition in a repeated oral exposure to dichlorvos.''
(Ref. 3 at 133). Specifically, it was the lowest LOAEL considering both
the human and animal studies and cholinesterase effects in adults and
juveniles. EPA's determination that the Gledhill study ``is
sufficiently robust for developing a Point of Departure for estimating
dermal, incidental oral, and inhalation risk from exposure to DDVP,''
was concurred in by the Human Studies Review Board, an independent
expert panel of scientists. (72 FR 68675).
The level of concern for the risk assessments relying on the
Gledhill study for the Point of Departure was expressed in terms of a
target MOE of 30. That value was based on an intraspecies uncertainty
factor of 10X and a FQPA safety factor of 3X. Although EPA concluded
that neither the data on pre- or postnatal toxicity or on exposure to
dichlorvos showed a need for a FQPA safety factor, EPA found that the
data deficiency with regard to the Gledhill study--namely, its lack of
a NOAEL--justified the retention of a 3X FQPA safety factor.
2. FQPA safety factor decision. In making a FQPA safety factor
determination, EPA follows a weight-of-the-evidence approach that
focuses on the three considerations explicitly noted in FFDCA section
408(b)(2)(C): the completeness of the toxicity database; the potential
for pre- and post-natal toxicity; and the completeness of the exposure
database. (Ref. 10 at iv). Each of those considerations is discussed
below.
i. Completeness of the toxicity database. In ruling on NRDC's
petition, EPA concluded that it had a complete toxicity database under
the pesticide data requirements in 40 CFR part 158. This included all
required data specifically pertaining to effects on the young--
developmental studies in two
[[Page 54416]]
species (rat and rabbit); a two-generation reproduction study in rats;
and a developmental neurotoxicity study in rats. EPA also had
comparative cholinesterase inhibition data in adult and juvenile rats.
EPA did not have data submitted pursuant to the Endocrine Disruptor
Screening Program, but for the reasons explained in its order denying
NRDC's petition, EPA has concluded that it has adequate data on
dichlorvos' endocrine effects for the purposes of its FQPA safety
factor decision. (73 FR 42697-42698).
In addition to these standard animal toxicity studies, the
dichlorvos registrant had submitted one toxicity study in humans, the
Gledhill study, that EPA had determined was in compliance with its
Human Research Rule. (40 CFR part 26). As discussed below, there is a
data deficiency issue with this study that is pertinent to the
completeness of the toxicity database consideration. Although this
study was conducted in adults, it is highly relevant to the protection
of infants and children because EPA has, for the reasons explained in
Units VII.B.1. and VII.C.1, selected the Gledhill study for identifying
a Point of Departure for as to several risk assessment scenarios for
all population groups, including infants and children. Thus, how EPA
addresses the data deficiency in the Gledhill study will directly
affect how it assesses risks to infants and children.
The Gledhill study was a repeat dose study measuring RBC
cholinesterase inhibition in control and dichlorvos- treated human
subjects. Only a single dose level (7 mg) was used in the study.
Cholinesterase inhibition in the treated subjects reached a level of 16
percent by day 18 of treatment (i.e., cholinesterase activity levels
declined to 84 percent of the pre-dose mean by day 18). As shown in
Table 2 below (reprinted from EPA's Data Evaluation Record of the
Gledhill study and the Gledhill study report), the statistical analysis
of the results of the Gledhill study shows a high level of statistical
significance (at the 1 percent level) \2\ for cholinesterase activity
levels both between controls and treated subjects and between pre- and
post-dosing cholinesterase levels for treated subjects for most days
post-dosing.
---------------------------------------------------------------------------
\2\ Statistical significance is a term used to describe observed
data that differ from the overall distribution of values by a level
that is unlikely to be due to random error. Statistical significance
is examined in terms of the probability of the observed differences
occurring. By convention, observed values that have a 5 or 1 percent
chance of occurring are treated as statistically significant, with 1
percent being the more rigorous standard. (Ref. 43).
Table 2--Results of the Gledhill study
--------------------------------------------------------------------------------------------------------------------------------------------------------
Placebo (n = 3) Dosed (n = 6)
----------------------------------------------------------------------------------------------------------------
Timepoint % pre-dose
Mean SD mean Mean SD % pre-dose mean
--------------------------------------------------------------------------------------------------------------------------------------------------------
Pre-dose............................... 18483.52 1346.91 100 17738.33 1713.50 100
Day 1.................................. 17930.00 1404.24 97 17628.33 1914.45 99
Day 2.................................. 18180.00 1564.7 98 16816.67* 1546.63 95
Day 4.................................. 18740.00 1771.13 101 16933.33** 1597.33 95
Day 7.................................. 18530.00 1888.36 100 16181.67** [dagger][dagger] 1759.48 91
Day 9.................................. 18460 1007.03 100 16708.33 2504.97 94
Day 11................................. 19210.00 1035.95 104 16036.67** [dagger][dagger] 1654.38 90
Day 14................................. 18490.00 1642.35 100 15333.33** [dagger][dagger] 1250.34 86
Day 16................................. 17706.67 2470.15 96 15191.67** [dagger][dagger] 1062.59 86
Day 18................................. 18260.00 2298.87 99 14855.00** [dagger][dagger] 1198.51 84
--------------------------------------------------------------------------------------------------------------------------------------------------------
* Statistically significant difference from pre-dose at the 5% level (paired t-test).
