March 13th, 2017
EPA-HSRB-17-1
Robert J. Kavlock, Ph.D.
Acting EPA Science Advisor
Office of the Science Advisor
1200 Pennsylvania Avenue, NW
Washington, DC 20460
Subject: January 25-26, 2017 EPA Human Studies Review Board Meeting Report
Dear Dr. Kavlock,
The United States Environmental Protection Agency (EPA or Agency) requested that the
Human Studies Review Board (HSRB) provide scientific and ethics reviews of five items:
•	A published study, Cimetidine-Carbaryl Interaction in Humans: Evidence for an
Active Metabolite of Carbaryl, authored by D. Gail May, Rebecca J. Naukam, J.
Reddy Kambam, and Robert A. Branch. Journal of Pharmacology Exposure
Therapy (1992) 262(3), 1057-1061;
•	An unpublished study: A randomized double blind study with malathion to determine
the residues of malathion dicarboxylic acid (DCA), malathion monocarboxylic acid
(MCA), dimethyl phosphate (DMP), dimethyl thiophosphate (DMTP), and dimethyl
dithiophosphate (DMDTP) in human urine;
•	A published study, Methylisothiazolinone contact allergy and dose-response
relationships, authored by Michael D. Lundov, Claus Zachariae, and Jeanne D.
Johansen. Contact Dermatitis (2011) 64, 330-336.
•	A published study, Methylisothiazolinone in rinse-off products causes allergic contact
dermatitis: a repeated open-application study, authored by K. Yazar, M. D.
1

-------
Lundov, A. Faurschou, M. Matura, A. Boman, J. D. Johansen, and C. Liden. British
Journal of Dermatology (2015) 173,115-122.
• A published study, An evaluation of dose/unit area and time as key factors
influencing the elicitation capacity of
methylchloroisothiazolinone/methylisothiazolinone (MCI/MI) in MCI/MI-allergic
patients, authored by Claus Zachariae, Anne Lerbaek, Pauline M. McNamee, John
E. Gray, Mike Wooder, and Torkil Menne. Contact Dermatitis (2006) 55,160-166.
The charge questions posed to the Board included assessing scientific and ethical aspects of these
studies viewed individually, as well as the proposed combined use of the three studies related to
methylisothiazolinone for a weight of evidence approach to risk assessment.
The Board's responses to the charge questions and detailed rationale and recommendations are
provided in the enclosed final meeting report.
Signed,
Liza Dawson, PhD
Chair
EPA Human Studies Review Board
2

-------
INTRODUCTION
On January 25-26, 2017, the United States Environmental Protection Agency's (EPA or Agency)
Human Studies Review Board (HSRB or Board) met to address the scientific and ethical charge
questions related to five items:
•	A published study, Cimetidine-Carbaryl Interaction in Humans: Evidence for an
Active Metabolite of Carbaryl, authored by D. Gail May, Rebecca J. Naukam, J.
Reddy Kambam, and Robert A. Branch. Journal of Pharmacology Exposure
Therapy (1992) 262(3), 1057-1061;
•	An unpublished study: A randomized double blind study with malathion to determine
the residues of malathion dicarboxylic acid (DCA), malathion monocarboxylic acid
(MCA), dimethyl phosphate (DMP), dimethyl thiophosphate (DMTP), and dimethyl
dithiophosphate (DMDTP) in human urine;
•	A published study, Methylisothiazolinone contact allergy and dose-response
relationships, authored by Michael D. Lundov, Claus Zachariae, and Jeanne D.
Johansen. Contact Dermatitis (2011) 64, 330-336.
•	A published study, Methylisothiazolinone in rinse-off products causes allergic contact
dermatitis: a repeated open-application study, authored by K. Yazar, M. D.
Lundov, A. Faurschou, M. Matura, A. Boman, J. D. Johansen, and C. Liden. British
Journal of Dermatology (2015) 173,115-122.
•	A published study, An evaluation of dose/unit area and time as key factors
influencing the elicitation capacity of
methylchloroisothiazolinone/methylisothiazolinone (MCI/MI) in MCI/MI-allergic
patients, authored by Claus Zachariae, Anne Lerbaek, Pauline M. McNamee, John
E. Gray, Mike Wooder, and Torkil Menne. Contact Dermatitis (2006) 55,160-166.
In addition, EPA requested that the HSRB respond to an additional charge question related to the
use of the three MI studies in a weight of evidence approach to establishing a point of departure
for risk assessment.
3

-------
REVIEW PROCESS
The Board conducted a public meeting on January, 25-26, 2017. Advance notice of the meeting
was published in the Federal Register as "Human Studies Review Board; Notification of a
Public Meetings" (EPA, December 19, 2016, 95982, Vol. 81, No.250).
Following welcoming remarks from Agency officials, the Board began its consideration of the
materials presented for review. On day one of the meeting, EPA staff began by presenting an
overview of physiologically - based pharmacokinetic (PBPK) modeling and its intended uses by
the Agency prior to the review of specific studies of cimetidine/carbaryl and malathion. On day
two of the meeting, Agency staff presented an overview of repeated open application test
(ROAT) studies prior to the review of the dermal exposure studies related to
methylisothiazolinone, as well as a description of the use of ROAT studies by the Agency for a
point of departure for risk assessment. These informational sessions provided the Board with
context for the scientific and ethical review of each study and help Board members understand
how the data would contribute to a larger process of risk assessment by the Agency.
On each day, the Board then heard presentations from EPA for each agenda item in sequence,
consisting of the Agency's review of scientific and ethical aspects of the studies, followed by
discussion from members of the HSRB who reviewed each of the studies for scientific and
ethical soundness. This Final Report of the meeting describes the HSRB's discussion,
recommendations, rationale and consensus in response to each charge question for each of these
items.
For each agenda item, Agency staff first presented their review of the science and the Board
asked the Agency presenters clarifying questions. The staff then described their review of the
ethical aspects and the Board asked clarifying questions about those. The HSRB solicited public
comments at that point. Following public comments, the Board read the Charge Questions under
consideration. The Board addressed the science charge questions first, followed by the ethics
charge questions. The Chair then called for a vote to confirm concurrence on a summary
statement in response to each charge question.
4

-------
For their evaluation and discussion, the Board considered the published papers, Agency
presentations on intended uses of the data and information from authors of the published articles,
the Agency's science and ethics reviews of the studies, other published papers and guidelines
relevant to the field, and public comments made at the meeting.
On day one, prior to the review of the first two items, EPA staff made a detailed presentation to
the board about the planned uses of data in PBPK modeling. Briefly, PBPK modeling allows
risk assessors to convert an externally applied dose from an animal toxicity study into an internal
dose at the target organ in humans. The models use biological and chemical properties that
determine the pharmacokinetic processes that take place in the human body: absorption,
distribution, metabolism and excretion. The aim of PBPK modeling is to make predictions about
human health risk based on the prediction of a tissue response from a specific dose.
Both of the pesticides discussed in the meeting, carbaryl and malathion, are AChE inhibitors, but
they have a different mechanism of action. The Agency plans to use plasma carbaryl
concentrations and AChE inhibition data from the carbaryl study to evaluate PBPK models. The
malathion study did not demonstrate AChE inhibition, so therefore AChE inhibition data from
the malathion study will not be used. The malathion study demonstrates a no observed adverse
effect level (NOAEL) which is used as a point of departure for risk assessment. The
pharmacokinetic (PK) data from the malathion study will be used, including the urinary
metabolites monocarboxylic acid (MCA) and dimethyl phosphate (DMP), and plasma malathion
and malaoxon. The plasma PK data will be used for making model predictions, and the urinary
metabolite data will be used for validation of the model. The PBPK models will be presented to
the FIFRA Scientific Advisory Panel for external peer review later in 2017.
HSRB review of a published article: Cimetidine-Carbaryl Interaction in Humans:
Evidence for an Active Metabolite of Carbaryl, authored by D. Gail May, Rebecca J.
Naukam, J. Reddy Kambam, and Robert A. Branch. Journal of Pharmacology Exposure
Therapy (1992) 262(3), 1057-1061
CHARGE TO THE BOARD AND BOARD RESPONSE
5

