August 7, 2007

Minutes of the 

United States Environmental Protection Agency (EPA) 

Human Studies Review Board (HSRB) 

June 27-29, 2007 Public Meeting

Docket Number: EPA-HQ-ORD-2007-0403

HSRB Web Site:  http://www.epa.gov/osa/hsrb/

Committee Members:	(See HSRB Members list – Attachment A) 

Dates and Times:  	Wednesday, June 27, 2007, 11:30 AM – 6:15 PM

			Thursday, June 28, 2007, 8:30 AM – 5:45 PM

			Friday, June 29, 2007, 8:30 AM – 3:00 PM 

(See Federal Register Notice – Attachment B) 

Location: 	EPA, One Potomac Yard (South Bldg.), 2777 S. Crystal Drive,
Arlington, VA  22202

Purpose: 	The EPA Human Studies Review Board (HSRB or Board) provides
advice, information, and recommendations on issues related to the
scientific and ethical aspects of human subjects research. 

Attendees: 	Vice Chair:		William S. Brimijoin, Ph.D.

Board Members: 	Alicia Carriquiry, Ph.D. 

Gary L. Chadwick, PharmD, MPH, CIP 

Janice Chambers, Ph.D., D.A.B.T. 

Richard Fenske, Ph.D., MPH

Susan S. Fish, PharmD, MPH

Suzanne C. Fitzpatrick, Ph.D., D.A.B.T. 

KyungMann Kim, Ph.D., CCRP

Michael D. Lebowitz, Ph.D., FCCP

Lois D. Lehman-Mckeeman, Ph.D.

Jerry A. Menikoff, M.D.

Sean M. Philpott, Ph.D.

Richard Sharp, Ph.D.

Meeting Summary:	Meeting discussions generally followed the issues and
general timing as presented in the meeting agenda (Attachment C), unless
noted otherwise in these minutes. 

Introduction and Identification of Board Members

Dr. William Brimijoin (Vice Chair, HSRB) introduced himself and stated
that he would serve as Chair for this HSRB meeting.  He welcomed Board
members, U.S. Environmental Protection Agency (EPA or Agency) staff, and
members of the public to the June 2007 HSRB meeting and acknowledged the
efforts of Dr. Paul Lewis (Designated Federal Officer [DFO], HSRB,
Office of the Science Advisor [OSA], EPA) and members of EPA’s Office
of Pesticide Programs (OPP) in planning and preparing for this meeting. 
At Dr. Brimijoin’s request, Board members introduced themselves.

Welcoming Remarks 

Dr. George Gray (Science Advisor, EPA) welcomed Board members and
conveyed EPA’s appreciation for their work in preparing for and
participating in the HSRB meetings.  He noted that the HSRB has
influenced how EPA uses the results of research involving human
subjects.  Dr. Gray welcomed members of the public and his EPA
colleagues and thanked Dr. Brimijoin for serving as Chair of this
meeting.  Dr. Gray introduced Ms. Susan Podziba, who will serve as
facilitator during the HSRB meeting.  Ms. Podziba is a public policy
mediator and has worked with a number of other EPA and federal advisory
committees. 

Dr. Gray provided a brief overview of topics to be addressed during this
meeting.  The topics included a redacted Confidential Business
Information (CBI) submission; Dr. Gray noted that he was pleased with
efforts to develop a framework that permits sound scientific and ethics
review of redacted CBI material at an open meeting.  He indicated the
science and ethics for protocols to measure exposure of occupational
pesticide handlers would be discussed.  The HSRB discussed plans to
implement such protocols at the June 2006 and April 2007 meetings, and
EPA appreciates the HSRB’s advice on these matters.

Mr. William Jordan (OPP, EPA) explained that because of scheduling
issues, Dr. Debbie Edwards (Director, OPP, EPA) was unable to attend
the morning session.  Mr. Jordan expressed EPA’s appreciation for the
HSRB’s work and looked forward to a productive meeting.  He outlined
three sets of topics for discussion.  The first topic involved two new
repellent efficacy protocols, one submitted by Dr. Scott Carroll
(Carroll-Loye Biological Research, Inc.) and the other by Insect Control
and Research, Inc. (ICR).  EPA also has obtained journal articles
describing four completed toxicology studies from a public literature
search performed in the course of gathering information for a risk
assessment of a registered pesticide active ingredient.  EPA seeks HSRB
input on these studies because the Agency has proposed to use them for
risk assessment activities for a re-registration review.  Lastly, the
HSRB will review major scientific and ethical issues in the design and
conduct of proposed pesticide handler exposure studies.

Meeting Administrative Procedures

Dr. Lewis welcomed Board members and thanked them for their efforts in
preparing for this meeting.  He welcomed members of the public and his
EPA colleagues.  As DFO, Dr. Lewis serves as liaison between the HSRB
and EPA and ensures that Federal Advisory Committee Act (FACA)
requirements—open meetings, timely announcements of meetings in the
Federal Register, and meeting materials made available at a public
docket—are met.  As DFO, he also works with the appropriate officials
to ensure that all applicable ethics regulations are satisfied.  Each
Board member has filed a standard government financial disclosure form
that has been reviewed by Dr. Lewis and the OSA Deputy Ethics Officer in
consultation with EPA’s Office of General Counsel to ensure that all
ethics disclosure requirements have been met.  Dr. Lewis reminded
participants that meeting times would be approximate and that public
comments would be limited to five minutes.

Because of the rescheduled start time for this meeting, review and
approval of the draft April 18-20, 2007 HSRB meeting report will be
performed during a teleconference to be scheduled in late July or August
2007.  A Federal Register notice will inform the public of the exact
time and date of the teleconference.

EPA Follow-up on HSRB Recommendations

Mr. Jordan reviewed EPA follow-up on HSRB recommendations from the April
2007 meeting.  Concerning the studies on the active ingredient IR3535 in
an aerosol formulation, EPA has accepted and agrees with the Board’s
recommendation and will rely on data from these studies in its review of
applications for registration of this product.  Regarding the field
mosquito efficacy protocol WPC-001, Dr. Carroll has revised the protocol
according to HSRB and EPA recommendations.  The field test will be
completed in July 2007, and Dr. Carroll will prepare his report and
submit it to EPA in time for review at the October 2007 HSRB meeting. 
The skin irritation study reviewed by the Board was deemed
scientifically useful and ethically acceptable; EPA agreed with the
Board’s conclusions and is using the data from this study in the
review of this product.  The HSRB had concerns that the skin sensitivity
study reviewed during the April 2007 meeting was not ethical.  EPA
agreed with this assessment and will not rely on the results of this
research.  EPA will require the product to carry a label instructing
users to stop using the product and seek medical attention if they
observe any skin irritation. 

The HSRB suggested revisions of the document describing best practices
for recruiting subjects for the Agricultural Handler Exposure Task Force
(AHETF) protocols; EPA will incorporate them into the document.  The
Board also endorsed EPA’s view that the research planned by the AHETF
and Antimicrobial Exposure Assessment Task Force (AEATF) will be useful
and provide important information for improving assessment of handler
risk while mixing, handling, or applying pesticides.  Further discussion
on topics related to the activities of these task forces is scheduled
during this meeting.

EPA Review of Carroll-Loye Protocol LNX-001

Introduction

Mr. John Carley (OPP, EPA) presented background information for
Carroll-Loye protocol LNX-001, which proposed a field test of mosquito
repellency for two conditionally registered formulations containing 20
percent picaridin as the active ingredient.  The Federal Insecticide,
Fungicide, and Rodenticide Act (FIFRA) allows EPA to conditionally
register such products; the condition for registration of these products
is efficacy studies.

LNX-001 is adapted from and similar to other Carroll-Loye protocols for
field mosquito repellency studies previously reviewed favorably by the
HSRB.  The initial submission met the standard of completeness as
defined in 40 Code of Federal Register (CFR) §26.1125; EPA’s science
and ethics review on May 24, 2007, was based on the initial protocol
submission.  Dr. Carroll proposed revisions to the protocol and
informed consent forms (ICFs), which were approved by the Independent
Investigational Review Board, Inc. (IIRB) of Plantation, FL on June 12,
2007, and provided to the HSRB on June 18, 2007.  Copies of the
IIRB-approved versions of the ICFs and a new IIRB-approval letter were
submitted to the docket on June 21, 2007.  Mr. Carley commended Dr.
Carroll’s responsiveness to HSRB recommendations, noting that the
original protocol was submitted before the April 2007 HSRB meeting and
revised quickly based on HSRB recommendations made at that meeting.

Scientific Considerations

Mr. Kevin Sweeney (OPP, EPA) provided the science assessment for
LNX-001.  The objectives of this protocol are to test the mosquito
repellent efficacy of the test material to satisfy a registration
condition imposed by EPA.  The test materials are lotion (EPA Reg. No.
39967-50) and pump spray (EPA Reg. No. 39967-50) formulations containing
20 percent picaridin.  The oral LD50 of picaridin is greater than 5,000
milligrams (mg) per kilogram (kg) and the dermal LD50 is greater than
2,000 mg/kg.  The study includes a dosimetry phase involving 10 subjects
to determine the typical consumer dose of each formulation for use in
efficacy testing.  The subjects are trained to aspirate landing
mosquitoes before they bite, using laboratory-reared, pathogen-free
mosquitoes.  The study is not blinded.

	

Ten subjects will be treated with each formulation and two untreated
control subjects will participate in each of two field trials.  The
untreated subjects are used to monitor mosquito pressure, and each
subject will be accompanied by two technicians who will aspirate landing
mosquitoes before they can probe or bite.  Both treated and untreated
subjects will be exposed to mosquitoes for 1 minute at a time every 15
minutes.  The duration of efficacy is calculated as the average time
from treatment to “first confirmed landing with intent to bite”
(FCLIBe).  The testing will be conducted at field sites in either the
California Central Valley or Southern California (depending on season). 
Expected wild mosquito populations include Aedes vexans, Ochlerotatus
melanimon, O. taeniorhynchus, Culex tarsalis, and C. pipiens.  Variables
to be measured include subject limb area and weight of test materials
delivered to subject limb (lotion) or gauze dosimeters (spray) for the
dosimetry phase.  For the efficacy phase, biting pressure must be
greater than or equal to one landing with intent to bite per minute. 
The test results will be analyzed by calculating the mean time to
FCLIBe, with standard deviation and 95 percent confidence interval;
depending on the results, other analyses also may be appropriate. 
Untreated controls will not be used for comparison of treatment means. 
Although the actual dose rate will not be known until the dosimetry
phase is complete, using a conservative estimated typical dose of 1 gram
(g) per 600 square centimeters (cm2), the Margins of Exposure (MOE) for
dermal toxicity is not expected to be less than and may be significantly
greater than 750.

The sample size of 10 reflects a compromise between financial and
scientific concerns; it also is recognized that sample size is difficult
to pre-determine without knowing the distribution of outcome values. 
EPA guidelines recommend 6 replicates, which has been widely regarded as
sufficient to show statistical significance at P<0.05.  Use of 10
replicates slightly improves accuracy in estimating the population mean;
however, each additional subject beyond 10 has a smaller affect on the
precision of the mean.  Nonetheless, EPA is reconsidering the issue of
sample size in light of the Board’s advice on this subject at the
January 2007 and April 2007 HSRB meetings.  EPA will apply any new
standards or requirements to future proposals; however, the Agency’s
current position is that a sample size of 10 treated subjects, which
exceeds the size specified in the current draft guidelines, is
acceptable for studies of this nature.

Deficiencies noted by EPA in review of this protocol include lack of an
explicit hypothesis; lack of an explanation for using untreated controls
in dosimetry; no information on diagnostic statistical tests for
normality, or on how non-normally distributed data will be analyzed; and
no justification for using Kaplan-Meier statistical analysis. 
Additionally, the exact locations of the four measured circumferences
for determining limb surface area during the dosimetry phase should be
recorded so that dosimeters can be placed in the same locations.  These
deficiencies were addressed in revisions to the protocol, which were
submitted in the June 18, 2007 amendment to EPA.

As revised in the June 18, 2007 amendment, this protocol is likely to
yield scientifically reliable information and produce important
information that cannot be obtained except by research with human
subjects.  The protocol has clear scientific objectives and an explicit
hypothesis; the study design should produce data adequate to achieve the
objectives and test the hypothesis.

Ethical Considerations

Mr. Carley provided the ethics review of LNX-001.  This study proposes
to test the mosquito repellent efficacy of two test formulations in the
field.  Both test formulations are conditionally registered; registrants
requested product-specific field efficacy testing to keep the products
on the market.  If demonstrated to be efficacious, these products
present a potentially attractive alternative to other available
repellents, some of which are considered unpleasant by many users.

Subjects will be recruited from among those who have participated in
previous Carroll-Loye Biological Research repellent efficacy tests
and/or have agreed or requested to be included in their volunteer
database.  The study excludes participants under the age of 18 years or
over the age of 55 years, students or employees of the investigator,
pregnant or nursing women, those sensitive to repellents or mosquito
bites, those in poor health or physical condition, and subjects unable
to speak and understand English.  Two “experienced” subjects will
serve as untreated controls in each field trial.  No eligible subjects
come from populations vulnerable to coercion or undue influence.  

Risks include possible irritation to the eyes if contacted by the
repellents and harm if swallowed, possible exposure to biting
arthropods, and possible exposure to arthropod-borne disease.  Risks
from the test materials have been minimized by excluding sensitive
candidates, closely monitoring the dosimetry phase, and having
technicians apply the repellent.  Risks from mosquito bites are
minimized by excluding sensitive candidates, training subjects to
aspirate mosquitoes before they have time to bite, and minimizing
exposure of skin.  Risks of disease are minimized by conducting the
field study in an area where mosquito-borne viruses have not been
detected for at least a month, by minimizing bites, and by testing
aspirated mosquitoes for pathogens.  The probability of risks is
characterized as extremely small because of low acute and chronic hazard
profiles of the products (although the lotion is a Toxicity Category II
eye irritant), research is designed to minimize exposures, training
subjects to aspirate landing mosquitoes before they have time to probe
or bite, and field testing in areas free of West Nile Virus (WNV) for at
least a month.

The primary direct beneficiary of these tests is the sponsor, and there
are no direct benefits to subjects.  If the materials are proven
effective and remain on the market, indirect beneficiaries will include
repellent users who prefer one of these products to other available
repellents.  EPA has found no reasonable opportunities to further reduce
risk while maintaining scientific robustness; the residual risks to the
subjects are very low and are reasonable considering the expected
societal benefits to repellent users.

The Plantation, FL IIRB reviewed and approved the protocol and informed
consent materials on April 5, 2007, and reviewed and approved the
amendments to the protocol and ICFs on June 12, 2007.  This IIRB is
independent of the sponsors and investigators and registered with the
U.S. Department of Health and Human Services’ (HHS) Office for Human
Research Protections (OHRP), but is not accredited by the Association
for the Accreditation of Human Research Protection Programs, Inc.
(AAHRPP).  There have been no changes in IIRB procedures since previous
submissions to EPA and EPA has determined that they meet regulatory
standards.  Concerning recruitment and informed consent, the description
of subject recruiting and consent processes is complete and
satisfactory, as supplemented by the amendments dated June 18, 2007.  An
IIRB-approved ICF for both treated and untreated subjects was included
in the original submission on April 10, 2007.  A mark-up of separate
ICFs for treated and untreated subjects and IIRB-signed final ICFs were
included in the supplemental submission on June 18, 2007.  Concerning
respect for subjects, the methods proposed for managing information
about prospective and enrolled subjects will generally protect their
privacy.  The June 18, 2007 amendment proposes to delete subject names
from data collection forms, thereby improving protection of subject
privacy.  Subjects will be free to withdraw at any time and will be
reminded of this at several points; medical care for research-related
injuries will be provided at no cost to the subjects.

This is a proposal for third-party research involving intentional
exposure of human subjects to a pesticide, with the intention of
submitting the resulting data to EPA under pesticide laws.  The primary
ethical standards applicable to this research are 40 CFR Part 26,
subparts K and L.  A point-by-point evaluation of how this protocol, as
submitted on April 10, 2007, addresses the requirements of 40 CFR Part
26, subparts K and L, and the additional criteria recommended by the
HSRB appears as Attachment 1 to the EPA Review of May 24, 2007.

Several deficiencies were noted in the EPA review of this protocol, but
these have been addressed in subsequent amendments to the protocol. 
Needed descriptions of how untreated controls will be recruited, and how
the process of informing them will differ from that used for treated
subjects were included in the June 18, 2007 amendments.  An erroneous
statement on page 7 claims that concentrations of the active ingredient
in test materials are lower than previously registered products;
however, the concentrations in question are actually higher and this
error still requires correction.  To further respect subject privacy,
data collection forms are reported to have been modified to delete
subject name.

	As modified, this protocol meets all requirements of 40 CFR §26.1111,
§26.1116, §26.1117, §26.1125, §26.1203, and all elements of National
Academy of Sciences (NAS) recommendations 5-1 and 5-2.  If further
revised to correct the remaining error, protocol LNX-001 will meet the
applicable requirements of 40 CFR Part 26, subparts K and L.

Public Comments

Dr. Scott Carroll of Carroll-Loye Biological Research, Inc., and Dr.
Ghona Sangha, 	Consultant, on behalf of Carroll-Loye Biological
Research, Inc.

Dr. Ghona Sangha has worked with picaridin since its development and
serves as a consultant to the registrant of the product.  She attended
this meeting to answer general questions concerning picaridin.

Dr. Carroll thanked the Board for its review of Carroll-Loye protocols
and ICFs.  He noted that in response to HSRB suggestions he has
segregated the protocols for treated and untreated subjects.  Dr.
Carroll noted that, concerning the specific protection of pregnant or
nursing women, these women could be included as untreated controls
because they would not be exposed to the products.  This issue was
raised with the IIRB, although few such women are likely
to participate.

Dr. Carroll stated that the issue of sample size should be discussed. 
His protocols currently use 10 subjects; historically, most
scientifically sound studies of DEET were performed using only 6
subjects.  He asked the Board for comments concerning statistical
arguments to address this issue.  Dr. Brimijoin noted that in his
experience with small group animal experiments to demonstrate an effect,
the primary aim was to determine whether a product was effective or not
and roughly how effective.  He asked whether this was the goal of the
repellent efficacy studies or if the protocol sought to determine
whether the results of the study would be applicable to the general
population or if there would be subjects who would find the products to
be ineffective or undesirable.  Dr. Carroll explained that the protocols
seek to determine a rough efficacy.  It is known that any product
effectiveness will vary between people; the goal of the protocol is
reasonable precision for determining a minimum protection time.

Dr. Brimijoin asked Dr. Sangha for information concerning the irritancy
or other toxicities of the 20 percent picaridin formulation, because
this concentration of picaridin is higher than that found in similar
products.  Dr. Sangha responded that although picaridin is new to the
United States, it has been used in Europe and Australia.  Picaridin was
developed based on chemical modeling to fit receptors found in the
mosquito.  Dr. Sangha stated that she served as Director of Toxicology
for Bayer and was responsible for toxicology testing.  Based on
toxicology data, picaridin is safe for use even up to the highest doses
tested, which were accompanied by only minor liver changes in rats. 
Dermal toxicity and teratology testing was performed using dermal
application, and all results indicated that picaridin was safe. 
Picaridin also has no neurotoxin potential and no cancer-causing
properties have been noted.  A complete packet of this information could
be available and picaridin meets EPA safety standards.

Picaridin itself is not a significant irritant; however, many
formulations contain ethanol, which can be irritating.  Picaridin is
efficacious compared to similar products and provides protection equal
to or better than DEET.  Picaridin also has a good skin feel and does
not react with plastics, unlike other DEET products.  Picaridin is not
significantly dermally absorbed; less than 2 percent of applied product
is absorbed by the skin.  Picaridin is safe for use on all people,
including children.  

Dr. Janice Chambers asked OPP representatives whether efficacy studies
are required for any new formulation of a substance.  Mr. Sweeney
responded that efficacy testing is required unless the formulation is
substantially similar to a previously tested formulation.  Dr. Chambers
inquired how efficacy testing was performed in children if a label
indicates that a product can be used on children.  Mr. Carley noted that
efficacy testing on children is not permitted.  Mr. Jordan clarified
that EPA expects a product that works well on adults will work equally
well on children and thus does not require efficacy testing on children;
adult data are used to evaluate efficacy.  He explained that EPA used
the data in the picaridin database described by Dr. Sangha to evaluate
the range of exposures and possible effects on those exposed, including
children.  Information on developmental toxicity is available in this
database, which EPA uses to determine whether young animals are more
sensitive to picaridin than older animals.  This information is used for
risk assessment activities regarding safety in children, along with
accounting for children’s different body surface areas and weights. 
Dr. Brimijoin commented that if children are known not to be more
sensitive to a compound, based on sound animal data, a 10-fold safety
factor also usually is applied.  Mr. Jordan explained that the Food
Safety Act requires additional safety factors to account for the higher
sensitivity of children.  This applies to pesticides and to repellents
only if the repellent is found on food.  When performing risk
assessments, EPA is mindful of the differences between exposure patterns
in children and adults.  If the data are insufficient to accurately
characterize risk, EPA will include uncertainty factors.

Board Discussion 

Scientific Considerations—LNX-001

Dr. Chambers opened the science review of LNX-001.  She thanked Mr.
Jordan for the memorandum framing the questions concerning this protocol
and commended Dr. Carroll for continually improving the clarity of his
materials.  With respect to methods, this protocol is similar to
previously reviewed protocols.  While the scientific criteria are
justified, the existing data are insufficient and new data are needed. 

Five deficiencies were noted in EPA’s review of this protocol.  Dr.
Carroll has addressed the lack of an explicit hypothesis in terms of
time length of efficacy; explicit hypotheses are unnecessary for this
type of research.  He added an explanation of untreated controls and
described modifications to the dosimetry protocol.  Dr. Chambers also
recommended that Dr. Carroll include limb surface area measurements in
the protocol, as noted by EPA.  Overall, the protocol is sound, similar
to previous protocols submitted by Dr. Carroll.

With regard to sample size, a balance must be struck between
practicality and risk; EPA should determine the appropriate sample size
for these protocols.  Although a larger sample size usually is preferred
and there is little risk involved in this particular protocol, future
protocols could involve increased levels of risk and perhaps expose
unnecessarily large numbers of subjects to risk if EPA does not provide
guidance.  Dr. Chambers concluded that the protocol is sufficiently
sound from a scientific perspective to assess the efficacy of these
20-percent picaridin formulations against mosquitoes.

Dr. Michael Lebowitz continued the scientific assessment of LNX-001. 
The protocol, with amendments, meets most HSRB scientific criteria;
however, the safety considerations of the protocol would be strengthened
by including an additional assay for pathogens in captured mosquitoes. 
The dosimetry phase also is a valuable addition to repellent testing
protocols.  Dr. Lebowitz agreed with Dr. Chambers regarding the
hypothesis for this research.  EPA statements concerning the
representativeness of subjects are correct, and Dr. Lebowitz agreed with
Dr. Chambers’ assessments of the sample size issue.  He expressed some
concern regarding statistical comparisons using the untreated controls
and the lack of positive or negative controls for the product matrix. 
Overall, the protocol is responsive to the HSRB charge.

Dr. Alicia Carriquiry commented on statistical issues related to this
protocol.  She clarified that the HSRB had not concluded that a sample
size of 6 to 10 was inadequate, but rather that it is impossible to know
whether this sample size is adequate or not.  EPA should develop a way
to weigh the risks and benefits of increasing sample size; however, EPA
should not set a specific sample size because not all protocols will
require the same sample size.  The appropriate sample size for a
protocol should meet criteria, such as power, and then be weighed
against the risks to subjects; the Board understands that in some cases
power may need to be sacrificed to mitigate risk.  Dr. Carriquiry added
that she appreciated Dr. Carroll’s improvements to the protocols.

Dr. KyungMann Kim commented on recurrent themes in these repellant
studies.  He specifically expressed concern about the definition of
“FCLIBe.”  The typical consumer will apply more repellant if bitten
and will not wait for a second “confirming” bite to determine
efficacy.  Dr. Kim also stated that the way the data are analyzed
presents problems.  For example, if no bites are received during the
course of the experiment, 15 minutes are added to the total time in the
field and that is the point at which the event (biting) is considered to
have happened.  Censored survival methods would provide a better way to
analyze this data, instead of using a normal distribution assumption for
data that is not likely to be normally distributed.  Dr. Brimijoin
inquired about the seriousness of Dr. Kim’s concern regarding the
basic measurement of FCLIBe.  Dr. Kim explained that at the October 2006
HSRB meeting, data points were presented in which the subject
experienced multiple landings, but these were not confirmed within 30
minutes by a subsequent landing and thus were not counted; Dr. Kim
considered this to be a flawed approach for measuring protection time. 
Once this issue is resolved, the Board should consider the appropriate
way to analyze censored observations.  It should be recognized that the
time to occurrence of an event does not follow a normal distribution, so
using the mean and standard deviation to analyze the data is
inappropriate.  

Dr. Kim added that it is unclear how the Board should address this
issue.  Measuring confirmed bites is an established method in the
repellency testing field because it is believed to reduce noise and
variability.  The Board should recommend that investigators preserve the
raw data so that it can be analyzed appropriately, even retrospectively;
investigators also should continue to provide the raw data to the Board
for review.  The use of FCLIBe may not reflect typical consumer use;
therefore, it is inappropriate to provide information derived using this
approach.  Dr. Chambers stated that Board members should consider the
need for consistency with past labels and the need to be fair across
past and present products.

Mr. Jordan agreed that Dr. Kim’s observations related to how best to
design these studies and interpret the resulting data were important. 
EPA has a long history of evaluating such studies, and against this
history, the Agency has developed labels in a particular way against a
backdrop of pesticide products.  Because of EPA’s investment and
discussions at HSRB meetings, EPA is revisiting established ways of
performing insect repellent efficacy studies and is in the process of
revising guidelines.  EPA is examining the evidence and considering
revision of guidelines concerning events indicating failure of efficacy
(landing, probes, bites, or FCLIBe) and also how to analyze truncated
data.  Any changes will be required only prospectively at first, then an
assessment of the significance of the changes will be made with respect
to labeling to determine whether re-testing of old products or
re-evaluation of old data is required.  Dr. Lebowitz noted that other
analyses could change the minimum protection information provided to the
consumer in what is likely to be a less conservative manner.  The goal
should be to protecting the consumer in the most conservative way.

Dr. Brimijoin summarized that the Board expressed general satisfaction
with protocol LNX-001 and thanked Dr. Carroll for his responsiveness to
Board recommendations.  Minor deficiencies in the protocol were
identified, but the protocol has many strengths, especially as amended. 
Dr. Carriquiry noted that statistical issues remain; determining a
specific sample size is not the issue, but rather developing a strong
basis for establishing sample size for a given protocol.  EPA needs to
develop a process based on criteria, such as power and risk, to advise
sponsors and registrants.

Dr. Brimijoin noted that Dr. Kim pointed out that the basic unit of
measure of failure of efficacy (FCLIBe) should be discussed to determine
if this is optimal.  Dr. Kim also raised questions concerning data
analysis, which should not rely simply on means and standard deviations
but instead should incorporate censored survival analyses.  Although
this problem is recognized, the benefit to ensuring consistency with
historical means of determining efficacy also was recognized to be of
interest for informing the consumer.  The HSRB commended EPA for
revisiting its data collection methods and considering proper
statistical treatment of this data.

Ethical Considerations—LNX-001

Dr. Sean Philpott opened the ethics discussion of protocol LNX-001.  He
deferred to his scientific colleagues regarding the scientific validity
of the study, noting that a study that lacks scientific validity cannot
be ethical.  He commended Dr. Carroll for his improvements to
the protocol.

This is a combined dosimetry and efficacy study enrolling a total of 20
subjects in both phases, and also includes two untreated controls.  The
untreated controls will be experienced field workers or frequent
participants in repellency studies.  The untreated controls will be used
to determine the ambient mosquito biting conditions in the field.  An
additional three alternative subjects have been enrolled in the case of
withdrawal of primary subjects and to protect subject privacy in the
case of a need to withdraw because of a previously undiscovered
pregnancy or other condition.  Dr. Philpott deferred to his colleagues
regarding whether the sample size would provide sufficient power.

Concerning compliance with 40 CFR Part 26, subparts K and L, Dr. Carroll
has submitted to EPA all information related to the conduct and review
of the investigation.  Concerning study design, the risks to subjects
are minimal and justified by societal benefits, including the efficacy
of picaridin and increased number of repellent choices for the public. 
The nature and likelihood of risks and side effects are clearly stated
in the ICFs and amended protocols.  The risks include reactions to
picaridin or other test materials, exposures to biting insects, and
exposure to arthropod-borne diseases.  Reasonable attempts to minimize
the risks have been made and a clear medical management plan was
provided.  Drs. Carroll and Sangha have provided toxicology data that
indicate that subjects are unlikely to be at risk for adverse reactions.
 Reactions to mosquito bites are usually mild and easily treated with
over-the-counter remedies.  Appropriate efforts have been made to
exclude those with a history of severe reactions to insect bites, and
clear plans to manage severe reactions are in place.  The protocol was
designed to minimize actual bites by using FCLIBe as an indication of
efficacy failure.  Exposure to biting insects also has been limited. 
The field tests are planned for areas where known disease have not been
detected by vector control agencies for at least 1 month.  Mosquitoes
collected during the field test will be analyzed using reverse
transcription polymerase chain reaction to test for pathogens, and plans
are in place to contact subjects if pathogens are detected.  These
precautions represent expected standards to ensure that subjects in
repellency studies are protected against exposure to arthropod-borne
diseases.

40 CFR Part 26, subpart L excludes pregnant and nursing women and
children from the subject pool; pregnancy tests will be administered the
day of the study to ensure exclusion of pregnant women.  Confidentiality
of subjects is protected and mechanisms are in place to minimize
potential coercion, such as excluding students and colleagues of the
investigator.  Subject compensation is at a level considered unlikely to
represent undue influence to participate.  With the provisions provided
in the amended protocol submitted to EPA on June 14, 2007, LNX-001
comports with 40 CFR Part 26, subparts K and L.

Dr. Richard Sharp agreed with Dr. Philpott’s assessment of the ethics
of LNX-001.  He noted that the ICFs should clarify that the
concentration of picaridin used in this protocol is higher than that
contained by previously registered products; the protocol will use a
20-percent formulation of picaridin and previously registered products
contain between 5 and 15 percent.  This is a substantial increase in the
amount of active ingredient and Dr. Carroll should specifically describe
this difference to potential subjects.  Dr. Jerry Menikoff agreed with
Drs. Sharp and Philpott and had no further comments.

Mr. Carley clarified the number of subjects and whether this was stated
incorrectly in the protocol.  Ten subjects will participate in the
dosimetry phase and will test both formulations.  It is not clear,
however, whether the same 10 subjects will participate in the efficacy
phase.  A total of 10 treated and 2 untreated subjects in each field
study results in a total of 34 subjects, not including alternate
subjects.  Therefore, the characterization of “at most” 34
participants is accurate because some subjects could participate in both
phases of the study.  Dr. Philpott agreed to check this value.  He
assumed the group would be the same because volunteers for the field
test would be trained in the laboratory to aspirate mosquitoes and
identify FCLIBe.  Dr. Carroll clarified that approximately 80 to 90
percent of the subjects participate in all phases of the trial;
therefore, the total number of subjects will be low.  He used the value
of 34 to try to account for the maximum number of participants, but
would consider changing this number to 37 or even 40 to accommodate
alternates.  All potential subjects receive training even if they do not
participate in the dosimetry phase of the study; Dr. Carroll offered to
state this more explicitly in the protocol.

Dr. Brimijoin summarized that Dr. Philpott had considered all categories
of risk and found a favorable risk-to-benefit ratio.  Appropriate
attempts were made to minimize risks.  This protocol also has
innovations that further minimize risk, such as post-test analysis of
captured mosquitoes for pathogens and notification of subjects if
pathogens are detected.  The HSRB has found this protocol to be ethical,
but recommends including the clarification recommended by Dr. Sharp to
emphasize to subjects the significant difference in the concentration of
active ingredient in this formulation compared to those in existing
products.

EPA Review of ICR Protocol G0590307001A044

Introduction

Mr. Carley provided background on protocol ICR A044, submitted on behalf
of the sponsor and ICR, Inc., by toXcel on April 12, 2007.  This
protocol proposes a field study in two locations of the
mosquito-repellent efficacy of two new un-registered formulations
containing picaridin.  The initial submission met the standard of
completeness as defined in 40 CFR §26.1125.  EPA’s Science and Ethics
Review on May 24, 2007, was based on the initial protocol submission.

In response to EPA’s Science and Ethics Review on May 24, 2007, toXcel
(representing the sponsor) submitted comments to the HSRB docket on June
18, 2007, noting three errors in EPA’s review and promising numerous
changes in the protocol and ICFs; however, toXcel’s comments to the
HSRB docket did not include all proposed language to accomplish the
promised changes.  Although much work is still required to address all
deficiencies noted in EPA’s Science and Ethics Review, EPA considers
this protocol ready for HSRB review.