** Statistically significant difference from pre-dose at the 1% level (paired t-test).
[dagger][dagger] Statistically significant difference between placebo and dose groups at the 1% level (t-test, based on repeated measures of analysis of
covariance).
(Refs. 34 and 35).
EPA found these statistical results to be sufficiently ``robust''
to support use of the Gledhill study as the Point of Departure. This
judgment was concurred on by the Human Studies Review Board. (Ref. 36).
The Board relied upon the following aspects of the study: The repeated
dose approach which allowed examination of the sustained nature of RBC
cholinesterase inhibition; robust analysis of RBC cholinesterase
inhibition both in terms of identifying pre-treatment levels and
consistency of response within and between subjects; and the
observation of a low, but statistically significant RBC cholinesterase
inhibition response. (Id. at 39). The HSRB concluded that ``[a]lthough
a study using a single dose level is not ideal for establishing a
LOAEL, there was general consensus that RBC cholinesterase is a well-
characterized endpoint for compounds that inhibit acetylcholinesterase
activity and therefore, because the decreased activity in RBC
cholinesterase activity observed in this study was at or near the limit
of what could be distinguished from baseline values, it was unlikely
that a lower dose would produce a measurable effect in RBC
cholinesterase activity.'' (Id. at 41).
There is one significant deficiency with the Gledhill study,
however. Because the study used a single dose level, and that dose was
found to cause an adverse effect on RBC cholinesterase activity, the
study does not identify a NOAEL. As discussed earlier, this type of
deficiency is incorporated and addressed as part of the FQPA safety
factor because it relates to the first consideration noted in FFDCA
section 408(b)(2)(C)--completeness of the toxicity database. (See Unit
III.B.2.vi.).
In deciding what level of safety factor is necessary to address
this data deficiency, EPA is guided by EPA science policy on use of
uncertainty factors, the scientific literature on safety factors, and
EPA prior practice with regard to FQPA safety factor decisions. EPA's
RfD policy recommends a default value of 10X for an uncertainty factor
addressing the lack of a NOAEL but makes clear that ``[t]he size of the
LOAEL-to-NOAEL uncertainty factor may be altered, depending on the
magnitude and nature of the response at the LOAEL.'' (Ref. 19 at 4-44).
Further, as discussed in Unit III.B.2.v, Dourson et al. concluded that
``[t]he data indicate that when faced with a LOAEL and not a NOAEL, the
choice of uncertainty factor should generally depend on the
[[Page 54417]]
severity of the effect at the LOAEL.'' (Ref. 9). In specific FQPA
safety factor decisions, the magnitude of the response has frequently
been an important consideration supporting use of a 3X FQPA safety
factor to address reliance on a LOAEL for the Point of Departure. (See,
e.g., 75 FR 22245, 22249, April 28, 2010 (selecting a 3X FQPA safety
factor for lack of a NOAEL where ``[t]he neurotoxic effects in this
study showed a good dose response which resulted in minimal effects on
motor activity and locomotor activity at the LOAEL.''); 74 FR 67090,
67094, December 18, 2009 (selecting a 3X FQPA safety factor for lack of
a NOAEL where ``[t]he gastric lesions (most sensitive effect) are due
to the direct irritant properties of endothall (i.e., portal effects)
and not as a result of frank systemic toxicity; the severity of the
lesions were minimal to mild; and there was no apparent dose-response
for this effect.''); 74 FR 53172, 53177, October 16, 2009 (``The
concern is low for the use of a LOAEL to extrapolate a NOAEL, given the
relatively insignificant nature of the effect (transient diarrhea seen
in the rat); the fact that diarrhea was only seen in studies involving
gavage dosing in the rat but not in repeat dosing through dietary
administration in rats, mice, rabbits, and dogs; the very high dose
level needed to reach the acute oral lethal dose (LD)50
(>5,000 milligrams/kilogram (mg/kg)), and the overall low toxicity of
azoxystrobin.''); 74 FR 26536, 26541, June 3, 2009 (selecting a 3X FQPA
safety factor for lack of a NOAEL where ``[t]he response was marginal
at the LOAEL.''); 72 FR 41224, 41228, July 27, 2007 (``The uncertainty
factor of 3X for use of the LOAEL instead of the NOAEL is considered
appropriate because an increased incidence and severity of epithelial
hyperplasia, hyperkeratosis and ulceration of the non-glandular region
of the stomach in females were seen in few animals and were minimal in
severity and observed in one sex only.''); 72 FR 33901, 33905, June 20,
2007 (``The 3X factor is considered to be protective because the
incidence of the effects at the lowest dose tested was only marginally
higher than the historical controls.''); 71 FR 71052, 71056, December
8, 2006 (``A 3x safety factor (as opposed to a 10x) for the lack of a
NOAEL in this critical study is adequate because the magnitude of the
response was low (low incidences without dose response) and the effect
of concern was seen in an unusual strain (Chinchilla) of rabbits and
not in the New Zealand strain commonly used in developmental toxicity
studies.'')).