-------
Charge to the Board:
Is the research described in the published article "Cimetidine-Carbaryl Interaction in Humans:
Evidence for an Active Metabolite of Carbaryl" scientifically sound, providing reliable data?
Board Response:
The Board concluded that the research described in the published article "Cimetidine-Carbaryl
Interaction in Humans: Evidence for an Active Metabolite of Carbaryl" is scientifically sound
and provides data that are reliable for the purposes proposed by EPA.
HSRB Detailed Recommendations and Rationale:
This study, performed in the early 1990's, involved in vitro tests of carbaryl, alone or together
with the histamine-2 (H2) receptor antagonist, cimetidine, using erythrocyte acetylcholinesterase
(AChE) activity as a pharmacodynamics endpoint of carbaryl-related bio-activity. In vivo study
in this work, in four male subjects, followed plasma carbaryl and erythrocyte AChE activity over
time following a single, oral dose. Subsequently the impact of pretreatment with a known
cytochrome P450 inhibitor on pharmacokinetics and pharmacodynamics in those same study
participants was assessed by pretreating them with 600 mg of cimetidine (divided into three, 8-
hour doses) daily for three days, followed by a single carbaryl oral dosing. The investigators
found that when cimetidine was pre-administered, that carbaryl peak plasma concentration was
higher, although not statistically significant; oral clearance was lower and statistically significant
(p<0.05); and plasma half-life was longer, but not statistically significant. These findings
suggested that metabolism of carbaryl was inhibited by cimetidine. Unexpectedly, the in vivo
studies suggested that, in spite of the increased body exposure to carbaryl caused by cimetidine
pretreatment, the inhibition of AChE by carbaryl was reduced in the pretreated individuals.
Authors concluded with a tenable hypothesis that carbaryl is metabolized (by a cimetidine-
inhibitable enzyme) to a more pharmacodynamically active metabolite.
The HSRB commented that one of the statements made in the EP A summary of the paper
appeared to be incorrect, namely, "The results of the in vitro assays confirmed that, at a
concentration of 10 or 100 (ig/mL, cimetidine did not impact the RBC AChE inhibition caused
by increasing concentrations of carbaryl." In fact, at 1 OOug/ml both cimetidine and carbaryl
6

-------
inhibit erythrocyte AChE, so the total observed inhibition is impacted. Furthermore, EPA's
statement implies that the combined effect is additive, but the graphs at least raise the possibility
that the effect may not be additive at 1 OOug/ml cimetidine. The graphs appear not to be
completely additive, and in any case the data are not conclusive on this point. The Board also
pointed out an inconsistency in the paper, in that the paper reports 4 human subjects, but reports
n=6 for in vitro studies. This discrepancy is not explained. However, these issues do not cause
the Board to be concerned about the internal validity of the study.
The Board opined that the fundamental study design seems sound, and the results suggest that
there is a real possibility of a drug-drug interaction between carbaryl and cimetidine. The second
element of the charge is related to provision of reliable data. The Board expressed confidence in
the analytical/experimental reliability of the results. With regard to the human population-level
reliability and generalizability of the results, the small number of subjects, the unclear process
for selecting study participants, and the fact that the putative underlying processes uncovered in
this study (activity of drug metabolizing enzymes) is subject to substantial individual variability
in humans, all taken together reduce the Board's confidence in the population-level reliability of
the data. In other words, it is not clear if the study participants, as a sample of the larger
population, are a representative sample providing generalizable data. Nevertheless, the study has
acceptable internal validity, in that reliable measurements were made and reasonable conclusions
drawn regarding the metabolism of carbaryl. The Board asserts that the data are of acceptable
scientific merit for EPA scientists to use for the stated purpose, namely, for PBPK modeling.
Comments on statistical aspects of the study
Cimetidine-carbaryl interaction in humans is investigated in study with four non-smoking, drug-
free normal males. In vitro assays, cimetidine did not impact the outcome, whereas with vivo
exposure, pre-treatment of cimetidine had an effect. The results support the hypothesis that the
main site of metabolism for carbaryl is the liver by drug-metabolizing enzymes that can be
inhibited by cimetidine.
The sample size of n=4 subjects seems small and it is unclear how the sample size was
determined. However, this may be a pilot study, as the purpose appears to be to investigate the
7

-------
pathway for these four individuals in depth and there is not as much concern about the
population-level inference.
Statistical methods used for the data analysis include regression analysis, ANOVA, and repeated
measures analysis. However, details are largely missing about the software package and
procedures used. It appears that data and the code for analyzing the data are not available,
making it a challenge to evaluate the quality of the data analysis. The statistical comments do
not imply that the data are unreliable for EPA's intended uses, since generalizability to a larger
population is not the main concern.
Ethics
Charge to the Board:
Does available information support a determination that the study was conducted in substantial
compliance with subpart Q of 40 CFR part 26?
Board Response:
While the research was conducted before implementation of 40 CFR 26 subpart K and L, the
information provided supports a determination that the studies were conducted in substantial
compliance with subparts K and L of 40 CFR Part 26.
HSRB Detailed Recommendations and Rationale:
40 CFR 26 subpart Q applies to studies initiated prior to the implementation of the human
subjects regulations at 40 CFR 26 subpart K and L; subpart Q requires that EPA not rely on data
from studies if there is clear and convincing evidence that the conduct of the research was
fundamentally unethical or deficient relative to prevailing ethical standards at the time the
research was conducted. Subpart K requires that studies initiated on or after April 7, 2006
involving intentional exposure of human subjects to a pesticide be reviewed and approved by an
institutional review board (IRB) that meets the membership and review criteria listed in that
subpart. As this study was conducted prior to 2006, subpart Q is the applicable standard.
Consistent with standards applicable at the time the research was conducted, the Vanderbilt
Institutional Committee for the Protection of Human Subject reviewed and approved of this
study. While we do not know if this Committee meets the criteria stated in 40 CFR 26 subpart K,
8

-------
there was independent review at a reputable research university, the HSRB believes this 1991
study complied with the spirit of independent review.
40 CFR 26 subpart K mandates studies minimize risk to subjects, equitably select subjects, seek
and appropriately document informed consent, make adequate provisions to ensure safety of
subjects, and protect the privacy of subjects and confidentiality of data.
Minimize Risk: There is not a great deal of information provided about steps the investigators took
to minimize participant risk. The dosages used in the study were below the level to cause symptoms
and were "not expected to cause any additional short term adverse events."
Equitable Selection: The investigator did not provide information about equitability of selecting
the 4 male participants in this study. It appears to be a convenience sample of individuals
affiliated with the University of Pittsburgh Center of Clinical Pharmacology, probably including
the principal investigator. The Board did discuss the inclusion of only men in this study and
noted the time period of the research (early 1990's), prior to the widespread recognition of the
importance of including women as well as men in biomedical research in order to obtain reliable
information about the effects of treatments or exposures in females. The Board agreed that,
given the time period of the study, inclusion of only male participants was consistent with typical
research practice, particularly for non-beneficial exposure studies.
Informed Consent: As the research was done 26 years ago, neither the investigators nor the
independent review board could not provide informed consent documentation, but the
investigator told EPA that participants 1) were told of the science and risks of the study, 2) were
sufficiently educated to understand the risks, and 3) signed informed consent documentation. The
investigator does not provide information about provisions to ensure safety, protect privacy or
confidentiality. While not ideal, the board does not feel the lack of this information violates the
intention of 40 CFR 26 subpart K, and does not provide clear and convincing evidence of
unethical conduct which would violate the substantive standard of subpart Q.
9

-------
40 CFR 26 subpart Q prohibits the EPA from relying on third-party research involving
intentional exposure to a pesticide of human subjects who are children or pregnant or nursing
women. The 4 participants were male, so this does not violate 40 CFR 26 subpart Q.
HSRB review of an unpublished study: A randomized double blind study with malathion to
determine the residues of malathion dicarboxylic acid (DCA), malathion monocarboxylic
acid (MCA), dimethyl phosphate (DMP), dimethyl thiophosphate (DMTP), and dimethyl
dithiophosphate (DMDTP) in human urine.
CHARGE TO THE BOARD AND BOARD RESPONSE
Charge to the Board:
Did the research on plasma levels of malathion and malaoxon, and urinary metabolites of
malathion, as described in the study reports "A randomized double blind ascending single oral
dose study with malathion to determine the No Effect Level on plasma and RBC cholinesterase
activity" and "Determination of residues of malathion dicarboxylic acid (DCA), malathion
monocarboxylic acid (MCA), dimethyl phosphate (DMP), dimethyl thiophosphate (DMTP), and
dimethyl dithiophosphate (DMDTP) in human urine," generate scientifically sound, reliable
data?
Board Response:
The Board concludes that the research on plasma levels of malathion and malaoxon, and urinary
metabolites of malathion, as described in the study reports "A randomized double blind
ascending single oral dose study with malathion to determine the No Effect Level on plasma and
RBC cholinesterase activity" and "Determination of residues of malathion dicarboxylic acid
(DCA), malathion monocarboxylic acid (MCA), dimethyl phosphate (DMP), dimethyl
thiophosphate (DMTP), and dimethyl dithiophosphate (DMDTP) in human urine," generated
scientifically sound data that are reliable for the uses proposed by the EPA.
10