The sponsor has asserted a claim of confidential business information
with respect to the identity of the sponsor and the concentration of the
active ingredient in each product.  A HSRB workgroup reviewed the
redacted version prepared by the submitter and concluded that the
redactions to protect the CBI from disclosure would not prevent the HSRB
from reviewing the protocol.

Scientific Considerations

Mr. Sweeney provided the science review of ICR A044.  The objective of
this study is to test the field efficacy of two unregistered mosquito
repellent aerosol formulations containing picaridin.  The test materials
include Product A and Product B, both of which are aerosol formulations
containing picaridin at concentrations within previously registered
ranges.  The oral LD50 is greater than 5,000 mg/kg and dermal LD50 is
greater than 2,000 mg/kg.  The study does not include a dosimetry phase;
instead, efficacy testing will be performed using the standard dose rate
of 1g/600 cm2, equivalent to 1.67 mg/cm2, with an expected MOE greater
than or equal to 926.  One treatment will be applied to a 250 cm2 area
on each forearm of each treated subject—Product A on the right arm and
Product B on the left arm.  Only the subjects will be blinded.

The study design calls for 10 subjects plus 2 alternates who will be
treated with both formulations, and 2 untreated control subjects,
selected by lot, in each of two field trials.  Untreated subjects are
included to monitor mosquito pressure; each subject is exposed for up to
5 minutes every 30 minutes, or until experiencing 5 mosquito landings. 
Treated subjects are exposed to mosquitoes for 5 minutes every 30
minutes.  Both treated and untreated subjects will move to a screenhouse
between 5-minute exposure periods.  To minimize the potential to remove
the repellent, treated subjects will not cover the treated skin between
5-minute exposure periods.  The duration of efficacy will be measured as
average time from treatment to “first confirmed bite” (FCB).  The
protocol calls for 1 day of testing lasting up to 14 hours; subjects
will spend a total of 6 days away from home, including travel days for
each site.

The field test sites are located in Savannah, GA, and Pine Island, FL. 
Expected wild mosquito populations include A. vexans, Psorophora ferox,
O. infirmatus, and O. taeniorhyncus.  Measured variables include subject
limb dimensions, mosquito pressure (landings greater than or equal to 1
per minute), whole body landing count, time of all bites, and time to
FCB.  The test results will be analyzed by calculating the mean time to
FCB (TFCB), and the standard deviation and 95 percent confidence
interval around the mean.  Untreated controls will not be used for
comparison of treatment means and results will not be compared between
the two formulations.

ICR’s rationale for sample size follows the EPA guideline recommending
six replicates.  An analysis by Rutledge and Gupta (1999) shows that a
sample size of 10 or 11 is needed to achieve, with 95 percent
confidence, a standard deviation not greater than 2 hours after 8 hours
of testing.  Thus, this study will involve 10 treated test subjects; two
additional alternate subjects will be treated to help ensure a minimum
number of 10 and will also aid in protecting the privacy of any subjects
who withdraw.

Several deficiencies were noted in the EPA review of this protocol.  The
protocol lacked an explicit hypothesis; the June 18, 2007 response
promised to add the hypothesis that test products will be effective for
8 to 12 hours.  Information concerning diagnostic statistical tests for
normality, or on how non-normally distributed data will be analyzed was
not included in the initial protocol.  The June 18, 2007 response
explained that statistical analysis will be limited to the calculation
of mean TFCB, standard deviation, and 95 percent confidence interval
around the mean, and that TFCB will be recorded as 12 hours for subjects
who do not experience repellency failure.  There was inadequate
justification for excluding “outlier” results, but the June 18, 2007
response promised that outlier measurements would be included in
determination of complete protection time (CPT).  A direct reference to
Good Laboratory Practice (GLP) regulations at 40 CFR Part 160 was
needed; the June 18, 2007 response promised to add references to
4 sections of 40 CFR Part 160 (GLP regulations) to the protocol.

The June 18, 2007 response failed to address questions of distributions
raised in the EPA review and the protocol has not yet provided
information on diagnostic statistical tests for normality, and on how
non-normally distributed data will be analyzed.  Concerning GLP, EPA has
asked for citation of all applicable sections in addition to the four
mentioned in the June 18, 2007 response.  Minimally, sections 160.33,
160.47, 160.51, 160.120, and 160.130 should be cited, and the best
solution would be to cite and promise compliance with “all applicable
requirements of 40 CFR Part 160.”

If amended as promised in the June 18, 2007 response, and if an
acceptable reference to GLP regulations and an adequate statistical
analysis plan are incorporated, this protocol is likely to yield
scientifically reliable information, produce important information that
cannot be obtained except by research with human subjects, and produce
adequate data to achieve the clear scientific objective.

Ethical Considerations

Mr. Carley presented the ethics review of ICR A044.  The proposed study
would test the mosquito repellent efficacy of two unregistered test
formulations in the field.  EPA requires formulation-specific efficacy
testing to register the products and demonstration of field efficacy for
these test products would contribute to making available potentially
attractive alternatives to other available repellents, some of which are
found unpleasant by users.

Subjects will be recruited from a database including previous subjects
of similar ICR tests as well as friends and colleagues of previous
subjects.  The subject pool is characterized as being “as
representative of potential repellent users as we are able to make
it.”  Untreated control subjects will be selected by lot the night
before the test, from the group of subjects who have traveled to the
test site.  Children or pregnant or nursing women are excluded, as well
as those sensitive to repellents or mosquito bites, those in poor health
or physical condition, those unable to speak and understand English,
adults over the age of 65 years, and permanent full-time employees of
ICR.  No subjects will be drawn from populations vulnerable to coercion
or undue influence.  This protocol differs from the Carroll-Loye
protocols in that the subjects live in different parts of the country,
some near the test sites and some far from the sites.  Because of this,
participation will involve 6 days away from home for testing and travel,
and subjects will be provided with airfare and hotel accommodations at
the test sites.  A group of 14 subjects are recruited, consented, and
then convene at the test location.  This includes 10 treated subjects,
two untreated controls, and two alternate subjects; subjects serve as an
untreated control only once.  In addition to airfare, lodging, and
payment for meals, subjects are compensated for participation; assuming
a 14-hour test day and payment over 6 days, compensation is expected to
be approximately $676.50 per subject.  Although this appears high enough
to warrant concerns about inducement to participate, most of those in
the volunteer database have participated in these types of studies
before; therefore, the compensation is considered unlikely to affect
their decisions to participate.

Potential risks arise from the repellents themselves; the materials are
Toxicity Category II/III for eye irritation (Material Safety Data Sheets
show “Warning” label).  The protocol contains misleading references
to Toxicity Category IV for certain effects of the active ingredient and
misleading references to beneficial effects of the inert ingredients
(i.e., skin benefits).  The risk of allergic or irritation response from
test materials are minimized by excluding sensitive candidates and
monitoring subjects closely.  The MOE for dermal toxicity for the
formulations is at least 1,000.

Risks of allergic or irritation responses to arthropod bites also exist.
 These can be reduced by excluding candidates with a history of severe
reactions to mosquito bites, minimizing the number of untreated control
subjects, exposing untreated controls only long enough to confirm
continued mosquito landing pressure, permitting intermittent exposure of
only a small area of treated skin, instructing subjects to move away
from mosquito-infested areas between exposure periods, covering treated
skin after efficacy breakdown, teaming subjects in pairs to watch each
other for landing mosquitoes, brushing away mosquitoes attempting to
bite with fewer than all six of their legs on the treated area of skin,
and treating subject (and staff) shoes with permethrin to repel ticks. 
Risk of reaction to insect bites could be further reduced by treating
landings as evidence of efficacy breakdown.

EPA requires demonstration of efficacy of repellency of mosquitoes known
to carry specific diseases, such as WNV, to substantiate related product
claims.  Thus, the sponsor must conduct field testing in areas
containing these potential vector species.  The principal carriers of
WNV are not common at the test site, and risks of arthropod-borne
diseases are reduced by conducting field tests in areas where WNV has
not been detected by the local Mosquito Abatement District for at least
a week.  Risks could be further reduced with an improved medical
management plan, provision for post-exposure follow-up, and by excluding
subjects over 55 years of age.

The primary direct beneficiary of this research is the sponsor and there
are no direct benefits to the subjects.  If these materials are proven
effective, indirect beneficiaries will include repellent users who
prefer this product to other repellents.  Concerning risk-benefit
ratios, opportunities have been identified to further reduce risk while
maintaining scientific robustness.  If these risks are minimized,
residual risks to subjects would be low.  The test materials are likely
to prove effective and, if minimized, risks to subjects are likely to be
reasonable in light of the expected societal benefits to repellent
users.

An independent ethics review was conducted by the Essex Institutional
Review Board (EIRB), Inc. of Lebanon, NJ.  This board reviewed and
conditionally approved the protocol on April 2, 2007, subject to
revision, and subsequently reviewed and approved amendments 1-8 on April
6, 2007, and reviewed and approved ICFs on April 9, 2007.  EIRB is
independent of the sponsors and investigators and is registered with
OHRP but does not hold Federal Wide Assurances (FWA) or accreditation by
AAHRPP.  EIRB procedures have been submitted directly to EPA under CBI
claims; EPA has determined they meet regulatory standards.  

Concerning the informed consent process, most subjects do not live near
the ICR laboratory; therefore, recruiting and consent processes must
rely heavily on telephone contacts and mailing of documents.  It is
unclear whether subjects reviewed the informed consent materials before
traveling to the test site.  EPA is working with ICR to ensure that
subjects were permitted to review the materials before traveling so that
they could make a fully informed  decision about participation.  The
submission of April 12, 2007, included an Institutional Review Board
(IRB)-approved ICF for each site meeting the requirements of 40 CFR
§26.1116 and §26.1117.  Inconsistent descriptions of the recruiting
and consent processes in the protocol and ICF still require
reconciliation.  Methods for managing information about prospective and
enrolled subjects have been proposed and will generally protect subject
privacy.  Subject names currently are included on data collection forms;
deletion would improve protection of subject privacy.  The procedure for
protecting the privacy of candidates discovered to be pregnant at the
test site needs refinement.  Subjects will be free to withdraw at any
time and medical care for research-related injuries will be provided at
no cost to the subjects.

This is a proposal for third-party research involving intentional
exposure of human subjects to a pesticide, with the intention of
submitting the resulting data to EPA under the pesticide laws.  The
primary ethical standards applicable to this research are 40 CFR Part
26, subparts K and L.  The protocol cites ICR policy of compliance with
HHS rules at 45 CFR Part 46 and the “EPA model rule;” HHS rules do
not apply to this study and the reference to the “EPA model rule” is
unclear.  A point-by-point evaluation of how the protocol as submitted
on April 12, 2007 addresses the requirements of 40 CFR Part 26, subparts
K and L and additional criteria recommended by the HSRB appears as
Attachment 1 to the EPA Science and Ethics Review dated May 24, 2007.

Deficiencies noted in the EPA ethics review of this protocol include
inclusion of subjects over 55 years of age; such subjects should either
be excluded or their inclusion should be justified.  The June 18, 2007
submission promises to cap participation at the age of 55 years if the
HSRB agrees with EPA that to do so is warranted and would not
scientifically compromise the study.  ICR also was asked to clarify
exclusion factors concerning employees—either exclude all employees
(i.e., full time and part time) of ICR or the sponsor, or justify not
excluding all of them and describe the specific safeguards in place. 
The June 18, 2007 submission also clarifies that all employees of the
sponsor, of ICR, and of any other interested organization will be
excluded; the protocol and ICFs still require amendment.

Other deficiencies pertain to the risk of mosquito bites.  This risk
could be further reduced by treating landings as evidence not only of
biting pressure but also of efficacy failure.  ICR was asked to explain
and justify the emphasis on distinguishing landings from probes, probes
from bites, and bites with six legs on treated skin from other bites. 
The June 18, 2007 response argues that bites are a “more reliable and
rigorous” endpoint than landings, treated subjects will typically
receive only two bites, TFCB is a recommended endpoint in EPA draft
guidelines, risks are sufficiently reduced by other means, and TFCB is a
necessary endpoint for “robust and reliable” test results.  The June
18, 2007 response does not provide justification for these fine
distinctions.  EPA also found that the risks of contracting an
arthropod-borne disease could be further reduced by providing for
post-exposure follow-up; ICR’s June 18, 2007 response promises to add
follow-up calls to subjects 2 weeks after testing, although the protocol
still requires amendment.  EPA recommended including a straightforward
commitment to conduct this research in compliance with 40 CFR Part 26,
subparts K and L; the June 18, 2007 response does not address this
comment.  ICR also was asked to provide a single, clear and
authoritative description of the process of informing candidates and
seeking their consent, and harmonize discrepant references to the
process; the June 18, 2007 response indicates that discrepancies will be
resolved through amendments to the protocol and ICFs.  Regarding
protection of subjects’ privacy, ICR promises that subject names will
be deleted from data collection forms in favor of subject identification
codes and initials.  EPA recognized that recruitment of alternate
subjects provides some opportunity for discrete withdrawal without
explanation, but may not be adequate to protect the privacy of subjects
who are found to be pregnant only after traveling to the test area and
participating in the evening-before-the-test activities; the June 18,
2007 response does not address this comment.

Concerning compliance with ethical standards, with EIRB approval of all
requested revisions, all requirements of 40 CFR §26.1111, §26.1116,
§26.1117, and §26.1125 would be met; requirements of 40 CFR §26.1203
are met.  All elements of NAS recommendation 5-1 and 5-2 are satisfied. 
EPA concludes that, if revised as requested, ICR A044 and the associated
ICFs will likely meet the applicable requirements of 40 CFR Part 26,
subparts K and L.  This research must not be initiated without EIRB
approval of the revised protocol and ICFs.

Dr. Lebowitz inquired if efficacy testing using first bite was developed
before WNV became prevalent.  Mr. Carley responded that this was true. 
Dr. Lebowitz noted that this demonstrated the need for revisions in
practices when new diseases become prevalent and human testing may lead
to infection.  He was uncertain how EPA or the Centers for Disease
Control and Prevention (CDC) should respond to this, but current public
health prudence suggests that this standard should be changed to prevent
possible infection.  Dr. Chambers asked what percentage of landings
result in bites; neither Mr. Carley nor Mr. Sweeney could provide an
answer.  Dr. Susan Fish requested clarification of the difference
between a bite with six legs contacting the skin versus fewer legs
making contact.  Mr. Sweeney explained that borders without repellent
exist on the edges of the treated areas.  Mosquitoes may land there and
bite; these bites are not considered proof of failure of efficacy
because the mosquitoes may be contacting areas not treated by the
repellent.  Dr. Chambers observed that the protocol states that such
bites would not be counted and used to determine efficacy breakdown.

Dr. Kim noted that this protocol increases exposure time to 5 minutes,
compared to 1 minute used in the Carroll-Loye protocols.  He inquired
whether this would have an impact on outcome, and if so, how the
outcomes from the different protocols could be reconciled.  Mr. Sweeney
explained that each bite is considered an event, and if more events
occur over time, a more accurate CPT calculation can be determined. 
More intervals and more endpoints will lead to more accurate data.  EPA
currently is working to reconcile these points as it addresses its
guidelines for repellency testing.  Dr. Carriquiry observed that
although the ICF indicated that subjects will receive a maximum of two
bites, subjects could receive more than two if the bites from mosquitoes
that do not have a six-legged landing are not counted.  Mr. Carley
agreed, and stated that this relates to whether a confirming event is
needed to confirm efficacy.  He explained that it is unlikely that a
subject would receive a large number of bites that are all
“unconfirmed.”  He agreed that the protocol and ICFs should contain
a more carefully worded statement about the likely number of bites a
subject may receive.  The documents also should clarify that although
subjects receive both products A and B, each independently-dosed arm is
considered to be part of a different experience.  Therefore, no fewer
than two bites will be needed to indicate product breakdown.  Problems
arise if no bites are received before the end of the exposure period,
and also if a subject is bit once at Hour 3 of the study but is not
bitten again.  It is difficult to accurately convey to potential
subjects the maximum number of likely bites.

Dr. Lebowitz asked Mr. Sweeney if he was aware of the reliability of
data concerning detection of arthropod-borne diseases at the test site
and whether lack of detection for 1 week was sufficient to reduce
potential exposure.  Mr. Sweeney agreed that EPA did not know whether 1
week without evidence of pathogens provided sufficient protection.  Dr.
Chambers observed that this protocol followed current guidelines for
efficacy testing more closely than the protocols submitted by
Carroll-Loye; Mr. Sweeney and Mr. Carley agreed.

Public Comments

Mr. Niketas Spero and Dr. Robin Todd of ICR

Mr. Niketas Spero (ICR) noted that ICR has a long history of performing
repellent efficacy studies which have received favorable review from
EPA.  He also explained that the ICR A044 protocol was designed using
current EPA guidelines.

The dosing application rate used in this study was based on guidelines
developed during the testing of DEET.  The application rate of 1.67
mg/cm2 provides complete coverage without run-off.  This application
rate has been used in multiple studies so that comparisons can be made. 
Concerning counting bites versus landings, current EPA guidelines call
for determining the TFCB; this guideline also has been used for many
studies.  Repellency failure, as measured by landing, can be
problematic.  Not all landings result in a bite and most consumers will
not notice a landing, but will notice mosquito probing or bites.  To
minimize risk, the protocol excludes those who are over-reactive to
insect bites, sensitive to repellents, older than 55 years of age, or
who are not in good health.  Subjects are exposed to mosquitoes for 5
minutes every 30 minutes and also are provided with protective gear so
that bites can occur only on the exposed area.  Concerning exclusion of
landings in which all six legs do not make contact, this is commonly
used to measure repellency and ensure that results are not confounded by
a mosquito contacting an area near, but not on, the area of repellent
application.  ICR depends on mosquito abatement agencies for information
concerning the lack of arthropod-borne diseases in the test areas and
tests only in disease free areas.  ICR believes that all possible
measures to reduce risk and gather sound data have been made.

Dr. Carriquiry noted that subjects are informed of their freedom to
withdraw from the study, but asked whether ICR would continue to pay for
the withdrawing subject’s hotel stay or to change their ticket to
return home.  Mr. Spero explained that the company would do either; the
subject could stay at the site at ICR’s expense or ICR would pay to
change the ticket.  Dr. Lebowitz inquired whether ICR would change its
testing plans if local abatement agencies found infected mosquitoes 5
kilometers from the test site or found a positive mosquito 2 or 3 weeks
prior, but not 1 week prior to the field study.  Mr. Spero noted that
the presence or absence of arthropod-borne pathogens can change quickly.
 He was unsure of the value of monitoring for the presence of pathogens
for an entire month.  Monitoring for 1 week prior to and during the test
likely minimizes the risk of disease transmission.  Dr. Lebowitz noted
that the protocol did not indicate what ICR would do in the event of a
positive test result and that if a test site is positive 2 weeks before
the study but negative 1 week before the study; the likelihood that
pathogens are present during the week of testing is high.  He also
questioned how close the test site was to monitor for pathogens, for
example, whether it was within the range of all mosquitoes that could
carry WNV.  Mr. Spero explained that monitoring is performed on a
county-wide basis; he agreed that ICR would consider monitoring for more
than 1 week.

Dr. Carriquiry questioned how the results obtained from the two
different test sites would be analyzed, particularly if the efficacies
at each site were significantly different.  Dr. Robin Todd (ICR)
responded that, when this has occurred in the past, ICR submitted the
data as two different reports to EPA.  EPA used the shorter measurement
of efficacy for labeling purposes.

Dr. Chambers inquired whether Dr. Todd or Mr. Spero knew what proportion
of landings resulted in bites; neither Dr. Todd nor Mr. Spero could
answer this.  Dr. Chambers also asked about the fraction of bites from
mosquitoes landing partly on the gauze rather than entirely on the skin.
 Mr. Spero explained that this was unknown because such mosquitoes would
be dislodged from the subject.  ICR does not want to evaluate mosquitoes
that have discovered a way to avoid the repellent but still bite.  He
added that over the course of an 8-hour day, the bandage may abrade away
the repellent.  Dr. Chambers questioned the number of times confirmed
bites are received using a 5 minute exposure every 30 minutes.  Mr.
Spero noted that this was a new procedure for ICR; in the past, subjects
were exposed continuously.  He added that once a subject receives two
bites, exposure of the treated site is halted.

Dr. Gary Chadwick questioned why two sites will be used for testing. 
Mr. Spero replied that EPA recommends testing at two different sites
populated by different genera of mosquitoes.

Board Discussion

Scientific Considerations—ICR A044

Dr. Chambers opened the scientific discussion of ICR A044.  She
commended the authors for a clearly written protocol.  She noted that
the need for data arising from this study is justified.  Several
deficiencies were noted by EPA.  The need for an explicit hypothesis is
not significant for this type of research, as discussed for the protocol
LNX-001.  Dr. Chambers declined to comment on issues of statistical
analysis.  She concurred with EPA on subject exclusions and deemed the
references to GLP to be reasonable.

Consistency is an issue for this protocol.  The guidelines define bites
as failure of efficacy, and the protocol is responsive to these
guidelines; however, the Board has heard that counting FCLIBe provides a
greater measure of safety.  Scientifically, landing with intent to bite
provides a more conservative estimate of efficacy and thus is more
protective for the consumer.  The HSRB requests that ICR consider
counting landings rather than bites.  Concerning dosing, this protocol
does not include a dosimetry phase but instead relies on established
guidelines for dosing.  Dr. Chambers recommended that a dosimetry study
be performed before the field study and that the dose established by
this work then be used.  Mosquitoes also should be collected during
testing and assayed for the presence of pathogens.  The Board
understands ICR’s rationale for counting only six-legged landings;
however, a more conservative measure of efficacy would count landings
with fewer than six legs.  ICR allows for collection of information on
the total number of landings, but the value of this activity is unclear.

Dr. Chambers concluded that, in general, this is a sound protocol, but
suggestions such as testing captured mosquitoes for pathogens and
performing a dosimetry study would benefit the protocol, along with
EPA’s suggestions.  EPA must decide whether landings or bites should
be counted in tests of repellent efficacy and should develop guidelines
reflecting this decision.

Dr. Lebowitz commended Mr. Carley and Mr. Sweeney on their reviews of
the protocol.  He stated his concerns about the adequacy of testing
using bites, and how public health law would impact EPA given current
WNV concerns.  He suggested that EPA search for policies or decisions
made by the CDC regarding this issue.  It may be prudent for EPA to
change its guidelines regarding the use of bites versus landings for
efficacy testing.  He commended the protocol’s use of 5-minute
exposure times every 30 minutes and providing a screened area for
subjects when not exposed.  Dr. Lebowitz agreed with EPA’s and Dr.
Chambers’ critiques of the protocol.

Dr. Carriquiry stated that proposing to compute the mean and confidence
interval of the data is inappropriate unless the data are normally
distributed.  Data should be collected and its distribution analyzed
before plans for analysis are made.  To state that the standard
deviation is estimated to be 2 with a 95-percent confidence interval
would imply a mean CPT of between 4 and 12 hours, which probably is not
what the investigators intended.  Dr. Carriquiry agreed with Dr.
Chambers about EPA’s need to decide whether to count bites or landings
as failure of efficacy.  The LNX-001 protocol would be at a disadvantage
if its data were compared to that of this protocol, because landings
with intent to bite likely occur before a bite, resulting in a
comparably poorer measure of CPT.  

Dr. Brimijoin agreed with the need for consistency across studies
regarding bites versus landings.  If the intent of these protocols is to
generate data for labeling purposes, this discrepancy is unfortunate. 
EPA should quickly decide which standard to use.  Using landings is
protective of subjects and may be as robust as bites, but is likely to
yield a different CPT.  Dr. Brimijoin acknowledged the questions raised
by Dr. Lebowitz regarding checking test site mosquitoes for pathogens
but considered the protocol to be adequate in this regard. 
Dr. Carriquiry raised important issues focusing on determining the
nature of the distribution of the data before deciding on the analyses
to perform.  The Board also recommended that dosimetry be performed
before testing is implemented.

Ethical Considerations—ICR A044

Dr. Philpott opened discussion of the ethical considerations of ICR
A044.  He agreed with Mr. Carley’s review and recommended that all Mr.
Carley’s suggested changes be made to bring the protocol into
regulatory and ethical compliance.

Regarding the issue of landings versus bites, from an ethical
perspective minimizing risk by using landings takes precedence unless
there is a scientific justification indicating that using bites is more
appropriate.  If mosquitoes are collected, tests to detect the presence
of pathogens are not difficult and should be performed.  The
investigators should provide more detailed procedures for monitoring
subjects post-test to ensure that subjects have not been infected. 
Other concerns relate to the provision of EIRB procedures described in
the minutes that were submitted as CBI; the Board must rely on Mr.
Carley to assure the adequacy of these procedures because the redacted
minutes provided to the Board were weak and difficult to assess. 
Because of this, it is difficult for the Board to determine whether the
EIRB was qualified and adequately considered all ethical issues.

With respect to study design, subject risks arise from exposure to the
test materials, bites from insects, and exposure to arthropod-borne
diseases.  Adequate stopping procedures are in place to minimize and
respond to reactions to test compounds.  Dr. Philpott expressed concern
that the materials were described as Toxicity Category IV, although the
materials are Toxicity Category II/III for acute or ocular exposure. 
Dr. Philpott assumed that these statements would be corrected based on
the June 18, 2007 response.  The issue of counting landings versus bites
is relevant with regard to reactions to bites; however, most mosquito
bites are minor and over-the-counter remedies are provided to
participants to combat minor reactions.  Exclusion of subjects with
previous reactions to bites or skin care products further mitigates
risk.  To minimize exposure to disease, the protocol proposes proceeding
only if no disease is detected for at least 1 week before field testing
occurs.  ICR believes that disease-free conditions for 1 month are
unnecessary; however, a negative result the week before testing does not
necessarily mean that pathogens are not present at the time of the field
test.  Establishing a temporal pattern of lack of disease in test areas,
perhaps through weekly monitoring for a month, would provide better
protection of subjects.

The ICFs discussed these risks with subjects by describing specific
diseases present in each area in 2006.  The forms also indicate that no
human cases of specific disease were detected; this is deceptive because
it refers to human disease, not to pathogens present in mosquitoes. 
Also, including the statement concerning conditions in 2006 is
misleading and may inappropriately minimize the participants’
perceptions of risk.  This statement should be deleted and replaced with
a detailed description of risk, symptoms of possible arthropod-borne
diseases, and plans for monitoring and follow-up.  The need to exclude
pregnant women raises issues of protecting confidentiality if a subject
tests positive for pregnancy after travel to the test site.  The HSRB is
unable to provide a solution for this, except perhaps use of a
self-conducted test before travel.

Dr. Philpott commended ICR on addressing the issue of exclusion of ICR
family members and part-time and full-time employees.  He recommended
that ICR add descriptions of the informed consent process.  Current
documents explain that subjects will initially be contacted by telephone
and states that the investigators will review the specifics of the
studies with the subjects, but it is unclear whether another discussion
of the risks and benefits of participation or methods of the study is
held.  Informed consent is a process and investigators should constantly
discuss risks and benefits with subjects.  Considering the selection of
untreated controls by lot, a clear discussion of the greater risk to
untreated controls should be provided in the ICF documents.  The lottery
approach to selecting controls also could be reconsidered.

Dr. Sharp noted that using landings as indications of efficacy failure
appears to be more appropriate and minimizes risk.  He asked whether the
need to travel to a distant site for testing would hurt the ability of
subjects to withdraw from the trial once they are at the site; this
could establish a potentially coercive situation.  Given the small
number of subjects, local recruitment might provide a better alternative
and would be better from an ethical perspective because of decreased
potential for coercion.

Dr. Menikoff agreed with Dr. Philpott concerning resolution of the issue
of landings versus bites.  The ICFs need revision, including
descriptions of how untreated controls are selected and a subject’s
chance of being selected as an untreated control.  Dr. Menikoff was less
concerned about coercion because ICR stated their willingness to pay for
subject accommodations or travel expenses even if the subject withdraws
from the study at the site.  Dr. Lebowitz added that the Board should
recommend participants be limited to those under 55 years of age.

Dr. Chadwick agreed with his colleague’s assessments.  He asked
whether a more conservative measure of CPT was in fact more protective
of consumers, because it might prompt consumers to apply the product
more frequently, thus exposing themselves to greater amounts of
chemicals.  He also echoed Dr. Sharp’s concerns about the preference
for flying participants to a distant site rather than recruiting
locally.  He stated that there is no apparent scientific justification
for the travel, and without this, there can be no ethical justification
of travel.  He added that although EPA prefers tests to be conducted in
different areas, in his opinion the two sites selected for this study
are not sufficiently different to meet this preference.

Dr. Chambers noted that the protocol indicates that landing mosquitoes
will be dislodged from untreated controls; therefore, risk to the
untreated controls is less than risk to treated subjects.  Dr. Philpott
agreed and reiterated that the risk to untreated controls must be
clarified.  Using experienced subjects with a greater appreciation of
the risks involved as untreated controls could be considered.  Dr.
Chambers also commented on the travel issue, noting that ICR chooses its
subjects from a pool of previous participants who have experience with
similar protocols.  Because these people are more cognizant of what is
required for these tests, they are less likely to be coerced to
participate than a naïve local volunteer.  She added that if the
mosquito species at the two sites are different, this justifies the use
of distant sites.  

Dr. Brimijoin asked the Board to indicate whether issues surrounding
travel should be considered major or minor.  Dr. Fish responded that she
understood Dr. Sharp’s concerns, but that Dr. Chambers’ point about
experienced subjects also should be considered.  Justification by the
investigators for use of distant test sites would be helpful.  The use
of experienced subjects could result in better data and such subjects
may be less open to undue influence because of their experience with
similar studies.  The transportation issue would therefore be less of a
concern.  Dr. Philpott clarified that two issues—travel and
compensation—could be considered undue inducement.  He speculated on
whether being at the site and staying in a hotel would coerce subjects
to remain in the study or to withdraw in hopes of participating in more
interesting activities at the site than the testing.  He stated that, in
his opinion, the Board should err on the side of assuming that these are
experienced volunteers and are likely to make sound judgments concerning
their participation.  Dr. Brimijoin asked Dr. Sharp to indicate whether
he considered the travel issue to be a significant problem and to
describe how the protocol could be revised to address the issue.  Dr.
Sharp responded that if ethics require maximizing a subject’s ability
to withdraw, testing at the distant sites could make this more difficult
and an alternate design (i.e., use of local volunteers) should be
considered.  The subjects also participate in the study for 6 days,
which is longer than if local volunteers were used.  Traveling to
distant sites also carries the risk of flying and of health issues
unrelated to the study arising in a subject far from home.  Using local
volunteers would maximize subjects’ ability to withdraw and minimize
other travel-associated risks.  Dr. Brimijoin noted that early versions
of a Carroll-Loye protocol described having subjects travel to a
potential distant test site; this would seem to imply a perceived
benefit in relying on experienced subjects or difficulty recruiting
suitable subjects at the test site.  The value of experienced subjects
versus the risks of travel must be considered.  Dr. Brimijoin concluded
that the Board’s consensus is that there is concern about the ability
of subjects to withdraw from the study held at a distant test site; ICR
should consider the advantages and disadvantages of using local
volunteers.

Mr. Carley clarified EPA’s requirement for testing in two areas. 
Investigators also may choose multiple sites based on the times of year
during which the study will be performed to maximize the presence and
activity of relevant mosquito species.  In response to a question from
Dr. Chadwick, Mr. Sweeney clarified that the Florida site is mangrove
swamp and the Georgia site is a more open habitat.  The mosquito species
present at each site are different.  Testing is not performed in ICR’s
home state of Maryland because there are insufficient sites with
consistently high biting pressure.  Additionally, many of these areas
are located in national or state parks and special permits would be
required to test at these sites.  The mosquito season is longer in the
southern states and better biting pressure exists.

Dr. Carriquiry asked whether, given EPA guidelines indicating use of
TFCB for efficacy studies, EPA could now request that investigators use
time to first landing instead.  Mr. Jordan clarified that the guidelines
are recommendations and are not requirements that companies or EPA must
follow.  Those involved in repellent testing understand that much
discussion is occurring around this issue and that EPA is currently
trying to determine the best choice of endpoint.  The Board’s opinion
appears to indicate that use of landings is ethically defensible,
scientifically acceptable, and results in a more conservative endpoint
and thus may be preferable to using bites.  If EPA concurs, this change
will be communicated to the relevant investigators.  Mr. Carley added
that any changes in guidelines would be implemented in a prospective
manner.

Dr. Philpott was in agreement with Mr. Carley regarding deficiencies in
the protocol.  The issue of the use of landings or bites as endpoint is
under consideration by EPA.  The Board expressed concerns that a more
detailed protocol to conduct analysis of the presence of pathogens in
the relevant mosquito populations and a plan for informing subjects if
disease is detected are needed.  Several Board members expressed concern
that the absence of arthropod-borne pathogens for 1 week prior to the
study is inadequate and a larger continued period of monitoring is
desirable.  Concerning informed consent, the investigators should focus
more fully on the consent process and ensure that continued discussion
and revisiting of risk with subjects occurs.  The ICF appears to
minimize risk because it refers only to human cases of disease and not
to positive mosquitoes; this statement should be changed to more clearly
reflect true risk.  ICFs for subjects and controls should be separated
because each group has different risks.  Issues of confidentiality
regarding pregnancy tests for subjects should be clarified.  Regarding
the requirement of travel to the test site, this may not be coercive but
does complicate subjects’ ability to withdraw.  Further justification
of this aspect of the study design or a different recruitment plan
(i.e., local volunteers) may be needed.