EPA's policy on cholinesterase inhibition provides important
guidance on characterizing the magnitude of a RBC cholinesterase
finding. The policy explains that cholinesterase activity data is
treated ``like most continuous endpoints (i.e., graded measures of
response such as changes in organ weight, hormone levels or enzyme
activity),'' in that ``no fixed generic percentage of change from the
baseline is considered to separate adverse from non-adverse effects.''
(Ref. 27 at 14). Given the continuous nature of the inhibition
response, ``OPP has used statistical significance, rather than a fixed
percentage of response from baseline, as the primary, but not
exclusive, determinant of toxicological and biological significance in
selecting Points of Departure.'' (Id.) Nonetheless, the policy advises
that, in examining what level of cholinesterase inhibition will be
judged an adverse effect, the level of inhibition must be critically
evaluated ``in the context of both statistical and biological
significance.'' (Id. at 37) (emphasis in original). Although the policy
notes that ``[n]o fixed percentage of change (e.g., 20% for
cholinesterase enzyme inhibition) is predetermined to separate adverse
from non-adverse effects,'' (Id.), it explains that ``OPP's experience
with the review of toxicity studies with cholinesterase-inhibiting
substances shows that differences between pre- and post-exposure of 20%
or more in enzyme levels is nearly always statistically significant and
would generally be viewed as biologically significant.'' (Id. at 37-
38). The policy recommends that ``[t]he biological significance of
statistically-significant changes of less than 20% would have to be
judged on a case-by-case basis, noting, in particular the pattern of
changes in the enzyme levels and the presence or absence of
accompanying clinical signs and/or symptoms.'' (Id. at 38). The policy
notes that similar or higher levels of cholinesterase inhibition are
used ``in monitoring workers for occupational exposures (even in the
absence of signs, symptoms, or other behavioral effects).'' (Id. at
31). For example, the policy points out that the California Department
of Health Services requires that workers exposed to toxic chemicals
such as organophosphate pesticides be removed from the workplace if
``red blood cell cholinesterase levels show 30% or greater
inhibition,'' and that the World Health Organization ``has guidelines
with the same RBC action levels (i.e., 30% or greater inhibition).''
(Id.). In conducting Benchmark Dose analyses for dichlorvos, as well as
other organophosphate pesticides, EPA generally has used a 10 percent
inhibition level as indicating an adverse effect for both RBC and brain
compartments given that both of these compartments were used for
developing Points of Departure. (Ref. 37 at I.B p.17). A close
examination of the cholinesterase inhibition data for dichlorvos,
however, has shown that, while both brain and RBC compartments have
similar levels of inhibition for acute or very short-term exposures,
for longer-term exposures brain cholinesterase inhibition is much less
sensitive than RBC inhibition and thus 20 percent RBC inhibition would
be adequately protective. (72 FR 68691; Ref. 38). RBC cholinesterase
inhibition is not itself an adverse effect; rather, it is used as a
surrogate for effects on the nervous system.
In the Gledhill study, the average level of RBC cholinesterase
inhibition of the final day of treatment was 16 percent. Although the
level of RBC cholinesterase inhibition was relatively low and not
accompanied by clinical signs, EPA concluded, contrary to the study's
author, that the 7 mg dose did produce an adverse effect. In reaching
this conclusion, EPA relied on the uniform nature of the results in the
subjects that showed a clear pattern of increasing response over time
and a high level of statistical significance in the differences in
cholinesterase inhibition both between treated and control subjects and
between pre-treatment and post-treatment of individual subjects.
Nonetheless, consistent with its cholinesterase policy and its
conclusions in regard to other dichlorvos cholinesterase data, EPA
found the magnitude of the change in cholinesterase levels to be
marginal. The Human Studies Review Board agreed both with EPA's
determination on adversity and the marginality of the response. As to
the marginality of the response, the Board specifically noted that
``because the decreased activity in RBC cholinesterase activity
observed in this study was at or near the limit of what could be
distinguished from baseline values, it was unlikely that a lower dose
would produce a measurable effect in RBC cholinesterase activity.''
(Ref. 36 at 41). Under EPA's cholinesterase policy, the level of
cholinesterase inhibition in the Gledhill study falls at the low end of
the scale of what might be considered an adverse effect and the policy
recommends a case-by-case inquiry into the adversity determination for
inhibition at this
[[Page 54418]]
level. Accordingly, EPA determined previously, and reaffirms in this
order, that a full 10X safety factor is not needed to address the lack
of a NOAEL in the Gledhill study. When a full order of magnitude of
additional protection (i.e. 10\1\) is unnecessary, EPA will generally
use a half of that value (i.e, 10\.5\ or approximately 3X) if that
value is protective. Here, EPA determined, and in this order reaffirms,
that the marginal nature of the cholinesterase response shows that a 3X
factor is safe.
In reaching its determination, EPA placed, and continues to place,
great weight on the view of the Human Studies Review Board. This Board
was created by EPA in response to a congressional mandate. (71 FR 6138
(February 6, 2006)). It is comprised of non-EPA scientists,
overwhelmingly from academia, who are specialists in the field of
bioethics, biostatistics, human health risk assessment, and human
toxicology. (73 FR 42690). The members of the Board at the time the
Gledhill study was considered are listed in Appendix 1 to EPA's prior
denial order. (73 FR 42713). The Board is charged with reviewing both
the ethics and scientific merit of intentional exposure human studies.