-------
HSRB Detailed Recommendations and Rationale:
The unpublished study report describes two studies. The first was a study in which 48 human
participants were administered placebo or malathion at one of 5 different doses ranging from 0.5
mg/kg to 15 mg/kg. The study was entitled "A randomized double blind ascending single oral
dose study with malathion to determine the no effect level on plasma and RBC cholinesterase
activity" and it was conducted by Inveresk Research, Elphinstone Research Centre, Tranent, EH
32 2NE, Scotland.1 In this study, the pharmacodynamic endpoints of plasma and erythrocyte
AChE inhibition were assessed. The second study was entitled "Determination of residues of
malathion dicarboxylic acid (DCA), malathion monocarboxylic acid (MCA), dimethyl phosphate
(DMP), dimethyl thiophosphate (DMTP), and dimethyl dithiophosphate (DMDTP) in human
urine," and was conducted by Pacific Toxicology Laboratories, 6160 Variel Avenue, Woodland
Hills, CA 91367.2 In this second study, malathion and its metabolites from urine and plasma
were analyzed.
As described above, the Agency plans to use PK data from the study for PBPK modeling
purposes. Urine and plasma collection in the study was designed in a manner typical of
pharmacokinetic studies, with pre-dosing and serial post-dosing samples collected. The design
of the studies, the analytical methods used, and the data produced by the studies appear to be of a
high scientific standard, and ideally suited to be used as inputs to aid the development and
refinement of PB/PK models.
Comments on statistical aspects of the study
This is a randomized double blind ascending single oral dose study with a total of 34 subjects
receiving treatments and 14 subjects in the control group. The sample size seems adequate for
drawing population-level inference, although it is unclear from the protocol how the sample size
was determined.
The statistical method used for the data analysis is a repeated measures analysis using SAS. It is
1	The study and protocol were identified as ICR 013177. March 20, 2000. MRID 45125602.
2	The Study and protocol were identified as PTL119801. October 11, 2000. MRID 45244601
11

-------
unclear which SAS procedure was used and which covariance structure was specified. This
information would be helpful to have.
In the repeated measures analysis, the main effects (dose level and time) and the interaction
terms are specified. However, it is not clear why gender is not included as a main effect at the
subject level. That is, it is unclear why were the data analyzed separately for male and female
subjects. The model diagnostics are also unclear. More details and justifications for a particular
choice (e.g., linear trend detection versus Bonferroni correction for multiple comparison) would
be helpful to have. In addition, it is not clear how much the analysis can leverage the robustness
of these methods against non-normality. Based on the description about the residuals, it seems
that there is some heavy-tail (i.e., non-normal) distribution in the error term. It may be interesting
to examine the robustness issue further.
Ethics
Charge to the Board:
Does available information support a determination that the study was conducted in substantial
compliance with subpart Q of 40 CFR part 26?
Board Response:
While the research was conducted before implementation of 40 CFR 26 subparts K and L, the
information provided supports a determination that the studies were conducted in substantial
compliance with subpart Q of 40 CFR Part 26.
HSRB Detailed Recommendations and Rationale:
As mentioned above, 40 CFR 26 subpart Q applies to studies initiated prior to the
implementation of the human subjects regulations at 40 CFR 26 subpart K and L; subpart Q
requires that EPA not rely on data from studies if there is clear and convincing evidence that the
conduct of the research was fundamentally unethical or deficient relative to prevailing ethical
standards at the time the research was conducted. Subpart K requires that studies initiated on or
after April 7, 2006 involving intentional exposure of human subjects to a pesticide be reviewed
and approved by an institutional review board (IRB) that meets the membership and review
12

-------
criteria listed in that subpart. As this study was conducted prior to 2006, subpart Q is the
applicable standard.
The Inveresk Research Independent Ethics Review Committee reviewed and approved of this
study. Based on information provided, the Committee meets the criteria stated in 40 CFR 26
subpart K and satisfies the standards of subpart Q, which is the applicable citation. The study
was conducted in accordance with the guidelines set out in Declaration of Helsinki, 1964 and its
1975, 1989, 1996 amendments. The HSRB believes this study complied with the spirit of
independent review. Amendments not reviewed by the Review Committee did not affect patient
safety.
40 CFR 26 subpart K mandates studies minimize risk to subjects, equitably select subjects, seek
and appropriately document informed consent, make adequate provisions to ensure safety of
subjects, and protect the privacy of subjects and confidentiality of data.
Minimize Risk: The study listed above minimized risk to subjects by recruiting participants who
were healthy (normal physical exam, laboratory data, and ECG), choosing doses not expected to
cause adverse effects, and not escalating doses if there were any significant inhibition of
cholinesterase. The study also contacted the participants' primary physician to determine if there
was a concern with that individual participating in this study.
Equitable Selection: There are no specific information regarding the population from which
participants were recruited, except that they were from the surrounding area and a generic
advertisement was used. The use of randomization to different dosage or placebo allowed for
equitable exposure to risk.
Informed Consent: The study physician explained the study and participants were allowed to ask
questions. Written information regarding malathion. Additionally, this document contained most
of the elements contained in a standard informed consent document (procedures, risks and
benefits, withdrawing from the study, confidentiality, and physician contact information).
13

-------
Safety: A physician provided oversight for the entire study. Participants resided in the study
clinic for 2 days and were monitored at outpatient visits 3, 5, and 13 days after dosing.
Privacy and Confidentiality: Each participant was assigned a random individual identification
number which protected privacy and confidentiality.
40 CFR 26 subpart Q prohibits the EPA from relying on third-party research involving
intentional exposure to a pesticide of human subjects who are children or pregnant or nursing
women. Materials submitted to EPA and HSRB indicate that no children or pregnant or nursing
women enrolled in this study. Pregnancy testing occurred up to 21 days prior and again the day
prior to the study. As a result, the study was conducted in substantial compliance with 40 CFR
part 26, subpart Q.
HSRB review of a published study Methylisothiazolinone contact allergy and dose-response
relationships, authored by Michael D. Lundov, Claus Zachariae, and Jeanne D. Johansen.
Contact Dermatitis (2011) 64, 330-336.
CHARGE TO THE BOARD AND BOARD RESPONSE
Charge to the Board:
Is the research described in the published article "Methylisothiazolinone contact allergy and
dose-response relationships" scientifically sound, providing reliable data?
Board response:
The board after discussion agreed that the research in the published article
"Methylisothiazolinone Contact Allergy and Dose-Response Relationships" is scientifically
sound, providing reliable data. In the use of this data for risk assessment to establish the
elicitation threshold for methylisothiazolinone, the Board wishes to highlight some relevant
exposure conditions or statistical concerns that should be considered when using this data.
14

-------
HSRB detailed recommendations and rationale:
In this study authors wanted to 1) investigate the eliciting doses of MI in a patch test and in a
ROAT for already sensitized subjects, 2) investigate whether phenoxyethanol influenced
reactivity to MI, and 3) test a model for an exposure conversion between ROAT and patch test.
The model will not be discussed here. For this study 11 test subjects (2 women and 9 men,
average age of 49.7 years) with previous positive reactions to MCI/MI or MI were included.
There were 14 controls (6 women and 8 men, average age 27.5) recruited into study. Subjects
and controls were patch-tested first with 12 decreasing doses of MI (60 down to 0.0105 |ig/cm2)
in 10% ethanol and 90% aqua, and then again with those same doses of MI in same mixture but
with 0.4% phenooxyethanol (9.26 |ig/cm2). Blanks consisted of just the 0.4% phenoxyethanol in
aqua and ethanol. Readings occurred on days 2, 3, 4 and 7 following occlusion for 2 days on the
subject's backs. Placement of doses was double blinded. For the ROAT arm of study, patients
applied 20 |il of lOOppm, 50 ppm and 5 ppm of MI mixture containing phenoxyethanol (using a
fixed micropipette) twice a day on 3 x 3 cm on the volar part of forearm resulting in
concentrations of 0.42, 0.21 and 0.021 |ig/cm2/day (this amount is double the per application
dose because the study involved two applications a day with no rinse off). Response were
determined on day 2, 3, 4, 7, 14, and 21. In this study, an 8-point response scale from 0 (no
reaction) to 8 (intensive erythema, infiltration and papules) was used for patch and ROAT.
Responses greater than 5 on the response scale triggering stopping exposures in the ROAT arm.
Results: The study reported no statistical difference in reactions for test subjects in patch test
with or without phenoxyethanol, or across threshold doses. Responses were seen at
concentrations starting at 1.47 |ig/cm2 (55%) in the patch, while there was no response to 60
|ig/cm2for controls in patch. In the ROAT study, for test subjects, 7/11 (64%) reacted to the
0.21 |ig/cm2/day, and 2/11 (18%) to the 0.021 |ig/cm2/day (this dose is twice the per application
dose due to twice daily exposure). Controls had no reaction in the ROAT study. Doses are
described as low 1.47 |ig/cm2 for MI in patch test (55%, which would mean a concentration in
cream of 49 ppm). The percentage of individuals who responded to 0.021 |ig/cm2/day in the
ROAT was 18%.
The HSRB expressed confidence that this study provides reliable and valid data. The Board
noted that sufficient information was provided by the authors that the study has the potential to
15