Mr. Carley clarified that, concerning separate ICFs for subjects and
controls, controls are not initially identified but instead are chosen
by lot just prior to commencement of testing.    Because landings are
considered the endpoint for controls and bites for treated subjects, Mr.
Carley recommend using landings as evidence of repellency breakdown; if
landings are sufficient to measure biting pressure, they also should be
sufficient to measure breakdown.

Dr Brimijoin summarized the ethical discussion by commenting that the
Board review had been thorough and largely agreed with Mr Carley’s
analysis of minor deficiencies that need remedying and were actually in
the process of being remedied.  Among these were the need for more
detailed protocols for conducting analysis of virally positive
mosquitoes and how the resulting information will be transmitted and
used.  Also mentioned in the summary were issues relating to the use of
multiple test sites and the travel involved, the need for separate
consent forms for control subjects, a recommended age limit of 55, and
the appropriateness of measuring landings versus bites. 

Day 2 Opening Remarks

Dr. Edwards expressed her appreciation for the Board’s work and its
importance to OPP.  She recognized the difficulty of the HSRB’s tasks;
discussing tests of potential toxins on humans is bound to be
controversial and complicated.  She commented that she would be
attending an interagency conference on migrant worker and farm worker
exposure to pesticides and the importance of protecting these workers to
the fullest possible extent.  Sound risk assessment requires using the
best science and ethics available and OPP is interested in ensuring that
it has the best possible data for its work.  Dr. Edwards also commented
on the importance of the Board’s work in helping EPA ensure the
scientific and ethical validity of research to test repellent efficacy,
which is important for public health.

Follow-up Comments

Mr. Jordan thanked the Board for the previous day’s discussions.  EPA
had no follow-up questions at this time.

Acrolein Inhalation Studies

Introduction

Mr. Carley provided background information on acrolein inhalation
studies by Weber-Tschopp et al (1977).  EPA seeks to use information
from this article to assess human risk resulting from potential acute
inhalation exposure to acrolein.  Acrolein is a strong irritant that is
used as a biocide to control growth of underwater plants, such as algae
and slime growth in irrigation canals; it also has modest use as a
pesticide.  Acrolein is a chemical intermediate in the production of
methionine, acrylic acid, and acrylates.  The article, Experimentally
Induced Irritating Effects of Acrolein on Man, (unpublished English
translation of “Experimentelle Reizwirkungen von Akrolein auf den
Menschen”) by Weber-Tschopp, A.; Fischer, T.; Gierer, R.; and
Grandjean, E. published in International Archives of Occupational and
Environmental Health 1977 (40): 117-130 contains three sub-studies that
will be discussed regarding EPA’s use of the information contained
therein for risk assessment studies.  This work was conducted in the
mid-1970s at the Institute for Hygiene and Occupational Physiology,
Swiss Federal Engineering College, Zurich.  The investigators were
active in studies of tobacco smoke and the study was funded by the Swiss
Association of Cigarette Manufacturers.  The investigators are no longer
active at the Institute.

The three sub-studies examined:  A) continuous exposure to acrolein
concentrations increasing over 40 minutes from zero to 0.60 parts per
million (ppm); B) 90-second exposures separated by 8-minute recovery
periods to concentrations increasing from 0.15 to 0.60 ppm; and C)
continuous exposure over 60 minutes to constant concentration of 0.30
ppm.  All sub-studies were conducted in a 30-cubic meter (m3) chamber
and the subjects were described as “healthy college students.” 
Subjective measures of annoyance were subject responses to questions
about air quality (“good”, “acceptable”, or “poor”), a wish
to leave the room (“no”, “don’t know”, or “yes”), and
perceived eye, nose, and throat irritation (1=not at all; 2=a little;
3=medium; 4=strong).  Objective measures of response were recorded only
for tests A and C and included measurement of eye blink rate for two
subjects in each group of three subjects and respiration rate and depth
for the remaining subject in each group.

Scientific Considerations

Dr. Abdallah Khasawinah (OPP, EPA) provided the scientific assessment of
the acrolein study.  Acrolein is an acrylic aldehyde that is highly
reactive and highly toxic by all routes.  It can cause significant
irritation and participates in lipid peroxidation and metabolism of
α-hydroxyamino acid.  Acrolein is highly volatile with a low boiling
point.  Because it is highly water soluble, acrolein is suitable for use
as a pesticide in irrigation canals.

Exposure to acrolein in the Weber-Tschopp experiments took place in a
large (30m3) climatic chamber.  Acrolein was vaporized and blown into
the chamber using a carrier gas, and measurements were taken to confirm
the acrolein concentration in the chamber.  Three subjects were exposed
at a time for Tests A and C and four subjects were exposed for Test B. 
Subjective irritation and annoyance were evaluated based on graded
responses to questions (Tests A, B, and C), blink rate (Tests A and C),
and respiratory rate (Tests A and C).  Respiratory rate was measured
during the entire test and the mean value for 3-minute periods was
reported.

Test A involved continuous, 40-minute exposure to increasing
concentrations of acrolein.  The test exposed 31 male and 22 female
subjects, in groups of three.  The initial acrolein concentration was 0
and rose continuously to 0.60 ppm after 35 minutes.  Subjective
annoyance/irritation questions were posed every 5 minutes and the blink
rate was measured for two subjects in each group every 5 minutes;
respiratory rate was measured continuously for the third subject.  The
results of Test A indicated that subjects experienced eye irritation at
0.09 ppm, nose irritation at 0.26 ppm, throat irritation at 0.43 ppm,
elevated blink rate at 0.26 ppm, and depressed respiration (by 25
percent) at 0.60 ppm.

Test B involved intermittent exposures to increasing concentrations of
acrolein.  It is unclear whether the same subjects who participated in
Test A also participated in Tests B and/or C.  In Test B, 17 male and 25
female subjects, in groups of four, experienced 1.5-minute exposures to
acrolein at 0 (control), 0.15, 0.30, 0.45, and 0.60 ppm.  The subjects
were permitted an 8-minute recovery period in a well-ventilated room
between exposures and were asked the subjective evaluation questions
after each exposure.  The results of Test B found that the
subjects experienced eye irritation at 0.30 ppm and nose irritation at
0.60 ppm, but no evidence of throat irritation.

For Test C, 21 male and 25 female subjects in groups of three were
exposed to a constant concentration (0.30 ppm) of acrolein for 1 hour. 
Measurements were taken initially (0 ppm control) and throughout the
exposure period as in Test A.  Subjective questions were asked every 5
minutes, eye blink rate was measured every 5 minutes (two of three
subjects), and respiration was continuously monitored (one of three
subjects).  Test C found that subjective irritation responses rose
sharply over 20 minutes and then reached a plateau.  The blink rate
doubled after 10 minutes and then plateaued, and the respiratory rate
decreased 20 percent after 40 minutes of exposure.

From these three tests, the authors concluded that the average threshold
of irritation sensations ranged from 0.09 ppm (eye irritation) to 0.30
ppm (respiration rate and throat irritation), with nasal irritation at
0.15 ppm.  Irritation was significantly stronger in the event of
continuous exposure and no adaptation to the effects was observed.

EPA also considered an animal study of sub-chronic acrolein exposure to
compare the sensitivity of different species to this irritant.  Feron,
et al (Integrated Risk Information System [IRIS] pg 27, 1978) exposed 20
hamsters, 12 rats, and four rabbits per sex per dose to different
concentrations of acrolein.  Concentrations of acrolein were 0, 0.4,
1.4, or 4.9 ppm and exposure was for 6 hours per day, 5 days per week,
over 13 weeks.  The results of this study found that the highest dose
(4.9 ppm) resulted in all animals closing their eyes, hamsters
experienced increased salivation and nasal discharge, and rabbits showed
evidence of sneezing and breathing difficulties.  Rats had bristling
hair; six of these animals died.  Depressed growth and diminished food
intake was observed in rats and animals at the mid- and high-level
doses.  Changes in organ-to-body weight ratios (lungs, hearts, kidneys,
adrenal glands) were observed at high doses, and hematological effects
were noted in female hamsters (increased red blood cells, packed cell
volume, hemoglobin, and lymphocytes; decreased neutrophilic leukocytes)
at high doses.  The lungs of the rats that died had evidence of
hemorrhage and collapsed dark reddish-purple areas.  Histopathological
findings in the respiratory tract indicated destruction and hyper- and
metaplasia of the epithelial lining accompanied by acute and sub-acute
inflammation in the nasal cavities of rats at all doses and in rabbits
and hamsters at 4.9 ppm; trachea and larynx of rats and hamsters at 4.9
ppm; and in the bronchi of rats and rabbits at this dose.  At the lower
dose (1.4 ppm), hamsters and rats fell asleep and rabbits sneezed;
there was decreased body weight gain and diminished food intake in rats
and rabbits.  No abnormal behavior was observed at 0.4 ppm, which was
thus determined to be a “minimal” lowest observed adverse effect
level (LOAEL), based on nasal effects in rats.

	The IRIS evaluation of this work determined a reference concentration
(RfC) for chronic exposure derived from the Feron et al study LOAEL of
0.4 ppm causing nasal effects in rats.  Adjusting for exposure duration
and to human equivalents, and applying a 1000X uncertainty factor, the
RfC is 0.00002 mg per cubic meter (m3).  A chronic RfC of 0.00002 mg/m3
is equivalent to 0.000009 ppm.  Starting from the Weber-Tschopp human
study, an Agency for Toxic Substances and Disease Registry (ATSDR)
evaluation derived an acute-duration Minimum Risk Level of 0.003 ppm
based on nose and throat irritation and decreased respiratory rate at a
LOAEL of 0.3 ppm and uncertainty factors totaling 100.  Comparison of
human and animal studies found that humans are more sensitive to the
irritating effects of acrolein and the minimal concentration for acute
irritating effects in humans (ocular) was lower (0.09 ppm for humans
versus 0.4 ppm for nasal irritation in rats, 1.4 ppm for hamsters, and
4.9 ppm for rabbits).

The Weber-Tschopp et al study provides the most suitable point of
departure for acrolein risk assessment activities because it involved
large numbers of healthy young subjects, used multiple exposures, and
confirmed exposure concentrations analytically within 10 percent. 
Shortcomings of this study include use of subjective evaluations, lack
of blind controls, and biased reactions; these are considered by EPA to
be minimal.  This work also determined that ocular effects are the most
sensitive indicator of irritation and are also seen in the animal
studies summarized in the ATSDR and IRIS reports.  The evidence shows
that humans are more sensitive to acrolein than experimental animals.

Based on this analysis, EPA concludes that 0.09 ppm (0.2 mg/m3) is the
minimal threshold for ocular irritation in humans and is the appropriate
point of departure for assessing acute exposures.  Uncertainty factors
of 10X for intraspecies variability and 3X for “minimal” threshold
are appropriate.

Dr. Lois Lehman-Mckeeman inquired how the odor threshold of 0.16 ppm was
determined.  Dr. Khasawinah responded that this was not specifically
stated and speculated that perhaps a middle value was selected.  Dr.
Lehman-Mckeeman questioned how acrolein usually is used, given that
these studies examined acute exposures.  Dr. Khasawinah explained that
EPA is assessing the effects of acrolein in water on bystanders,
including agricultural workers, because of potential evaporation.  Ms.
Becky Daison (OPP, EPA) explained that acrolein is injected below the
water surface in canals, but there is evidence of ambient
concentrations.  EPA’s concern is the effects of this on workers, who
can spend between 15 minutes and 8 hours applying acrolein to water;
exposure during set-up and take-down also is an issue.  EPA also is
concerned about use of acrolein in canals near residential areas because
of potential evaporation and off-gassing.  EPA is concerned about the
duration of exposure during the application process and speculates that
most canals would be treated on 26 days over the course of a year,
exposing workers to short-term intermittent exposures 3 to 4 times a
year.  In response to a question from Board member Dr. Suzanne
Fitzpatrick, Ms. Daison explained that the maximum level of acrolein in
water is 15 ppm, but this value depends on the plants present in the
water and physical parameters, such as flow rate.  Dr. Brimijoin
inquired how the data are used, given the discrepancy between acute
human exposures with a low LOAEL and higher sub-chronic exposure limits
in animals.  Ms. Daison explained that the durations of exposure in the
human studies more closely mimic real-life use of acrolein, so these
values are more appropriate to use when modeling exposure.

Ethical Considerations

Mr. Carley provided the ethics review of the Weber-Tschopp et al study. 
The EPA ethics review on May 22, 2007 considered the publication
Weber-Tschopp, A.; Fischer, T.; Gierer, R.; and Grandjean, E. (1977)
Experimentally Induced Irritating Effects of Acrolein on Man
(unpublished English translation of “Experimentelle Reizwirkungen von
Akrolein auf den Menschen”) International Archives of Occupational and
Environmental Health 1977 (40): 117-130.  This is a report of
third-party research involving intentional exposure of human subjects
that measured toxic endpoints.  The research was not conducted with the
intention to submit it to EPA under the pesticide laws.  EPA retrieved
the article from the published literature; it has not been submitted by
a regulated entity.  Because the document was not submitted to EPA, 40
CFR §26.1303, which requires documentation of ethical conduct of
studies submitted after April 6, 2006, does not apply.  40 CFR
§26.1602(b)(2) requires HSRB review, 40 CFR §26.1703 forbids EPA
reliance on research involving intentional exposure of pregnant or
nursing women or children, and 40 CFR §26.1704 forbids EPA reliance on
pre-rule research if there is “clear and convincing evidence” that
its conduct was fundamentally unethical or significantly deficient
relative to standards prevailing when it was conducted.

The stated purpose of this research was to characterize the relative
involvement of acrolein in the effects of air pollution caused by
cigarette smoke.  The authors concluded that acrolein was not a
significant contributor to the irritancy of cigarette smoke.  They
reported threshold irritancy effects for pure acrolein at measured
concentrations.  Irritancy information shows humans to be the most
sensitive species and is of potential value to EPA in defining an
endpoint for assessing risk to humans from exposure to acrolein when it
is used as a pesticide.

The subjects participating in this research were described only as
healthy college students.  Roughly equal numbers of males and females
were enrolled in each of three tests (Test A:  31 males and 22 females;
Test B:  17 males and 25 females; Test C:  21 males and 25 females). 
The report does not describe how subjects were recruited, any
relationship between the investigators and the subjects; subject age and
reproductive status; the sequence in which Tests A, B, and C were
conducted; the time lapse between sub-studies; whether any subjects
participated in more than one sub-study; or whether subjects were
compensated for participating.  The primary risks to subjects were eye,
nose, and throat irritation from exposure to acrolein, a known strong
irritant.  The report does not describe the qualitative nature or
likelihood of risks, the probable duration or reversibility of effects,
or any steps taken to reduce risks to the subjects.  The research
offered no direct benefits to subjects, but societal benefits included
improved understanding of the threshold irritation effects of acrolein
in humans, confirmation of the U.S. Threshold Limit Value level for
8-hour exposure, and evidence that the Occupational Safety and Health
Administration (OSHA) limit for peak exposures was set too high.  Based
on the report, it is unclear what benefits were anticipated and
insufficient information is available to assess the risk-benefit
balance.

The report does not describe ethics oversight, independent ethics
review, or any standard of ethical conduct.  The report also does not
indicate whether or how participants were informed or how their
voluntary consent to participate was obtained.  The privacy of the
subjects was not compromised in the published report.  A common
subjective response measure was reported as a “wish to leave the
room;” during the 60-minute continuous exposure to 0.3 ppm acrolein,
50 percent of subjects had a “wish to leave the room” after 10
minutes, rising to 72 percent after 20 minutes, but the report does not
indicate whether any exposures were terminated early because of a
subject’s expressed wish to leave the inhalation chamber.

No prevailing standard of ethical conduct was identified in the report. 
The prevailing standard thus is assumed to be the Declaration of
Helsinki (DoH), 1975, which prevailed in medical research in 1977, but
may not have been considered applicable to this research.  The basic
principles of the DoH state that the design of an experimental protocol
should be clearly formulated and evaluated by an independent committee;
biomedical research involving human subjects cannot be performed unless
the importance of the objective is in proportion to the risk to the
subject; every such project should be preceded by careful assessment of
predictable risks compared to foreseeable benefits to the subjects or to
others; and any potential research subject must be adequately informed
of the aims, anticipated benefits, and potential hazards and discomforts
of the study, and should be clearly informed of their right to abstain
from participation or withdraw consent to participate at any time.  For
non-therapeutic research, the DoH states that the subjects should be
volunteers and the research should be discontinued if judged to be
harmful.  The interest of science and society should never take
precedence over considerations related to the well-being of the
subjects.

Ethical concerns indicate a lack of evidence that the protocol was
evaluated by an independent committee; however, this is typical for
published human studies from this time period and does not constitute
“clear and convincing evidence” of failure to adhere to this
regulation.  The report also does not indicate whether the importance of
the objective was in proportion to the inherent subject risk or whether
the project was preceded by assessment of risks in comparison to
benefits; again, clear and convincing evidence of deliberate failure to
adhere to these regulations is lacking.  Other ethical concerns based on
the published report include a lack of information on recruitment and
informed consent procedures and whether exposures were ended early
because of a subject’s wish to leave the inhalation chamber.  Although
the investigators were aware of new U.S. standards for occupational
safety (0.1 ppm for an 8-hour total weight average [TWA] and 0.3 ppm
short-term exposure limit [STEL] for up to 15 minutes at a time, up to 4
times per day).  If the rate of increasing concentration in sub-study A
was constant over 35 minutes from zero up to 0.6 ppm, then the
concentration was greater than the 15-minute STEL for about 23 minutes
out of the 40 minutes in that sub-study.  In test C, exposure for 60
minutes at 0.3 ppm was at the maximum permitted by the U.S. (4 by
15-minute a day STEL).  So it was within it, but the quota for the day
was consumed in that one hour. 

40 CFR §26.1703 forbids EPA to rely on research involving intentional
exposure of pregnant or nursing women or children; the subjects in this
research were described as college students, and were thus likely to
have been at least 18 years old.  Roughly half the subjects were female,
but the report does not indicate their reproductive or nursing status. 
When evidence concerning subject age and reproductive status is both
absent and unobtainable, it is EPA’s policy that §26.1703 does not
prohibit reliance on a study.  40 CFR §26.1704 forbids EPA to rely on
pre-rule research if there is “clear and convincing evidence” that
its conduct was fundamentally unethical or significantly deficient
relative to standards prevailing when it was conducted.  Very little is
known about the ethical conduct of this research, and information to
clarify ethical questions is not available.  However, in the absence of
“clear and convincing evidence” there is no regulatory barrier to
EPA’s reliance on this study.  

The Agency has concluded that this study contains information sufficient
for assessing human risk resulting from potential acute inhalation
exposure.  The HSRB was asked to comment on whether the study is
sufficiently sound, from a scientific perspective, to be used to
estimate a safe level of acute inhalation exposure to acrolein, whether
there is clear and convincing evidence that the conduct of the study was
fundamentally unethical, and whether there is clear and convincing
evidence that the conduct of the study was significantly deficient
relative to the ethical standards prevailing at the time the research
was conducted.

Dr. Philpott noted that one stated societal benefit pertained to
evidence that the OSHA limits were set too high and asked whether this
indicates that the current limit is too high and EPA is seeking to lower
it.  Mr. Carley explained that the authors concluded that the OSHA
limits were too high and claimed that checking these values was a
societal benefit; however, this idea may have developed after completion
of the study and may not have influenced the initial analysis of risks
and benefits.  Dr. Philpott inquired how EPA would use information from
this study.  Mr. Carley responded that the information would be used as
a point of departure for assessing risk to bystanders and workers of
intermittent short-term inhalation exposure to acrolein.  Dr. Brimijoin
asked if EPA was attempting to establish a point of departure where none
exists or to refine an existing point of departure value.  Ms. Daison
answered that using this data would result in a more protective value
than those permitted by existing OSHA rules.

Dr. Sharp asked whether it was appropriate for the HSRB to consider EPA
use of this material.  Dr. Brimijoin questioned what EPA’s response
would be if the HSRB found a completed study to be scientifically
excellent and indicative of grossly unprotective existing regulations,
but obviously deficient relevant to applicable ethical standards.  Mr.
Jordan answered that current regulations allow EPA to use unethical
studies to establish more stringent standards for human health
protection.  As part of this, the views of the HSRB must be sought and
time must be provided for public comments concerning EPA reliance on
potentially unethical studies.  EPA must demonstrate that the
information is crucial to improve public health, that changes in
standards cannot be justified without using the unethical study, and
must publish its discussions and findings concerning these matters.  In
the case of the acrolein studies, it is not clear whether use of this
information would result in lowering of standards.  EPA currently is
discussing appropriate uncertainty factors to use in its risk assessment
and concluded that the study provided an appropriate point of departure.
 This combined with the uncertainty factor leads to a lower reference
dose and to more stringent regulatory controls on the use of the
product.  This meeting fulfills EPA’s requirement to obtain the views
of the HSRB before use of such information.

Public Comments

Dr. Brimijoin invited oral public comment on the Weber-Tschopp et al
Human Study.  No oral public comments were received.

Board Discussion

	Scientific Considerations—Acrolein

Dr. Lebowitz opened the scientific discussion of the acrolein studies. 
He stated that he was familiar with this research because in 1981, he
chaired an EPA advisory committee on pollution from tobacco smoke
research in which these reports were used as evidence.  Dr. Lebowitz
complemented OPP’s review of the studies.  The study measured ocular
irritation objectively and also noted other signs and symptoms of
exposure.  Noticeable irritation occurred within seconds at 0.3 ppm. 
High airborne concentrations can result in increasingly severe
irritation, which is an effect of acrolein’s high solubility.  Because
of potential exposure to the general population through the atmosphere,
OSHA recommends a limit of 0.1 ppm in a room.  The ATSDR toxicity
profile shows that comparable cellular changes in the human nasal
passages and initiation of human irritation occur at the same level of
acrolein exposure.

	Concerning the Weber-Tschopp study, Dr. Lebowitz emphasized the careful
measurement and reproducibility of the acrolein concentration within the
test chamber as a strength of the study.  The chamber itself was
well-designed and operates in a defined manner.  The chemical analysis
in the report was appropriate and accurate.  Healthy subjects were used,
although the degree of subject overlap among the three studies is
unknown.  Stopping rules likely were in place and intermittent exposure
was sufficient to cause acute effects and allow determination of a
LOAEL.  The recorded measurements—eye blink, respiratory rates—were
appropriate.  Dr. Lebowitz agreed with the weaknesses of the study
outlined by Dr. Khasawinah.  Dr. Lebowitz concluded that, in his
opinion, the study contains information sufficient for assessing human
risk arising from potential acute inhalation exposure to acrolein.

Dr. Lehman-Mckeeman continued the discussion by agreeing that the study
does provide scientific data that can be used to determine a point of
departure for acrolein risk assessment activities.  Dr. Lehman-Mckeeman
stated that she did not agree with EPA’s assessment that humans are
more sensitive than animals.  The studies used different measurements;
human indications of distress were subjective and it is unknown whether
rats experience distress earlier than can be noted by a human observer
and before histopathological lesions are evident.  The numerical values
derived from the data are meaningful and it is a strength that these
values are consistent with current OSHA standards.  Dr. Lehman-Mckeeman
concluded that the data have been generated in a scientifically valid
manner and can be used by EPA.

Dr. Kim stated that his conclusions were similar to those of Drs.
Lebowitz and Lehman-Mckeeman.  The study is reasonable and sound enough
to use its data as a point of departure for EPA risk assessment
activities.  He noted that the studies appear to have a large sample
size, but there is no justification for choosing those sample sizes. 
False negative rates cannot be assessed, nor can the significance of the
differences in reactions noted at different concentrations of acrolein;
however, a lack of significance is not a lack of effect.  Dr. Kim noted
that the study participants were close in age and asked whether there is
any evidence indicating that human sensitivity is age-dependent.  Dr.
Lebowitz answered that the age of the participants (likely between 18
and 26 years) was typical for human exposure studies.  He added that the
authors chose exposure levels based on EPA evidence concerning ambient
exposure at 3, 6, or 9 parts per billion.  In the presence of tobacco
smoke, indoor air concentrations of acrolein can be significantly higher
than this.  Dr. Lebowitz informed the Board that the researchers had
received funding from a tobacco company and also that an independent
review of the research had been performed.

Dr. Fish asked Dr. Lebowitz to comment on any stopping rules applicable
to this research.  Dr. Lebowitz remarked that the independent scientific
review group always considered this issue and insisted on stopping rules
for safety.  The stopping rules for this study were adequate for
stopping the research for indications of asthma, serious reactions, and
acute morbidity.  The issue of whether irritation occurred at or before
the levels indicated in the study must consider cultural issues and also
the year in which the study was conducted, with respect to typical
exposure to tobacco smoke containing acrolein.  Dr. Lebowitz added that
when studies are performed outside of the United States, differences in
perceptions of risk and whether applicable rules were sufficient must be
considered.

Dr. Brimijoin summarized by thanking Dr. Lebowitz for his insights into
the science and the investigators who performed this research.  The work
was considered to be innovative and of high quality, with a strong study
design and validation of test concentrations.  The study was
well-controlled and gave reliable and usable results.  Dr.
Lehman-Mckeeman concurred with Dr. Lebowitz and observed that humans
and animals likely have similar sensitivity to acrolein.  Dr. Kim also
agreed, but noted that the statistical analysis used had no
justification for sample size; thus the low end of the effects cannot be
accurately determined.  There was strong consensus among Board members
that the study was scientifically reliable. 

	Ethical Considerations—Acrolein

Dr. Fish opened the ethics assessment by noting that there is much that
is unknown about this study.  The number of individual subjects used
could range from 56 to 141 and how many may have participated in more
than one test is unknown.  There also is little information about the
subjects themselves; whether any were students or employees of the
investigators; how they were recruited; and what they were told about
risks, their freedom to withdraw, or the informed consent process.  The
description of expressing a “wish” to leave is problematic because
it does not necessarily indicate that a subject would leave; a subject
may have “wished” to leave but did not actually leave.  There is no
information concerning compensation, possible undue influence, or the
applicable version of the DoH.  Dr. Lebowitz stated that the study was
reviewed independently for science, but did not indicate whether an
independent ethics assessment had been performed.  There was no
justification for the sample size, which raised questions as to whether
more subjects than necessary were placed at risk.  Another issue
concerns whether the study should have been stopped when indications of
a wish to leave (after 10 and 20 minutes) became evident.  The subjects
were tested in groups of three and it is unclear whether the
investigators analyzed the data as it was obtained to determine
irritation levels or if the analysis was not performed until all
subjects had been tested.  Thus, it is unknown whether the study should
have been stopped sooner, as accumulating evidence suggested irritation
was occurring.  Despite this, Dr. Fish indicated that she agreed that
stopping rules for subject safety likely were in place.  Dr. Fish
concluded that there is not clear and convincing evidence that the
studies were unethical or deficient relative to existing standards.

Dr. Sharp stated that review of this research was challenging because of
the lack of documentation.  He noted that there was no clear evidence
that vulnerable subjects were adequately protected and appropriately
recruited, but this is not grounds for concluding that the study did not
comply with existing standards.  Concerning the ability to opt out of a
study, it is disconcerting to note that 72 percent of subjects expressed
a wish to leave the chamber at 20 minutes but did not withdraw from the
study.  A further explanation of this would be helpful, but does not
indicate that the study was not compatible with existing standards. 
Statements concerning independent review by external advisory panels
generally were not included in publications in 1975, thus, the lack of
this information in the manuscript is not evidence that such a review
did not take place.  

Dr. Sharp said that the study was significantly deficient relative to
prevailing ethical standards, particularly DoH Part I-9.  This standard
calls for evaluation of the importance of the objective relative to
subject risk.  In this work, acrolein was characterized as “highly
toxic” and there was no intent of the research to provide therapeutic
benefit or diagnostic results, and no benefit at all to the subjects. 
Intentional exposure to this highly toxic substance thus constituted an
inappropriate risk-to-benefit ratio.

Dr. Philpott commented that he was less sure of the ethics of this study
than Dr. Fish, but not as concerned as Dr. Sharp.  Part I-9 of the DoH
is a topic of significant debate concerning its exact meaning.  The
importance of the objective relative to subject benefit is subjective. 
Dr. Philpott’s main concern was with the informed consent process;
the way EPA’s human studies rule is written with respect to completed
studies is complicated.  Dr. Philpott agreed with Dr. Fish that there
was no clear and convincing evidence that the study was significantly
unethical or deficient relative to existing standards.  He would
strongly recommend to EPA as they decide whether or not to use this
information that they carefully consider whether use of this information
would lead to more protective standards and consider whether the
information from the animal studies would suffice for its risk
assessment work.

Dr. Lebowitz clarified that any information he provided on the
laboratory, its investigators, informed consent, compensation, or
recruitment was based largely on hearsay.  He was not a reviewer of this
particular study.  He noted that he reviewed later studies with similar
designs, but these reviews did not include a thorough ethical review as
performed by the HSRB.  His statement concerning independent review was
factual.  Concerning the number of students, Dr. Lebowitz noted
speculation on overlap of subjects among studies was based on
information he had been provided with in previous reviews but is not in
EPA’s documents.  He also added that stopping rules were always
required by the sponsors of this research and the university at which
the research was performed.

Dr. Chadwick expressed his discomfort with reviewing a study for which
significant information is lacking.  The Board does not know when the
study was performed and thus does not know which version of the DoH
applies; the 1964 DoH was significantly revised in 1975.  The DoH was
written by physicians with them as its primary target, thus the document
may not apply to non-medical facilities.  Dr. Kim clarified that the
stopping rule referred to safety of the subjects, and asked whether
“stopping” meant the study stopped or the subjects left the chamber.
 Dr. Lebowitz remarked that the investigation of that particular subject
was stopped, but the study itself did not come to an end.  He added that
Weber was the primary investigator who observed the subjects at all
times.  In response to a question from Mr. Carley, Dr. Lebowitz
clarified that none of the investigators were medical doctors, but
medical doctors participated in the review and screening of subjects for
similar studies, although he could not state definitively whether such
screening was performed for this study.

Dr. Fish asked the Board to discuss the meaning of “highly toxic” as
related to acrolein exposures that took place in this study.  Dr.
Lebowitz said that higher airborne concentrations (between 2 and 5 ppm)
result in increasing irritation over the entire respiratory tract. 
Relative to this, the doses used in this study were not considered
highly toxic.  Dr. Lehman-Mckeeman noted that acrolein is an aldehyde
and a reactive moiety that can do damage at the point of contact;
because of this, there are minimal systemic effects associated with
acrolein.  Evidence of this is in the animal studies, in which
inhalation resulted in damage primarily to the respiratory system
(changes in adrenal gland weight likely were due to the stress of the
experiments).  She summarized that acrolein is highly irritating, has an
odor threshold, and manifests toxicity at the point of contact.  Dr.
Brimijoin asked Dr. Sharp to comment on his statement that the study is
significantly deficient with respect to ethical standards of the time
because of the failure of the study’s objectives to justify subject
risk.  Dr. Sharp remarked that his view had not changed; acrolein may
not cause systemic damage, but the harm posed to the subjects by
irritation is not offset by the purported societal benefit.  Dr.
Philpott commented that the timing of OSHA regulations regarding
acrolein exposure and the demands of DoH on such experiments should be
considered in the Board’s ethics assessment.  He added that Board
members appeared to have different opinions concerning the ethics of
this research, and perhaps should strongly suggest to EPA that the
Agency’s decision on whether or not to use this data rest on whether
the information will result in significantly increased protection.  Dr.
Menikoff reminded Board members that it is inappropriate to conclude
that there is clear and convincing evidence that the work was unethical
or did not meet the relevant standards of the time if documentation was
not available.

Dr. Brimijoin summarized that several Board members believed that the
study was not fundamentally unethical nor was it significantly
deficient.  Dr. Sharp believed the study to be fundamentally unethical
because of an inappropriate risk-to-benefit balance.  Dr. Chadwick
raised questions concerning whether the DoH would apply to this work,
and if not, what impact this would have on the views of Board members
who believe the study to be ethical.  If the net effect of using this
information is to increase human protection, EPA has mechanisms in place
to permit it to use the data.

Dr. Richard Fenske said that the DoH likely was the prevailing standard
at the time, and it is difficult to believe that the study did not
include medical supervision because it did involve humans; there was
likely to have been some degree of oversight from the medical community.
 Mr. Carley acknowledged that application of DoH Part I.4 required a
judgment call, driven by the available facts.  He suggested to Dr. Sharp
that it would be helpful if Dr. Sharp could identify the available
evidence that led him to conclude that the provisions of Part I.4 were
not met.  Dr. Sharp answered that the definition of acrolein as highly
irritating and known to be toxic at high concentrations was not
outweighed by any of the societal benefits documented in this report.