Its proceedings are conducted in public and it accepted three rounds of
public comment on review of the Gledhill study: (1) Written comment
submitted prior to its open meeting on dichlorvos; (2) oral comments at
the open meeting; and (3) oral comments at a telephone conference on
its proposed decision. (73 FR 42692). No comments were submitted prior
to the Board's review suggesting that the cholinesterase response was
greater than a marginal response and no meaningful comments were
submitted to the Board or EPA, following release of the proposed and
final Board opinions, contesting the conclusions of this independent
and expert scientific panel on this point. The Board's conclusion with
regard to the marginality of the cholinesterase inhibition effects in
the Gledhill study are strongly supportive of EPA's choice of a 3X
factor to address the lack of a NOAEL in the Gledhill study. After all,
the Board concluded that ``it was unlikely that a lower dose would
produce a measurable effect in RBC cholinesterase activity.'' (Ref. 36
at 41). Use of a 3X factor is protective because it represents a choice
of not simply of any lower dose (decreasing the dose by 10 percent fits
this criterion) but of a significantly lower dose than that in the
Gledhill study for estimating risk (by applying a 3X factor EPA was
essentially dividing the dose by a factor of 3).
The court suggested in its opinion that EPA had not conducted an
adequate statistical analysis to determine the accuracy of the 16
percent cholinesterase inhibition figure and thus had no basis for
making a conclusion ``with any level of precision [as to] the magnitude
of the cholinesterase inhibition.'' \3\ 658 F.3d at 218. Although EPA
scientists and the scientists on the Human Studies Review Board,
including the three biostatisticians, found the statistical analysis
sufficient to support their conclusion on the marginality of the
cholinesterase effect, EPA agrees that a precision analysis, i.e., the
calculation of confidence intervals, conveys valuable information on
the plausible range in which, within a certain degree of probability,
the true value lies. Accordingly, EPA has calculated the confidence
intervals for the mean cholinesterase inhibition levels. (Ref. 39). For
the days 14, 16, and 18 which had average cholinesterase inhibition
levels of 14 percent, 14 percent, and 16 percent, respectively, this
calculation shows a 95 percent confidence that average inhibition is
between 9- and 18 percent, 9- and 19 percent, and 8- and 24 percent,
respectively. Because these ranges of RBC cholinesterase inhibition
consistently fall at the low end of what might be found to be a
statistically and biologically significant effect on RBC cholinesterase
activity, EPA reaffirms its conclusion that the RBC cholinesterase
inhibition seen in the Gledhill study was marginal.
---------------------------------------------------------------------------
\3\ The court stated that EPA had found the Gledhill study to
``have had sufficient statistical power to detect a cholinesterase
inhibition greater than 0, [but] EPA did not explain whether the 9-
person study (six dosed subjects, 3 placebo subjects) had sufficient
power to determine with any level of precision the magnitude of the
cholinesterase inhibition.'' (Ref. at 218) (emphasis added). To
clarify, EPA did not do a ``statistical power'' calculation because
statistical power is a way of determining the probability of whether
a study would detect an effect of a given size if such an effect is
there to find. The concern is that a study may indicate that there
is no effect when, in fact, the study missed the effect because it
had a low probability of finding it (i.e., the study gives a false
negative). Because the Gledhill study identified the positive effect
it was looking for (cholinesterase inhibition), EPA dismissed NRDC's
arguments regarding statistical power as irrelevant. (73 FR 42704-
42706). What EPA's statistical analysis of the Gledhill study did
show was that there was a statistically significant difference (at
the level of 1 percent) in cholinesterase inhibition between control
and treated subjects and between pre- and post-dosing for treated
subjects on most days of treatment. That is, the differences in
cholinesterase inhibition between controlled and treated subjects
and between pre- and post-dosing of treated subjects were very
unlikely to have been due to chance.
---------------------------------------------------------------------------
Finally, the determination to retain a FQPA safety of 3X for
assessments for which the Point of Departure was selected from the
Gledhill study is also supported by two BMD analyses on the dose levels
causing cholinesterase inhibition in animals performed in conjunction
with the IRED. As explained earlier, BMD analysis is preferred by EPA
to the NOAEL/LOAEL approach of selecting a Point of Departure from
studies because all of the data from a study can be used in deriving a
dose response curve. (Ref. 23). In the absence of the Gledhill study,
these analyses would substitute for the LOAEL in the Gledhill study for
selection of the Point of Departure for short- and intermediate-term
risk assessments because they define the most sensitive effect for
these exposure durations. The first of these analyses is a BMD analysis
of comparative cholinesterase studies conducted in adult and juvenile
rats. (This BMD analysis is discussed in more detail immediately below
in the section on ``pre- and post-natal toxicity.'') The lowest BMDL
from that analysis (focusing on pooled historical controls) is 0.38 mg/
kg/day. (Ref. 42). The second BMD analysis is an analysis of the
cholinesterase inhibition results of the subchronic toxicity rat study.