-------
be repeated and validated. The study objectives are stated; an acceptable control group is used;
and the investigators used measures to reduce bias in the study by conducting a double blind
study and by using an acceptable reading scale for responses. The investigators used
methodologies to help assure adherence to study protocols (e.g., instructions to subjects,
micropipette for application), and response rates related to application rates/concentrations were
reported and related back to the objectives. These study characteristics provide support for the
Board's conclusion that the data are scientifically sound.
The Board further recommended that EPA consider the following issues that may affect
elicitation thresholds in the study:
1)	Five sets of ROAT bottles were weighed to determine application conformance by
subjects, however beyond stating adherence, no exact weights were reported and
compared between control and test subjects.
2)	Accumulated doses are mentioned in this study and compared to the patch results (again
patch studies are occluded). This is done to build their model comparison between patch
and ROAT. EPA may want to consider for risk assessment purposes the relationship of
ROAT studies looking at elicitation threshold and potentially to relate back to induction
threshold or even irritant thresholds that avoid induction and therefore elicitation points
in participants. These relationships are only mentioned as a potential 'model' to explore
elicitation thresholds, where induction of participants can be avoided or where human
subjects for studies who have been induced are hard to find. Authors suggest repeated
lower doses elicited quicker responses based on their comparison curves between patch
and ROAT using equivalent ROAT doses, and an algorithmic relationship between the
two is derived.
3)	Although it is established that phenoxyethanol is not a sensitizer based on lack of
response in subjects when phenoxyethanolwas applied alone, this does not necessarily
preclude that it could act synergistically or antagonistically with MI. For the ROAT arm,
MI was not applied alone and therefore phenoxyethanol should not be entirely dismissed
as having any vehicle effects when comparing to other studies.
4)	Mi-allergic subjects were previously tested at concentrations of up to 2000 ppm.
Confirmatory patch testing was also conducted at this level. These are potentially high
16

-------
induction threshold levels and the relationship between induction thresholds and
elicitations thresholds are not well understood; thus the patch testing itself could have
contributed to elicitation thresholds, although this is not known.
5) The average days to respond in the ROAT are not specifically reported and might be of
interest. In addition, two of the 11 subjects in the ROAT did not follow the application
scheme. It is not clear how this affected the statistical analysis.
Comments on statistical aspects of the study
Standard logistic regression analysis was used to estimate the dose-response relationship in the
patch tests. The ED50 and 95% CI were estimated from the dose response curve. Comparison
between the patch test reactions with or without phenoxyethanol was performed with Wilcoxon's
ranked sums test, and correlations between the individual threshold doses were investigated by
Spearman's ranked correlation. Differences in reactions to the same doses in the patch test and
ROAT were investigated with McNemar's test.
Based on the information related to statistical analysis provided in the article, the statistical
validity of patch test results is questionable due to the following reasons: (1) because same
subjects were used for with and the without treatments in the patch tests, the observed responses
derived from treatments with and without phenoxyethanol were usually correlated, and this
correlation was ignored in the dose-response models; and (2) the use of the statistically
significant Spearman's ranked correlation to conclude that treatments with and without
phenoxyethanol produced similar rankings was inappropriate; (3) the Agency's statistical
validation confirmed the statistical results reported in the paper (ED50 and 95% CI were
estimated from the dose response curve) was non-reproducible.
In summary, although there were flaws in the statistical analysis, the data presented in this article
could be considered as scientifically sound and providing reliable data but limited to weight of
evidence based applications when combined with other two reviewed papers.
17

-------
Ethics
Charge to the Board:
Does available information support a determination that the study was conducted in substantial
compliance with subparts K and L of 40 CFR part 26?
Board response:
The portions of the EPA regulations regarding the conduct of research with human subjects, 40
CFR part 26 subpart K and L, do not apply because the research was not conducted or supported
by EPA, nor was it conducted with the intention to submit the results to EPA. EPA identified this
study through a review of the public literature; the published article or any results of this research
were not independently submitted to EPA.
The study was conducted in Denmark; therefore, the U.S. Federal Insecticide, Fungicide and
Rodenticide Act (FIFRA) did not apply to the conduct of this research, but does for EPA use of
data generated. The available information supports a determination that the study is in substantial
compliance with subpart Q of 40 CFR Part 26 (Ethical Standards for Assessing Whether to Rely
on the Results of Human Research in EPA Actions). Assuming scientific validity, the study may
be used in actions under FIFRA (7 U.S.C. 136 et seq.) and section 408 of the Federal Food, Drug
and Cosmetic Act (FFDCA).
HSRB Detailed Recommendations and Rationale:
In accordance with 40 CFR 26.1703, the research did not involve intentional exposure of any
pregnant or nursing female subjects or any children. All of the subjects in this study were adults.
Pregnancy and nursing were exclusion criteria for both test and control subjects. Female subjects
were not tested for pregnancy before participation; they were asked if they were breastfeeding or
could be pregnant. Given the minimal risk nature of the study (no subject had any severe or
widespread reaction) and that patch testing during pregnancy or lactation is not known to be
harmful, this seems adequate.
In accordance with 40 CFR 26.1704, the study was conducted under procedures at least as
protective as those in subparts K and L of 40 CFR Part 26. The protocol and the published article
state that the study was conducted according to the principles in the Declaration of Helsinki.
18

-------
Documents provided to the EPA indicate that the study was reviewed and approved by the
Capital Region of Denmark (the ethics committee with jurisdiction) on May 17, 2010. The ethics
committee reviewed the study prior to initiation, amendments and reports with standards similar
to those required by EPA (Danish Act on Research Ethics Review of Health Research Projects
and the Ministerial Order No 806 of 12 July 2004 on Information and Consent).
The informed consent materials contained adequate information for the subjects (n=25) about the
risks, discomforts and benefits of participation, and of their right not to participate or to
withdraw (n=0). The risk-benefit ratio was determined to be acceptable and risks were
minimized appropriately and were justified by the potential societal benefits associated with
gathering data to determine the dose-response relationship in persons already sensitized to or
allergic to the test article.
The selection of study participants appeared to be equitable and no children or pregnant women
were enrolled. No conflicts of interest were declared. There is no evidence that the conduct of the
research was fundamentally unethical or deficient relative to ethical standards and minimized the
risk of harm to subjects.
HSRB review of a published study, Methylisothiazolinone in rinse-off products causes
allergic contact dermatitis: a repeated open-application study, authored by K. Yazar, M. D.
Lundov, A. Faurschou, M. Matura, A. Boman, J. D. Johansen, and C. Liden. British
Journal of Dermatology (2015) 173,115-122.
CHARGE TO THE BOARD AND BOARD RESPONSE
Charge to the Board:
Is the research described in the published article "Methylisothiazolinone in rinse-off products
causes allergic contact dermatitis: a repeated open-application study" scientifically sound,
providing reliable data?
Board response:
The Board after discussion agreed that the research in the published article
"Methylisothiazolinone in Rinse-Off Products Causes Allergic Contact Dermatitis: A Repeated
Open-Application Study" is scientifically sound, providing reliable data.
19

-------
HSRB Detailed Recommendations and Rationale:
In the use of this data for risk assessment to establish the el i citation threshold for
methylisothiazolinone, the board wishes to highlight some relevant exposure conditions or
statistical concerns that should be considered when using these data.
The objective of this study was to determine if currently allowed concentrations of MI in rinse-
off products can cause allergic dermatitis. In this study consisting of 19 MI allergic subjects
(mean age 40 years) and 19 control non-MI allergic subjects (mean age 37 years) participated. In
one arm of the study 9 MI allergic subjects and all controls applied a 100 ppm MI soap five
times a day to a 5 cm x 10 cm area of the ventral side of the forearm, resulting in a concentration
of 0.48 |ig/cm2 per application and a daily concentration of 2.4 |ig/cm2. In the other arm of the
study 10 MI allergic subjects and all controls applied a 50 ppm MI soap five times a day to a 5
cm x 10 cm area to the ventral side of the forearm resulting in a concentration of 0.24 |ig/cm2 per
application and a daily concentration of 1.2 |ig/cm2. Controls used soaps with no MI.
Confirmatory patch test for MI subjects was 6 solutions of MI soap from 60 down to 0.48
|ig/cm2, and vehicle control of aqua and 16% petrolatum. Control subjects only received the 60
|ig/cm2 concentration soap. Patch tests were applied to back, occluded for 2 days, with readings
on day 4 only. Readings for the ROAT were conducted once a week unless a reaction was
experienced. Trained nurses and researchers conducted the reading for the patch, and researchers
conducted readings for the ROAT. The study used an 8-point response scale from 0, (no
reaction) to 8 (intensive erythema, infiltration and papules) for patch and ROAT. The
investigators also used the Fisher scale (>25% area of application). Threshold concentration for a
positive reaction in this study was a visible reaction of 1 on the scale on day 4 (if there was a
continuous reaction). Subjects received instruction manuals and an instruction video on the
application methods. The area of skin is moistened, soap applied and spread evenly and allowed
to dry for 25s (some absorption into the skin is expected), area is rinsed and patted dry, and then
a moisturizer (supplied by team) is applied once daily to area. Study participants were instructed
not to use any of their own products during the study.
Results: 10/10 (100%) of participants who used the lOOppm (0.48 |ig/cm2/application)
developed a reaction to MI soaps within 4-11 days (average 7.3 days); while 7/9 (78%)
developed a reaction to the 50 ppm soaps within 5-21 days (average 8 days). Controls had no
20