Dr. Lebowitz explained that information from this study was applicable
and considered during development of the Clean Air Act.  Data from this
and similar research was instrumental in developing tobacco smoke
exposure regulations and laws that banned tobacco smoke in many places. 
The societal benefits of these activities, in the forms of occupational
and societal regulations and worker protection standards, have been
substantial by reducing tobacco smoke exposure to many people.  This and
other studies were considered scientifically sufficient given the
standards of the time to justify placing regulations on tobacco smoke. 
Dr. Sharp remarked that assessing the benefits of what are post-hoc
results is complicated.  He was not convinced that at the time the study
was conducted, such benefits clearly outweighed subject risk. 
Dr. Brimijoin concluded that the Board’s report would indicate that
although a majority of members did not find the research to be
significantly ethically deficient, at least one Board member disagreed
with this assessment.

EPA Science and Ethics Assessments of Three Published Clinical Studies
on 4-Aminopyridine (4-AP)

Introduction

Mr. Carley presented background information on three clinical studies of
4-AP.  4-AP is a chemical intermediate in the production of certain
pharmaceuticals and agrochemicals.  It has been used as an experimental
drug for treatment of spinal cord injury (SCI), multiple sclerosis (MS),
and has orphan drug status for treatment of Guillain-Barre Syndrome. 
4-AP is a fast potassium channel blocker, improves axonal conduction in
demyelinated nerve fibers and is an antagonist to non-depolarizing
neuromuscular blocking agents.  4-AP also has pesticidal use as a bird
repellent (Avitrol).

	Three 4-AP clinical studies were considered in this review, including
the following:

Van Diemen, H., et al. (1993) 4-Aminopyridine in Patients with Multiple
Sclerosis:  Dosage and Serum Level Related to Efficacy and Safety. 
Clinical Neuropharmacology vol. 16 No. 3, pp. 195-204.  MRID 47093603.

Segal, J., et al. (1999) Safety and Efficacy of 4-Aminopyridine in
Humans with Spinal Cord Injury: A Long-Term, Controlled Trial. 
Pharmacotherapy 19(6):713-723, 1999.  MRID 47093602.

Grijalva, I., et al. (2003) Efficacy and Safety of 4-Aminopyridine in
Patients With Long-Term Spinal Cord Injury: A Randomized, Double-Blind,
Placebo-Controlled Trial.  Pharmacotherapy 23(7):823-834.  MRID
47093601.

The study by Van Diemen et al was conducted at the Free University
Hospital in Amsterdam in the early 1990s.  It was a randomized,
placebo-controlled, double-blind crossover study involving 70 patients
with MS.  4-AP was administered by both oral and intravenous (IV)
routes.  The authors had previously reported a possible therapeutic
effect of 4-AP on patients with MS; this paper reports the relationship
between dosage, serum level, efficacy, and side effects of 4-AP in the
same patients.  The study by Segal et al was conducted at the Veterans
Administration (VA) Medical Center in Long Beach, CA, in the late 1990s.
 This medical center had a registered IRB and holds a FWA from OHRP. 
The study was a randomized, active-treatment-controlled,
partially-blinded trial performed on 21 adult outpatients with traumatic
SCI of at least 2 years’ duration.  4-AP was orally administered.  The
article reports improvements in the patients’ condition with no
significant associated drug toxicity, and thus indicates thresholds for
side effects useful to EPA in assessing hazard.  The study by Grijalva
et al was conducted in 1999-2000 at the Instituto Mexicano del Seguro
Social (IMSS) in Mexico City.  It was a randomized, placebo-controlled,
double-blind crossover study in 27 adult outpatients with traumatic SCI
of at least 1.5 years’ duration.  4-AP was orally administered and
this report also details improvement in patients’ condition with no
significant associated toxicity.

Scientific Considerations

Dr. Khasawinah presented the science assessment of the three trials. 
4-AP is soluble in water and in organic solvents.  It is highly toxic
with an LD50 of 3.7 to 20 mg/kg.  Accidental ingestion of approximately
60 mg (0.86 mg/kg for a 70-kg adult) of 4-AP resulted in rapid onset of
symptoms such as weakness, dizziness, dyspnea, profound thirst, and
combative behavior.

The Van Diemen et al study enrolled 43 women and 37 men with MS, between
the ages of 23 and 68 years.  Phase 1 of the study involved IV
administration of 1mg 4-AP in 1 to 2 minutes over 20 minutes for the
first 4 hours and then 2.5 mg in 1 to 2 minutes for 20 minutes
afterward, to a maximum dose of 0.5 mg/kg.  The infusion duration was 60
to 260 minutes.  Phase II involved oral dosing of 4-AP as
nonenteric-coated capsules or placebo for 12 weeks, beginning 1 week
after Phase I.  Dosing started at 10 to 20 mg/day in 2 to 4 divided
doses, increasing by 5 to 15 mg/day at week 2 and 6 (or week 14 and 18)
to a maximum of 0.5 mg/kg/day.

Serum levels of 4-AP were measured in subjects who received IV
infusions.  The mean level 80 minutes after beginning infusion was 38.1
nanograms (ng) per milliliter (mL) (range = 24 to 45 ng/mL), 63.4 ng/ml
at the end of the infusion (range = 26 to 86 ng/mL), and 36.6 ng/mL
120 minutes after the end of the infusion (range = 22 to 58 ng/mL). 
Side effects included paresthesis in the infusion arm (observed at a
minimum dose of 1 mg), perioral paresthesis (1 mg), dizziness and
light-headedness (9 mgs), dizziness/lightheadedness and nausea/vomiting
(10 mgs), dizziness/light-headedness and a feeling of restlessness (9
mgs), and headache (20 mgs).  All side effects were reversed within
2 hours after cessation of infusion; seven patients reported no side
effects.

The oral phase of 4-AP administration resulted in a mean dosage of 31.2
mg per day (mg/day) (range of 10 to 50 mg/day; 0.17 to 0.55 mg/kg/day)
in 2 to 4 doses.  Average 4-AP serum levels were 53.6 ng/mL (range of 7
to 107 ng/mL) and increased 1.3 ng/mL/mg per day.  Side effects included
paresthesias/dysesthesias in 15 patients at a minimum oral dose of
5 mg/day; dizziness/light-headedness in 36 patients at 5 mg/day; gait
instability in 11 patients at 5 mg/day; nausea or vomiting in nine
patients at 5 mg/day; restlessness/anxiety in four patients at 5 mg/day;
abdominal pain in five patients at 10 mg/day; and obstipation in one
patient at 25 mg/day.  Side effects were reported after 30 to 45
minutes and resolved within 2 to 5 hours of 4-AP administration. 
Fourteen patients required dose reduction and three withdrew due to side
effects.  No side effects were reported by 15 patients, but one had
significant electroencephalography (EEG) changes.  More pronounced side
effects were observed if the oral 4-AP was administered when the subject
had an empty stomach.  The authors concluded that the patients showed
statistically significant improvement in smooth pursuit gait when
receiving 4-AP by either through IV or oral administration.  Improvement
was proportional to 4-AP serum level, with serum levels of 60 ng/mL and
above providing the best results.  The minimum oral LOAEL was
established as 5 mg/day.

The Segal et al study was a randomized, open label, dosage blinded trial
with active-treatment-control.  It enrolled 21 patients with SCI of at
least 2 years (14 tetraplegic and seven paraplegic).  4-AP dosages were
administered in increments of 2, 5, or 10 mg for 2 weeks as immediate
release encapsulated 4-AP.  Patients were divided into three groups. 
Group A had six patients who were dosage-blinded and 4-AP naïve;
dosage was titrated to 30 mg/day for 3 months.  Group B consisted of
five dosage-blinded, 4-AP naïve patients who received 6 mg/day in
divided doses for 3 months.  Group C had 10 patients who were
dosage-cognizant, had received 4-AP for more than 1 year before the
study, and received 30 mg/day for 3 months during this study.

Segal et al reported no clinically significant adverse effects.  Side
effects included nervousness, giddiness or dizziness, and
gastrointestinal upset as mild abdominal cramps or nausea.  Side effects
were transient, self-limited, or disappeared with changes in dosage or
timing of drug ingestion with meals.  Some patients reported enhanced
mood.  No seizure or seizure-like activity was observed or reported by
patients or caregivers at any time or at any dosage.  Serially acquired
EEG, electrocardiogram, biochemical and hematologic profiles, and
urinalyses remained within normal ranges.  Clinically meaningful
improvements (as defined by American Thoracic Society criteria) were
seen at 1 month, and persisted at 3 months in subjects receiving 4-AP at
30 mg/day compared to the low-dose control group receiving 6 mg/day. 
The authors concluded that 4-AP is a potentially toxic drug with a
narrow therapeutic index, but significant toxicities were not observed,
probably because of careful patient selection and individualized dosing
regimens.  This work established a LOAEL of 5 to 10 mg/day.

The study performed by Grijalva et al enrolled 27 patients with
long-term SCI.  Patients received 4-AP by oral dose of 5 mg/day, which
was increased by 5 mg/week to a maximum dose of 30 mg/day; 2 capsules
were given every 8 hours for a total of 6 capsules/day over 12 weeks. 
Patients were divided into two groups.  Group A received 4-AP for 12
weeks while Group B received placebo; the groups switched after 12
weeks.

A complete safety evaluation was performed every 4 weeks.  Two patients
withdrew from Group A, for reasons unrelated to 4-AP.  One patient in
Group B had a moderate adverse reaction (arterial vasospasm).  Reported
adverse reactions included dry mouth (dosage range of 5 to 30 mg/day);
dizziness (5 to 30 mg/day); nausea (15 to 30 mg/day); gastritis (10 to
30 mg/day); paresthesia (30 mg/day); arterial vasospasm (20 mg/day);
side effects included insomnia, anxiety, headache, cramps, memory
alterations, increased saliva viscosity, bitter taste in mouth, and
global pinching pain (5 to 30 mg/day).  These side effects (except for
arterial vasospasm) were also observed in patients receiving placebo. 
Diaphoresis, abdominal distention, abdominal pain, phosphenes,
hyperphagia, and itchy eyes also were reported by seven patients who
received placebo and none of the patients receiving 4-AP.  There was
some confusion about interpretation of side effects data concerning
whether the effects were noted in patients receiving placebo before or
after they received 4-AP.  Adverse reactions occurred within the first
week of receiving 4-AP and usually resolved within 1 to 2 hours.  Some
mild signs of liver effects were observed but resolved during or shortly
after discontinuation of treatment.

The authors concluded that 4-AP was beneficial and provided positive
gains in motor function, sensation, and independence; these effects were
observed in more patients (69 percent) receiving 4-AP than patients
receiving placebo.  The authors also state that patients should be
monitored for peripheral vasospasm.  This work sets a LOAEL for adverse
effects of 5 mg/kg.

EPA’s summary assessment of these studies found that 4-AP serum levels
were proportional to dose, short-lived, and declined after exposure
termination.  There is a therapeutic value of 4-AP associated with its
continued use.  All adverse effects occurred after 4-AP intake and were
proportional to dose.  The three studies are complimentary; Girjalva et
al is the most detailed and Van Diemen et al provides important
information on serum 4-AP levels.  The studies helped EPA determine a
minimal LOAEL for short- to long-term exposures of 5 mg/day, equivalent
to 0.07 mg/kg/day.  

Dr. Brimijoin asked Dr. Khasawinah if he had noted any deficiencies or
questionable aspects of these studies from a scientific perspective. 
Dr. Khasawinah answered that the studies were well-controlled and the
patients were closely monitored.  The reported adverse effects may have
been an underestimation of the true adverse effects, which could lead
EPA to underestimate risks associated with 4-AP.  When birds eat corn
treated with 4-AP, they show severe reactions.  EPA would use the LOAEL
from this assessment as a baseline for risk assessment.

Dr. Lebowitz asked how the adverse effects were used to determine the
LOAEL.  Dr. Khasawinah explained that the minimal dose that produced
any effect was used.  Dr. Fenske noted that the weight of evidence
document stated that there was no data on acute, sub-chronic, oral, or
inhalation toxicity, no information on developmental or reproductive
toxicity and no information on potential carcinogenicity.  He asked how,
if this is true, this product could be registered.  Mr. Jordan explained
that EPA data requirements depend on the use pattern of a product.  For
a product that may get into the food supply, very extensive data are
required.  If the use pattern indicates that human exposure
opportunities will be limited, fewer toxicity studies are required.  The
4-AP use pattern indicates that humans are not likely to be exposed
through food, water, or most work practices, therefore, less data were
needed.  4-AP also is an older product that was registered years ago at
a time of less demanding data requirements and thus has limited toxicity
data.  For the purposes of current risk assessment activities, EPA is
attempting to find the best data available to permit ongoing use of 4-AP
as described by current regulations.

Dr. Fitzpatrick asked why 4-AP was not anticipated to enter the food
supply if it was used on crops or could be accidentally ingested by
animals.  Mr. Jordan explained that it is unknown whether 4-AP sprayed
on crops is taken up by the plants.  Mr. Ray Kent (OPP, EPA) commented
that this has been an issue in the past because 4-AP-treated corn
designed to keep birds away from crops was distributed throughout the
fields and could lodge in growing corn, but because of differences in
4-AP distribution, this is no longer an issue.  Mr. Jordan noted that
because of potential risk to non-target species, EPA required separate
studies of 4-AP on different bird species and fish to evaluate
differences in effect.  

Ethical Considerations

Mr. Carley provided EPA’s ethics assessment of the three studies.  All
three articles were considered in EPA’s ethics review on May 22 and
25, 2007.  The articles were reports of research involving intentional
exposure of human subjects in clinical trials of experimental
pharmaceuticals.  All three studies pre-date EPA’s rule.  Two (Segal
and Grijalva) were subject to the Common Rule, the third (Van Diemen) to
the rules in place in the Netherlands in 1993.  The research was not
conducted with the intention to submit it to EPA under the pesticide
laws and was not submitted by a regulated entity; instead, EPA retrieved
the articles from the published literature.  Since the documents were
not submitted to EPA, 40 CFR §26.1303, which requires documentation of
ethical conduct of studies submitted after April 6, 2006, does not
apply.  40 CFR §26.1602(b)(2) requires HSRB review.  40 CFR §26.1703
forbids EPA reliance on research involving intentional exposure of
pregnant or nursing women or children, and 40 CFR §26.1704 forbids EPA
reliance on pre-rule research if there is “clear and convincing
evidence” that its conduct was fundamentally unethical or
significantly deficient relative to standards prevailing when the
research was conducted.  

The Van Diemen article reports the relationship between dosage, serum
level, efficacy, and side effects of 4-AP in 70 patients with MS
participating in a drug trial.  This information is of potential value
to EPA in defining endpoints for assessing risk to humans from exposure
to 4-AP when it is used as a pesticide.  

Subjects included in the Van Diemen trial were 43 women and 37 men with
MS for periods from 2 months to 25 years, ages 23 to 68 years.  The
subjects did not have hepatic or renal disease or any history of
epilepsy.  The reproductive or nursing status of the women was not
reported, nor was the means by which subjects were recruited.  All 70
subjects participated in the first (IV) phase of the research; one did
not participate in the second (oral) phase for reasons unrelated to this
study.  Risks to subjects included potential serious side effects
reported in earlier studies of clinical use of 4-AP to treat MS.  This
study used a rising-dose design and close monitoring to explore the
relationship between dosage, serum level, efficacy, and safety of 4-AP
used to treat patients with MS.  All subjective side effects were
registered; as they increased, dosing was lowered or discontinued.  All
side effects were monitored to resolution and blood chemistry was
extensively monitored, as were cardiac, hepatic, and renal function.

Concerning the benefits of this research, the efficacy measures
described in this article implied that the research offered therapeutic
benefits to subjects.  Societal benefits include further insights into
potentially effective treatment of MS and improved understanding of
thresholds for side effects of 4-AP, including pain, paresthesia,
dizziness, headache, gait instability, nausea, restlessness/anxiety,
abdominal pain, and obstipation.  The information provided in this study
suggests that benefits were sufficient to justify the risks to
individual subjects. 

The Van Diemen protocol was approved by the Ethical Committee of the
Free University Hospital, Amsterdam, but the report identifies no
prevailing standard of ethical conduct.  The DoH (1989) was assumed to
be the relevant standard, and there is no indication that conduct of the
study was deficient relative to DoH standards.  Informed consent was
obtained from all patients before they were accepted into the study; no
further details are provided.  The privacy of subjects was not
compromised in the published report.

The purpose of the Segal study was to determine the effects of long-term
administration of 4-AP on sensorimotor function in humans with
long-standing SCI.  The article reported improvements in the patients’
condition with some minor side effects but no significant associated
drug toxicity.  This information is of potential value to EPA in
defining endpoints for assessing risk to humans from exposure to 4-AP
when it is used as a pesticide.

Subjects included 18 men and 3 women with traumatic SCI of at least 2
years’ duration.  Ten subjects had previously been exposed to 4-AP in
a short-term test; the remaining 11 had never been exposed to 4-AP.  One
additional subject began the study but did not complete it for reasons
unrelated to the study.  Pregnant women were excluded, but the means by
which subjects were recruited were not reported.  Concerning risks to
subjects, previous studies had shown that 4-AP in the dose range
administered could be effective in treating SCI with relatively few side
effects.  The subjects were extensively monitored before initiation of
treatment and at intervals during the research.  Although the study was
conducted on an outpatient basis, investigators maintained daily
telephone contact with subjects.  All side effects were monitored to
resolution.

Efficacy measures demonstrated that the research offered therapeutic
benefits to subjects.  Societal benefits include further insights into
potentially effective treatment of SCI and improved understanding of
thresholds for side effects of 4-AP.  The information available suggests
that the benefits were sufficient to justify the risks to individual
subjects.

Concerning independent ethics review, all subjects “provided
institution-approved, written informed consent,” and the report
identified no standard of ethical conduct.  The VA hospital at which the
research was conducted had a registered IRB, held a FWA from OHRP, and
is subject to the VA Common Rule.  Informed consent was obtained from
all patients, but no further details were provided.  The privacy of
subjects was not compromised in the published report.  The prevailing
standard of ethical conduct was the Common Rule and 21 CFR Parts 50 and
56; there is no indication that conduct was deficient relative to these
standards.

The purpose of the Grijalva study was to examine the efficacy and safety
of 4-AP and to document sensorimotor changes after discontinuation of
the drug in patients with long-term SCI.  The article reports
improvements in the patients’ condition with some minor side effects
but no significant associated drug toxicity.  This information is of
potential value to EPA in defining endpoints for assessing risk to
humans from exposure to 4-AP when it is used as a pesticide.

Twenty-one men and 4 women, ages 23 to 48 years, with traumatic SCI of
at least 1.5 years’ duration completed the study; two others withdrew
for reasons unrelated to the study and three withdrew late in the
study, but their data were included.  Before enrolling, each participant
underwent a comprehensive clinical evaluation.  Pregnant or nursing
women were excluded but the means by which subjects were recruited were
not reported.

Previous studies had shown that 4-AP in the dose range administered
could be effective in treating SCI with relatively few side effects. 
The subjects were extensively tested before initiation of treatment and
at intervals during the research.  All but one of the side effects
reported were mild, but some effects not previously reported were noted.
 All side effects were monitored to resolution.

As with the other two studies, potential benefits include efficacy
measures that demonstrated potential therapeutic benefits to subjects. 
Societal benefits included increased information on a potentially
effective treatment for SCI and improved understanding of the threshold
for side effects associated with 4-AP use.  The information provided by
Grijalva et al suggested that the benefits of the study justified the
risks to individual subjects.

This research was initiated after acceptance by both the local research
committee of the hospital and the National Research Council of the IMSS.
 The IMSS has an IRB (#3566) registered with OHRP, and holds a FWA
(#4956).  No standard of ethical conduct was cited in the article. 
Concerning informed consent, patients were fully informed in writing and
verbally, and provided signed consent; no further details were reported.
 The privacy of subjects was not compromised in the published report. 
The prevailing standard of ethical conduct for this study was the Common
Rule; ethical conduct was more completely reported than in most
published studies and there were no indications that conduct was
deficient relative to applicable standards.

EPA had reached the following conclusions based on the Agency’s review
of the three clinical trials of 4-AP.  40 CFR §26.1703 forbids EPA to
rely on research involving intentional exposure of pregnant or nursing
women or children.  To this end, all subjects in the three studies were
over 18 years of age, pregnant women were excluded from all studies, and
nursing women were excluded from the Grijalva study.  When evidence
concerning subject age and reproductive status is both absent and
unobtainable, it is EPA’s policy that §26.1703 does not prohibit
reliance on a study.

40 CFR §26.1704 forbids EPA to rely on pre-rule research if there is
“clear and convincing evidence” that its conduct was fundamentally
unethical or significantly deficient relative to standards prevailing
when it was conducted.  Despite minor gaps in documentation of ethical
conduct for all three studies, there are no indications that any of the
studies failed to meet applicable standards of ethical conduct.  Thus,
there is no regulatory barrier to EPA’s reliance on these studies.

Dr. Menikoff noted that in the event of a lack of information, the Board
has assumed that previously conducted studies were conducted ethically
and asked if this was the accepted approach.  He also asked if
additional data, such as protocols or ICFs could be obtained. 
Mr. Carley responded that if nothing in the published reports suggest a
serious deficiency, it is appropriate to accept the reports at face
value.  In this case, Mr. Carley did not believe there were any
questions that warranted contacting the authors.  

Public Comments

	Dr. Brimijoin invited oral public comment on the three published
clinical studies on 4-AP.  No oral public comments were received.

Board Discussion

	Scientific Considerations—4-AP

Dr. Brimijoin opened discussion of scientific considerations for EPA’s
use of the data from these studies to derive a point of departure for
estimating risk to humans from exposure to 4-AP.  He noted that 4-AP is
a fast potassium channel blocker, which leads to a prolonged activation
potential, increased nerve transduction, and increased neurotransmitter
release at a wider range of terminal types.  4-AP also does not
over-stimulate receptors.  Concerning the scientific strengths for use
of this data as a point of departure for estimating human risk from 4-AP
exposure, there are advantages and disadvantages to using clinical
trials data rather than animal studies to assess exposure.  The weight
of evidence document showed that extensive animal data exists; however,
this data was not collected in a way that would help EPA
determine low-level doses associated with low or no side effects. 
Thus, a need for more and better data exists.

Because the goals of these studies were to demonstrate a clinical effect
and analyze safety with respect to therapeutic dose, they were not
designed to determine the minimal amount of 4-AP that produces an
effect.  This presents a weakness for using the data to determine LOAEL.
 Dr. Brimijoin noted that the studies appear to have varying degrees of
quality as clinical trials, but fail to demonstrate more than a very
small, questionable therapeutic effect.  The studies have major
weaknesses in design and outcomes.  The Segal et al study is not
placebo-controlled, only dose-blinded.  Powerful placebo effects were
observed in this study with respect to side effects.  Because of this,
it is difficult to determine effects attributable to 4-AP itself.  Dr.
Brimijoin continued by noting that the data in the Grijalva study are
confusing.  The report states that adverse effects were reported by 56
percent of treated patients.  Based on the number of adverse effects
reported, he assumed that the effects listed in the placebo arm were
occurring during placebo treatment, but instead appear to refer to
events that occurred while a patient previously treated with 4-AP was
receiving placebo.  Dr. Kim explained that data on adverse effects were
collected over the entire course of the study, which included both a
treatment phase and a placebo phase.  Some patients received treatment
first and then placebo, others first received placebo, followed by
treatment.  He noted that it does not appear that adverse effects
occurred during treatment with placebo, but the report does not
specifically state when the adverse effects occurred.  

Dr. Brimijoin summarized that the Grijalva report is unclear concerning
which adverse effects were treatment-related.  The study cannot be used
to establish the level of 4-AP dose at which effects were noticed
because of the unclear study design and report.  The Segal report
focused on treatment-related effects.  This report noticed nervousness,
giddiness, dizziness, and gastrointestinal upset as the most frequent
side effects, but these effects were not linked to the 4-AP dose level. 
The data from this report seem to indicate that the drug is clinically
safe within certain narrow limits, but provide weak evidence upon which
to base conclusions about a reference dose.  The Van Diemen study does
contain information that would tend to support EPA’s conclusion that a
LOAEL can be derived.  The study was blinded and contained good detail
about treatment-related side effects.  The report does not establish a
dosage low enough to be used to establish a no observed adverse effect
level (NOAEL).  A total oral dose of 5 mg/day was associated with mild
but definite discomfort, but no associated blood chemistry changes. 
This study alone or in the context of the other studies provides
information on dose that could be used to establish a LOAEL of
approximately 0.07 mg/kg as a reference dose.

Dr. Fitzpatrick continued the scientific discussion.  The scientific
question asked by these studies concerned clinical endpoint, not safety.
 Existing data on 4-AP was sparse, but animal data could be used to
determine dose.  4-AP is an acute neurotoxicant and the reported side
effects address this.  None of the three studies contain justification
for dosing; this has implications for consent, particularly for the
patients who received three-quarters of the previously reported toxic
dose of 60 mg.  Dr. Fitzpatrick noted that the patients participating in
these studies could be considered members of a vulnerable population,
because there are no other potential treatments for their conditions. 
Additionally, they may not have recognized that the trials were not
designed to benefit them directly.  The results of this study also are
not likely to be generalizable to the general population.

All three studies focused on subjective clinical endpoints. 
Additionally, patients with too many side effects could be excluded from
the trial, which could underestimate side effects.  The investigators
claim that the side effects are minor, which may be the perception of a
population with few other options for treatment, but perhaps not of a
healthy population.  A lack of individual subject data for assessment of
toxicity also is a weakness.  It is difficult to conclude that all three
studies could be used to establish a true LOAEL.  In addition, this data
would give an oral (or IV) LOAEL and it is unlikely that this is how the
general public would be exposed to 4-AP.

Dr. Fitzpatrick expressed some surprise that the Segal study, conducted
under the auspices of the Food and Drug Administration (FDA), received
an Investigational New Drug designation with so little animal or human
data.  She questioned whether FDA truly supervised the study or whether
Segal perhaps had additional information that was not reported in the
article.  She also clarified that “orphan drug status” means that
the drug is used on less than 100,000 people per year but has no
implications for efficacy.  She concluded that it would be difficult to
support use of this data for establishing a LOAEL.

Dr. Fenske reiterated that it is important to understand that these
studies were conducted to evaluate efficacy.  Regarding safety, it
appears that the subjects might be willing to endure strong side effects
in hope of a treatment or cure for their conditions.  The Segal study
used different doses, and it is difficult to conclude whether the higher
dose was associated with stronger side effects.  The Grijalva study used
the lowest dose (5 mg/day) but nonetheless reported adverse effects. 
This report does not clearly describe the number of patients who
experienced side effects, or whether some experienced more than one, and
does not indicate which side effects occurred at which dose level.  In
the Van Diemen study, side effects were observed in patients receiving
4-AP both orally and by IV administration.  Side effects were observed
in those receiving the minimum dose of 5 mg and some patients may have
experienced multiple side effects.  Giddiness, dizziness, and nausea are
symptoms of central nervous system toxicity.  Dr. Fenske stated that he
would support EPA’s use of the data as a point of departure for risk
assessment scenarios, but would be reluctant to endorse 5 mg/day as a
LOAEL, given that side effects were observed at this dose and given the
steep dose-response relationship observed for 4-AP.

Dr. Kim noted that each study had flaws, particularly with reporting
adverse effects.  He also was concerned that the risk-benefit ratio may
be different in patients compared to healthy volunteers.  He took issue
with non-reporting of toxicity for adverse effects treatment.  The
agents were used in a FDA setting, which may have different assessments
of risk based on the severity of the disease the study attempts to
treat.  He agreed that the nature of safety data reporting was casual
and not rigorously defined.  He also noted that it is always difficult
to confirm that improvement is related to therapy.  It also will be
difficult to determine whether EPA applied appropriate uncertainty
factors because of differences in the risk-benefit ratio for patients
versus healthy volunteers.

Dr. Lehman-Mckeeman stated that she had reservations about the overall
utility of the data.  In this case with actual evidence of exposure, it
is ironic that EPA may not be able to use this data.  She agreed that it
is difficult to use these studies to assess toxicity because their
primary objective was demonstrating efficacy.  Dr. Lehman-Mckeeman also
expressed concern with the toxicity data with respect to whether
patients with SCI or MS accurately reflect events that would occur in a
normal population.  She said that she was inclined to think that this is
an unusual subpopulation, which could relate to the perceived
under-reporting of adverse effects.

Other confusing aspects of these reports include the low effect level
based on total daily dosage.  The daily dose was stated as 5 mg/day, but
the drug was administered two to four times per day, and it is unclear
whether adverse effects were observed after administration of the first,
second, third, or fourth dose.  Patients in the Grijalva report appeared
to need several weeks of treatment before an effect was observed; these
subjects received lower doses at repeated points during the day.  Dr.
Brimijoin agreed that this was an important observation, given the short
half-life of 4-AP.

Dr. Lehman-Mckeeman continued assessment of the Grijalva report.  This
study reported adverse effects during the wash-in placebo period.  The
report appeared to indicate that 14 of 21 patients experienced 26
possible reactions, and this included effects experienced both during
treatment with 4-AP and during the placebo phase.  This would make it
impossible to determine adverse reactions specifically related to 4-AP
treatment.  Information from a data and safety monitoring board or
analysis of the raw data would be necessary to determine when side
effects occurred.  Dr. Kim noted that most reports of this type are
conservative with respect to reporting treatment-emergent adverse
effects and the Grijalva report described all events that occurred over
the entire 26 weeks of the trial.  Because of this, adverse effects may
actually be over-reported. 

Dr. Brimijoin summarized that much of the Board’s criticism focused on
the issue of the studies’ value as a record of toxicity level and safe
doses.  This arises primarily because these studies were not designed to
test for a toxic endpoint but rather to determine if 4-AP has a
therapeutic effect at a safe level (accompanied only by mild toxicity). 
There was confusion related to the Girjalva report concerning the dose
associated with treatment-induced toxicity and whether there was clear
evidence of difference in toxicity between the treatment and placebo
arms.  Additionally, subjective side effects may have been
under-reported because of a highly motivated subject population seeking
treatment options for serious conditions (SCI and MS).

4-AP also has a steep dose-response curve, which could mean that doses
quickly become toxic as they rise.  The collective data from the studies
indicate that a dose of between 5 and 10 mg/day is not associated with
dramatic toxicity.  EPA would like to call this dose the LOAEL and use
it as a point of departure.  Several Board members have spoken against
specifically relying on this data for a point of departure.  Although
Dr. Brimijoin stated that he initially believed 5 mg/day would be a
defensible point of departure, after hearing discussion from other Board
members he now would not enthusiastically endorse this.

Dr. Fenske stated that the science was sound and the studies well
conducted.  The side effect reporting was not systematic.  If EPA
believes it is important to use this data to establish a LOAEL, Dr.
Fenske would not prevent it, but neither would he strongly endorse doing
so.

Dr. Kim inquired whether EPA had other resources to use to establish a
LOAEL for 4-AP.  Dr. Khasawinah responded that there is little data from
the literature upon which to rely, which is why all three studies were
presented; EPA believed the three would complement each other.  He
stated that taken together, the studies seem to indicate that 5 mg/day
would be an acceptable LOAEL and EPA also can include uncertainty
factors for more protection.  EPA traditionally includes a 10-fold
uncertainty factor but could increase this for this case.  Dr. Lebowitz
disagreed about the lack of data on 4-AP, describing other clinical
trials that used a sustained-release form of the drug and suggested that
EPA may wish to analyze these studies to gain more clarity on this
issue, given the Board’s reluctance to endorse use of the three
studies presented at this meeting.  Dr. Brimijoin commented that peak
serum levels would be different for a sustained-release form of 4-AP. 
Dr. Fenske noted that the weight of evidence documents referred to
animal studies that determined a LOAEL of 3 ppm and asked if EPA could
use these studies.  Dr. Khasawinah explained that these studies had
limited information.  Dr. Fenske asked why human data was needed if a
NOAEL was available from animal studies.  Dr. Khasawinah replied that
the animal studies were not considered acceptable.  He added that the
studies using the sustained release form of 4-AP also measured efficacy
and reported side effects in a manner similar to that used by the three
studies presented at this meeting.

Dr. Brimijoin stated that 5 mg/day probably was in the correct range for
a LOAEL, but the design and reporting of the studies does not allow the
Board to conclude this with a high level of confidence.  He endorsed
EPA’s consideration of this data with some reluctance.  Because of the
uncertainty of the data, EPA should treat the Board’s approval as
provisional and should incorporate uncertainty factors in its
development of a reference dose for 4-AP.

	Ethical Considerations—4-AP

Dr. Menikoff opened the ethics discussion.  He stated that because these
are completed studies, the Board is restricted in its analysis and must
assume, in the absence of clear and convincing evidence of unethical
conduct, that the studies were ethically conducted.  He agreed with Mr.
Carley’s ethics assessment of the studies.  He noted that obtaining
ICFs, IRB documentation, or protocol information for review of
previously published work would be useful, if the article was published
recently enough to make this possible.  