(Ref. 40). There, the BMDL was calculated as 0.4 mg/kg/day. The only
other potential animal study for use in selecting a Point of Departure
for short- and intermediate-term exposures, the subchronic
neurotoxicity study, had a significantly higher LOAEL (7.5 mg/kg/day)
and produced percentage inhibition levels consistent with, or lower
than, the other animal cholinesterase studies. (Ref. 41). A 100X safety
factor to address interspecies extrapolation and interspecies
variability would be used with these BMDLs if they were chosen as
Points of Departure. No additional FQPA factor would be needed for the
same reasons that a FQPA factor was not applied to the other
assessments relying on animal data. (72 FR 68694-68695). Reliance on
the BMD analyses for the Point of Departure with a 100X safety factor
produces a level of concern that is comparable to using the Gledhill
study for the Point of Departure with a 30X safety factor. This is most
easily seen if alternative RfD/PADs are calculated using the BMD
analyses from the comparative cholinesterase studies and the subchronic
study and from the LOAEL in the Gledhill study. With Gledhill study,
the LOAEL of 0.1 mg/kg/day would be divided by 30 (10X for intraspecies
and 3X for FQPA) yielding a RfD/PAD of 0.0033 mg/kg/day. With
[[Page 54419]]
the BMD analyses, the BMDL of 0.38 mg/kg/day or 0.4 mg/kg/day would be
divided by 100 (10x for interspecies and 10X for intraspecies) for a
RfD/PAD of 0.0038 mg/kg/day or 0.004 mg/kg/day, respectively. The
similarity of these results, whether extrapolating from the animal or
human data, provides extra confidence in EPA's FQPA safety factor
decision. Additionally, EPA notes that reliance of the Gledhill study
produces a marginally lower and thus more protective level of concern.
Thus, the completeness of the toxicity database consideration
indicates that an additional safety factor of no greater than 3X is
needed to protect the safety of all populations, including infants and
children, due to a data deficiency in the Gledhill study. This decision
is consistent with EPA policies on RfD selection, the FQPA safety
factor, and cholinesterase inhibition, and with the scientific
literature on safety/uncertainty factors. It is also consistent with
long-established practice in making FQPA safety factor decisions in
circumstances where a LOAEL-to-NOAEL extrapolation is necessary.
Finally, EPA's scientific conclusions underlying this determination
have been concurred in by the Human Studies Review Board, an
independent panel of scientific experts in the field of toxicology and
bio-statistics.
ii. Pre- and post-natal toxicity. There was no evidence for
increased susceptibility of rat and rabbit offspring to prenatal or
postnatal exposure to dichlorvos. In both rat and rabbit developmental
studies, no developmental effects were observed. In the rat
reproduction study, the parental/systemic NOAEL/LOAEL was 2.3/8.3 mg/
kg/day, which was identical to the reproductive/offspring NOAEL/LOAEL.
The developmental neurotoxicity study showed evidence of sensitivity in
one parameter, auditory startle amplitude. However, there are no
residual concerns for sensitivity from this parameter because the
effects in pups were seen at a dose well above the Points of Departure
upon which EPA is regulating and a clear NOAEL for the effect (again,
well above the Points of Departure) was identified.
In addition, EPA evaluated the relative sensitivity of adult and
juvenile animals to cholinesterase inhibition from dichlorvos exposure
using a Benchmark Dose (BMD) analysis. For dichlorvos, EPA did a BMD
analysis of the rodent toxicity studies for adult and juvenile
cholinesterase inhibition (in both brain and RBC) in acute and repeated
dose scenarios. (Refs. 3 at 129; 42). EPA analyzed for a BMD showing a
10 percent inhibition of cholinesterase. EPA found similar results for
BMDs and BMDLs for cholinesterase inhibition in both the acute and
repeated dose scenarios for compartments (brain or RBC), sex, and age.
In other words, this analysis indicated that there was no significant
sensitivity difference with regard to cholinesterase inhibition between
adults and juveniles.
These data showing a lack of sensitivity of juvenile animals
relative to adults indicate a low level of concern that the
intraspecies factor applied to the Point of Departure from the Gledhill
study will fail to protect infants and children. Therefore, the
potential pre- and post-natal toxicity consideration, by itself,
indicates that risks to infants and children can be safely assessed
absent an additional safety factor.
iii. Completeness of the exposure database. EPA has extensive data
for estimating human exposure levels to dichlorvos. Although NRDC
objected to portions of EPA's dietary exposure assessment, after a
careful re-analysis of that assessment EPA concluded that its
dichlorvos exposure estimate from food, if anything, overstates
dichlorvos exposure given the many conservatisms retained in the food
exposure assessment and dichlorvos' documented volatility and rapid
degradation. (73 FR 42699; 72 FR 68686). Further, EPA concluded that
drinking water exposure to dichlorvos was also likely to have over-
estimated exposure because of conservative assumptions. (72 FR 68679-
68680). A similar conclusion was reached as to residential exposure to
dichlorvos after EPA revised this assessment taking into account
concerns raised by NRDC. (72 FR 68691). Thus, the completeness of the
exposure base consideration, by itself, also does not indicate a need
for an additional safety factor to protect infants and children.
3. Conclusion. The FQPA safety factor provision requires EPA to
presumptively retain an additional 10X safety factor for the protection
of infants and children. EPA may apply a different factor only if
reliable data show that factor to be safe. Under EPA policy, EPA
considers whether the additional FQPA safety factor is warranted taking
into account the other safety factors being applied.