-------
reaction to MI soaps and neither subjects nor controls had reaction to the Mi-free soap. There
was statistical significance for the 100 ppm and 50 ppm soap compared to controls (Fisher test).
In the patch test, some MI -allergic subjects (3/19) reacted to doses as low as 0.48 |ig/cm2. All
Mi-allergic subjects reacted to the 15 |ig/cm2level.
The HSRB expressed confidence that this study provides reliable and valid data. The Board
noted that sufficient information is provided by the authors that the study has the potential to be
repeated and validated. The study objectives are stated; an acceptable control group is used; and
the investigators used measures to reduce bias in the study by conducting a double blind study
and by using an acceptable reading scale for responses. The investigators used methodologies to
help assure adherence to study protocols (e.g., instructions to subjects, soap bottles with pump
rates), and response rates related to application rates/concentrations are reported and related back
to the objectives. These study characteristics provide support for the Board's conclusion that the
data are scientifically sound.
The Board further recommended that EPA consider the following issues that may affect
elicitation thresholds in the study:
1)	There is a need to clarify locations (anatomical) of application across studies.
Anatomical areas will differ in their sensitivity due to Stratum Corneum/Viable epidermis
thickness and other structural differences resulting in varying local reactions. If cosmetics
using MI are applied to various parts of the body, the most sensitive areas should be
tested/considered in a risk assessment.
2)	It is not clear why readings occurred on day 4 as opposed to day 2 and day 4 for the patch
test.
3)	Detailed information is provided in this study on where the MI and liquid soap
preparation was obtained and chemical content of all applied test and control products.
This will allow further analysis of the potential role of vehicle effects on MI elicitation
thresholds.
4)	Because this is a rinse-off study, the application, daily dose or accumulated may not be
reflective of actual dose that remains on the skin, in the skin or absorbed into the skin
(total dose compared to the ROAT). The actual LOAEL in this study may in fact be
lower due to removal of the product following rinse activities, and pat and dry activities.
21

-------
5)	The control soap in the MI studies was preserved with methyl and ethyl paraben.
Although there was no reaction to control soap in the ROAT, it might be important to
note the presence of these preservatives.
6)	MI for rinse off products can be found in liquid soaps and shampoos. Exposure rates will
depend on customary uses of the products. This study tested liquid soap at 5 times a day
and makes a statement that this may underestimate usage rate. There is no study
reference for usage rates (volumes applied and areas of application), although a
conservative approach (testing at higher doses and for higher usage rates) is appropriate
in order to establish a safety factor. In the risk assessment, some questions arise in
comparing MI studies with various application rates and therefore applied dose, or
loading rates.
7)	Vehicles effects for those who applied the optional provided lotion are not discussed.
Comments on statistical aspects of the study
The study goal was to determine if MI concentrations of 100 ppm and 50 ppm have the potential
to elicit an allergic response in previously sensitized individuals. The treatment group consisted
of 19 subjects with dermatitis and contact allergy to MI. Assignment to treatment group was
determined by patch testing at 2000 ppm prior to start of the study. Ten subjects were assigned
to 100 ppm dosages, and 9 subjects to 50 ppm. There was no indication of how the assignments
were determined. The control group consisted of 19 subjects without contact allergy to MI as
determined by no allergic reaction to patch testing prior to the start of the study. The
investigators used Fisher's Exact Test to compare differences in the proportion of allergic
responses for the two groups. This test would have been appropriate if underlying assumptions
were met, however there was a problem related to study design, because there was no 50ppm
control group. For MI of 100 ppm there was significant difference (p = 5.0 x 10"8) based on 10
allergic and 10 controls. The authors also claim there was a significant difference for an MI of
50 ppm (p = 3.0 x 10"5), but the control subjects were not tested at 50 ppm. The apparent
assumption of the investigators is that the fact that subjects had no reaction at 100 ppm of MI for
a control subject implies they would have no reaction at 50 ppm of MI. This might be a
22

-------
reasonable guess, but for statistical purposes, the design of the study should have included a 50
ppm control group.
The investigators used McNemar's test for equality of proportions for comparison of pre-study
patch test and ROAT on MI allergic subjects. For 10 allergic subjects at 100 ppm of MI there
was a significant difference (p = 0.00195). None reacted to the patch test but all 10 reacted to the
ROAT testing. The problem with the use of this test, however, is that the p-value is for two
sided alternative (difference in either direction) but the authors' conclusion is that ROAT
reactivity is higher (one sided conclusion). For all 19 allergic subjects,3 there was a significant
difference (p = 0.000122) between the reaction to the patch test at 100 pm and the combined 100
ppm and 50 ppm reactions to the ROAT testing.
Finally, the investigators used Kendall's tau-B for an association for the allergic subjects
between a positive MI patch test threshold and the threshold for ROAT. Two subjects with no
ROAT reaction were not included. Kendall's tau-B is the difference between the number of
concordant and discordant pairs of observations, which is not quite the same as the usual
correlation coefficient, although the authors describe this as a test for correlation. The results
show that Kendall's tau-B = 0.381 and p= 0.062; the p value from SAS is 0.0569 which indicates
a computational difference, but does not change the overall finding that the test was not quite
statistically significant.
In summary, the chosen statistical analyses are appropriate but some of the conclusions have
what might be considered as minor flaws because the underlying assumptions are not met.
Specifically, the lack of a 50 ppm control group, the one-sided conclusion from a two sided test,
and the [problem with the Kendall's tau-B test] somewhat weaken the statistical analysis.
3 The Board noted a typographical error in the paper: the n = 49 for ROAT at 50 ppm should be n
= 19.
23

-------
Ethics
Charge to the Board:
Does the available information support a determination that the studies were conducted in
substantial compliance with subparts K and L of 40 CFR Part 26?
Board response:
The portions of the EPA regulations regarding the conduct of research with human subjects, 40
CFR part 26 subpart K and L, do not apply because the research was not conducted or supported
by EPA, nor was it conducted with the intention to submit the results to EPA. EPA identified this
study through a review of the public literature; the published article or any results of this research
were not independently submitted to EPA.
The study was conducted in Sweden and Denmark; therefore, the U.S. Federal Insecticide,
Fungicide and Rodenticide Act (FIFRA) did not apply to the conduct of this research, but does
for EPA use of data to set a human dermal sensitization endpoint/point of departure in its risk
assessment for methylisothiazolinone.
The available information supports a determination that the study is in substantial compliance
with subpart Q of 40 CFR Part 26 (Ethical Standards for Assessing Whether to Rely on the
Results of Human Research in EPA Actions). Assuming scientific validity, the study may be
used in actions under FIFRA (7 U.S.C. 136 et seq.) and section 408 of the Federal Food, Drug
and Cosmetic Act (FFDCA).
In accordance with 40 CFR 26.1703, the research did not involve intentional exposure of any
pregnant or nursing female subjects or any children. All of the subjects in this study were adults.
Pregnancy and nursing were exclusion criteria for both test and control subjects. Female subjects
were not tested for pregnancy before participation; they were asked if they were breastfeeding or
could be pregnant. Given the minimal risk nature of the study (no subjects had any severe or
widespread reaction) and that patch testing during pregnancy or lactation is not known to be
harmful, this seems adequate.
In accordance with 40 CFR 26.1704, the study was conducted under procedures at least as
protective as those in subparts K and L of 40 CFR Part 26. The protocol and the published article
24

-------
state that the study was conducted according to the principles in the Declaration of Helsinki.
Documents provided to the EPA indicate that the study was reviewed and approved by regional
ethics review boards in Stockholm, Sweden and Capital Region, Denmark (the ethics committees
with jurisdiction) in July 2013. The ethics committees reviewed the study prior to initiation with
standards similar to those required by EPA (Swedish Act on the Ethical Review of Research
Involving Humans; Danish Act on Research Ethics Review of Health Research Projects; and the
Ministerial Order No 806 on Information and Consent).
The informed consent materials contained adequate information for the subjects (n=38) about the
risks, discomforts and benefits of participation, and of their right not to participate or to
withdraw (n=2). The risk-benefit ratio was determined to be acceptable and risks were
minimized appropriately and were justified by the potential societal benefits associated with
gathering data to determine if products preserved with allowed concentrations of the test article
have the potential to elicit allergic contact dermatitis in previously sensitized individuals.
The selection of study participants appeared to be equitable and no children or pregnant women
were enrolled. No conflicts of interest were declared. There is no evidence that the conduct of the
research was fundamentally unethical or deficient relative to ethical standards and minimized the
risk of harm to subjects.
HSRB review of a published study, An evaluation of dose/unit area and time as key factors
influencing the elicitation capacity of methylchloroisothiazolinone/methylisothiazolinone
(MCI/MI) in MCI/MI-allergic patients, authored by Claus Zachariae, Anne Lerbaek,
Pauline M. McNamee, John E. Gray, Mike Wooder, and Torkil Menne. Contact Dermatitis
(2006) 55,160-166.
CHARGE TO THE BOARD AND BOARD RESPONSE
Charge to the Board:
Is the research described in the published article "An evaluation of dose/unit area and time as key
factors influencing the elicitation capacity of methylchloroisothiazolinone/methylisothiazolinone
(MCI/MI) in MCI/MI-allergic patients" scientifically sound, providing reliable data?
25