Dr. Menikoff continued that these studies were unusual in the context of
determining LOAEL because their primary objective was to measure a
therapeutic effect.  As Dr. Fitzpatrick noted, the risk-benefit ratio
for this population of patients with serious conditions may be different
than that for the general population.  MS and SCI are serious conditions
with few treatment options.  There is good evidence based on
pharmacological data that 4-AP could be beneficial.  The studies also
indicate that the side effects associated with treatment are likely to
be minimal and reversible.  It is appropriate to conclude, therefore,
that the risk-benefit ratio was appropriate.  All three studies claimed
appropriate informed consent, which the Board must assume in the absence
of clear and convincing evidence that contradicts this.  All studies
also were reviewed by an IRB or its equivalent.  Given the available
information, it is difficult to conclude that these studies were
fundamentally unethical or that there is clear evidence of deviation
from prevailing ethical standards.

Dr. Brimijoin summarized that the Board concluded there was no clear and
convincing evidence that the studies were unethical or failed to meet
prevailing ethical standards.

Overview of AEATF and AHETF Research Programs

Mr. Jordan provided an overview of the AHETF and AEATF.  Three protocols
from the AHETF were presented at a previous HSRB meeting.  At the
meeting, the Board indicated that the materials presented were not
sufficient for a substantive evaluation to determine the scientific
reliability of the data that would be generated or to determine if the
protocols met ethical standards.  The Board raised a number of
scientific and ethical questions for EPA to consider.  This led to a
January 2007 meeting of the FIFRA SAP at which information related to
scientific issues raised by the HSRB were discussed.  The information
presented at the April 2007 HSRB meeting focused on the social value of
the proposed research, i.e., had EPA demonstrated the need to collect
new data on agricultural handler exposure.  The Board agreed with
EPA’s assessment of the need for new data to assess handler risk. 
Discussion at the April 2007 HSRB meeting also addressed broad issues
applicable to all protocols that the HSRB will review in the future.

The goal of AHETF and AEATF discussions at this meeting was to ensure
that the task forces and the HSRB are ready to present and review
materials and protocols meaningfully, namely that the task forces will
be able to provide all information necessary for a meaningful review of
protocols by the HSRB.

	Mr. Jordan described several broad issues related to the charge
questions in this meeting’s transmittal memorandum.  The issues
included the following:

An overview of task force materials, to ensure that terms are understood
and used correctly;

Balancing of risks and benefits in this research to ensure that the
benefits justify the risks to participants and to ensure that this
question is properly framed;

The potential for underestimation bias, which reflects certain aspects
of the methods used to collect pesticide residues on handlers, including
passive dosimetry, hand rinse, and face wipe;

Biomonitoring as an alternative, or in addition to, passive dosimetry;

Quality assurance (QA) and control procedures developed by the task
forces to ensure that results are scientifically reliable;

Task force proposals to use purposive diversity sampling as an
alternative to stratified random sampling;

The number of monitoring units and how to cluster these units in terms
of the precision to be achieved;

Repeated measurements that would enable characterization of the degree
of within-worker variability;

Ethics issues including risk minimization and recruitment issues,
especially of non-native English speakers.

	EPA seeks advice on scientific issues that will ensure the data are
scientifically reliable and whether the planned approaches will affect
reliability of results and that the results will be of sufficient value
to justify subject risk.  The Board should consider the value of this
data compared to existing data in the Pesticide Handler Exposure
Database (PHED).  EPA also asks whether the proposed research designs
use its resources efficiently and provide a significant
return-on-investment.  EPA wishes to consider the value of increasing
funding to collect additional data.  The Board should consider whether
EPA is cognizant of relevant risks for this work and if information is
available and correct for making decisions concerning risks
and benefits.

Overview of Handler Research

Mr. Carley described documents provided by the task forces and the
planned designs of the databases.  The task forces provided documents
describing standard operating procedures (SOPs); AEATF has not completed
its SOPs but has included a sample protocol including a description of
the informed consent process.

Each task force has developed governing documents that provide a
description of the programs at the highest level.  Each task force has a
number of different scenarios; for example, the AHETF has 37 different
scenarios.  A scenario is described as a specific pesticide handling
situation defined by data gathered under common properties.  The
governing documents have statistical design and recruiting information
at a general level.  Design and recruiting processes will be made
specific to each scenario.  Each scenario could be viewed as a distinct
research project; scenarios will have specific research plans, and will
collect and analyze data.

Data to populate each scenario will be collected in clusters.  A cluster
is a set of related handler data from each scenario.  Within each
cluster will be monitoring units (MUs).  A MU is not a subject but
instead should be thought of as a string of data.  The AHETF calls for
five MUs in each of five clusters to populate a scenario.  The AEATF
plans to have six MUs in three clusters.  This may not work for all
scenarios and more appropriate designs can be developed if needed.

Mr. Carley described how a proposal for a specific field study would fit
into the program, contribute to the database, and be presented to EPA
and the HSRB.  A field study is a convenient group of MUs with a related
protocol and final report, both of which will be reviewed by EPA.  MUs
will fall within a single cluster, which can be defined generally as a
location.  Within a cluster, MUs can represent different types of
handling, for example mixing/loading or application using different
types of equipment.  The data gathered within such a cluster may apply
to different scenarios.  The task force will write the final report
after data have been entered into a cluster and then into a scenario. 
For the AEATF database, all MUs required to fill a cluster are part of
different field studies.  When the design of each scenario is fulfilled,
scenario-level analysis can begin.

The HSRB review will have to consider each scenario design and its
appropriateness.  When a completed study is reviewed, the Board will
consider whether the study fulfills the requirements of its protocol and
also contribute to the scenario-specific design.  The task forces will
not perform scenario analysis without all necessary data, but may
reconsider the scenario design when analyzing a completed field study.

Dr. Fitzpatrick questioned whether a MU would be assigned to one or
multiple scenarios.  Mr. Carley replied that he would need to discuss
this with task force leaders.  Dr. Fenske described a cluster as a space
and temporal event, describing the location at which a field study would
be performed and asked whether more than one field study would
contribute to a cluster.  Mr. Carley added that it was unclear how best
to consider a field study involving, for example, 10 to 12 MUs that
could contribute to multiple scenarios.  Such a situation could be
considered two clusters that occur in time or space or could be
considered one cluster that feeds multiple scenarios.  Dr. Fenske
remarked that data from a field study likely would contribute to
multiple scenarios.  He asked whether one field study would take place
at one cluster, or if multiple field studies could occur at one cluster.
 Mr. Carley noted that either situation could arise.  For example, if
there are fewer usable MUs from a field study, data from another field
study may be needed to fulfill the statistical design.  

Dr. Fenske inquired whether it was correct to assume that the first
protocol the Board reviews will have multiple scenarios because of
differences between mixer/loader and applicator exposures.  Dr. Chambers
questioned if scenario design would be related to the protocols. 
Mr. Carley explained that the Board will receive information on
scenario design along with the protocol.  The challenge for the review
schedule is that, in addition to developing protocols, a presentation of
scenario design also must be developed.  Dr. Chambers asked whether the
protocol for the field studies may have multiple scenarios, and, if this
were the case, would different sampling strategies be described in the
same field study.  Mr. Carley responded that some field studies would
have multiple scenarios and different sampling strategies.  For example,
if the study includes mixer/loader and applicator exposure, there will
be two scenario design documents and one protocol.  These data may feed
into a single cluster.  Dr. Brimijoin thanked Mr. Carley for his
explanation of the task force designs and noted that analysis of
specific protocols will probably help clarify any remaining questions.

Risks and Benefits of Handler Research

Mr. Carley presented his review of the risks and benefits associated
with the exposure monitoring programs.  He stated that he was not asking
the Board to reach a conclusive judgment regarding risk and benefits at
this time.  He continued by stating that at the level of the Governing
Documents, conclusive weighing of risks and benefits is impossible,
because knowledge of the risks associated with a specific field study
and of the benefits specific to a specific scenario are required.  Risks
to subjects can be characterized in general terms, and potential ways of
minimizing risks and potential benefits of the research can be
identified.

Categories of subject risk described, in descending order of risk, by
the AHETF include heat-related illness, exposure to surrogate chemicals,
risk associated with scripted activity, psychological risk, risk from
exposure to detergent rinses, and the background risks of agricultural
work.  Risk categories described by the AEATF include chemical risks
related to use of a surrogate antimicrobial chemical and exposure to
alcohol/water face and hand rinse solutions.  Physical risks may arise
from heat stress or exaggeration of normal activities.  Psychological
risks also will be considered.

Heat-related risk arises from the increased requirement for additional
layers of clothing (passive dosimetry) and also possibly from scripted
activities.  This risk rises with the heat index.  Heat-related risks
can be minimized by training, the presence of onsite medical staff,
provision of liquids and shade, close observation, monitoring the heat
index, and developing clear stopping rules.  In addition to stopping
rules, “non-starting rules” that will prohibit commencement of study
based on the heat index also will be developed.  With these precautions,
residual risk is likely to be acceptably low.  Concerning risk from
surrogate chemical exposure, the use of registered pesticides as
directed are unlikely to cause unreasonable adverse effects.  Ensuring
that workers follow all label and Worker Protection Standard (WPS)
requirements for protective clothing and personal protective equipment
(PPE) will minimize risk.  A residual risk is that of irritation in the
event of a spill.

AHETF “scripting” may increase the length of the workday, or the
amounts of active ingredient handled, or call for use of less familiar
equipment.  This risk can be minimized by the same methods used for
risks of heat stress and active ingredient exposure, with the addition
of the opportunity to practice with any less familiar equipment.  AEATF
“exaggerating activity” may lead to fatigue or heat stress, which
can be minimized by careful study-specific design and by close
observation of subjects.  If these precautions are implemented, residual
risks are likely to be low. 

Two types of psychological risks have been identified, namely risks
associated with taking a pregnancy test and risk of embarrassment while
changing clothes to put on passive dosimetry garments.  Both these risks
can be minimized by actions to ensure privacy.  If these procedures are
implemented, residual risks are likely to be low.

Some degree of risk is associated with use of hand rinse or face wash
solutions.  AHETF expects to use 0.01 percent detergent in water and
AEATF proposes using 50 percent isopropyl alcohol (IPA) in water.  The
risk associated with these solutions is primarily irritation.  This can
be minimized by supervision by a technician when rinsing near the eyes,
close observation for signs of irritation, and a clear stopping rule. 
Any residual risks are likely to be low.

The background risk of agricultural work is primarily of injury arising
from possible increased use of less familiar equipment, and also
dizziness or confusion attributable to early heat-related illness and
exposure to other chemicals.  The choice to apply chemicals, in addition
to those under investigation in the protocol, is not under the control
of the investigators, but instead is a decision made by the owner of the
field.  This cannot be comprehensively addressed in protocols but could
be minimized through good agricultural practices.  The task force will
ensure that compounds required by the owner will not interfere with
analysis of results.  It also should be noted that the additional
chemicals might not be pesticides but instead could be substances, such
as fertilizers.

There is no direct benefit of this research for the subjects, but there
are possible indirect benefits if the research leads to better safety
standards.  The potential societal benefits of knowledge likely to be
gained should be assessed primarily at the scenario level, taking into
account existing data, the appropriateness of normalization factors, and
scenario-specific sampling designs.  The benefits of improved estimates
of handler exposure and potentially improved protection of workers can
be realized only when the scenario-level design is fulfilled.  Potential
benefits to AHETF members include conduct of monitoring programs that
keep them in regulatory compliance.  For growers and landowners, the
provision of free product may more than offset the disadvantages and
inconveniences of cooperating in the research.

When weighing risks and benefits, the task force monitoring programs
will include only studies designed to fill established needs for new
exposure data.  Scenario-specific designs will include full discussions
of anticipated societal benefits.  Study proposals will include
discussions of the balance between study-specific minimized residual
risks and anticipated benefits at the study level (if any) and at the
scenario level.  EPA believes the benefits of this research are likely
to outweigh the minimized risks to subjects.  The Board is asked to
consider whether the Governing Documents, in addition to the described
information on risks and risk minimization, provide an adequate basis
for assessing whether the risks of a particular study are justified by
the expected benefits of the proposed research, and if not, to describe
additional information that should be provided to an IRB, EPA, and/or
the HSRB.

Dr. Chambers questioned whether additional PPE would be recommended if
it was found that exposure was higher than expected and whether this
could be assessed before the scenario-level analysis was performed.  Mr.
Carley explained that each field study will generate a report that will
be reviewed by the HSRB.  The statistics-driven analysis of the
association of active ingredient handled with exposure will not be
performed until the scenario-level data requirements are fulfilled. 
There are also statutory regulations that require prompt submission of
any evidence of unexpected adverse effect; EPA expects that any
unexpectedly high exposure will be brought quickly to its attention.

Public Comment

Dr. Richard Collier on behalf of AHETF

Dr. Richard Collier (AHETF) clarified that for the AHETF studies, a
single MU will be assigned to only one scenario.  One MU represents a
person who is being monitored and all aspects of his work practices and
the environmental conditions associated with a given activity.  In most
cases, a field study will contribute data to only a single cluster;
however, this distinction may not be important.  Given a study with
mixer/loader activities, as well as an MU performing application
activities, this would represent parts of two clusters but would be
assigned to one scenario.  A study can be thought of as activities
performed in the field and the protocol that defines these activities. 
Dr. Collier added that the plan is for the HSRB to review documents that
define one scenario, one clusters, and one study protocol within that
scenario at the October 2007 HSRB meeting.

Dr. Fenske questioned whether a given field study was likely to include
people performing different activities (such as mixing, loading, and
application) and if data from these activities would apply to different
scenarios but would still be part of the same cluster.  Dr. Collier
explained that these would be considered different clusters because they
describe different activities.  From a statistical perspective, to
analyze cluster-level effects, this single study would involve more than
1 cluster.  Dr. Fenske clarified that he believed this study would have
two types of MUs whose information is entered into different parts of
the database.  Dr. Lebowitz inquired whether data applicable to two
scenarios could be gathered from the same site at the same time and
whether this would constitute conduct of two studies or protocols or
just one.  He asked whether a protocol would be developed for each type
of work.  Dr. Collier responded that such a study would involve two sets
and types of MUs.  The AHETF views this as one study, because a study is
an activity performed in a field in a particular place at a particular
time.  Dr. Fenske commented that this appeared to be analogous to doing
indoor air pollution studies in a single home in which data would be
collected on both adults and children.  The sampling may be performed
differently for each group and the data may be sent to different bins
for analysis, but this would be considered one field study.

Board Discussion

Dr. Fenske opened discussion of the charge questions.  He noted that EPA
has been responsive in supplying governing documents for task force
work, but despite this there continues to be confusion around
terminology such as “MU.”  Concerning risks and benefits of handler
research, Dr. Fenske agreed with Mr. Carley’s assessment and the
description in the reports by the two task forces.  He commended Mr.
Carley and the task forces for identifying potential hazards or risks
and developing ways to minimize these risks.  Dr. Fenske stated that the
range of risks described was comprehensive and each description of ways
to minimize risk was thoughtfully performed.  He stated that this
analysis builds confidence that the studies will be carefully conducted
with attention to potential risks. 

Concerning benefits, Dr. Fenske agreed that only indirect benefits would
accrue to participants in the form of better data leading to better
protection policies.  He also agreed with Mr. Carley’s consideration
of the benefits to growers and landowners in the form of free product
compared to the inconvenience of participating.  Dr. Fenske added that
another potential benefit for growers is better data that could lead to
less over-regulation of chemicals that may not need to be strongly
regulated.  The data gathered in these studies could help ensure that
products are not inappropriately removed from the market.  Dr. Fenske
noted that under FIFRA, benefits of this work can apply to users, but
not to the manufacturers.  Mr. Carley agreed that analysis of risk and
benefit related to registration decisions was appropriate.  He noted
that his analysis examined only the benefits attributable to conducting
the studies, which will help keep task force members in regulatory
compliance.  Dr. Fenske continued that he agreed that the primary
societal benefit was that of better science to improve risk assessment
and public policy.  He concurred that for these studies, benefits
significantly outweighed risk.

	Mr. Paul Hamey (Consultant to the HSRB) provided the perspective of a
grower in the United Kingdom.  He agreed with EPA’s analysis of risks
and benefits.  He added that although the governing documents indicate
no direct benefit to the participants, one direct benefit could be
feedback on individual performance in the context of a study that could
improve handlers’ practices and decrease exposure.  He asked for
comments on appropriate actions if a study director noticed an
individual operator using unsafe practices.  For example, during a
series of studies by a United Kingdom task force, it was noticed that
handlers commonly used compressed air to clear build-up from equipment;
this generates aerosols and increases exposure.  In response to this,
the task force developed and provided a poster to increase awareness of
this issue and promote safer means of cleaning equipment.

He stated that the governing documents provided a good discussion of
risk and added two additional risks:  mechanical injury arising as a
result of collapse of a hydraulic system and electrical injury through
equipment contact with liquids.  If a study participant is asked to use
unfamiliar equipment because of scripted activity, the risks of either
of these events increases.  The governing documents discuss requiring
familiarity with equipment and provide for practice with unfamiliar
equipment, but do not consider existing standards for using new
equipment; in the United Kingdom, training and a formal test are
required.  Mr. Hamey commented on the heat-related illness mitigation
procedure of encouraging participants to drink more water and noted that
this may increase exposure by the oral route.  Drinking while working
with hazardous substances contradicts governing document statements
regarding safe practices when drinking at work.

Dr. Chambers stated that her review was similar to Dr. Fenske’s, and
acknowledged the strengths of the governing document.  She noted that a
conceptual problem was estimating potential risks of exposure to
chemicals which would require use of a MOE and asked how the MOE could
be calculated.  Mr. Jordan clarified that all chemicals used in task
force studies have been subject to EPA risk assessment using existing
exposure data in the PHED.  The task force uses the PHED value as an
assumption about the degrees of exposure occurring in a scenario and
applies that to the amount of active ingredient handled.  Dr. Chambers
thanked Mr. Jordan for the clarification and asked that a review of the
studies’ structure, with respect to MUs, clusters, and scenarios, be
presented at the October 2007 HSRB meeting.

Dr. Lebowitz commended EPA on the thorough documentation of its
analysis.  He noted that risks related to financial concerns and job
security were not discussed; he was informed that this would be
discussed during the meeting.  Dr. Lebowitz questioned whether
agricultural handlers would learn or be reminded of best practices
during the course of the studies, noting that observation by study
personnel may discover that some handlers are using inappropriate
methods for handling pesticides.  He also asked how exposure to families
upon workers’ return to their homes would be assessed and whether risk
to bystanders of accidental exposure during pesticide application would
be considered.

Dr. Philpott inquired whether either task force had considered whether
there would be unacceptable levels of risk associated with the potential
for heat-related illnesses.  He noted that the heat index would be used
to create stopping rules, but the heat index makes assumptions of
physical attributes such as height, weight, gender, ethnicity, movement,
and clothing.  Those participating in passive dosimetry studies may be
wearing extra layers of clothing and some participants may have higher
activity levels in direct sun than is assumed by the heat index.  These
considerations should be used to develop additional safety factors.  He
noted that the report categorizes extreme danger as a heat index of 130
degrees Fahrenheit (oF) or higher, but at a heat index of 105oF, the
risk of sunstroke, heat cramps, and heat exhaustion increases.  A lower
heat index may need to be considered to trigger the stopping rule.

Dr. Popendorf (Consultant to the HSRB) questioned why compensation was
not considered as a benefit.  Dr. Brimijoin noted that the HSRB had
specifically been asked not to consider this.  Dr. Popendorf continued
that another benefit to participants could be providing an option for
participants to learn about their exposure levels, which could result in
decreases in unsafe practices.  He added that clarification of
growers’ benefits also is needed.

Dr. Brimijoin summarized that the Board was impressed with the progress
and efforts of EPA in preparing these documents.  Concerning the
discussion of potential additional benefits, he commended EPA for
raising the issue of benefit to landowners being mitigated by the
inconvenience of participation.  There is a danger of landowner benefit
resulting in coercion of workers to participate.  Dr. Brimijoin
acknowledged Mr. Hamey’s contributions regarding potential educational
opportunities through identification of unsafe practices at the
individual and site levels.  Plans should be developed to capture this
potential benefit.  Regarding scripted activities, it will be
advantageous for workers to use familiar equipment as much as possible. 
Other issues raised included the wisdom of urging increased fluid
consumption to mitigate heat-related risks because this could increase
risk of pesticide ingestion.  Dr. Chambers raised questions concerning
error calculations that were answered by EPA.  Dr. Lebowitz raised
questions concerning minimizing exposure to families and assessing risk
to bystanders.  Dr. Philpott suggested that EPA reconsider heat index
levels used to set stopping rules.  Dr. Popendorf mentioned knowledge
of individual exposure levels as a potential benefit for participants.

Addressing Potential Sources of Underestimation Bias

Mr. Jeffrey Evans (OPP, EPA) presented EPA’s analysis of potential
sources of underestimation bias.  The task forces will rely on whole
body dosimeters (WBD), hand rinse/washing, and face/neck wipes to
determine exposure.  EPA has analyzed whether concern for breakthrough
and/or collection/removal efficiencies of passive dosimetry techniques
warrants adjustments to field measurements.  EPA sought the advice of
the FIFRA SAP in January 2007 on a number of topics related to the
potential for underestimation bias of passive dosimetry, including
comparison of passive dosimetry and biological monitoring studies,
analysis of hand measurements in the current PHED, and a review of
available literature related to hand rinse techniques.

Comparison of whole body passive dosimetry and biological monitoring
suggested a lack of systematic bias between methods; both methods
yielded similar results, suggesting that both give reasonable estimates
of exposure.  Thus, passive dosimetry is not likely to substantially
underestimate exposure that may be caused by dermal absorption during
sampling or breakthrough of dermal dosimeters.  The SAP concluded that
bias may exist, but bias between dermal exposure and biological
monitoring could not be detected in part because of the statistical
uncertainty inherent in exposure data.  The SAP also noted that passive
dosimetry can generate data that can be used to develop predictive
estimates of exposure for a number of different scenarios and
activities.  The SAP suggested that biological monitoring could be a
useful check on passive dosimetry or to measure WBD breakthrough, but
declined to require biological monitoring in a protocol.  EPA agreed
with the overall SAP conclusions and described some disadvantages to
biological monitoring including additional cost, logistical
considerations (number of days required for metabolites to clear), and a
lack of acceptable biomonitoring methods for all surrogates.

The SAP also assessed potential underestimation bias resulting from
determining exposure by hand washing.  The literature indicates that
hand wash/rinse performance can be influenced by the chemical properties
of the pesticide, such as solubility, octanol/water partition
coefficient, or formulation type; residence time on the skin before hand
rinsing is performed; type of solvent used to rinse the hands (e.g.,
alcohol, soap and water); concentration of the chemical on the skin
(microgram/cm2); duration of the exposure monitoring period; and nature
of the residue (whether exposed to pesticide concentrates, dilute
sprays, or field residues).  The literature also indicates that hand
rinse removal efficiency values from studies involving human subjects
ranged from approximately 70 to 90 percent for many of the surrogate
compounds selected by the AHETF (with the exception of chlorpyrifos,
which has an efficiency of approximately 20 to 40 percent).  An AEATF
hand rinse efficiency study indicates up to 90 percent efficiency for
didecyl dimethyl ammonium.

Face/neck wipes were not specifically discussed at the SAP meeting nor
are they presented in the 1987 Subdivision U, Agency guidelines. 
Face/neck wipes have been widely used since the late 1980s or early
1990s and are an option in Organization for Economic Co-Operation and
Development (OECD) guidance; series on Testing and Assessment No. 9. 
Investigators generally use the same solvents or surfactants for both
hand rinses and face/neck wipes.  Exposure to the head/face and neck is
expected to be very low for the majority of exposure scenarios planned
by the AHETF.  An exception to this can be seen in a series of existing
AHETF open cab airblast studies, in which both face/neck wipe and patch
data (outside and inside chemical resistant headgear) were collected to
determine potential exposure for these areas.  Face/neck wipes will be
treated the same as hand washes when making corrections.

The SAP was equivocal about the need to correct the results from hand
washing for its efficiency at recovering pesticides from skin.  The SAP
would accept a rinse validation study if it can decrease the uncertainty
at a reasonable cost and be done within approved human subjects study
guidelines.  Some panel members recommended using modeling to adjust
hand exposure and offered as an example an algorithm based on some of
the literature cited by the Agency in the SAP documentation.  Others
noted weaknesses in both modeling and validation study approaches and
raised concerns about potential confounding by field conditions (i.e.,
effects of repetitive rinsing that may change the skin’s absorption
rate).

EPA has concluded that for the proposed AHETF studies, the contribution
of hand exposure is expected to be minimal because all subjects will be
wearing chemical resistant gloves (CRG) during all operations.  AEATF
intends to collect data based on individuals not wearing gloves
(consumer products).  For most scenarios, significant exposure to head,
face, and neck is expected to be low.  A series of options for
considering this has been proposed to AHETF.  Biological monitoring can
be included as a check for potential breakthrough or other losses when
using surrogates that have well-established methods.  The use of cotton
gloves beneath the CRG and the use of hat patches when measuring head,
face, and neck exposures for scenarios having the potential for overhead
exposures (e.g., open-cab airblast) could be considered.  Conditions
should be established for correcting hand rinses and face/neck wipes. 
EPA has proposed a set of conditions for consideration by both task
forces, namely, if measured exposures from hands, face, and neck
contribute less than 20 percent of total exposure, no action is
required, but if measured exposure contribution represents between 20
and 60 percent of total exposure, an automatic 50 percent adjustment can
be made or a validation study can be submitted.  If measured exposure
contribution is greater than 60 percent, a validation study is required.
 Because validation studies involve intentional dosing of human
subjects, review of these studies by the HSRB will be required.  

AHETF has concluded that no correction is needed for any potential
method bias because of reasonable congruence in exposure estimates
between studies based on biological monitoring and those using passive
dosimetry.  AEATF has concluded that no correction is needed in studies
where individuals will not be wearing gloves and the hand correction
factor is reasonable.  EPA has concluded that substantial
underestimation by WBD is unlikely and the Agency believes that the hand
wash, face/neck wipe options are the most appropriate for correction of
potential underestimations using these techniques.

The Board was asked to consider whether EPA had identified the relevant
scientific and practical considerations affecting the choice to include
biomonitoring and appropriately characterized the limitations on the
scientific usefulness of the resulting data if no biomonitoring is
conducted.  If not, the HSRB should describe other considerations that
should bear on a decision to conduct biomonitoring in addition to WBD. 
The HSRB also is asked to consider whether EPA had appropriately
characterized the limitations of the scientific usefulness of a handler
database that does not include data characterizing the efficiency of
residue removal procedures.

Dr. Chambers inquired how exposures of 20 percent or less would be
assessed if there is insufficient exposure data to determine this.  Mr.
Evans replied that the analysis will have to rely on old data, and that
decisions concerning this matter will be made as the data accumulates. 
Dr. Lehman-Mckeeman questioned if the biomonitoring samples would be
taken from urine.  Mr. Evans responded that this was correct.  She
inquired whether the availability of well-established methods would be
the only criterion for determining if biomonitoring could be used to
check for breakthrough and asked whether the feasibility of collecting
samples based on the half-life or absorption of a compound should also
be considered.  Mr. Evans explained that these parameters also would be
considered; if analysis of a certain compound required sequestering
participants for several days, biomonitoring of that compound would
probably not be performed.

Dr. Popendorf noted that if skin recovery data was required, human
exposure protocols would be needed; however, to simply determine
recovery, in vitro substitutes for human skin are available, and use of
cadaver skin also could be considered.

Quality Control and Quality Assurance in AHETF and AEATF Occupational
Handler Monitoring

Mr. Jeff Dawson (OPP, EPA) presented Agency activities concerning
quality control (QC) and quality assurance (QA) procedures for the AHETF
and AEATF protocols.  He opened by defining QA as a program for the
systematic monitoring and evaluation of the various aspects of a
project, service, or facility to ensure that standards of quality are
being met and QC as operational techniques and the activities used to
fulfill and verify requirements of quality.

EPA uses GLPs under FIFRA and the Toxic Substances Control Act; GLPs
were implemented by EPA in 1989.  Key elements for occupational
monitoring include documentation throughout each phase, specific
requirements and formats for protocols, establishment of SOPs for all
phases, audit of each study phase and general issues (e.g., QC
maintenance records), establishment of personnel training records,
establishment of lines of communication and responsibility, and
definition of the fines and criminal penalties possible for
non-compliance.  Field investigators and laboratories are required to
work under GLPs.  A number of SOPs pertaining to QA and QC were
developed as a collaborative effort involving both task forces, EPA,
California EPA Department of Pesticide Regulation (CDPR), and the
Canadian Pest Management Regulatory Agency (PMRA).  The SOPs are
essentially identical in scope and verbiage in many cases; thus, AHETF
SOPs were presented as an example.  The SOPs address all elements of
occupational study conduct, including administration, protocol
development, the field phase, analysis, and reporting.  The AHETF
governing document has 11 SOP chapters.

SOP 1 pertains to administration and covers organizational structure,
responsibilities, inspections, and external communication.  The study
director has overall responsibility and ensures protocol compliance,
addresses addenda and deviations, and takes corrective action in
response to unforeseen circumstances.  The principal investigator (PI)
provides a direct line of communication between a facility and the AHETF
study director.  Ethics training also falls under this SOP.  The QA unit
(QAU) is an independent entity that reports directly to the task force
manager and study director.

SOP 2 covers initiation, development, and content of protocols that
dictate study conduct.  Protocols should link GLP requirements and SOPs.
 This SOP also establishes a protocol approval process, and each
protocol will specifically reference the appropriate SOPs.  Amendment
and deviation processes are part of this SOP.  

SOP 3 describes development of SOPs, including format; preparation and
approval; review, revision, and retirement; distribution; and
consideration of contractor SOPs.  

SOP 4 describes study report content, format, and submission
requirements.

SOP 5 describes the scope and responsibilities of the QAU.  The QAU, as
described in SOP 1, is a separate third party that reports to AHETF
management.  The QAU oversees GLP-required elements, such as
communication with management and recordkeeping.  Facility inspections
and protocol, data, and report audits are the responsibility of the QAU.

SOP 6 establishes archive processes for records and specimen/sample
storage.  This SOP describes requirements for raw data storage, QAU
audit storage, formulated product or standard materials, and
confidential worker information.  

SOP 7 oversees test, reference, and control substances, including
shipping, receiving, storage, and tracking of test materials.  Also
included in this SOP are labeling and tracking requirements (such as
chain of custody forms), container retention, and the AHETF test
substance definition (“All registered pesticides that may be used on
an AHETF study for the sole purpose of providing detectable residues in
the determination of a pesticide exposure profile”).

SOP 8 pertains to matrix samples, including sample media, sample
collection, storage and labeling, and QC procedures.  For each
monitoring method, pre-field preparation and in-field sample procedures
will be addressed, as well as QC elements (for example, positive WBD
controls to evaluate losses due to conditions during the sample
collection period).  This SOP also will review workers’ personal
apparel for compliance with WPS and develop systematic sample tracking
codes.

SOP 9 covers analytical detection, data corrections, and paper and
electronic elements.  Concerning analytical elements, this SOP requires
corrections to be proposed based on field recovery estimates. 
Analytical elements include the limit of quantitation (LOQ), which is
the lowest level fortified for a matrix in a study; limit of detection
(LOD) is 0.3X LOQ.  If mean field recovery is less than the LOQ,
one-half of the LOQ will be used for corrections.  This SOP also defines
methods for recording data, including GLP error codes, and provides
details for authenticating copies and portable document format
preparation.

SOP 10 defines field study operations, including processes for equipment
calibration and field/worker data collection.  This SOP describes
methods for air samples (e.g., primary standards), application equipment
(e.g., flow meters), worker observation criteria (e.g., motions, out of
scope activities such as not wearing PPE), environmental conditions
(e.g., wind speed, direction, temperature, humidity, etc.), and sample
shipment.  

SOP 11 covers human subject management, including ethics training,
recruitment, language requirements, and informed consent.  Worker
recruitment, language issues, pregnancy status, hazard information,
unanticipated adverse events, and heat stress management (National
Oceanic and Atmospheric Administration and National Weather Service
criteria) also are covered by this SOP.  SOP 11 helps ensure ethical
compliance and identifies Agencies and rules pertaining to this issue.

Development of these SOPs represents a collaborative effort involving
the task forces and multiple federal and other agencies.  The results of
recent SAP and HSRB reviews have been incorporated into SOP development.
 The scope of the SOPs includes the task forces and associated
contractors, and requirements of the SOPs are dictated in part by GLP
requirements.

In collaboration with officials in the Canadian PRMA and the CDPR, EPA
has worked with the task forces as they developed a set of SOPs to
ensure the data resulting from their proposed research is of high
quality.  The task force SOPs reflect current, state-of-the-art methods
for QA and QC in the collection, storage, and analysis of analytical
samples.  Therefore EPA believes that the resulting data will be of very
high quality.  The Board is asked to evaluate whether the SOPs are
adequate to ensure high quality data results from the proposed research,
and, if not, to describe other QA or QC procedures that need to be
addressed.