For the Gledhill-based risk assessments, EPA has applied a 10X
intraspecies safety/uncertainty factor to account for the potential for
variable sensitivity among humans. EPA has not applied an interspecies
factor in these risk assessments because the Point of Departure is
drawn from a study in humans, not laboratory animals. (See Unit
VII.B.2). Thus, the precise question under the FQPA safety factor
provision for dichlorvos is whether EPA should retain the presumptive
additional 10X factor for the protection of infants and children or
whether there are reliable data showing that a different additional
factor will, in conjunction with the 10X intraspecies factor, protect
the safety of infants and children. As the above discussion of the all-
important FQPA safety factor considerations indicates, there are (1)
reliable data from animal studies on adult/juvenile sensitivity showing
that the standard 10X intraspecies factor will be protective of
potential pre- and post-natal toxicity to infants and children; (2)
reliable data on human exposure to dichlorvos demonstrating that an
additional safety factor is not needed to protect infants and children
due to exposure concerns; and (3) reliable data with regard to the one
toxicity data deficiency identified to show that a 3X additional factor
will be protective of all human populations, including infants and
children, as to the only toxicity data completeness issue. Therefore,
EPA reaffirms its selection of a 3X FQPA safety factor for Gledhill-
based assessments.
D. Conclusion
For all of the reasons set forth above, EPA denies NRDC's objection
to the use of a 3X FQPA safety factor for assessments relying on the
Gledhill study for a Point of Departure. Based on the revised
explanation provided in this order, EPA concludes, like it did in the
July 23, 2008 order, that a 3X additional safety factor will protect
the safety of infants and children. Because this revised explanation
addresses the court's reason for finding portions of the July 23, 2008
order to be arbitrary and capricious, EPA has not otherwise reopened or
reconsidered that prior order.
VIII. Statutory and Executive Order Reviews
This action denies an objection to a denial of a petition to revoke
tolerances, is in the form of an order and not a rule. (21 U.S.C.
346a(g)(2)(C)). Under the Administrative Procedure Act (APA), orders
are expressly excluded from the definition of a rule. (5 U.S.C.
551(4)). Accordingly, the regulatory assessment requirements imposed on
a rulemaking do not apply to this action, as explained further in the
following discussion.
A. Executive Order 12866 and Executive Order 13563
Because this order is not a ``regulatory action'' as that term is
defined in Executive Order 12866 entitled ``Regulatory Planning and
Review'' (58
[[Page 54420]]
FR 51735, October 4, 1993), this action is not subject to review by the
Office of Management and Budget (OMB) under Executive Orders 12866 and
13563 entitled ``Improving Regulation and Regulatory Review'' (76 FR
3821, January 21, 2011).
B. Paperwork Reduction Act
This action does not contain any information collections subject to
OMB approval under the Paperwork Reduction Act (PRA), 44 U.S.C. 3501 et
seq.
C. Regulatory Flexibility Act
Since this order is not a rule under the APA (5 U.S.C. 551(4)), and
does not require the issuance of a proposed rule, the requirements of
the Regulatory Flexibility Act (RFA) (5 U.S.C. 601 et seq.) do not
apply.
D. Unfunded Mandates Reform Act; and Executive Orders 13132 and 13175
This order denies an objection to a denial of a petition to revoke
tolerances; it does not alter the relationships or distribution of
power and responsibilities established by Congress in the preemption
provisions of section 408(n)(4) of FFDCA. As such, the Agency has
determined that this action will not have a substantial direct effect
on States or tribal governments, on the relationship between the
national government and the States or tribal governments, or on the
distribution of power and responsibilities among the various levels of
government or between the Federal Government and Indian tribes. Thus,
the Agency has determined that Executive Order 13132 entitled
``Federalism'' (64 FR 43255, August 10, 1999) and Executive Order 13175
entitled ``Consultation and Coordination with Indian Tribal
Governments'' (65 FR 67249, November 9, 2000) do not apply to this
order. In addition, this order does not impose any enforceable duty or
contain any unfunded mandate as described under Title II of the
Unfunded Mandates Reform Act (UMRA) (2 U.S.C. 1531-1538).
E. Executive Orders 13045, 13211 and 12898
As indicated previously, this action is not a ``regulatory action''
as defined by Executive Order 12866. As a result, this action is not
subject to Executive Order 13045, entitled ``Protection of Children
from Environmental Health Risks and Safety Risks'', (62 FR 19885, April
23, 1997) and Executive Order 13211 entitled ``Actions Concerning
Regulations That Significantly Affect Energy Supply, Distribution, or
Use'', (66 FR 28355, May 22, 2001). In addition, this order also does
not require any special considerations under Executive Order 12898
entitled ``Federal Actions to Address Environmental Justice in Minority
Populations and Low-Income Populations'' (59 FR 7629, February 16,
1994).
F. National Technology Transfer and Advancement Act
This action does not involve any technical standards that would
require Agency consideration of voluntary consensus standards pursuant
to section 12(d) of the National Technology Transfer and Advancement
Act (NTTAA), (15 U.S.C. 272 note).
IX. Congressional Review Act
The Congressional Review Act, 5 U.S.C. 801 et seq. does not apply
because this action is not a rule as that term is defined in 5 U.S.C.
804(3).
X. References
1. Natural Resources Defense Council. (February 1, 2008).
Objection to the Order Denying NRDC's Petition to Revoke All
Tolerances for Dichlorvos (DDVP), and Request for Public Evidentiary
Hearing.