-------
Board Response:
The Board after discussion agreed that the research in the published article "An evaluation of
dose/unit area and time as key factors influencing the elicitation capacity of
methylchloroisothiazolinone/ methylisothiazolinone (MCI/MI) in MCI/MI allergic patients" is
scientifically sound, providing reliable data.
In the use of this data for risk assessment to establish the elicitation threshold for
methylisothiazolinone, the Board highlighted some relevant exposure conditions or statistical
concerns that should be considered when using this data.
The authors evaluated the allergic response of MCI/MI patients to the application of a mixture of
MCI/MI where authors wanted to determine the elicitation response to a mixture of MCI/MI and
the influence of time and concentration (i.e., the exposure scenario) on elicitation thresholds.
This study enrolled 29 eczema patients aged 18 and over who had a previous response to 100
ppm MCI/MI in a patch test, where response was confirmed for 25 with a diagnostic patch test to
100 ppm MCI/MI. There were healthy 10 volunteers enrolled with no response to 100 ppm
MCI/MI in a patch test. The dose of MCI/MI for diagnostic patch was 3 |ig/cm2 (determined
from applying 15|il of 100 ppm on a 0.5 cm2 area). A controlled patch using deionized water
with 10% ethanol was applied for two days with readings at days 2, 3, 4, and 7. The design of
the double blind placebo ROAT was for subjects to apply 2 drops (estimate volume of
0.05668|il) of a MCI/MI mixture, twice a day for 4 weeks to 9 cm2 area of the volar part of
forearm, and allow to dry before putting on clothing. ROAT 1 was a total dose of 0.025 |ig/cm2
per day (4 drops over the course of a day using a concentration of 2 ppm MCI/MI over a 9 cm2
area). ROAT 2 was conducted 4 weeks after the end of ROAT 1, and also for 4 weeks. ROAT 2
was a total dose of 0.094 |ig/cm2 per day (4 drops over the course of a day using a concentration
of 7.5 ppm MCI/MI over a 9 cm2 area). Response was graded on a five-point scale (negative,
doubtful, weakly positive, moderately positive, or strongly positive) for outcomes of erythema,
papules and vesicles. Test bottles were weighed before and after use, which helps to confirm
overall dose and daily dose over the 4-week period. Average daily dose across subjects (those
who had negative or positive responses and from ROAT 1 and ROAT 2) and controls was
comparable according to the authors, although no significance testing was reported on that

-------
Seven out of 25 (28%) subjects showed a positive response in ROAT 1 and the average number
of days to response was 16.5 (reported as non-significant compared to vehicle control). Fourteen
out of 25 (56%) subjects showed a positive response in ROAT 2 and average days to response
was 12.1 days (reported as significant compared to vehicle). The difference between ROAT 1
and ROAT 2 responses was statistically significant. The majority of reactions were classified as
weak or moderate; the 10 non-allergic controls did not show reactions in the ROAT arms of the
study.
The HSRB expresses confidence that this study provides reliable and valid data. The Board
noted that sufficient information is provided by the authors that the study has the potential to be
repeated and validated. The study objectives were stated; an acceptable control group was used;
and the investigators used measures to reduce bias in the study by conducting a double blind
study and by using an acceptable reading scale for responses. The investigators used
methodologies to help assure adherence to study protocols (e.g., instructions to subjects, bottles
for application), and response rates related to application rates/concentrations are reported and
related back to the objectives. These study characteristics provide support for the Board's
conclusion that the data are scientifically sound.
The HSRB recommended that EPA consider the following scientific issues that may affect
elicitation responses.
1) Application to same area, following a rinse out period might be expected to result in
more rapid sensitivity for the ROAT 2 arm. However, EPA staff contacted the study
team to determine whether the same area was used for ROAT 2; study team
confirmed that a different area was used in the second application. Given the number
of positive reactions and the days to reaction in ROAT 1 versus ROAT 2, this
outcome is in fact reported. This may indicate a growing sensitivity to exposures of
MCI/MI mixtures, for especially sensitive individuals. However, because a higher
dose of MCI/MI was used for ROAT 2, it is not possible to determine whether
repeated application, or higher dose, or both factors played a role. It is unclear what
the outcome would have been if the same dose or even lower was used in ROAT 2
arm compared to ROAT 1. In this study, 5 of the 7 who showed positive responses to
ROAT 1, showed positive responses also to ROAT 2, with reactions in first week.
27

-------
The HSRB suggests that ROAT 2 may not offer any additional information on
exposure conditions and threshold responses, given the higher dose used.
2)	The mixture of MCI/MI is not confirmed although authors mention a typical mixture
of 3:1. If this assumption is used to determine the concentration of MI, the ROAT 1
results is (0.025 |ig (MCI/MI)/cm2 /day/ 4) x 1 = 0.00625 |ig (MI)/cm2/ day and for
ROAT 2 = (0.094 |ig (MCI/MI)/cm2 /day /4) x 1 = 0.0235 |ig (MI)/cm2/day.
Although authors state that they believe that the elicitation threshold is in the
proximity of 0.025 |ig (MCI/MI)/cm2/ day (based on response in ROAT 1 to daily
applications of this concentration), they found no significance for this level compared
to control. In addition, this would indicate a dose of 0.00625 |ig (MI)/cm2/day of MI
alone (28% responded), assuming all the response is attributed to the MI in the
mixture. However, it may not be possible to disaggregate the response to the
individual components in the mixture. In Lundov et al., 2011, page 334, MCI is
mentioned as a more potent sensitizer than MI. This is not a study to determine the
sensitively of allergic patients to MI but to a mixture of MCI/MI. It is possible that
MCI can act to increase or decrease the activity of MI (i.e., synergistic or antagonistic
effects). The potential synergy or antagonism is not tested in this study but may be an
important consideration depending on whether a determination is to be made of MI in
specific mixtures, for specific application types. EPA may want to clarify if a
distinction between MCI and MI is necessary in this risk assessment determination.
3)	Authors have not confirmed what other ingredients are in the MCI/MI mixtures that
might act as sensitizers or influence elicitation responses. Ethanol is mentioned an
ingredient.
4)	This study, compared to those by Lundov et al., 2011 and Yazar et al., 2015 used
eczema patients as study subjects (although none with active eczema). This could
result in higher response rates, or higher sensitivities for subjects, because these
subjects might be more sensitive to lower concentrations of MI or MCI/MI mixtures.
However, the various skin conditions (and their population prevalence) that influence
elicitation thresholds are not well understood.
5)	Age or sex of participants is not reported to determine demographic representation.
Greater skin sensitivities might be expected for the very young and in the elderly.
28

-------
Sensitivity might also be greater on areas like the face or scrotum. In this study the
forearm is used. In an elicitation response, greater sensitivity may be related to
thickness of the stratum corneum and access to the viable epidermis. Thickness of
layers of the skin vary over the body, and also by age and sex.
6)	There is some suspicion that elicitation response (i.e., threshold) may be related to
thresholds for induction. Full history of each participant is not known, and the study
does not report whether these participants were already induced/sensitized and at
what level, or if the confirmatory patch at 100 ppm is their threshold for induction.
This is a lower level confirmatory patch than used in the other studies compared here.
7)	Response was based on a grading scale or weak, moderate or strong for outcomes of
erythema, papules and vesicles. It is not clear is this is a slightly different reading
scale than that used in the two other studies under review, in which an 8-point reading
scale is mentioned.
Comments on statistical aspects of the study
The study goal was to determine the effect of concentration and time on elicitation capacity of
MCI/MI in sensitized subjects using ROAT testing. The treatment group consisted of 29 eczema
subjects; the number of males and females is not specified. Assignment to treatment group
required at least one prior positive reaction to patch test with MCI/MI concentration of 100 ppm.
The control group consisted of 10 healthy subjects with no allergic reaction to pre-study patch
test at 100 ppm. In the study, a pre-study patch test followed by two 4 week stages of ROAT
testing separated by a 4-week washout period. Stage 1 was ROAT with an MCI/MI
concentration of 2 ppm (0.025 (j,g/cm2); stage 2 was ROAT with an MCI/MI concentration of 7.5
ppm (0.094 (j,g/cm2). The investigators used Fisher's Exact Test which was appropriate for
comparing the proportion of subjects in the allergic group having a positive reaction to the
control group proportion. A logistic regression was used for the allergic group subjects to
compare the proportions of positive reactions in the stage 1 ROAT and stage 2 ROAT. A subject
effect was included in the analysis. The problem with the logistic regression analysis is that
apparently subjects were not considered as random effects. Also, with a random subject effect,
significance of the results using SAS GLIMMIX depended on the method used to fit the logistic
model (p = 0.0407 with Residual pseudo-likelihood; p = 0.4936 with Laplace). The investigators
29