Public Comment

Dr. John Ross of Infoscientific, Inc.  on behalf of CropLife America

Passive dosimetry methods, such as air monitoring and dermal monitoring
using clothing dosimeters (inner and/or outer), hand washes, or
face/neck wipes, have been established with national and international
input from experts.  Biomonitoring is not viable for a generic
database and cannot be performed for many compounds because of a lack of
pharmacodynamic/pharmacokinetic data.  A recent study (Ross et al.,
2007) demonstrates that concurrent passive dosimetry versus
biomonitoring shows that passive dosimetry neither under- nor
over-predicts exposure.  A graph comparing concurrent passive dosimetry
with biomonitoring using results from a number of studies showed neither
an over- nor underestimation bias.

Hands can represent a significant proportion of total dermal exposure. 
A combination of soap, water, and mechanical agitation is the primary
method of hygiene for removal of pesticide residues on hands and very
effectively removes most compounds.  There is a great deal of GLP
available to regulators to determine the removal efficiency of
pesticides from skin, including more than 80 dermal absorption studies
in rats with soap and water removal and more than 10 human/monkey
dermal absorption studies with soap and water.  Most of this data was
generated using radio labeled pesticides to ensure accountability.  Data
from more than 20 human/monkey dermal absorption studies of different
pesticides indicate, on average, less than 10 percent absorption
following an 8-hour exposure.  Data from rat studies show that what is
adsorbed is frequently absorbed, although a few compounds have bound
skin residues (adsorbed) that are 2 to 3 times greater than absorbed. 
Quality hand removal efficiency studies show that more than 90 percent
of a compound on the hands can be removed.  

Hand wash methods can overestimate exposure.  Most measured residues
would slough or wash off the hand during the work period, so only a
fraction of the amount applied is dermally absorbed.  In addition,
pesticide frequently is adsorbed to dirt on the hands, reducing
bioavailability.  Pesticide washed off in the first half of the day has
no opportunity for absorption, but may be counted as if it did in some
protocols.  Any task that requires gloves also reduces hand exposure
10-fold or more compared to an ungloved hand wash.

If hand wash removal efficacy studies are planned, investigators must be
aware that applying a dose of a substance to clean hands may not
simulate agricultural work conditions.  Under these conditions, hand
exposure is intermittent and does not occur as a bolus, which is typical
of a hand wash removal study.  Removal efficiency also is related to
concentration, and a worker may be exposed to both dilute and
concentrated pesticides.  Use of radio labeled pesticides would insure
accountability of total dose, but the risks of this approach may
outweigh the benefits.

Dr. Ross concluded that the weight of evidence indicates that dermal
removal efficiency is adequate.  There are questions about the
applicability of short-term removal efficiency (0.5 hour) to typical
worker removal time (2 to 8 hours).  Without reason to believe that
recovery may be compromised (e.g., high reactivity, polarity or
lipophilicity), a removal efficiency study is unjustified.  If hands
represent 50 percent exposure and 10 percent is lost due to
adsorption/absorption, underestimation is 5 percent and is negligible.

Dr. Lehman-Mckeeman requested clarification of the graph presented by
Dr. Ross.  Dr. Ross explained that each point on the graph represented
an individual.  The graph has data from eight or nine different
chemicals from 14 different studies and more than 24 different exposure
scenarios.  Some of the studies are worker re-entry studies, others
assessed exposures during application.  A number of the compounds are
the same as those that will be used in AHETF studies.

Dr. Philpott addressed Dr. Ross’ argument that without reason to
believe recovery was compromised, an efficiency study is unjustified. 
He noted that Mr. Evans had mentioned that one of the compounds in
question (chlorpyrifos) had an unpredicted low rate of recovery and
asked whether Dr. Ross would have predicted this.  Dr. Ross answered
that approximately 20 percent of the points in the graph represent
chlorpyrifos chemicals.  The assertion that the recovery rate was only
10 to 20 percent is questionable because the experiments did not use
radio labeled material.  Dr. Lebowitz inquired whether oral exposure
(i.e., ingestion of pesticide-contaminated foods) would affect bias and
how this could be determined if biomonitoring is not performed.  Dr.
Ross responded that the data on the graph were largely generated using
worker exposures.  In most cases, any contribution of exposure from diet
was subtracted using background levels described by the CDC National
Biomonitoring Program.  Handlers’ dermal exposure is at least an order
of magnitude higher.  Dr. Fenske noted that the data on chlorpyrifos
came from studies he had performed and that in these experiments, all
mass except for that stuck to the skin was collected.

Dr. Ray McAllister of CropLife America

Dr. Ray McAllister presented comments from a regulatory perspective
concerning the issue of adjusting hand and face exposure by a removal
efficiency coefficient.  He reiterated that data show that the guideline
passive dosimetry methodology in question does not have a systematic
bias nor does it underestimate exposure.  Because of this, he questioned
the need for a correction factor to apply to data generated by guideline
studies.  EPA has proposed either arbitrary adjustment factors that
vastly increase hand/face exposure estimates or requiring removal
efficiency studies whenever the hand/face exposure estimate is 20
percent or greater of the total exposure; this is likely to result in a
required efficiency study.  He suggested that the HSRB should ask EPA to
define such studies (i.e., what loading factor should be applied to a
human subject’s hand, how long should the substance remain on the skin
before removal, and whether pipette administration of a pesticide to
hands over a 5-second period is similar to the dynamics of deposition
and removal over a full day of agricultural work activities).  Because
of these questions, data may be uninterpretable, leading to the
conclusion that intentional human exposure in such studies would be
unethical on grounds of inadequate scientific validity.

It is possible and preferable to determine the qualitative potential for
high hand or face exposure from the use pattern.  For example, if a
pesticide concentrate is handled with unprotected hands, hand exposure
would be expected to be a high percentage of the total exposure.  If
this is demonstrated in a study, the expected result (i.e.,
proportionally high hand exposure) should not be penalized by applying
an adjustment factor.  If the data do not demonstrate what was expected,
further investigation is warranted.  This is the opposite of what EPA
has proposed.

Exposure assessment for determination of risk incorporates various
uncertainty factors unique to each assessment.  These factors involve
pesticide use information, robustness of the toxicology database, and
specific laws under which the data are being evaluated.  Dr. McAllister
stated that EPA’s recommendation of an across-the-board adjustment
factor without consideration of these unique factors will introduce
overestimation to the exposure assessment and will diminish the accuracy
of the assessment.  He agreed that there may be unique circumstances
that call for adjustments for hand rinse efficiency studies, but each
regulatory agency should determine the need on a case-by-case basis, not
by an across-the-board decision by any one regulatory agency or the
HSRB.

Dr. McAllister noted that AHETF and AEATF followed EPA guidelines that
are consistent with international OECD guidelines and were developed in
an open process involving multiple public meetings.  Any considerations
of changes to the guidelines must involve public debate on the impact on
the methodology as a whole and on the overall risk assessment process. 
He contended that the HSRB meetings do not provide such a forum.  He
concluded by stating that requiring a hand/face correction factor will
not ensure extra protection for handlers but may generate an
overestimation of exposure that could lead to higher-tier human exposure
studies that would not have been planned if the initial assessment had
been more accurate.  The current methodology provides adequate accuracy
without adjustments.

Day 3

Follow-up from Previous Day’s Discussion

Mr. Jordan thanked the Board for the advice given in previous sessions
and stated that EPA had no further questions at this time.

AEATF and AHETF Research Programs (continued)

Board Discussion

	Addressing Potential Sources of Underestimation Bias

Dr. Fenske presented information related to potential sources of
underestimation bias for the AEATF and AHETF research programs.  He
noted that biomonitoring is a potential means for evaluating the
accuracy of passive dosimetry, but disagreed that Ross et al (2007), as
presented during the previous day’s session, made a solid case for
validation of passive dosimetry using biomonitoring.

Hand wash and skin wipe techniques underestimate true dermal exposures;
the question is by how much.  The nature of the chemical in question and
the removal method will affect the degree of underestimation. 
Interception (capture of the analyte before it reaches the skin) can
overestimate exposure, removal (after skin contact) underestimates
exposure, and visual inspection (using dyes or fluorescent compounds) is
typically most useful for qualitative evaluation and worker education.

The term “passive dosimetry” is not intuitive because it measures
exposure, not dose; correction factors are needed to determine dose. 
Passive dosimetry also can include a mixture of techniques (i.e.,
interception and removal), which makes dosimetry protocols complex. 
Although the Ross et al (2007) article claims that passive dosimetry
methods used in the studies analyzed in this article have not been
validated, the AHETF Human Research Monitoring Program states that there
are validated passive exposure monitoring dosimetry techniques that will
be used in the AHETF field study program and that “basic passive
dosimetry methodology has long been accepted as a standard, reproducible
procedure that provides accurate and reliable data and does not
underestimate exposure.”  Dr. Fenske noted that this is an
overstatement because some passive dosimetry techniques are standard
consensus techniques, but have not been validated.  OECD guidelines
state that “it is not possible to evaluate the accuracy of any
procedure.  The best that can be achieved for a hand wash or hand rinse
method is a laboratory validation of the efficiency of recovery of
material from the hands of human volunteers.”  

Dr. Fenske presented a graph illustrating the correlation between
passive dosimetry measurements of chlorpyrifos exposure compared to
biomonitoring results.  The graph showed that passive dosimetry
underestimated the biomonitoring results in a systematic manner.  A
similar analysis of atrazine showed a systematic over-prediction of the
biomonitoring-based dose.  To accurately estimate dose, input factors
(such as dermal absorption and excretion fraction, which are specific
for different chemicals) are needed in addition to dosimetry; these
factors are not standardized or vetted.  Additionally, it can be
difficult to compare passive dosimetry results because of differences in
techniques used to determine the amount of a substance on the skin. 
Biomonitoring also presents difficulties; for example, 88 percent of a
dose of atrazine does not appear as an excreted metabolite.

Because EPA has proposed hand wash efficiency studies, standard
procedures for these studies should be developed.  Dr. Fenske presented
an example of such a procedure.  He described the mass balance
calculation (mass removed by hand wash divided by mass transferred to
the hand equals efficiency) to determine the amount of exposure.  He
noted that for chlorpyrifos, the residence time of the substance affects
the removal efficiency, which can be a problem in the field given that
handlers do not frequently or consistently wash their hands; higher
amounts of chlorpyrifos would be absorbed in such a situation.  A
comparison of hand wash and hand wipe methods to determine exposure to
the apple thinner azinphosmethyl showed that the calculated exposure
rates differed based on the methods used.  Use of gloves resulted in an
overestimation by approximately 2.4 fold, hand wash captured 68 percent
of the true exposure, and hand wipe captured 10 percent of the true
exposure.

Dr. Fenske noted that most protocols assume “best practices” by
workers, but this is unlikely to be the case, which inevitably will lead
to bias.  Constraints on best practices that will affect exposure
include label compliance constraints and the possible occurrence of
behaviors prohibited by the label; protective clothing constraints, such
as the assumption that clothing is in good condition and is in proper
use at all times; and equipment constraints, including the assumptions
that the equipment is in good condition, properly calibrated, and is
used to properly reduce the probability of accidents or need for repair.

Other concerns include the effect of observation on real-life use. 
Handlers will be aware that their performance is under observation,
which could change the way they work; however, behavior tends to
normalize with multiple observations.  Motivational bias may also occur
as handlers attempt to meet the expectations of the study director. 
Workers with good health and safety practices also are more likely to
volunteer and workers with poor practices may avoid participation.  The
study duration may impact behavior; if the study lengthens the workday,
fatigue may result in less attention to safety during equipment cleanup
and repair, and this exposure may not be captured by the study.  The
goal of the task forces is to create a distribution of exposures
reflective of true exposure; however, because of the above-mentioned
constraints, the high end of exposure may be truncated because of extra
precautions.

Dr. Popendorf thanked Dr. Fenske for his presentation, and agreed with
his conclusions.  He reiterated the challenges of generating recovery
data, but noted that these may not be as large as anticipated because
useful recovery data could be obtained by using substitutes for human
skin.  Although human skin is not a simple membrane, washing only
involves the top layer, so true human skin is not needed.  Concerning
the problem of low recovery of a substance during washes, wipes are less
efficient than washing and more variable.  Dr. Popendorf recommended
that adjustment for wipe and wash recovery be added to the protocols.

Dr. Chambers commented that, given the uncertainties of passive
dosimetry, performing biomonitoring appears logical; however, the SAP
concluded that biomonitoring also is uncertain and that good
pharmacokinetic/pharmacodynamic data is available for only a few
compounds.  Biomonitoring also requires a longer monitoring period and
that workers not be exposed to a compound for a certain length of time
prior to test days.  She asked about the recommendation that workers
wear cotton gloves beneath the CRGs and whether this would limit
mobility and increase accidents. 

Dr. Fitzpatrick requested the percentage of exposure that occurs on the
hands and face compared to the body.  Dr. Fenske explained that it is
difficult to generalize and depends on the application method.  Using an
airblast open cab application process would result in a relatively high
level of exposure and a relatively large amount of deposition on the
face.  The kind and effectiveness of protective clothing also will
affect exposure.  For example, workers wear gloves, but often take them
off; answering cell phones has become a particular problem in this
regard.  This results in contamination of the hand, and after
re-gloving, an occlusive atmosphere develops that may enhance
uptake/absorption of substances on the hands.  Dr. Fenske added that
EPA’s request for HSRB input on the need for additional data should be
discussed at a future meeting.  

Dr. Brimijoin noted that the high degree of uncertainty inherent in
passive dosimetry is frustrating.  He said that the question to address
is whether there are clearly identifiable procedures that can be
included, concurrent with or after data collection, to allow reasonable
interpretation of the collected data.  Dr. Fenske agreed with Dr.
Brimijoin, but noted that rodent data often is extrapolated to humans
using uncertainty factors.  He agreed with Mr. Evans that there is
minimal potential for breakthrough when WBD garments are concerned.  The
issue of concern is that exposed skin—face, neck, and hands—are
significant contributors to total dermal exposure.  He commended EPA for
its focus on these important issues.  Dr. Fenske did not recommend human
hand wash efficiency studies but instead recommended that EPA apply
uncertainty factors when necessary.  If data indicates a significant
discrepancy in exposure, hand wash studies can be performed after data
are collected.  He concluded that EPA has identified the main problems
and proposed strategies to mitigate these problems.  Further discussion
of exposure analysis will be needed.

Dr. Brimijoin summarized that the use of passive dosimetry instead of
biomonitoring is correct.  The plans for this work generally are correct
although certain details need to be confirmed before final
interpretation of the data.  Dr. Chambers noted that issues of
over-estimation and under-estimation of dose are compound specific.  She
commented that Dr. Popendorf had suggested effective and simple
experiments that could help clarify recovery rates.  She suggested that
if the task forces can identify compounds whose doses are systematically
underestimated, perhaps experiments with these compounds could be
avoided and surrogates with the same active ingredient used instead.

Dr. Lehman-Mckeeman opened discussion of whether EPA has identified the
relevant scientific and practical considerations affecting the choice of
a sample selection strategy.  She stated that the lack of biomonitoring
is intellectually unsatisfying, but there is no easy answer to this
question.  EPA has not, however, characterized the limitations of the
utility of the dataset if biomonitoring data is not collected.  The
Board must discuss whether or not to advocate biomonitoring and whether
it can be performed in the context of these studies.  She reiterated
Dr. Chambers’ comments that this work will use a group of surrogate
chemicals.  If biomonitoring of these chemicals appears technologically
feasible and would provide usable data, it should be considered. 
Determining the appropriate analytical methods for a given compound and
its kinetic properties will impact decisions on feasibility and will
contribute to collection design.  The abundance of a compound’s
metabolites, how many must be analyzed, and the ability to detect
metabolites will also impact this decision.  

Dr. Lehman-Mckeeman noted that there is the potential to generate a
great deal of unusable data using biomonitoring.  Additionally,
biomonitoring will provide chemical-specific exposure information, but
not use-specific information.  Currently, biomonitoring is performed
using rodents to gather initial data, which then is used to develop an
implied dose for risk assessment; thus, excluding biomonitoring from
task force protocols is consistent with current EPA risk assessment
activities.  She concluded that it is questionable whether biomonitoring
would be feasible and generate useful, valid data.  The idea of using
biomonitoring to detect break-through is a minor issue.

Dr. Lebowitz expressed his concern about the lack of information on
biomonitoring and metabolites for the compounds in question.  He noted
that pharmacokinetic/pharmacodynamic data is constantly being generated
and may permit analysis of more compounds in the future.  EPA could
continue to pursue and gather this data.  For a variety of reasons, in
particular the contribution of multiple routes of exposure to total
dose, measurement of whole body dermal and inhalation exposure would
underestimate exposure and would not be scientifically reliable. 
Dr. Lebowitz stated that the goal of this work is to ensure exposures
are minimized, but the protocols would underestimate aggregate and total
exposure.  Evaluation of data provided by AHETF comparing dermal
dosimetry and biomonitoring indicates that biomonitoring detects higher
total exposure levels, which may be chemical specific.  Dr. Lebowitz
concluded that the task forces have not addressed all the scientific
considerations of biomonitoring.

Dr. Brimijoin clarified that Dr. Lebowitz suggested that in certain
exposure scenarios, exposure by other than dermal routes could become
important and thus using only dermal monitoring methods would seriously
underestimate exposure, and that Dr. Lebowitz had suggested a focus on
biomonitoring efforts to compare representative compounds in one set of
workers amongst scenarios to develop corrections for the dosimetry data.
 Dr. Lebowitz remarked that he had concluded that for surrogates for
which metabolic and excretion data is available, a set of workers could
participate in biomonitoring protocols sufficiently prior to and after
exposure because dermal exposure usually contributes only a small
fraction of total exposure to these compounds.  By subtraction, an
estimate for these surrogates could be developed and this would provide
researchers with more accurate and reliable information on adsorption. 
Dr. Lebowitz advocated selective application of biomonitoring protocols.

Dr. Popendorf stated that his overall opinion was that biomonitoring was
not needed and could result in unnecessary complications.  The AHETF
governing documents propose 10 chemicals.  It is reasonable for the
task forces to review the biomonitoring capabilities of those 10
chemicals and comment on the threshold and viability of performing
biomonitoring as Dr. Lebowitz has described.  Dr. Popendorf added that
he was comfortable with the ability of passive dosimetry garments to
prevent other routes of exposure, although face and hand exposure could
allow some absorption.  The compounds proposed have high thresholds for
biomonitoring and probably will not be easily detected, particularly if
passive dosimetry garments provide good interception.  The number of
chemicals examined will need to be reduced if biomonitoring is required
and biomonitoring also would be limited to workers who did not have
prior exposure to a chemical, resulting in selection bias of
participants.  Biomonitoring also will not help resolve issues related
to uncertainty factors needed to determine exposure from a generic
database.  Dr. Popendorf concurred that further documentation of the 10
experimental chemicals is needed, but remained unconvinced that this
would result in discovery of more viable options for biomonitoring.

Dr. Brimijoin concluded that the Board does not appear to be in favor of
imposing biomonitoring across all protocols.  Without biomonitoring, the
data will still be valid; however, if biomonitoring could inform the
data and if there are instances in which limited studies could be
performed in parallel, the Board would consider recommending such
studies.  EPA could examine the surrogate chemical list and determine
which have robust analytical methods and known kinetics and could be
used in a biomonitoring study.

Mr. Jordan agreed with Dr. Lebowitz’s observation that aggregate
exposure occurs from many routes is correct; however, EPA data covering
many chemicals shows that occupational exposure is significantly higher
than that which occurs by other routes (such as through ingestion of
food or water or exposure to residential pesticides).  The goal of the
task forces is to develop a generic database to use for estimating
exposure for a wide variety of chemicals.  Passive dosimetry allows EPA
to identify the portion of the body in a given scenario that receives
exposure.  Methods to reduce risk usually involve specific protective
equipment and this data would help inform decisions regarding the use of
such equipment.  Mr. Evans added that the break-through zone around a
respirator will be monitored by a well-established method. 
Dr. Chambers agreed that these protocols seek to determine occupational
worker exposure and thus disagreed with Dr. Lebowitz’s call for
biomonitoring, which would analyze exposure from food and other sources,
not occupational exposure.  She stated that she would prefer that extra
resources be applied to fix variability and uncertainty related to
passive dosimetry itself, rather than to biomonitoring.  Dr. Fenske
agreed with Dr. Chambers.  The goal of this project is to assess dermal
and inhalation exposure.  Advising EPA to impose biomonitoring could be
perceived as onerous by sponsors.  He summarized that these are large,
labor intensive studies, as evidenced by the QA/QC parameters.  Adding
biomonitoring would result in a completely different study.  He also
noted that, despite his criticism of the Ross et al (2007) article, the
article contains important information that the Board should consider. 
The article carefully culled information from many studies that seeks to
address the issue Dr. Lebowitz raised.  The article found 14 studies
with sound biomonitoring and skin exposure measurements that contain
information that could be useful to EPA.  He concluded that
biomonitoring was not an initial part of the proposal and he did not
wish the Board to leave the impression that it was considering requiring
biomonitoring.

Dr. Popendorf noted that the Ross article did show that correlation was
better for average experimental values.  This reinforces the idea that
passive dosimetry, biomonitoring, and inhalation monitoring are
comparable.  He agreed that biomonitoring was not needed. 
Dr. Brimijoin summarized that passive dosimetry was favored.  The Board
would not recommend biomonitoring, but the Board report can include Dr.
Lebowitz’s reasons for including it.  There is substantive evidence
that the proposed methods are appropriate.  Dr. Lebowitz stated that
after hearing discussion from other reviewers, he declined to include
his views about biomonitoring in the report.

	QA/QC Controls

Dr. Lehman-Mckeeman opened discussion of the proposed QA/QC procedures
by commenting on how the extensive number of procedures underscored the
efforts to develop an appropriate infrastructure and noted that the SOPs
create a new level of sophistication for this field of research.  She
commented that some information that should (but did not appear) in SOPs
could be found in the governing documents, which indicates that these
issues had been considered.  She addressed the area of administration
and noted that the SOPs in this area were reasonable, with the exception
that specifics concerning the training of PIs and study observers were
not clear.  Study observers working in the field will make decisions
concerning whether clothing dosimeters are worn correctly or whether a
worker shows signs of heat-related illness.  The relationship of study
directors to other personnel needs clarification given that the study
directors may not physically observe the study.

Concerning issues related to data quality and sample integrity and
consistency, some of the SOPs provide less information than the
governing documents.  What constitutes a “good” sample is not clear
in the SOP; for example, whether and how worker compliance with the
protocol or variation in parameters such as airflow or temperature,
would affect collection of data.  Dr. Lehman-Mckeeman commented that
spiking samples onsite to create analytical standards is a good
approach, but clarification and description of the process is needed,
because this will affect the quality of the results.  She reiterated
that the definition of a “good” sample and clarification of
characteristics or changes that would disqualify a sample from the
dataset is needed.  She concluded that the SOPs were generally of high
quality and will assure the quality of samples, with some minimal
changes.

Dr. Fitzpatrick agreed that the SOPs were sound, but quite high in
number.  It will be difficult for study participants, including
investigators, to remember these and thus procedures for assessing
compliance (both intentional and unintentional) must be defined.  The
SOPs indicate that scheduled inspections will be used to check for
compliance; unscheduled inspections also should be considered.  Dr.
Fitzpatrick noted that the questions related to QA inspections are
adequate for ensuring consent, determining that forms are signed and
inclusion criteria followed; however, they do not provide a good
assessment of the consent process.  Observation of the consent process
also creates difficulties because people may change their behaviors if
they know they are being observed.  Dr. Fitzpatrick concluded that
training and re-training is the best way to ensure compliance. 
Inspectors should observe deviations from protocols and use this
information to develop additional training needed to constantly improve
the quality of the research.  She also noted that casual conversations
with investigators and study participants may reveal noncompliance
without the need to ask about this specifically.  Whistleblowers also
can be a source of information concerning compliance.  A reporting plan
to facilitate whistleblowers confidentially contacting EPA or the
governing organization should be developed.

Dr. Chambers agreed with Drs. Lehman-Mckeeman’s and Fitzpatrick’s
assessment.  She added that EPA can best judge whether all necessary
issues have been addressed and considered that the task forces’ desire
to collect high quality data will help ensure compliance with
appropriate standards.

Mr. Hamey questioned whether photographs could be included in reports. 
Dr. Fenske explained that photographs and videos could be included.  Dr.
Popendorf noted that the AHETF documentation assuming half of the LOQ as
the LOD was unusual and recommended that they cite reasons for using
this procedure; if this approach is not citable, the AHETF should use
the approach described by the AEATF for this calculation.

Design of Scenario-Level Sampling

Mr. David Miller (EPA, OPP) presented EPA’s design of scenario-level
sampling strategies.  He defined the target population considered during
the design of scenario-level sampling strategies as the set of all
possible handler-days in which scenario-specific tasks are performed. 
The protocols will involve approximately 1.1 million handlers and
approximately 2 million handler days.  Two approaches were considered
for gathering probability samples—the simple random sample and the
complex probability sample.  Complex probability sampling is more
typical for these types of projects.  For example, National Health and
Nutrition Examination Survey (NHANES) used complex probability sampling
to sample a representative U.S. population.  Complications associated
with complex probability sampling include its high cost, the absence of
a sampling frame, and issues related to selection bias.  Selection bias
is a particular issue for these protocols because the studies will only
use volunteers, increasing chances of bias.

Because of these issues, the task forces have considered two alternative
sampling strategies, purposive representative sampling and purposive
diversity sampling (PDS).  Purposive representative sampling captures a
small sample of handler-days that is a “miniature” of the target
population, with respect to important factors concerning range and
extent of exposure.  PDS captures a small sample of handler-days which
are diverse with respect to factors related to range and extent of
exposure.  The task forces propose that PDS is more likely to reflect a
broad range of heterogeneous conditions than probability sampling.  PDS
can be diversified on the amount of active ingredient handled, the
individual (MU), location and time, and other factors (such as equipment
type, crops, rates, and micro-location).  Site selection will emphasize
more common conditions and the task forces will be required to provide
rationale and/or justification for selection of sites or site conditions
based on diversity criteria.

PDS permits a non-random sample to perform at least as well as a small,
same-sized probability sample.  It provides greater assurance of
obtaining a sample that reflects a broad range of conditions and makes
it less likely that high or low end exposure conditions would be missed.
 Augmenting scenario data with new clusters in the future will be
straightforward, and conditions of interest will be easier to target. 
Experts including Drs. Leslie Kish and R. Whitmore contend that, in
survey sampling, judgment is preferable to probability sampling unless
the number of clusters exceed 10 to 20 and that a probability-based
sample is not necessary if data for only a small sample (20 or fewer
persons) can be collected for reasons of cost.  In these cases, expert
judgment and prior knowledge can be used to ensure that the sample units
are representative.

Given this, it is acknowledged that PDS is not a probability-based
sample and can only be used to establish a surrogate distribution of
exposures.  Surrogate distribution cannot be equated to actual
distribution in a target population using pure statistical sampling
theory; however, PDS can capture major aspects of an actual
distribution.  Results using this type of sample are not expected to be
substantially different from those derived using a small, same-sized
cluster random sample.  PDS also is considered to be adequate for
practical regulatory purposes.

The SAP has expressed concern with the proposed purposive nature of
sample selection because PDS assumes underlying random selection can be
used to estimate sample sizes.  Appendix C of the SAP report provides
discussion of potentials for bias and an alternative stratified
approach.  The SAP expressed concern that use of a non-probability
sample would essentially preclude consideration of appropriate weighting
to estimate distributional parameters including means, standard
deviations, upper percentiles, etc.  Thus, the SAP recommended an
informal approach for identifying top factors and for assigning
probability weights to approximate frequencies.

In response to the SAP, the task forces have attempted to address this
concern given the constraints regarding available data and resources;
EPA will evaluate the supporting data and documentation that will be
submitted by the task forces to support their approach.  However, given
the unique aspects of this monitoring program and its relatively small
size, OPP believes PDS is adequately representative of the target
population and can be used to develop exposure assessments of
occupational handler populations.

Statistical Justification of Number of Clusters and Monitoring Units

The number of clusters and MUs must permit collection of sufficient data
for each handler scenario to meet specific minimum or ‘benchmark’
adequacy requirements.  Sufficient data will permit calculation of the
arithmetic mean, geometric mean, and 95th percentile of the normalized
exposure distribution accurate to within K-fold of the true (underlying)
parameters with 95 percent confidence.  To meet this objective, the task
forces have proposed to collect data using a cluster approach.  For this
approach, groups of individuals within a location will be sampled using
a nested design.  Sampling multiple individuals within a given location
can provide economies of scale and improve efficiency because
individuals within a cluster tend to be more similar to each other than
individuals in different clusters, given study effects such as
protocols, study personnel, weather, etc.  OPP agrees with the task
forces’ conclusion that cluster sampling is the most efficient way to
obtain the necessary data and is more appropriate than other sampling
designs.

The design must consider the number of clusters that must be sampled and
the number of individuals in each cluster needed to generate an estimate
of the arithmetic mean, geometric mean, and 95th percentile of the
distribution that are within benchmark accuracy goal limits.  This will
depend on the Intra-class Correlation Coefficient (ICC), shape of the
distribution, and the “spread” (variance) of the distribution.  The
K-factor is the ratio between the estimated parameter based on a sample
and the true (or actual) factor as determined by simulation.  Estimates
of the geometric mean, arithmetic mean, and 95th percentile of an
underlying log normal distribution should be within 3-fold of the true
values at least 95 percent of the time.

The AHETF used simulations to create a table of K values that lists the
corresponding number of clusters, number of individuals per cluster, and
associated K value for each of a variety of assumed coefficient of
variance (CV) and ICC values to help determine the number of clusters
needed.  The AHETF performed simulations using Monte-Carlo techniques to
estimate values of K for the 95th percentile and arithmetic mean under
various assumed conditions (ICC = 0.1, 0.2, 0.3, 0.4, 0.5 and CV = 1.5,
2.0, 2.4, 2.9, 3.5).  For geometric mean, the AHETF used analytical
solutions rather than simulations.  Based on these approaches and the
SAP’s support of the task forces’ general approach to estimating K
values, OPP agrees with the use of K=3 as a reasonable benchmark
accuracy objective.

Sensitivity analysis examined how ICC and CV affect K across various
design configurations.  K is relatively insensitive to ICC and CV in the
selected five MUs in five clusters for one scenario design.  It also was
determined that there were no alternate configurations that provide
practically equivalent K values that would reduce both cost and the
number of individuals needed; some alternate configurations produced a
similar K at substantially higher cost with fewer individuals.

Within-Worker Variability

Dr. Jonathan Cohen (ICF International) presented the AEATF’s
assessment of sampling approaches, cluster design, and optimal sample
design.  AEATF faces the same issue for its sampling approach as the
AHETF.  Stratified random sampling and probability sampling would be
impractical and inefficient.  Thus, the AEATF also will use a PDS
approach.

Clusters are slightly different for the AEATF, because antimicrobial
users work primarily indoors; thus, location and environment are
expected to have minimal impact.  Work also will occur under similar
environmental conditions with respect to surface types, temperature,
humidity, and air exchange rates.  For the AEATF, a cluster includes a
building (or complex), a time period, and research staff and
participants.  Surrogates for confounding factors are room size,
construction materials, loading levels (dirtiness), and research staff
behavior differences.

There was no available data on an antimicrobial ICC; no appropriate
studies with tests on multiple subjects at the same location/time period
could be identified.  The ICC is expected to be low, and given the
average ICC of 0.3 from AHETF outdoor agricultural scenarios, the AEATF
has decided to use an ICC of 0.3 as the upper bound for ICC for
antimicrobial scenarios.

Four studies were used to estimate CV.  These studies were Chemical
Manufacturers Association studies on exposure by mopping, wiping, and
aerosol, and a study on aerosol exposure from the PHED.  A graph
illustrating the distribution of normalized dermal exposure showed that
the means were different, but within each study, the relative
variability was approximately the same.  The PHED study had a lower CV
than the other three studies.  Analysis of the pooled CV of the studies
demonstrated that the relative variabilities were the same, and
estimates for a pooled CV of 1.42 and a geometric standard deviation
(GSD) of 2.86 were derived.  The CV of 1.42 is lower than that of the
AHETF CV of 2.4, as expected.

Issues similar to those considered by AHETF for sample design were
considered for AEATF, namely the number of clusters and the number of
individuals from each cluster that need to be sampled to obtain
estimates of the arithmetic mean and 95th percentile of the distribution
that are within K-fold of the true estimate, with 95 percent confidence.
 Using Monte-Carlo simulations, it was determined that an ICC of 0.3 and
CV of 1.42 gave a K value of 3 when three clusters and six MUs per
cluster were used.

The current occupational exposure approach used in the PHED estimates a
single unit exposure from the PHED database for both single-day and
multiple-day exposures; these are considered to be “central
tendency” estimates and repeated measures (replicates) are
“counted” as different (distinct) measures.  For example, one person
performing a task three times would be counted as three individuals. 
OPP recognizes that the distribution of single (one-time) measures on
different people cannot directly be used to estimate distribution of
long-term exposures and that within-worker and between-worker variance
must be considered.  Multiple measures on the same individual also
cannot be treated as independent measurements and expected (day-to-day)
correlation of measurements within individuals must also be considered.