2. Natural Resources Defense Council. (June 2, 2006). Petition
of Natural Resources Defense Council To Conclude Special Review,
Reregistration and Tolerance Reassessment Processes and To Revoke
All Tolerances and Cancel All Registrations for the Pesticide DDVP.
3. Office of Prevention, Pesticides and Toxic Substances, EPA.
(June 2006). Interim Reregistration Eligibility Decision for
Dichlorvos (DDVP). Available from: http://www.epa.gov/oppsrrd1/reregistration/REDs/ddvp_ired.pdf.
4. Lu, F. and Sielken, R. (1991). Assessment of safety/risk of
chemicals: inception and evolution of the ADI and dose-response
modeling procedures. Toxicology Letters 59, 5-40.
5. Schueplein, R. (2002). Pesticides and Infant Risk: Is There a
Need for an Additional Margin of Safety. Regulatory and
Toxicological Pharmacology. 31, 267-279.
6. Lehman, A. and Fitzhugh, O. (1954). Hundredfold margin of
safety. Quarterly Bulletin of the Association of Food and Drug
Officials of the United States. 33-35.
7. Food and Agriculture Organization. (1965). Evaluation of the
toxicity of pesticide residues in food. Joint report of the FAO
working party on pesticide residues and the WHO Expert Committee on
Pesticide Residues FAO Meeting Report PL/1965/10/1, WHO/Food Add./
27.65, Rome. Stoner, H. (1964). The Concept of acceptable Daily
Intakes of Pesticides for Man. Food and Cosmetics Toxicology. 2,
457-466.
8. International Programme on Chemical Safety. (2005). Chemical-
specific adjustment Factors for Interspecies Differences and Human
Variability: Guidance Document for use of DATA in Dose/
Concentration-Response Assessment. Available from: http://whqlibdoc.who.int/publications/2005/v9241546786_eng.pdf.
9. Dourson, M., Felter, S., and Robinson, D. (1996). Evolution
of Science-Based Uncertainty Factors in Noncancer Risk Assessment.
Regulatory Toxicology and Pharmacology. 24, 108-120.
10. Office of Pesticide Programs, U.S. EPA, Office of Pesticide
Programs' Policy on the Determination of the Appropriate FQPA Safety
Factor(s) for Use in the Tolerance Setting Process (February 28,
2002). Available from: http://www.epa.gov/oppfead1/trac/science/determ.pdf.
11. Barnes, D. and Dourson, M. (1988). Reference Dose (RfD):
Description and Use in Health Risk Assessments. Regulatory
Toxicology and Pharmacology. 8, 471-486.
12. National Research Council. (2009). Science and Decisions:
Advancing Risk Assessment. (The National Academies Press).
13. Burin, G. and Saunders, D. (1999). Addressing Human
Variability in Risk Assessment--The Robustness of the Intraspecies
Uncertainty Factor. Regulatory Toxicology and Pharmacology. 30, 209-
216.
14. Renwick, A. and Lazarus, N. (1998). Human Variability and
Noncancer Risk Assessment--An analysis of the Default Uncertainty
Factor. Regulatory Toxicology and Pharmacology. 27, 3-20.
15. Dourson, M. and Stara, J. (1983). Regulatory History and
Experimental Support of Uncertainty (Safety) Factors. Regulatory
Toxicology and Pharmacology. 3, 224-238.
16. Renwick, A.G. (1991). Safety factors and establishment of
acceptable daily intake. Food Additives & Contaminants. 8, 135-150.
17. Dourson, M., Charnley, G., and Scheuplein, R. (2002).
Differential Sensitivity of children and Adults to Chemical
Toxicity. Regulatory Toxicology and Pharmacology. 35, 448-467.
18. National Research Council. (1993). Pesticides in the Diets
of Infants and Children. (National Academy Press).
19. EPA. (2002). A review of the Reference Dose and Reference
Concentration Processes. EPA/630/P-02/002F.
20. U.S. EPA. (1994) Methods for derivation of inhalation
reference concentrations and application of inhalation dosimetry.
EPA/600/8-90/066F.
21. U.S. EPA. (1999) Toxicology Data Requirements For Assessing
Risks Of Pesticide Exposure To Children's Health: Report of the
Toxicology Working Group of the 10X Task Force [April 28, 1999
draft]. Available from: http://www.epa.gov/scipoly/sap/meetings/1999/may/10xtx428.pdf.
22. Dourson, M., Knauf, L., and Swartout, J. (1992). On
Reference Dose (RfD) and its underlying toxicity data base.
Toxicology and Industrial Health 8(3), 171-189.
23. U.S. EPA. (2012). Benchmark Dose Technical Guidance
Document. EPA/100/R-12/001.
24. FIFRA Science Advisory Panel. (March 19, 2002). Methods Used
to Conduct a Preliminary Cumulative Risk Assessment for
Organophosphate Pesticides. Final Report from the FIFRA Scientific
Advisory Panel Meeting of February 5-7, 2002 Available from: http://www.epa.gov/scipoly/sap/meetings/2002/february/final.pdf.
[[Page 54421]]
25. FIFRA Science Advisory Panel. (April 15, 2005). Final report
on N-Methyl Carbamate Cumulative Risk Assessment: Pilot Cumulative
Analysis. Final Report from the FIFRA Scientific Advisory Panel
Meeting of February 2005. Available from: http://www.epa.gov/scipoly/sap/meetings/2005/february/minutes.pdf.