-------
also used Kaplan-Meier survival (no positive reaction) estimates over time to a positive reaction
were calculated from the stage 1 and stage 2 data. The problem with this analysis is that the
stage 1 and stage 2 are not independent because they come from the same set of subjects. In
summary, the chosen statistical analyses might be appropriate but there are issues with whether
or not the authors' underlying assumptions were reasonable.
Ethics
Charge to the Board:
Does available information support a determination that the study was conducted in substantial
compliance with subpart Q of 40 CFR part 26?
Board Response:
The HSRB concludes that the available information supports a determination that the Zachariae
et al. (2006) study was conducted in substantial compliance with subpart Q of 40 CFR part 26.
HSRB Detailed Recommendations and Rationale
The Agency's rules at 40 CFR part 26 subpart Q that are applicable to this review include the
following:
§26.1703: Except as provided in §26.1706, EPA must not rely on data from any research subject
to this subpart involving intentional exposure of any human subject who is a pregnant woman
(and therefore her fetus), a nursing woman, or a child.
And,
§26.1704: EPA must not rely on data from any research subject to this section if there is clear
and convincing evidence that: (1) The conduct of the research was fundamentally unethical (e.g.,
the research was intended to seriously harm participants or failed to obtain informed consent); or
(2) The conduct of the research was deficient relative to the ethical standards prevailing at the
time the research was conducted in a way that placed participants at increased risk of harm
(based on knowledge available at the time the study was conducted) or impaired their informed
consent.
30

-------
§26.1703:
The manuscript indicates (p. 161-162) that subjects were at least 18 years of age and that
pregnant or lactating women were excluded from study participation. EPA follow-up with the
author noted that there was no pregnancy testing to confirm non-pregnant status.
§26.1704:
Ethics Committee Review: The manuscript (p. 162) indicates that a local ethics committee had
reviewed and approved the study and that it was conducted according to Good Clinical Practice
guidelines. The author states that he no longer has copies of the ethics committee approval since
the study is more than a decade old. GCP standards (adopted June 1996) has clear standards for
ethics committee membership and review processes which are consistent with Agency rules at 40
CFR part 26, Subpart K. GCP also requires compliance with the ethical standards of the
Declaration of Helsinki, including risk minimization, informed consent of human subjects, and
protection of vulnerable populations.
Informed Consent: The manuscript also indicates that informed consent from all subjects was
obtained. Denmark's "Ministerial Order No 806 of July 2004 on Information and Consent at
Inclusion of Trial Subjects in Biomedical Research Projects" provides robust guidelines for
informed consent standards in this trial. There is no evidence to suggest that this study was in
violation of these standards. As with ethics committee approval records, the author did not have
a copy of the consent form available when requested by EPA.
Risk Minimization: Principal risk associated with this research involves allergic reaction to the
MCI/MI material. This risk was minimized by providing subjects with instructions to visit the
institution's dermatology center in the event of an allergic reaction outside of the study's
scheduled monitoring visits. The author does not recall any adverse events that required medical
treatment for subjects.
31

-------
HSRB Review Of The ROAT Studies Considered Together
Charge to the Board:
When considered all together, do the three studies described in Lundov et al., Yazar et al., and
Zachariae et al., provide a scientific weight of evidence in support of establishing a point of
departure for the determination of an elicitation threshold for methylisothiazolinone (as
identified by the Lundov et al., study) for use in risk assessments?4
Board Response:
The HSRB board agrees that when considered all together, the three studies described in Lundov
et al., Yazar et al., and Zachariae et al., do provide a scientific weight of evidence in support of
establishing a point of departure for the determination of an elicitation threshold for
methylisothiazolinone (as potentially identified by the Lundov et al., study) for use in risk
assessments.
4 For all three articles EPA has provided in English the following materials: Zachariae et al.,
2006: Data Evaluation Record by two EPA reviewers, Ethics review by an EPA ethics Review
officer, Ethics Questions completed by Dr. Zachariae, and the Ministerial Order 806 for
Biomedical Research from the Danish Ministry of Interior and Health; Lundov et al., 2011: Data
Evaluation Record by two EPA reviewers, Ethics Review by an EPA ethics review officer,
Ethics Questions completed by Dr. Johansen, Ethical Application to the Danish National
Bioethics Committee, the Ministerial Order 806 for Biomedical Research from the Danish
Ministry of Interior and Health, and Danish Act on Research Ethics Review; Yazar et al., 2011:
Data Evaluation Record by two EPA reviewers, Ethics Review by an EPA ethics review officer,
Ethics Questions completed by Dr. Yazar, Ethical Application to the Danish National Bioethics
Committee, Danish Approved Amendments, Swedish Ethical Application and Approval,
Swedish Approved Amendments to the Ministerial Order 806 for Biomedical Research from the
Danish Ministry of Interior and Health, and Danish Act on Research Ethics Review, Swedish
Ethical Review Act
32

-------
HSRB Detailed Recommendations and Rationale:
For all three studies EPA provides a summary of the study and findings and conclude that these
studies do provide a weight of evidence to support the derivation of an elicitation point of
departure for MI. For the Lundov study, the Lowest Observed Adverse Effect Level (LOAEL) is
mentioned as 0.0105 |ig/cm2 for MI, where for Zachariae et al., 0.025 |ig/cm2 for MI (mixture of
MCI/MI) is mentioned, and for Yazar et al., 0.24 |ig/cm2 for MI (rinse-off product containing
MI) is mentioned. EPA has evaluated all studies with respect to "Guidance for Considering and
Using Open Literature Toxicity Studies to Support Human Health Risk Assessment (USEPA,
2012)" and based on 16 screening criteria determined that all three studies are appropriate for
quantitative use in establishing a point of departure for risk assessment. There was a finding for
Zachariae et al., 2006 that for criteria #13: Adequate data provided on the chemical tested.. .that
not enough information was provided. However, this lack of information does not prevent the
study from being found to be scientifically sound and providing reliable data. There is a
comparison of the Zachariae et al., and the Yazar et al., studies to the Lundov et al., et al., where
Lundov et al., have established the (lowest) LOAEL of 0.0105 |ig/cm2. These are all considered
acceptable/non-guideline, non-registrant submitted studies.
EPA recognizes the small number of participants in each study, problems with demographic
representation or reporting, lack of knowledge on the history of MI exposure for participants, and
in some cases issues with adequate data for repeatable statistical analysis. In addition, EPA finds
that the studies do not provide enough information to determine or construct a dose-response
curve (Lundov et al with 3 applied concentration points is lacking in that regard). EPA also
recognizes the potential influence of vehicles on the elicitation threshold for MI, and for some
studies not enough information has been provided regarding the vehicles or their potential for
influencing response. However, the intent is to establish weight of evidence in deriving an
LOAEL for elicitation thresholds across the three studies. Because immune responses in
individuals can vary so greatly, understanding or even repeating elicitation responses can be
challenging. Elicitation thresholds are a relatively new area research for preservatives, and the
influence of exposure conditions is also not well understood.
The HSRB recommended that the EPA take into account the following scientific issues when
considering risks assessment based on these data. There is variability across the three studies in
33

-------
types of applications and methods of applications. The objectives of the studies also vary. It
recognized that sensitization to MI products might occur at lower product concentrations than
currently allowed, however the limit or elicitation point is difficult to narrow down from these
diverse studies, and for a particular product. Typical usage rates and application rates for the
general public or even for the high-risk individual (high usage rate) is not discussed, making it
more difficult to associate practical and likely dosage rate with a general and reasonable
elicitation threshold. Care must be taken to compare appropriate doses/loadings across studies,
where product type and application method must also be considered. In a risk assessment,
appropriate/comparable doses might best be represented by daily doses. For the Lundov study,
0.021 |ig/cm2/day can be compared with 0.025 |ig/cm2/ day MCI/MI in the Zachariae study
(where MI is only Vi of this), and for comparison with the 1.2 |ig/cm2/day (note rinse-off may
greatly lower this level) in the Yazar study. The statistical significance reported across these
doses, however, was not all strong, and the association of response to concentration in a product
may be variable. Another option is to compare accumulated applied dose to an elicitation
threshold across MI ROAT studies.
In establishing LOAEL and applying reasonable safety factors it is important to understand how
consumers use these products in terms of number of applications, volume per application, and
area of application. Determining an elicitation threshold and applied safety factors are dependent
on understanding consumer behavior and the level of guidance or use recommendations that will
or can be provided. For manufacturers, a |ig/cm2/day must be translated to ppm in the product
(or particular products) with consideration for vehicle effects and usage patterns.
There may be a rationale to further study the concentration levels of MI or MCI/MI to be
allowed in a variety of consumer products, and to examine the extent to which individuals have
become sensitized (induced) to MI. Worst-case tests should be performed taking into account
sensitive areas, sensitive individuals, chemicals in the products that either promote absorption in
the skin or retention in the skin.
Another potential risk assessment approach is to look at minimum levels of MI (or other
products) needed for preservation. If it can be established that preservation of a product can be
achieved at lower concentrations and with other established safer products, then this option needs
to be explored. For example, Lundov et al mention previous studies showing that 5 ppm MI with
34