OPP recognizes that distribution of single-day exposures can only be
used to directly estimate single day exposures and long-term average
(arithmetic mean) exposures.  Distribution of single-day exposures can
be used to estimate distribution of multiple-day average exposures if
assumptions are made with respect to within-worker correlation (Rww). 
If Rww is 0, single-day distribution can be used for longer-term average
exposures, and repeated independent samples should be drawn.  This
assumption underestimates long-term average exposures at the high end of
the exposure distribution and overestimates low end exposure.  If Rww is
1, single-day distribution is equivalent to longer-term average
exposure.  A Rww of 1 assumption overestimates long-term average
exposures at the high end of the exposure distribution and
underestimates the low end of the distribution.  AHETF’s preliminary
literature search and available data suggests that Rww will be
approximately 0.5 to 0.9.

The majority of the SAP believed it was appropriate to de-emphasize
within-worker variability (repeated measures) and instead use available
resources to add clusters and increase sample size.  The SAP noted that
conducting repeated measurement would constrain the eligibility of
handlers, and thus introduce selection bias.  A minority of panel
members believed that repeated measure would provide an opportunity to
capture measures of Rww.

Given the unique aspects of this monitoring program and its relatively
small size, OPP concluded that PDS provides a sample that is adequately
representative of the target population and can be used in developing
assessments of exposure to occupational handler populations.  OPP agrees
with the task forces’ conclusion that cluster sampling is the most
efficient way to obtain the necessary data and is more appropriate than
other sampling designs.  OPP also believes it is appropriate to
de-emphasize within-worker variability (repeated measures), and
resources should be used to add clusters and increase sample size.
Therefore, OPP will not require the task forces to perform repeated
measurements to assess within-worker variability.

The HSRB was asked to determine whether EPA has identified the relevant
scientific and practical considerations affecting the choice of a sample
selection strategy.  The Board also was asked to comment on its
agreement, or disagreement, with EPA that the task forces should provide
scenario-specific information about the availability of data to identify
significant variables potentially influencing exposure and about the
feasibility of developing a sampling strategy to address those variables
quantitatively.  The Board was asked whether EPA had appropriately
characterized the limitations on the scientific usefulness of the
resulting data attributable to the choice of the sampling strategy.

The HSRB also was asked to comment on any additional information needed
by the Board to assess the adequacy of the justification for the number
of clusters and MUs in specific AHETF and AEATF study proposals, and
whether EPA had appropriately characterized any limitations on the
scientific usefulness of a database that doe not include repeated
measures.

Dr. Carriquiry questioned Mr. Miller how the 95 percent confidence
interval was computed.  Mr. Miller explained that he used layered
regression and a mixed model.  Dr. Kim stated that he had issues with
the justification for non-probability sampling, especially the assertion
of absence of frame and selection bias.  In any survey, it is difficult
to determine sampling frame, thus, absence of frame is not a good
justification for non-probability sampling.  Dr. Kim added that it is
difficult to measure the accuracy of estimated parameters for both forms
of sampling.

Public Comment

Dr. Larry Holden of Sielken and Associates Consulting, Inc.

Dr. Larry Holden (Sielken and Associates Consulting, Inc.) clarified
that the purpose of using a surrogate distribution is to develop an
estimate of sample size.  Using non-probability sampling, convenient
samples and numbers of samples are collected.  The idea behind the
approach detailed in the presentation was to use a surrogate model to
develop a reasonable sample and sample size.  Clusters were used because
previous studies showed that it was effective in avoiding the tendency
to use a small sample.

PDS will be used to counteract the tendency to collect conditions that
are too similar, and thus counteract the cluster effect.  Because
workers must volunteer to participate, it was believed that the sampling
frame was not consistent with all possible risk.  Dr. Holden referenced
Appendix B of the SAP report and noted that given the small sample size
and the accuracy desired, random sampling would not provide
significantly better data.  He acknowledged that it is possible to
include the volunteer nature of the participants in the probability
design, but the nature of recruiting may affect selection bias.

Board Discussion

Design of Scenario-Level Sampling

Dr. Fenske opened the Board discussion of the charge questions and
acknowledged that the task forces had adequately explained the
scientific and practical considerations of their study designs and
demonstrated a solid understanding of the methodologies in question.  He
referenced the comparison of purposive representative sampling and PDS,
noting that a problem is not knowing the characteristics of the handler
days within a scenario, which will make developing a scaled model of
this challenging.  He continued that it was not clear who had decided
which experimental factors were most important to consider.  Dr. Fenske
recommended that a broad range of the amount of active ingredient
handled be included in a scenario.  He added that the meaning of
“diversify on an individual” also was unclear.  Dr. Fenske agreed
that location and time define a cluster, and that different types of
clusters are needed, but there likely will not be many unique clusters
and the criterion “different” is unclear, as is how “other”
factors, such as crop and equipment type, will be weighed.  Mr. Miller
explained that the source of all the decisions presented at this meeting
will be available in the protocol that will be presented to the Board in
October 2007.

Dr. Fenske requested the meaning of “diversify on the individual.” 
Mr. Miller replied that this was an attempt to recruit and analyze
exposure for a broad range of individuals performing pesticide
application, based on language, age, and experience.  Dr. Fenske
commented that the importance of these factors must be addressed.  Mr.
Miller confirmed that this would be discussed within the task forces and
likely will be evident in the protocols.

Mr. Hamey stated that the practical and scientific issues had been well
identified; however, he believed that the statement in the governing
documents regarding the conclusion that PDS and purposive representative
sampling will, because of the small sample size, describe the true
distribution equally well, had not been resolved.  He also speculated on
the idea of diversity, commenting that in the United Kingdom, diversity
should be based on farm size because this correlates with different
sizes of equipment used, and different training procedures, behaviors,
and application details.

Dr. Fenske continued that, unlike in the United Kingdom, U.S. workers do
not need to be certified to apply pesticides and instead work under the
supervision of a certified applicator.  There are important differences
in training by a supervisor versus certification training.  Any attempts
at diversity should include an appropriate number of non-certified
applicators.

Dr. Lebowitz referred to the AEATF governing document (page 70) and
asked Dr. Cohen to comment on the discussion regarding quantification of
the impact of ignoring clusters and treating data as a simple random
sample.  Dr. Cohen responded that he believed exposure estimates would
be developed based on random sample assays, and then compared to
estimates using clustering.  Dr. Kim noted that any analysis should
follow an established design.  In this case, given a point estimate, the
mean will be unbiased but variation could cause an anti-conservative
estimate of the standard error.  Dr. Cohen stated that, if AEATF
clustering is too low, adding more variables could result in worse data
than would be generated using a random sample.  Mr. Miller added that
the AEATF design indicates that ICC must be considered which can get
complicated.  The goal of the analysis was to determine if clustering
made a difference, and if so, how much a difference.  Dr. Popendorf
noted that protocols will likely address clustering and it will likely
have less of an effect than was described in the presentation. 

Concerning the second charge question, Dr. Lebowitz began by stating
that scenario-specific information is needed, including the identity of
significant variables influencing exposure to active ingredients and
statistical sampling strategies needed to meet the criteria of providing
useful data.  Information on the relationship between scenario-specific
exposure and a representative scenario of the target population also
would be useful.  He noted that the AEATF will not analyze scenario data
in terms of exposure collected characteristics that would differ between
scenarios.  The governing documents describe some of the relevant
variables, but these are not characterized as causing significant or
even mild effects on exposure.  These include variables such as
temperature, humidity, air flow rates, equipment, amount of chemical
used, and dilution rates, as well as behavior of individual MUs (i.e.,
use of PPE, smoking, gum chewing, etc).  The AEATF appears to be
focusing on variables they consider essential and that would help
delineate an exposure scenario.

Dr. Lebowitz continued, saying that variables for inclusion and
exclusion criteria need to be recorded.  The SOPs call for a description
of measuring instruments and how they will be used, but are described in
a generic way.  In general, information on the characteristics of
variables, how they are collected and used, and their handling, should
be detailed in the governing documents or in an SOP.  The task forces
also presented information on known variability in previously collected
data, which will help design the study and determine population size. 
Dr. Lebowitz concluded that the task forces should provide
scenario-specific information to the Board.

Dr. Popendorf commented that the AHETF governing documents indicate that
they will provide scenario-specific information, but he could not find
whether AEATF would provide this information.  Scenario-specific
information can be a useful tool, as has been found for previous
studies.  For example, the amount handled may be the primary variable,
but there are other parameters that also drive exposure.  Standard field
notes or standardized descriptions of good and bad practices may help to
identify these parameters.  The parameters then can be triaged on a
subjective basis, based on observations; this may help determine which
practices might lead to higher rates of exposure.

Dr. Lehman-Mckeeman agreed that the task forces should provide
scenario-specific information to identify significant variables.  Dr.
Brimijoin summarized that Board agreed more information to identify
significant variables was needed and asked for comments on ways to
capture important variables.

Dr. Fenske questioned whether the Board’s decision on this issue meant
that EPA will ask the task forces to provide a ranking of critical
parameters, based on expert judgment.  A diversity of MUs is desirable,
but decisions must be made concerning the importance of some variables
(i.e., training, compared to others).  He suggested that a
point-by-point discussion of variable importance, with expert rationale,
be held.  Dr. Popendorf agreed, but stated that he was not sure this
conversation should be required before field studies begin.  Ranking of
variables can be developed during the observation phase of the studies
and field notes can be used to categorize exposures.  This information
could be entered into the database and would be useful for future
analyses.  Dr. Fenske countered that a structured list of variables to
observe would be needed beforehand to collect this information through
field notes.  Dr. Lebowitz suggested that the Board should emphasize a
need for both task forces to develop a set of questions and a checklist
to be used to gather this data.  Dr. Brimijoin noted that many potential
variables influence exposure and concluded that as much effort as
possible should be made to identify ahead of time the variables most
likely to be important, and a checklist to use in the field should be
developed.

Dr. Carriquiry opened discussion of the limitations of the scientific
usefulness of the data attributable to the choice of sampling strategy. 
She disagreed that PDS was the best alternative for this work and stated
that it was untrue that PDS gives better results than random sampling. 
PDS may be preferable for a unique group of people for which there is no
sampling frame, but this is an easy sampling exercise.  She commented
that Mr. Miller’s and Dr. Cohen’s presentations implied that many
factors impact exposure.  If this is the case, she recommended that EPA
not accept the assertion that a group of experts can identify all
possible factors and determine the effect of this on the PDS approach.

Dr. Carriquiry noted that PDS may be useful in applied social science
studies where the main objective is description; PDS is useful for
qualitative, not quantitative analyses.  PDS is inappropriate for
quantitative studies because it is impossible to measure the error for a
surrogate distribution.  The surrogate distribution itself could be a
good representative of the true exposure distribution or could be
significantly different, and there is no way of checking assumptions
when PDS is used.

Dr. Carriquiry addressed the sample size argument that PDS sometimes
generates better results than random sampling if the sample is small and
the “universe” also is very small and very well-known.  In the case
of the handler exposure studies, the “universe” is unknown, which
completely negates the advantages of PDS.  Sample sizes must be very
small, on the order of eight to 10, before non-random sampling is
preferable to random sampling.

Dr. Carriquiry suggested a way to use random sampling without greatly
increasing the cost of the studies.  She noted that the most expensive
part of the studies is gathering dosimetry data (approximately $20,000
per person per day) and also travel for study personnel.  A compromise
would be to systematically pick clusters or locations, which could be
defined as counties, states, or contiguous counties.  Within a location,
a list of farms could be easily developed based on information from
county extension offices, which have information on crop type, size, and
location of the farms.  The farms then would be stratified by size and
crop type and then one to three farms within a strata picked randomly
(depending on the number of observations in a cluster).  After this,
handlers at a farm would be randomly chosen.  This would be a more
defensible approach, especially if the data is intended for use for
regulatory issues.  She concluded that she supports the study, except
for the use of PDS; however, this is an easy issue to correct and using
the proper sampling approach will significantly improve the quality of
the data.

Dr. Brimijoin stated that Dr. Carriquiry’s analysis and suggestions
would constitute the Board’s recommendation in its report at this
time.

Dr. Dallas Johnson (Consultant to the Board) agreed with Dr. Carriquiry
concerning sampling.  He noted that certain factors, such as amount of
active ingredient handled, and equipment type can be controlled and
would be part of the fixed effect side of the model.  Factors, such as
location, time, and crop type, are less controllable and thus fall on
the random side of the model.  Purposely picking factors on the fixed
side of the model may be acceptable, but purposively picking from the
random side is problematic.  He agreed that there is no way to determine
the confidence interval (CI) if PDS is used.  Dr. Johnson speculated
that if the variances and estimated standard errors from the data are
overestimates of what would be determined using random sampling, PDS
exposure estimates would be conservative because of the wider CIs.

Dr. Carriquiry agreed with Dr. Johnson on the issue of fixed effects and
the number of factors.  Some of these must be accounted for, but at some
point factors must be sacrificed to avoid significantly increasing the
sample size.  The question is whether to poorly gather data for many
scenarios or reduce the number of scenarios by one or two and focus more
resources on generating sound data from the remaining scenarios.  Good
scenario data will allow extrapolation across multiple scenarios.

Dr. Chambers commented that scenarios constitute different activities,
so caution should be taken in reducing the number of scenarios.  She
asked whether the design of five MUs in a cluster, and five clusters per
scenario changed Dr. Carriquiry’s conclusions regarding PDS versus
random sampling.  Dr. Carriquiry responded that her conclusions would
not change because the relevant sample number for a scenario is 25,
which is large enough to justify random sampling.

Dr. Popendorf presented graphs depicting the differences in the two
sampling approaches.  The K value is driven by EPA and its approach that
it will be consistent with the amount of substance handled.  He assumed
that EPA has found this to be a successful approach from a regulatory
perspective.  If a log normal distribution is examined on a log scale,
the distribution becomes normal.  If this graph represented the amount
handled, the distribution would appear normally distributed and mean and
variation could be determined.  The purpose of PDS is to predefine
categories (i.e., amount handled) and then try to equalize this across
categories, which could generate an acceptable mean; however, the
standard deviation and 95th percentile values will be in error if the
amount handled is not a primary driver of the distribution.  This would
have implications for use of this data for regulatory purposes.  Dr.
Carriquiry agreed with this analysis.  Mr. Miller agreed with the
comments and that PDS is controversial.

Mr. Miller requested clarification of the random sampling strategy
described by Dr. Carriquiry.  He explained that needing to include
handling of a certain amount and certain chemical in a given scenario
had played a role in the decision to use PDS.  Dr. Carriquiry explained
that a compromise between random and purposive sampling can be reached. 
Locations can be picked systematically (or purposively) to have defined
information on the types of crops and size of the farm.  Within a
location, a more random approach to sampling could be used.  Issues,
such as chemical identity and handled amount, are correlated with
characteristics of the location (i.e., large producers use different
types of equipment than smaller growers and also may use different
application procedures or chemical types).  Although it may require more
legwork to determine these parameters, several Board members noted that
this information should be easily available from extension offices or
clearinghouses.

Dr. Kim commented that the unwillingness of the sponsors to create a
sampling frame is problematic.  Scenarios will dictate the sampling
frame.  He stated that this issue needs to be considered or it will be
difficult to draw meaningful conclusions from the study.  Dr. Brimijoin
concluded that the Board believed a random sampling approach, in whole
or in part, would provide better data.

	Statistical Justification of the Number of Clusters

Dr. Kim opened discussion of the statistical justification of the number
of clusters.  If the assumption is that the surrogate sample represents
probability sampling, the sampling described in the governing documents
is adequate and the documents also provide a thorough discussion of the
justification for the number of clusters and MUs based on sensible
assumptions.  Especially for the AHETF, the design of five MUs per
cluster and five clusters per scenario is justified because of the
analysis performed using existing data from the PHED, ICC development,
and sensitivity analysis to determine the influence of clustering.  He
commended the AHETF’s mathematical derivation of sample size, rather
than basing sample size on traditional approaches.  He noted that the
AEATF also provided good justification of the design of three clusters
with six MUs per cluster based on a GSD of 2.4 and an ICC of 0.3.  His
only caveat was to caution the task forces to be attentive to specific
conditions that may require scenario-specific adjustments.  He concluded
that the cluster design was justified, but agreed that use of PDS was
problematic.

Dr. Fenske argued that although the task forces believe they can
identify and rank risk factors in terms of their impact on exposure and
apply this to the study to get an appropriate distribution of exposure,
there is no way to prove if the results of this exercise are correct. 
He stated that although it is appropriate for task forces to trust in
their own knowledge base to rank risk factors, there is no way to
determine whether a truly representative miniature sample has been
created.  He recommended that EPA seriously investigate the feasibility
of random sampling within a defined sample frame.  He noted that Dr.
Carriquiry had described a feasible and practical plan for doing this.

Dr. Fenske concluded that field studies are similar to epidemiological
studies in that there is little control over conditions.  He suggested
that each study chose one goal around which they optimize what is
controllable.  He cautioned against weakening the design developed to
achieve the primary objective in hopes of addressing a secondary
objective.  The planned AHETF and AEATF activities include a survey that
will create usable exposure distributions for a number of scenarios. 
The numbers will be used to inform risk assessment activities for many
chemicals and scenarios.  Attempts to answer questions concerning
amounts handled by changing or restricting the study design will require
assumptions that a sufficient range of active ingredient handled can be
examined.  He recommended optimizing the representativeness of scenario
data to optimize exposure, and agreed that random sampling would be
preferable.

Dr. Johnson agreed with Drs. Fenske and Kim.  He stated that the AHETF
had provided sound justification of its five clusters with five MUs
design and commended the sensitivity analysis.  Based on this, Dr.
Johnson concluded that he would not include more than five MUs per
cluster, but instead would increase the numbers of clusters to reach 25
MUs per scenario.  Dr. Johnson was less pleased with the AEATF’s
design of six MUs in each of three clusters, stating that five or more
clusters would be better.  AEATF should consider changing its study
design in this manner.

Dr. Kim recommended that in terms of the numbers of clusters needed for
the AEATF, as data accumulates in the database, a better estimate of ICC
could be developed.  Dr. Cohen clarified that an ICC of 0.3 was
considered the upper bound estimate.  If the ICC is as low as 0, the K
value will be 2, which is better.  It is true that in the document, the
idea is to use three clusters as a default; once data is available to
estimate a new ICC, cluster design can be re-visited.  In response to a
question from Dr. Johnson, Dr. Cohen clarified that the CV had been
derived from data from four existing studies, but there was no data to
support derivation of the ICC.

Dr. Lebowitz inquired why a cluster design was needed for AEATF studies.
 Mr. Miller replied that the original suggestion from the AEATF was to
monitor more individuals in one location.  EPA objected to this for fear
of being unable to obtain information on the ICC.  This led to the
development of the three MUs per each of six clusters design.  Dr.
Carriquiry noted that there was an expectation of less between-cluster
variability in the AEATF than in the AHETF.  The different numbers of
clusters responds to this, and thus use of fewer clusters in the AEATF
design is justifiable.  Dr. Johnson commented that estimating variance
using only three clusters was somewhat weak.  Depending on the cost,
increasing the number of clusters would be optimal.  Dr. Brimijoin
concluded that the cluster size of three was acceptable to the Board and
noted Dr. Johnson’s disagreement.

	Within-Worker Variability

Dr. Carriquiry opened discussion of within-worker variability.  She
noted that she would like to recommend replicates, because these can be
used to estimate within-person variance relative to between-worker
variance and exposure and gives a better estimate of the usual daily
exposure of workers in a scenario.  However, another way to determine
within-worker variability is to gather replicate measures of a
sub-sample of the same workers and averaging this value over all
workers.  This approach was used by NHANES in a survey to gather
replicate measurements of a food intake in a sub-sample of survey
respondents.  If replicate measures are not taken, the tails of the
exposure distribution may extend too far.  If only one day’s exposure
distribution is used to determine the 95th percentile level of exposure,
this estimate will be too high.  This would actually be protective of
workers because it will overestimate exposure and perhaps trigger
stricter protective regulations.  In this case, the main exposure
estimates will be accurate with or without replicates.

Dr. Carriquiry acknowledged that this issue was recognized in the SAP
report.  The report proposes to borrow information from another source,
assume a certain within-worker variability, and correct the distribution
using external estimates of variance.  It is permissible to do this, but
extra estimates of within-worker variability must be chosen carefully
because they can have a significant impact on the tails of the
distribution.  In the absence of replicates, EPA should be conservative
and use 1-day exposure without correction as the true 10-day exposure,
even though this results in overestimation of the distribution at the
tails.  EPA has recognized this problem and it is not a significant
limitation.  Errors can be easily estimated in the absence of replicate
measures.  Dr. Fenske suggested using data from PHED, which has many
repeated measures performed on the same work force and could be used to
inform variability issues for which there are no samples in the AHETF
database.

Dr. Johnson agreed that if the choice is between including many workers
with one measure each or few workers with many measures, EPA should opt
for many workers with one measurement.  This does not place limits on
the database, although the ability to separate within-worker and
between-worker variance is lost.

	Dr. Lebowitz inquired if within-subject variability would contribute to
the ability to generalize the exposure estimates made from the data. 
Dr. Carriquiry responded that under the given resource constraints, a
limited total number of observations can be made.  If it were possible
to have a larger sample size, replicates of some workers would provide
uncertainties between estimates.  In this case, the estimate of the
95th percentile will be biased because of having only one observation
per worker, but this could be corrected using external information.

In response to questions from Mr. Miller, Dr. Carriquiry clarified that
EPA was interested in the distribution representative of one day’s
observations.  Given one observation per worker, the ideal approach
would be to have every worker observed over many days, and then take the
average of these observations to determine the distribution.  She
continued by reiterating that the estimates at the tails of the
distribution will be too long because of additional noise that cannot be
eliminated without replicates; thus, estimates at the 95th percentile of
exposure will be higher than the true exposure.  The overall result of
this will be protective of workers, and estimates at the center of the
distribution will not be biased.  Dr. Popendorf noted that slide 47 of
the Agency’s presentation described this effect.

Mr. Jordan clarified that EPA performs several types of risk assessment
for worker exposure including 1-day risk (exposure in a single day of
work) and short-intermediate term risk (4 to 5 days to 3 to 4 weeks of
exposure).  The value used for exposure components can differ between
these two levels of exposure because over multiple days, an individual
handler will receive more or less exposure, based on his activities. 
Without data to characterize within-worker variability, use of a single
day’s exposure data will overstate the multiple day risk and the 95th
percentile of exposures for multiple days will be greater than that of
one day.  The 95th percentile of a single day measured by the task
forces will correspond to the 95th percentile of 1-day exposures. 
Without repeated measures, EPA acknowledges that measures of
multiple-day exposure cannot be developed without making some
adjustments.  

Dr. Johnson noted that the measure of 1-day exposure should include both
between-worker and within-worker variation.  An advantage of having
repeated measures is that the two components can be estimated and
parameter estimates that depend on the variance of these components can
be made.  Dr. Carriquiry noted that for the distribution sought by EPA,
within-worker variance is noise and between-worker variance is needed. 
There are external sources of information to measure within-worker
variance (i.e., PHED data).

Subject Recruitment and Enrollment

Introduction

Mr. Carley presented EPA’s evaluation of subject recruitment and
enrollment issues.  His presentation focused on the AHETF monitoring
program, which is expected to be more complicated than that of the
AEATF.  The AHETF recruiting process begins with recruitment of
cooperating growers or landowners (hereafter known as cooperators). 
Potential cooperators are identified; screening and selection begins
after final protocol approval.  Any proposed deviations from the final
protocol must be justified.  After this, cooperators identify a
potential worker pool.  The Study Director (SD) presents the proposed
research to the workers as a group and then seeks informed consent from
individual workers.

Mr. Carley provided detail on the grower recruitment process.  Scenario
designs specify the needed characteristics of a cluster.  Based on this,
the Local Site Coordinator (LSC) suggests potential sites and
cooperators to the SD.  The LSC also identifies needed equipment and may
be involved in the management of samples, although this is under
consideration.  The SD develops the protocol without final
identification of the site(s).  The protocol is then reviewed (i.e., by
IRB, CDPR, EPA, and HSRB), and after revision and final approval of the
protocol by the IRB, the LSC recruits potential cooperating growers. 
The SD meets with potential cooperators to determine that the grower
meets the requirements specified in the scenario design (i.e.,
appropriate crop, equipment, skilled workers, and use of surrogate
chemicals), agrees to allow research on his/her property, promises
access to workers for recruitment, and promises not to influence
workers’ decisions.  Based on these criteria, the SD selects the study
cooperator(s) and adds any needed specifications to the protocol.

To recruit workers, the SD meets with workers as a group; the grower
will not attend this meeting.  The SD explains the research to the
workers and provides IRB-approved recruitment materials.  The SD also
may show IRB-approved videos illustrating the study procedures and use
of dosimeters.  After soliciting expressions of interest by individual
workers, the SD meets individually with interested candidates to review
informed consent documents, answer any questions, and solicit consent to
participate.  Because the SDs are primarily English-speaking and many of
the workers are expected to primarily speak Spanish, the assistance of
bilingual workers may be needed.

Preliminary EPA concerns include whether the study risks and benefits
can be fully assessed without knowing where a study will be conducted;
some study sites may employ a Spanish-speaking population of workers, so
it will be difficult to judge if the informed consent process is
adequate.  Also, although EPA has been promised study-specific
recruiting plans, the Agency questions whether such plans can be
developed before cooperators are finally identified.

EPA’s overarching concerns in recruiting, screening, and obtaining
consent include ensuring equitable subject selection, fully informed and
voluntary choice, and respect for subjects.  Equitable subject selection
must ensure representativeness, which hinges on the scenario-level
design and the choice of appropriate clusters and cooperators.  The
Agency considers the lists of appropriate inclusion/exclusion factors
(i.e., over 18 years of age, in good health, not nursing or pregnant,
and of normal intelligence to be appropriate).  The exclusion factors
also include those who habitually wear more than the typical PPE (these
workers would not be representative of the general worker population)
and also LSCs and employees of the SD.  One area in which EPA foresees
difficulties is the exclusion of parties with an interest in the
research.  The Agency may call for more extensive exclusion in this
regard for a particular study because of possible undue interest in the
data or influence on other subjects.

The involvement of vulnerable populations in this research also is a
concern of EPA.  Several vulnerable populations have been defined intask
force governing documents.  Language proficiency and dependent
relationships with the growers/employers are areas of specific concern. 
To help mitigate these issues, the recruiting strategy will be defined
for each with input from the community.  Study personnel are encouraged
to develop ties to the community to help develop the recruiting
strategy.

Fully informed choice requires the capacity to make decisions; the SD
will assess this in the informed consent interview.  The recruitment
process has been designed with the assumption that English and Spanish
will be the main languages spoken by workers.  If the worker speaks only
Spanish, a translation of the ICF performed by the Western IRB will be
given to the worker, and an interpreter will be provided if the SD
speaks only English.  Literacy also is expected to be an issue.  If a
worker is illiterate, the ICFs will be read to him and a witness will be
required.  An interpreter will be needed if the worker speaks only
Spanish and the SD only English; a witness also will be needed if the
worker also is illiterate in Spanish.  The interpreter in this situation
could be an employee of the grower or a person chosen by the subject. 
The SD must be certain that the interpreter can provide an adequate
description of the research and ICF.  The interpreter is not considered
to be part of the research team and will not sign the ICF.  The
witnesses must be fluent in the language in which the informed consent
discussion takes place.

EPA has concerns that the processes for English-speaking and
Spanish-speaking workers are not equivalent.  Depending on the
language(s) spoken by the SD and the worker, two or three parties may be
involved in the informed consent process.  The presence of an
interpreter may inhibit workers from asking questions.  In addition, the
interpreter may not understand the research well enough to translate
accurately, and it would be difficult for a non-Spanish speaking SD to
determine this.  The SD also will be unable to confirm worker
understanding.  The ability of study observers to communicate with
workers also will be impaired by language differences.

The complexity of the consent materials also presents a challenge.  The
AEATF wipe protocol informed consent was provided to Board members as an
example containing all required elements in an accessible form and
language.  A description of the development of these materials is
provided in the governing documents because preparation of appropriate
consent materials is an essential element of the study-specific
recruiting plan and different processes have been proposed for different
languages.  The wipe protocol also provides an example of how risks and
benefits will be communicated.  Confirming the subject’s full and
complete understanding of the protocol and risks and benefits is the
responsibility of the SD, which will be difficult if the worker speaks
only Spanish.

Fully voluntary choice also will require management of dependent
relationships; the nature of the relationships between workers and
growers will be site specific.  Cooperators must promise not to
influence workers and employees of interested entities will be excluded.
 To minimize peer pressure, no specific discussion of informed consent
will occur until a worker expresses interest.  Informed consent
discussions will be held in private, and pregnancy tests and discussions
of results also will be held privately.  Real alternatives to
cooperation will be provided by growers and explained by informed
consent documents.

An AHETF Employer Promise has been developed.  This promise calls for
the employer to allow AHETF to recruit any employees with applicable
training and experience in the tasks involved in the study, as
determined by the SD.  The employer acknowledges possible benefits, but
promises to neither encourage or discourage employees to participate in
the study, and promises that an employee’s decision to participate,
not to participate, or to withdraw from participation in the study will
have no impact on his/her employment status or pay.  Employees who
decide not to participate, who withdraw from participation, or who
complete participation in less than a typical work shift will be offered
alternative work at their usual pay to complete their usual work shift
and employees will receive their normal pay for days they participate in
the study.  Mr. Carley acknowledged that this promise had been
thoughtfully developed and covered most of his concerns.  He expressed
some concern that field investigators may need to be more involved in
recruitment and consent processes to alleviate some of the
responsibilities of the SD.

Mr. Carley described the AEATF consent process, which contains two
“fundamentally different” paradigms compared to the AHETF.  The
studies performed under the AEATF will be conducted either at an active
worksite, which will require supervisor agreement similar to that
described for the AHETF and will require interested workers to meet with
the SD at the workplace, or will recruit subjects to work at a site not
under their supervisor’s control.  In this case, flyers will be posted
at workplaces and interested workers will be instructed to contact the
Field Coordinator by telephone.  An IRB-approved script will be used to
explain the study and confirm workers’ interest; interested workers
will then meet with the SD at the SD’s office.

As part of respect for subjects, incentive payments ($20 for
participating in an informed consent interview and $80 for beginning
exposure monitoring) have been proposed.  These fees are subject to
refinement based on community input.  To ensure privacy and
confidentiality, no records of candidates who do not qualify or consent
to participate will be kept, data will be collected by subject code,
linkage of names and addresses to subject codes will be securely
maintained, and subjects will not be identified in reports or in
databases.  Given the potential benefit of providing feedback on work
habits and exposure to subjects, if a subject is interested in his or
her own data, they should be provided with the data.  The task forces
have been encouraged to incorporate a description of this process in the
SOP governing privacy of subject identity.

The Board was asked to consider whether the governing documents and
associated SOPs of the AHETF and AEATF research programs included
comprehensive and appropriate protections for human subjects of the
research and, if not, to describe overlooked issues.  The handling of
language differences was specified as an area requiring further
refinement and the Board was asked to determine whether EPA has
overlooked any other areas in need of revision.

Public Comments

	Dr. Brimijoin invited oral public comment on subject recruitment and
enrollment issues.  No oral public comments were received.

Board Discussion

Dr. Menikoff opened the discussion and complimented AHETF and EPA for
their thorough review of the recruiting process and incorporating Board
suggestions from previous meetings into the process.  He agreed with Mr.
Carley that only relatively minor refinements are needed.  Given the
vulnerability of the subjects, the involvement of worker representatives
could be emphasized more strongly.  Including subject advocates and
representatives will strengthen protection for workers and the governing
documents should include a discussion of this issue.  Dr. Menikoff
advised the task forces to consider the readability of consent
materials, because many potential subjects may have low literacy.

Dr. Chadwick noted that the AEATF documents describing the review
process had a table of contents that listed specific chapters describing
sampling, subject recruitment, and IRB review, but the pages in the
document read, “left intentionally blank.”  He found details on
these issues in Chapter 2 and thus advised AEATF to change “left
intentionally blank” to “described in Chapter 2.”  He also
complimented the task forces on their efforts and inclusion of
recommendations from HSRB and other agencies.

Dr. Chadwick specifically mentioned the age limit placed on pregnancy
testing (50 years) in the AEATF documents and recommended that an
exclusion based on menopause would be more appropriate than an age-based
exclusion.  AEATF also could consider testing all women, which is the
plan of the AHETF.  Dr. Chadwick also asked that the documents describe
more completely how documents, such as those related to pregnancy
testing, would be destroyed (i.e., shredded) to protect privacy.

Dr. Chadwick requested clarification on language issues.  For example,
AEATF states that if more than 15 percent of the sample is non-English
speaking, translated documents will be provided.  AHETF offers to
provide all documents in English and Spanish; this is the better
approach.  Other Agency guidelines require that ICFs be presented in the
potential subjects’ native language.  Regarding witnesses and
interpreters, Dr. Chadwick recommended that witnesses not serve as
interpreters and cautioned the task forces to be aware of the
qualifications of the interpreters.  He agreed that including more
bilingual study investigators would be valuable.