26. FIFRA Science Advisory Panel. (October 13, 2005). Final
report on Preliminary N-Methyl Carbamate Cumulative Risk Assessment.
Final Report from the FIFRA Scientific Advisory Panel Meeting of
July 29-30, 2005. Available from: http://www.epa.gov/scipoly/sap/meetings/2005/august/minutes.pdf.
27. Office of Pesticide Programs, U.S. EPA. (2000). The Use of
Data on Cholinesterase Inhibition for Risk Assessments of
Organophosphorous and Carbamate Pesticides. Available from: http://www.epa.gov/oppfead1/trac/science/cholin.pdf.
28. Natural Resources Defense Council. (July 30, 2010).
Petitioner's Brief, NRDC v. U.S. EPA, No. 08-3771-ag (2d Cir.).
29. U.S. EPA, Respondent's Brief. (November 18, 2010). NRDC v.
U.S. EPA, No. 08-3771-ag (2d Cir.).
30. Office of Chemical Safety and Pollution Prevention, U.S.
EPA. (August 8, 2012). Memorandum from Ray Kent to Melanie Biscoe,
``Lists of chemicals for which human studies were either: Approved
by the Human Studies Review Board, or the basis for RfDs or RfCs in
IRIS.''
31. Office of Prevention, Pesticides and Toxic Substances, U.S.
EPA. (June 15, 2006). Report of the Food Quality Protection Act
(FQPA) Tolerance Reassessment and Risk Management Decision (TRED)
for Ethephon. Available from: http://www.epa.gov/oppsrrd1/REDs/ethephon_tred.pdf. Regulatory Toxicology and Pharmacology.
32. Office of Prevention, Pesticides and Toxic Substances. (June
19, 2007). U.S. EPA, Memorandum from Feleica Fort to Tom Myers,
``Methomyl. Acute, Probabilistic Aggregate Dietary (Food and
Drinking Water) Exposure and Risk Assessments for the Reregistration
Eligibility Decision.''
33. Office of Prevention, Pesticides and Toxic Substances, U.S.
EPA. (June 25, 2008). Memorandum from Elissa Reaves and Anna Lowit
to Karen Santora, ``Mode of Action, Eye Irritation, and the Intra-
Species Factor: Comparison of Chloropicrin and MITC.''
34. Office of Pesticide Programs, U.S. EPA. (March 24, 1998).
``Data Evaluation Report: Dichlorvos: A Single Blind, Placebo
Controlled, Randomized Study to Investigate the Effects of Multiple
Oral Dosing on Erythrocyte Cholinesterase Inhibition in Healthy Male
Volunteers.''
35. Gledhill, A.J. (1997). Dichlorvos: A Single Blind, Placebo
controlled, Randomised Study to Investigate the Effects of Multiple
Oral Dosing on Erythrocyte Cholinesterase Inhibition in Healthy Male
Volunteers.
36. EPA Human Studies Review Board. (May 15, 2006). Minutes of
the United States Environmental Protection Agency (EPA) Human
Studies Review Board (HSRB) April 4-6, 2006 Public Meeting.
37. Office of Pesticide Programs, U.S. EPA. (June 2002). Revised
Cumulative Risk Assessment of the Organophosphorus Pesticides.
Available from: http://www.epa.gov/pesticides/cumulative/rra-op/.
38. Office of Prevention, Pesticides and Toxic Substances, U. S.
EPA. (November 16, 2007). Memorandum from Ray Kent to Robert
McNally, Dichlorvos (PC 084001). Additional characterization of
inhalation risk posed by use of dichlorvos-containing resin strips.
DP332823.
39. Office of Chemical Safety and Pollution Prevention, U.S.
EPA. (August 9, 2010). Memorandum from Bayasid Sarkar to Ray Kent,
Precision analysis of Gledhill study for litigation of DDVP.
40. Kent, R., Office of Pesticide Programs, U.S. EPA. (April 5,
2006). Dichlorvos: WOE Comparison of Human and Animal Studies for
Single Chemical Assessment and OP Cumulative Assessment.
41. Office of Pesticide Programs, U.S. EPA. (June 20, 1994).
Memorandum from Brigid Lowery to Jocelyn E. Steward, Dichlorvos
(DDVP). Review of Subchronic Neurotoxicity Study in Sprague-Dawley
Rats.
42. Office of Prevention, Pesticides and Toxic Substances, U.S.
EPA. (June 9, 2006). Memorandum from Anna Lowit to Ray Kent,
Benchmark Dose analysis of cholinesterase inhibition in neonatal and
adult rats (MRID no. 46688914) following exposure to DDVP.
43. National Research Council, Reference Manual on Scientific
Evidence 249-252 (3rd ed. 2011).
List of Subjects in 40 CFR Part 180
Environmental protection, Administrative practice and procedure,
Agricultural commodities, Pesticides and pests, Reporting and
recordkeeping requirements.
Dated: August 29, 2012.
Steven Bradbury,
Director, Office of Pesticide Programs.
[FR Doc. 2012-21844 Filed 9-4-12; 8:45 am]
BILLING CODE 6560-50-P