Dr. Sharp complimented the task forces on the documents.  He noted that
when specific studies are reviewed, a recurrent theme will be protecting
workers from undue influence by growers.  Although this is mentioned,
the task forces should have a management plan in place for each study
that will ensure workers’ right to decline to participate in the
study.  For example, if a grower has 50 workers who are all eligible to
participate, perhaps only 25 of the 50 would be selected to participate,
without revealing to the grower which workers wished to participate but
were not selected and which workers declined to participate.

Concerning the translation processes, Dr. Sharp noted that the task
forces proposed that materials be translated by the IRB, but speculated
that more materials will likely need to be translated than an IRB may be
willing to translate.  Dr. Chadwick informed Dr. Sharp that Western IRB
specifically provides a translation service.  Dr. Sharp continued with
his discussion of language issues and commended the task forces on their
thorough discussion of these issues.  He stated that ensuring that all
volunteers have access to all documents (ICFs, pamphlets, etc.) in both
English and Spanish is important.  He agreed with Mr. Carley on the need
to employ bilingual study staff.  Volunteers may have questions both
before and after undergoing the informed consent process and there will
be a continuing need for bilingual staff during the course of the study.
 Dr. Sharp expressed concern about the phrase “translators of
convenience,” which could mean co-workers or family members.  Using
such translators is discouraged in clinical research because these
people are unlikely to be familiar enough with the research to properly
inform the subject.  Dr. Sharp was supportive of the task forces’ plan
to include an impartial third-party witness for subjects who are unable
to read.  A clarification of the procedures used to recruit such
witnesses is needed; Dr. Sharp noted that it would be inappropriate for
the translators to also serve as witnesses.  Another option is to hire
dedicated consent monitors or research subject advocates who are
specifically trained to determine whether consent is fully voluntary. 
Dr. Sharp also recalled Dr. Fenske’s comment concerning the likelihood
of encountering workers who speak languages other than Spanish or
English and suggested that EPA consider if it would be appropriate to
restrict eligibility to only English and/or Spanish speakers, unless
this would introduce bias.

Dr. Fitzpatrick agreed that the potential for bias needs to be addressed
if EPA considers restricting the studies to only English and/or Spanish
speakers.  She supported EPA’s proposal to employ Spanish-speaking
investigators to assist with consent processes.  She inquired how the
task forces would perform QA on consent processes that include
interpreters, since the interpreters do not sign the ICF, thus there
would be no evidence of use of an interpreter.  She suggested that EPA
ask for a certified translation (translation forward and back to the
original language of the document) to ensure accuracy.  

Dr. Fitzpatrick addressed the medical compensation, noting that the task
forces agreed to compensate for reasonable medical costs if subjects do
not have insurance.  She stated that “reasonable medical costs” need
to be defined more specifically, for example, do these include initial
treatment, long-term care, or missed days of work.  She also noted that
subjects self-report their levels of health—most people tend to report
themselves as healthy—and asked how the task forces would handle
injury that occurred because of an unreported health condition.
Dr. Fitzpatrick acknowledged the task forces’ plans for adverse event
reporting and inquired how decisions to modify ICFs would be made if new
risks are identified as a result of adverse event reporting.  She added
that ICFs also should inform subjects that if new information related to
risk arises as a result of the study; subjects will be given this
information so they can re-evaluate their consent to participate.  She
commented that if the task forces plan to permit subjects access to
their own data, the task forces also must provide assistance with
interpretation of the data.  Dr. Fitzpatrick concluded by noting that
subjects will be asked to show a driver’s license or workers card to
prove immigration status and asked if the task forces have plans in
place to protect the privacy of illegal workers.  She also asked, while
acknowledging that EPA has no control over this request, why California
requested review of the studies by the Florida IRB.

Mr. Carley clarified for Dr. Fitzpatrick that because these are largely
one-day studies, the issue of withdrawing consent because of adverse
events is minor.  Dr. Fitzpatrick requested task force response to this
matter if an adverse event potentially affecting participation occurred
at another site.  Mr. Carley considered this to be unlikely.

Dr. Popendorf noted that the records of subjects who decline to
participate are destroyed, but said that this data may be useful for
populating a distribution of conditions if a random sampling approach is
used.  He expressed concern that if EPA continues with its proposed PDS
approach, future users of the database may not be aware of this and may
assume a random sampling approach was used; to prevent this, the
database should contain a statement indicating that PDS was used.  Dr.
Carriquiry remarked that she believed EPA will adopt a randomized
sampling strategy, which also will help protect workers from coercion. 
A list of possible workers would be provided to investigators, but the
grower will not know which workers were selected or which declined to
participate.

Dr. Brimijoin summarized that the Board was impressed with the
consideration of the recruitment process and complimented the level of
detail and careful thinking.  He stated that it is likely these studies
will achieve a high standard of scientific and ethical performance. 
Board members described some areas for improvement, including involving
worker representatives early in the recruitment process and describing
this in the governing documents; reconsideration of privacy protection
and exclusion factors for pregnancy tests; ways to manage and reduce
potential grower influence on subject participation; attention to
language issues and Board endorsement of the plan to involve bilingual
investigators with appropriate knowledge and expertise; technical points
concerning relying on translators of convenience; identification of
appropriate witnesses; clarification of covered medical costs in case of
adverse events; issues of confidentiality for illegal workers; and plans
for ICFs to alert subjects to newly discovered risks.

	Dr. Lewis thanked Dr. Brimijoin for serving as chair of this meeting. 
He also thanked the Board members for their efforts and his EPA
colleagues for their preparation and the presentations given at this
meeting.  He stated that a Federal Register notice will be published to
inform the public about the availability of the Board’s report for
this meeting.  Dr. Lewis also stated that review of the April 2007
report would be conducted by teleconference and that a Federal Register
notice would be published to inform the public of this event.

Dr. Brimijoin adjourned the meeting.

Respectfully submitted:

Paul I. Lewis, Ph.D.

Designated Federal Officer

Human Studies Review Board

United States Environmental Protection Agency

Certified to be true by:

William S. Brimijoin, Ph.D.

Vice Chair

Human Studies Review Board

United States Environmental Protection Agency

NOTE AND DISCLAIMER:  The minutes of this public meeting reflect diverse
ideas and suggestions offered by Board members during the course of
deliberations within the meeting.  Such ideas, suggestions, and
deliberations do not necessarily reflect definitive consensus advice for
the Board members.  The reader is cautioned to not rely on the minutes
to represent final, approved, consensus advice and recommendations
offered to the Agency.  Such advice and recommendations may be found in
the final report prepared and transmitted to the EPA Science Advisor
following the public meeting.

Attachments 

Attachment A 		List of HSRB Members 

Attachment B 		Federal Register Notice Announcing Meeting 

Attachment C 		Meeting Agenda 

 

Attachment A

EPA HUMAN STUDIES REVIEW BOARD MEMBERS 

Chair

Celia B. Fisher, Ph.D. *

Marie Ward Doty Professor of Psychology

Director, Center for Ethics Education

Fordham University

Bronx, NY 

Vice Chair

William S. Brimijoin, Ph.D. **

Chair and Professor 

Molecular Pharmacology and Experimental Therapeutics

Mayo Foundation

Rochester, MN 

Members

Alicia Carriquiry, Ph.D.

Professor 

Department of Statistics

Iowa State University

Ames, IA 

Gary L. Chadwick, PharmD, MPH, CIP

Associate Provost

Director, Office for Human Subjects Protection

University of Rochester

Rochester, NY 

Janice Chambers, Ph.D., D.A.B.T.

William L. Giles Distinguished Professor

Director, Center for Environmental Health Sciences

College of Veterinary Medicine

Mississippi State University

Mississippi State, MS 

Richard Fenske, Ph.D., MPH 

Professor

Department of Environmental and Occupational Health Sciences

University of Washington

Seattle, WA 

Susan S. Fish, PharmD, MPH

Professor, Biostatistics & Epidemiology

Boston University School of Public Health

Co-Director, MA in Clinical Investigation

Boston University School of Medicine

Boston, MA 

Suzanne C. Fitzpatrick, Ph.D., DABT

Senior Science Policy Analyst

Office of the Commissioner

Office of Science and Health Coordination

U.S. Food and Drug Administration

Rockville, MD 

KyungMann Kim, Ph.D., CCRP

Professor and Associate Chair

Department of Biostatistics & Medical Informatics

School of Medicine and Public Health

University of Wisconsin-Madison

Madison, WI 

Kannan Krishnan, Ph.D. *

Professor

Département de santé environnementale et santé au travail

Faculté de médicine 

Université de Montréal

Montréal, QC  Canada

Michael D. Lebowitz, Ph.D.

Research Professor of Medicine & Epidemiology/Public Health

University of Arizona

Tucson, AZ 

Lois D. Lehman-Mckeeman, Ph.D. 

Distinguished Research Fellow, Discovery Toxicology

Bristol-Myers Squibb Company

Princeton, NJ 

Jerry A. Menikoff, M.D.  

National Institute of Health

Office of Human Subjects Research

Bethesda, MD 

Sean M. Philpott, Ph.D., M.Bioethics

Policy and Ethics Director

Global Campaign for Microbicides

Program for Appropriate Technology in Health

Washington, DC  

Richard Sharp, Ph.D. 

Assistant Professor of Medicine

Center for Medical Ethics and Health Policy 

Baylor College of Medicine 

Houston, TX 

* Not in attendance

**Served as Chair at meeting

	

Attachment B

Federal Register Notice Announcing Meeting

Human Studies Review Board; Notice of Public Meeting   

[Federal Register: June 6, 2007 (Volume 72, Number 108)]

[Notices]

[Page 31323-31325]

From the Federal Register Online via GPO Access [wais.access.gpo.gov]

[DOCID:E7-10859]

-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

[EPA-HQ-ORD-2007-0403; FRL-8322-7]

Human Studies Review Board; Notice of Public Meeting

AGENCY: Environmental Protection Agency (EPA).

ACTION: Notice.

-----------------------------------------------------------------------

SUMMARY: The U.S. Environmental Protection Agency's (EPA or Agency)
Office of the Science Advisor (OSA) announces a public meeting of the
Human Studies Review Board (HSRB) to advise the Agency on EPA's
scientific and ethical reviews of human subjects' research.

DATES: The public meeting will be held from June 27-June 29, 2007
approximately from 8:30 a.m. to approximately 5:30 p.m., Eastern Time.

    Location: Environmental Protection Agency, Conference Center—Lobby
Level, One Potomac Yard (South Bldg.), 2777 S. Crystal Drive, Arlington,
VA  22202.

    Meeting Access: Seating at the meeting will be on a first-come
basis. To request accommodation of a disability please contact the
person listed under FOR FURTHER INFORMATION CONTACT at least 10 business
days prior to the meeting, to allow EPA as much time as possible to
process your request.

    Procedures for Providing Public Input: Interested members of the
public may submit relevant written or oral comments for the HSRB to
consider during the advisory process. Additional information concerning
submission of relevant written or oral comments is provided in Unit I.D.
of this notice.

FOR FURTHER INFORMATION CONTACT: Any member of the public who wishes
further information should contact Paul Lewis, EPA, Office of the
Science Advisor, (8105R), Environmental Protection Agency, 1200
Pennsylvania Ave., NW, Washington, DC 20460; telephone number: (202)
564-8381; fax: (202) 564 2070; e-mail address:   HYPERLINK
"mailto:lewis.paul@epa.gov"  lewis.paul@epa.gov . General information
concerning the EPA HSRB can be found on the EPA Web site at   HYPERLINK
"http://www.epa.gov/osa/hsrb/"  http://www.epa.gov/osa/hsrb/ .

ADDRESSES: Submit your written comments, identified by Docket ID No. 

EPA-HQ-ORD-2007-0403, by one of the following methods:

    Internet:   HYPERLINK "http://www.regulations.gov" 
http://www.regulations.gov : Follow the on-line instructions for
submitting comments.

    E-mail:   HYPERLINK "mailto:ORD.Docket@epa.gov"  ORD.Docket@epa.gov
.

    Mail: Environmental Protection Agency, EPA Docket Center (EPA/DC),
ORD Docket, Mailcode: 28221T, 1200 Pennsylvania Ave., NW, Washington, DC
 20460.

    Hand Delivery: The EPA/DC Public Reading Room is located in the EPA
Headquarters Library, Room Number 3334 in the EPA West Building, located
at 1301 Constitution Ave., NW, Washington DC. The hours of operation are
8:30 a.m. to 4:30 p.m. Eastern Standard Time (EST), Monday through
Friday, excluding Federal holidays. Please call (202) 566-1744 or e-mail
the ORD Docket at   HYPERLINK "mailto:ord.docket@epa.gov" 
ord.docket@epa.gov  for instructions. Updates to Public Reading Room
access are available on the Web site (  HYPERLINK
"http://www.epa.gov/epahome/dockets.htm" 
http://www.epa.gov/epahome/dockets.htm ).

    Instructions: Direct your comments to Docket ID No.
EPA-HQ-ORD-2007-0403. EPA's policy is that all comments received will be
included in the public docket without change and may be made available
online at   HYPERLINK "http://www.regulations.gov" 
http://www.regulations.gov , including any personal information
provided, unless the comment includes information claimed to be
Confidential Business Information (CBI) or other information whose
disclosure is restricted by statute. Do not submit information that you
consider to be CBI or otherwise protected through   HYPERLINK
"http://www.regulations.gov"  http://www.regulations.gov  or e-mail. The
  HYPERLINK "http://www.regulations.gov"  http://www.regulations.gov 
Web site is an “anonymous access” system, which means EPA will not
know your identity or contact information unless you provide it in the
body of your comment. If you send an e-mail comment directly to EPA,
without going through   HYPERLINK "http://www.regulations.gov" 
http://www.regulations.gov , your e-mail address will be automatically
captured and included as part of the comment that is placed in the
public docket and made available on the Internet. If you submit an
electronic comment, EPA recommends that you include your name and other
contact information in the body of your comment and with any disk or
CD-ROM you submit. If EPA cannot read your comment due to technical
difficulties and cannot contact you for clarification, EPA may not be
able to consider your comment. Electronic files should avoid the use of
special characters, any form of encryption, and be free of any defects
or viruses.

SUPPLEMENTARY INFORMATION:

I. Public Meeting

A. Does this Action Apply to Me?

    This action is directed to the public in general. This action may,
however, be of interest to persons who conduct or assess human studies,
especially studies on substances regulated by EPA or to persons who are
or may be required to conduct testing of chemical substances under the
Federal Food, Drug, and Cosmetic Act (FFDCA) or the Federal Insecticide,
Fungicide, and Rodenticide Act (FIFRA). Since other entities may also be
interested, the Agency has not attempted to describe all the specific
entities that may be affected by this action. If you have any questions
regarding the applicability of this action to a particular entity,
consult the person listed under FOR FURTHER INFORMATION CONTACT.

[[Page 31324]]

B. How Can I Access Electronic Copies of This Document and Other Related
Information?

    In addition to using regulations.gov, you may access this Federal
Register document electronically through the EPA Internet under the

“Federal Register” listings at   HYPERLINK
"http://www.epa.gov/fedrgstr/"  http://www.epa.gov/fedrgstr/ .    

    Docket: All documents in the docket are listed in the

  HYPERLINK "http://www.regulations.gov"  http://www.regulations.gov 
index. Although listed in the index, some information is not publicly
available, e.g., CBI or other information whose disclosure is restricted
by statute. Certain other material, such as copyrighted material, will
be publicly available only in hard copy. Publicly available docket
materials are available either electronically in   HYPERLINK
"http://www.regulations.gov"  http://www.regulations.gov  or in hard
copy at the ORD Docket, EPA/DC, Public Reading Room. The EPA/DC Public
Reading Room is located in the EPA Headquarters Library, Room Number
3334 in the EPA West Building, located at 1301 Constitution Ave., NW,
Washington DC. The hours of operation are 8:30 AM to 4:30 PM EST, Monday
through Friday, excluding Federal holidays. Please call (202) 566-1744
or email the ORD Docket at   HYPERLINK "mailto:ord.docket@epa.gov" 
ord.docket@epa.gov  for instructions. Updates to Public Reading Room
access are available on the Web site (  HYPERLINK
"http://www.epa.gov/epahome/dockets.htm" 
http://www.epa.gov/epahome/dockets.htm ).

    EPA's position paper(s), charge/questions to the HSRB, and the
meeting agenda will be available by early June 2007. In addition, the
Agency may provide additional background documents as the materials
become available. You may obtain electronic copies of these documents,
and certain other related documents that might be available
electronically, from the regulations.gov Web site and the HSRB Internet
Home Page at   HYPERLINK "http://www.epa.gov/osa/hsrb/" 
http://www.epa.gov/osa/hsrb/ . For questions on document availability or
if you do not have access to the Internet, consult the person listed
under FOR FURTHER INFORMATION CONTACT.

C. What Should I Consider as I Prepare My Comments for EPA?

			

    You may find the following suggestions helpful for preparing your
comments:

    a. Explain your views as clearly as possible.

    b. Describe any assumptions that you used.

    c. Provide copies of any technical information and/or data you used
that support your views.

    d. Provide specific examples to illustrate your concerns and suggest
alternatives.

    e. To ensure proper receipt by EPA, be sure to identify the docket

ID number assigned to this action in the subject line on the first page
of your response. You may also provide the name, date, and Federal

Register citation.

D. How May I Participate in this Meeting?

    You may participate in this meeting by following the instructions in
this section. To ensure proper receipt by EPA, it is imperative that you
identify docket ID number EPA-HQ-ORD-2007-0403 in the subject line on
the first page of your request.

    a. Oral comments. Requests to present oral comments will be accepted
up to June 20, 2007. To the extent that time permits, interested persons
who have not pre-registered may be permitted by the Chair of the HSRB to
present oral comments at the meeting. Each individual or group wishing
to make brief oral comments to the HSRB is strongly advised to submit
their request (preferably via email) to the person listed under FOR
FURTHER INFORMATION CONTACT no later than noon, Eastern Time, June 20,
2007 in order to be included on the meeting agenda and to provide
sufficient time for the HSRB Chair and HSRB Designated Federal Officer
(DFO) to review the agenda to provide an appropriate public comment
period. The request should identify the name of the individual making
the presentation, the organization (if any) the individual will
represent, and any requirements for audiovisual equipment (e.g.,
overhead projector, LCD projector, chalkboard). Oral comments before the
HSRB are limited to five minutes per individual or organization. Please
note that this limit applies to the cumulative time used by all
individuals appearing either as part of, or on behalf of an
organization. While it is our intent to hear a full range of oral
comments on the science and ethics issues under discussion, it is not
our intent to permit organizations to expand these time limitations by
having numerous individuals sign up separately to speak on their behalf.
If additional time is available, there may be flexibility in time for
public comments. Each speaker should bring 25 copies of his or her
comments and presentation slides for distribution to the HSRB at the
meeting. 

    b. Written comments. Although you may submit written comments at any
time, for the HSRB to have the best opportunity to review and consider
your comments as it deliberates on its report, you should submit your
comments at least five business days prior to the beginning of the
meeting. If you submit comments after this date, those comments will be
provided to the Board members, but you should recognize that the Board
members may not have adequate time to consider those comments prior to
making a decision. Thus, if you plan to submit written comments, the
Agency strongly encourages you to submit such comments no later than
noon, Eastern Time, June 20, 2007. You should submit your comments using
the instructions in Unit I.C. of this notice. In addition, the Agency
also requests that person(s) submitting comments directly to the docket
also provide a copy of their comments to the person listed under FOR
FURTHER INFORMATION CONTACT. There is no limit on the length of written
comments for consideration by the HSRB.

E. Background

    A. Topics for Discussion. The HSRB is a Federal advisory committee
operating in accordance with the Federal Advisory Committee Act (FACA) 5
U.S.C. App.2 section 9. The HSRB provides advice, information, and
recommendations to EPA on issues related to scientific and ethical
aspects of human subjects research. The major objectives of the HSRB are
to provide advice and recommendations on: (a) Research proposals and
protocols; (b) reports of completed research with human subjects; and
(c) how to strengthen EPA’s programs for protection of human subjects
of research. The HSRB reports to the EPA Administrator through EPA’s
Science Advisor. 

    The June 27-29, 2007 meeting of the Human Studies Review Board will
address scientific and ethical issues surrounding:

    • A research proposal from Carroll-Loye Biological Research to
evaluate the efficacy of two conditionally registered products
containing picaridin in repelling mosquitoes in the field.

    • A research proposal from Insect Control & Research, Inc. to
evaluate the efficacy of two unregistered products containing picaridin
in repelling mosquitoes in the field.

    • A completed study measuring the effects on human subjects of
acute inhalation exposure to acrolein. Acrolein is an active ingredient
used in biocides in agricultural and industrial water supply systems and
is currently undergoing reregistration.

    • Three completed clinical studies of the efficacy and side
effects of 4-aminopyridine when used as a therapeutic agent to treat
neurological

[[Page 31325]]

Symptoms in patients with either spinal cord injury or multiple
sclerosis, 4-aminopyridine is an active ingredient used in bird
repellents that is currently undergoing reregistration.

   • Extensive background materials concerning research to quantify
the level of exposure received by people who mix, load, and apply
pesticides. These materials, which were prepared by the Agricultural
Handlers Exposure Task Force and by the Antimicrobial Exposure
Assessment Task Force, generally explain the scope of the research
programs being proposed by the Task Forces and describe the general
scientific framework for conducting the research. In addition, each Task
Force has provided Standard Operating Procedures which will guide the
conduct of the studies. 

    The Board may also be reviewing draft HSRB reports for subsequent
Board approval. Finally, the Board may also discuss planning for future
HSRB meetings. 

    b. Meeting Minutes and Reports. Minutes of the meeting, summarizing
the matters discussed and recommendations, if any, made by the advisory
committee regarding such matters will be released within 90 calendar
days of the meeting. Such minutes will be available at   HYPERLINK
"http://www.epa.gov/osa/hsrb/"  http://www.epa.gov/osa/hsrb/  and  
HYPERLINK "

http://www.regulations.gov"  

http://www.regulations.gov . In addition, information concerning a Board
meeting report, if applicable, can be found at   HYPERLINK
"http://www.epa.gov/osa/hsrb/"  http://www.epa.gov/osa/hsrb/  or from
the person listed under FOR FURTHER INFORMATION CONTACT.

Dated: May 31, 2007.

Kevin Teichman,

Acting EPA Science Advisor

[FR Doc. E7-10859 Filed 6-6-07; 8:45 am]

BILLING CODE 6560-50-P

Attachment C

UNITED STATES ENVIRONMENTAL PROTECTION AGENCY

HUMAN STUDIES REVIEW BOARD (HSRB)

PUBLIC MEETING

June 27-29, 2007*

ONE POTOMAC YARD

ARLINGTON, VA

HSRB Web Site: http://www.epa.gov/osa/hsrb/

Docket Telephone: (202) 566-1752

Docket Number: EPA-HQ-ORD-2007-0403

BOARD REVIEW OF ITS DRAFT APRIL 18-20, 2007 HSRB MEETING REPORT,
ORIGINALLY SCHEDULED AT THE BEGINNING OF THIS MEETING, MAY BE
RESCHEDULED EITHER LATER DURING THE MEETING OR AT A SUBSEQUENT
TELECONFERENCE.  IN ADDITION, TODAY’S MEETING MAY START EARLIER THAN
LISTED ON THE AGENDA*  

Wednesday, June 27, 2007

11:30 a.m.	Convene Meeting and Identification of Board Members

			William Brimijoin, Ph.D. (HSRB Vice Chair)

11:40 a.m.	Welcome

			George Gray, Ph.D. (EPA Science Advisor) 

11:50 a.m.	Opening Remarks

			Debbie Edwards, Ph.D. (Director, Office of Pesticide Programs, [OPP])

12:00 p.m.	Meeting Administrative Procedures

		Paul Lewis, Ph.D. (Designated Federal Officer [DFO], HSRB, OSA, EPA) 

12:05 p.m.	EPA Follow-up on HSRB Recommendations

		Mr. William Jordan (OPP, EPA)   

Carroll-Loye Picaridin Mosquito Repellency Protocol LNX-001 

12:15 p.m.	Science and Ethics of Carroll-Loye Protocol

		Mr. Kevin Sweeney (OPP, EPA) and Mr. John Carley (OPP, EPA) 

1:00 p.m.	Lunch  

2:00 p.m.	Public Comments 

2:30 p.m.	Board Discussion 

a. 	If the proposed research described in Protocol LNX-001 from
Carroll-Loye Biological Research is revised as suggested in EPA’s
review, does the research appear likely to generate scientifically
reliable data, useful for assessing the efficacy of the test substances
for repelling mosquitoes? 

b. 	If the proposed research described in Protocol LNX-001 from
Carroll-Loye Biological Research is revised as suggested in EPA’s
review, does the research appear to meet the applicable requirements of
40 CFR part 26, subparts K and L?  

ICR Picaridin Mosquito Repellency Protocol

3:30 p.m.	Science and Ethics of ICR Protocol

		Mr. Kevin Sweeney (OPP, EPA) and Mr. John Carley (OPP, EPA) 

4:15 p.m.	Public Comments 

4:45 p.m.	Break

5:15 p.m.	Board Discussion   

a. 	If the proposed research described in ICR’s proposed picaridin
protocol is revised as suggested in EPA’s review, does the research
appear likely to generate scientifically reliable data, useful for
assessing the efficacy of the test substances for repelling mosquitoes? 

b. 	If the proposed research described in ICR’s proposed picaridin
protocol is revised as suggested in EPA’s review, does the research
appear to meet the applicable requirements of 40 CFR part 26, subparts K
and L?  

6:15 p.m.	Adjournment  

Thursday, June 28, 2007

8:30 a.m.	Convene Meeting

		Steven Brimijoin, Ph.D. (HSRB ViceChair)

8:40 a.m.	Follow-up From Previous Day’s Discussion

			Mr. William Jordan (OPP, EPA)  

Acrolein 

8:50 a.m.	Acrolein

Abdallah Khasawinah, Ph.D. (OPP, EPA) and Mr. John Carley (OPP, 

EPA) 

9:45 a.m.	Public Comments 

10:15 a.m.	Break

10:30 a.m.	Board Discussion 

a.  	The Agency has concluded that this study contains information
sufficient for assessing human risk resulting from potential acute
inhalation exposure.  Please comment on whether the study is
sufficiently sound, from a scientific perspective, to be used to
estimate a safe level of acute inhalation exposure to acrolein.  

b. 	Please comment on the following:

(1) Is there clear and convincing evidence that the conduct of the study
was fundamentally unethical?

(2) Is there clear and convincing evidence that the conduct of the study
was significantly deficient relative to the ethical standards prevailing
at the time the research was conducted?

4-Amino Pyridine 

11:30 a.m.	4-Amino Pyridine

Abdallah Khasawinah, Ph.D. (OPP, EPA) and Mr. John Carley (OPP, 

EPA) 

12:15 p.m.	Public Comments 

12:30 p.m.	Lunch 

1:30 p.m.	Board Discussion  

a.	The Agency’s weight-of-evidence (WOE) document for 4-aminopyridine
describes the study design and results of three clinical trials
(Grijalva et al. 2003, Segal et al. 1999, and Van Diemen et al. 1993). 
The WOE document also discusses the Agency’s conclusion that these
studies provide sufficient information to establish a point of departure
for the assessment of the risk to humans resulting from all potential
durations of exposure to 4-AP.  Please comment on whether the studies
are sufficiently sound, from a scientific perspective, to be used to
derive a point of departure for estimating risk to humans from exposure
to 4-AP.

b.	Please comment on the following:

	(1) Is there clear and convincing evidence that the conduct of any of
the clinical studies was fundamentally unethical?

	(2) Is there clear and convincing evidence that the conduct of any of
the clinical studies was significantly deficient relative to the ethical
standards prevailing at the time the research was conducted?

2:30 p.m.	Break

AEATF and AHETF Research Programs 

2:45 p.m.	Introduction – EPA Presentation

William Jordan (EPA, OPP)

2:50 p.m.	Overview / Risks and Benefits of Handler Research – EPA
Presentation

Mr. John Carley (OPP, EPA)

3:30 p.m.	Public Comments   

3:45 p.m.	Board Discussion  

Risks and Benefits of Handler Research

1. Will the Task Forces’ Governing Documents considered in conjunction
with the additional study- and scenario-specific information specified
above provide an adequate basis for assessing whether the risks of
conducting a particular study are justified by the expected benefits of
the proposed research?  If not, what additional information should be
provided for an IRB, EPA, and the HSRB?

5:00 p.m.	Addressing Potential Sources of Underestimation Bias/QA and QC
Controls 			– EPA Presentation

		Mr. Jeff Dawson (OPP, EPA)  

5:30 p.m.	Public Comments 

5:45 p.m.	Adjournment 

 

Friday, June 29, 2007

8:30 a.m.	Convene Meeting

		Steven Brimijoin, Ph.D. (HSRB Vice Chair)

8:40 a.m.	Follow-up From Previous Day’s Discussion

		Mr. William Jordan (OPP, EPA)  

AEATF and AHETF Research Programs: Addressing Potential Sources of
Underestimation Bias; QA and QC Controls (continued)

8:50 a.m.	Board Discussion

Addressing Potential Sources of Underestimation Bias

1. Has EPA appropriately characterized the limitations on the scientific
usefulness of a handler database that does not include data
characterizing the efficiency of residue removal procedures?  If not,
what limitations have been overlooked?

2. Has EPA identified the relevant scientific and practical
considerations affecting the choice to include biomonitoring, and has
EPA appropriately characterized the limitations on the scientific
usefulness of the resulting data if no biomonitoring is conducted?  If
not, what other considerations should bear on a decision to conduct
biomonitoring in addition to WBD?

QA/QC Controls

1. Do the Task Forces’ Standard Operating Procedures appear adequate
to ensure that the data resulting from the proposed research will be of
high quality?  If not, what other Quality Assurance or Quality Control
procedures need to be addressed?

10:00 a.m.	Break    

10:15 a.m.	Design of Scenario-Level Sampling; Statistical Justification
of Number of 			Clusters; and Monitoring Units and Within Worker
Variability – EPA 			Presentation

Mr. David Miller (OPP, EPA)

11:00 a.m.	Public Comments   

11:15 a.m.	Board Discussion 

Design of Scenario-Level Sampling

With regard to the AHETF and AEATF plans to conduct their proposed
handler research using purposive diversity sampling strategies: 

1. Has EPA identified the relevant scientific and practical
considerations affecting the choice of a strategy for sample selection? 
If not, what other considerations should bear on the choice?

2. Does the HSRB agree with EPA that the Task Forces should provide
scenario-specific information about the availability of data to identify
significant variables (other than AaiH) potentially influencing exposure
and about the feasibility of developing a sampling strategy to address
those variables quantitatively?   If not, what additional information is
needed?

3. Has EPA appropriately characterized the limitations on the scientific
usefulness of the resulting data attributable to the choice of the
sampling strategy?  If not, what has EPA overlooked?

Statistical Justification of Number of Clusters

1. What additional information, if any, would the HSRB need to assess
the adequacy of the justification for the number of clusters and number
of MUs in specific AHETF and AEATF study proposals?

Within-Worker Variability

1. Has EPA appropriately characterized the limitations on the scientific
usefulness of a database that does not include repeated measures?  If
not, what limitations has EPA overlooked?

12:15 p.m.	Lunch  

1:15 p.m.	Subject Recruitment and Enrollment Issues – EPA Presentation

Mr. John Carley (OPP, EPA)

1:45 p.m.	Public Comments   

2:00 p.m.	Board Discussion 

1. Does the Board agree that the Governing Documents and associated SOPs
of the AHETF and AEATF research programs include comprehensive and
appropriate protections for human subjects of the research?  If not,
what has been overlooked?

2. In singling out the handling of language differences as an area
requiring further refinement, has EPA overlooked other areas in need of
revision?  If so, what?

3:00 p.m.	Adjournment

	Steven Brimijoin, Ph.D. (HSRB Vice Chair) and Paul Lewis, Ph.D. (DFO,
HSRB, OSA, EPA) 

r

”

Ó

í

k

l

G

H

r

r

–

¸

Ó

î

k

l

ž

 h©

 hP

2

3

	

@

R

W

k

s

z

~

…

†

¡

°

ê

ò

#

%

-

M

O

P

Q

 h

Ê

Ë

 hÅ

h¨ 

 h3

b

c

ༀꂄᄅ킄㜂$␸䠀$葞֠葠ː摧灒±᠀

  h

 h

萏Ũ萑ﺘ葞Ũ葠ﺘ摧灒±

ༀꂄᄅ킄㜂$␸䠀$葞֠葠ː摧灒±ጀ

* Please be advised that agenda times are approximate and subject to
change. For further information, please contact the Designated Federal
Officer for this meeting, Paul Lewis via telephone: (202) 564-8381 or
email:   HYPERLINK "mailto:lewis.paul@epa.gov"  lewis.paul@epa.gov .

 PAGE   

  PAGE  2  of   NUMPAGES  101 