-------
0.4% phenoxyethanol was adequate to preserve standard cosmetic creams (where
phenoxyethanol rarely causes allergic response in subjects). Ultimately lower and safer levels of
MI applied to protect MI allergic patients will need to satisfy those acceptable preservation
levels.
In addition, in the risk assessment process, it would be of interest to relate elicitation thresholds
back to effective concentration inducing a stimulation index of 3 (EC3 values). This is a
measure used in the LLNA (local lymph node assay) to test whether a chemical is a sensitizer
and potentially how potent.5 MI has an EC3 value of 0.4%6 (Yazar et al., 2015 pg. 116) and
considered a strong sensitizer.7
In conclusion, there may not be full confidence that these three studies overall definitely
establish the point of departure for an elicitation threshold in MI allergic patients, but that there
may be a point of departure for an elicitation threshold in MI allergic patients. The studies do
provide reliable data for the specific exposure conditions investigated and will assist in the risk
assessment process for establishing appropriate and practical elicitation thresholds. The Board
suggests that EPA carefully consider relationships among the studies and their findings, where
vehicle effects, inductions thresholds, application methods, and average days to elicitation
response all play a role.
In addition, the HSRB recommended that EPA consider the following scientific issues:
1) Different population subgroups may have different sensitivities. Mi-sensitized
individuals are the main focus for these ROAT elicitation studies. However, there may be
other individuals affected by low levels of MI in products, such as those with eczema and
sensitive skin problems. Data on eczema and contact dermatitis would be useful in a risk
assessment and related to their influence on both induction and elicitation thresholds.
5 World Health Organization, 2008 "Skin Sensitization in Chemical Risk Assessment" Harmonization
Project Document 5, found at http://www.inchem.org/documents/harmproj/harmproj/harmproj5.pdf
7 Leiva-Salinas and Silvestre (2013) "Update on Allergic Contact Dermatitis due to
Methylchloroisothiazolinone/Methylisothiazolinone and Methylisothiazolinone" Actas Dermosifiliofr
105(9): 840-846.
35

-------
Cumulative effects from multiple products containing MI or mixtures of MCI/MI will
also need to be considered.
2)	In a risk assessment, longer follow up of MI allergic subjects may be needed to determine
a safe level. Is it possible that a very low dose (below those tested in any of these
studies) could produce a response in 3 or 4 months of continued use (beyond the times
used in these ROAT studies)?
3)	The HSRB also recommended considering warning labels on MI products and whether
they adequately advise consumers about the risk of skin sensitization.
4)	Finally, the HSRB recommended careful conversion of dosing calculations to determine
how product concentration and usage patterns determine actual dosing applied in
practice. In the Zachariae et al article, ppm must be converted to % (converted to jug/g)
and combined with the volume applied and area of skin to determine a |ig/cm2, of MI. It
is also necessary to know the g/|il of the product (entire product) to make an assessment
of |ig/cm2 of MI, where volume application of products is used. In the Lundov study, the
weight of the product is not reported, but there was an application of 4.2 mg
cream/cm2/day reported so this also corresponds to a total of 4.2 mg/cm2 x 9 cm2 = 37.8
mg/day, so 40|il of the product (two applications over the entire 9 cm2 area), telling us
the weight of the product 37.8 mg/40|il or 0.945 mg/|il (i.e., weight of product per
volume). Also, 100 ppm MI is equivalent to 100 |ig Ml/g in cream, therefore one
application of this concentration would mean (100 |ig Ml/g cream x 0.945 mg/|il x 20 |il
x lg/100mg)/9cm2 = 0.21 |ig Ml/cm2. There are other perhaps more direct ways to make
this calculation, but this agrees with the 0.21 |ig Ml/cm2 reported for the highest dose of
MI applied in the ROAT (Table 1), of the 100 ppm MI product. The Board
recommended that EPA confirm how concentrations are determined and eliminate any
confusion about concentrations applied in these studies.
36

-------
References
Paradis et al., 2014 "Developing a Framework for the Quantitative Risk Characterization of
Dermal Sensitizers in the Workplace" 3M Company, St, Minnesota, found at
https://schc.memberclicks.net/assets/meetings/spring2014/schc_poster_developing_framework.p
df
EPA, 2004, FIFRA Scientific Advisory Panel Background Document: Proposed Hazard
Identification Methodology for Assessment of Dermal Sensitization Risk, Washington, D.C.
Anderson and Meade, 2014 "Potential Health Effects Associated with Dermal Exposure to
Occupational Chemicals" Environmental Health Insights, 8(S1).
World Health Organization, 2008 "Skin Sensitization in Chemical Risk Assessment"
Harmonization Project Document 5, found at
http ://www. inchem. org/ documents/harmproj /harmproj /harmproj 5. pdf
Ezendam, J., Braakhuis, H., and Vandebriel (2016) "State of the Art in Non-Animal Approaches
for Skin Sensitization Testing: From Individuals Test Methods Towards Testing Strategies",
Arch Toxicol., 90:2861-2883.
Leiva-Salinas et al., (2014) "Update on Allergic Contact Dermatitis due to
Methylchloroisothiazolinone/Methylisothiazolinone and Methylisothiazolinone" Actas
Dermosifiliogr, 105(9): 840-846
Zang et al., (2017) "Prediction of Skin Sensitization Potency Using Machine Learning
Approaches", J. Appl. Toxicol. Doi 10.1002/jat3424
Griem et al., 2003 "Proposal for a Risk Assessment Methodology for Skin Sensitization Potency
Data", Regulatory Toxicology and Pharmacology, 38: 269-290
Alves et al., 2015 "Predicting Chemically-Induced Skin Reactions. Part II: QSAR Models of
Skin Permeability and the Relationship between Skin Permeability and Skin Sensitization",
Toxicol Appl Pharmacol 284(2):273-280.
Steiling (2016) "Safety Evaluation of Cosmetic Ingredients Regarding their Skin Sensitization
Potential", Cosmetics, 3: doi 10.3390/cosmetics3020014.
Kimber et al., (2002) "Allergic Contact Dermatitis", International Immunopharmacology, 2: 201-
211.
Jaycock (1998) "Risk Assessment of Contact Allergens", American Journal of Contact
Dermatitis, 9(3): 155-161.
37

-------
Dearman and Kimber (2003) "Factors Influencing the Induction Phase of Skin Sensitization",
American Journal of Contact Dermatitis, 14:188-194.
Gerberick and Robinson (2000) "A Skin Sensitization Risk Assessment Approach for Evaluation
of New Ingredients and Products", American Journal of Contact Dermatitis, 11:65-73.
38

-------
US ENVIRONMENTAL PROTECTION AGENCY
HUMAN STUDIES REVIEW BOARD
Chair
Liza Dawson, Ph.D., Research Ethics Team Leader, Division of AIDS National Institutes of
Health (NIH), National Institute of Allergy and Infectious Disease (NIAID), Bethesda, MD
Vice Chair
Edward Gbur, Jr., Ph.D., Professor, Agricultural Statistics Laboratory, University of Arkansas,
Fayetteville, AR
Members
Jennifer M. Cavallari, Sc.D., C.I.H., Assistant Professor, Division of Occupational and
Environmental Medicine, Department of Community Medicine, University of Connecticut
Health Center, Farmington, CT
Gary L. Chadwick, Pharm.D., M.P.H., C.I.P., Senior Consultant, HRP Consulting Group, Inc.
Fairport, NY and Professor Emeritus, Bioethics and Humanities, School of Medicine and
Dentistry, University of Rochester, Rochester, NY
Alesia C. Ferguson, MS., MPH., PhD., Associate Professor, Department of Environmental and
Occupational Health, College of Public Health, University of Arkansas Medical Sciences,
Little Rock, AR
George Fernandez, Ph.D., Statistical Training Specialist, SAS Institute, Statistical Training and
Technical Services, Sparks, NV
Kyle L. Galbraith, Ph.D., Research Integrity Officer, University of Illinois at Urbana-Champaign
Office of the Vice Chancellor for Research, Champaign, IL
Jewell H. Halanych, M.D., Assistant Professor, Internal Medicine Residency Program,
Montgomery Regional Campus, University of Alabama at Birmingham, Montgomery, AL
39

-------
Walter Thomas Klimecki, D.V.M., Ph.D., Associate Professor and Department Head (Interim),
Department of Pharmacology and Toxicology, College of Pharmacy, University of Arizona,
Tucson, AZ
Randy Maddalena, Ph.D., Physical Research Scientist, Indoor Environment, Lawrence Berkeley
National Laboratory, Berkeley, CA
Jun Zhu, Ph.D., Professor of Statistics and of Entomology, Department of Statistics, University
of Wisconsin - Madison, Madison, WI
Consultant to the Board:
Kendra L. Lawrence, Ph.D., BCE, PMP, Health Sciences Product Manager, US Army Medical
Materiel Development Activity, Ft. Detrick, MD
Human Studies Review Board Staff
Jim Downing, Executive Director, Human Studies Review Board Staff, Office of the Science
Advisor, United States Environmental Protection Agency, Washington, DC
40

-------
NOTICE
This report has been written as part of the activities of the EPA Human Studies Review Board, a
Federal advisory committee providing advice, information and recommendations on issues
related to scientific and ethical aspects of human subjects research. This report has not been
reviewed for approval by the Agency and, hence, the contents of this report do not necessarily
represent the view and policies of the Environmental Protection Agency, nor of other agencies in
the Executive Branch of the Federal government, nor does the mention of trade names or
commercial products constitute a recommendation for use.
In preparing this document, the Board carefully considered all information provided and
presented by the Agency presenters, as well as information presented by public commenters.
This document addresses the information provided and presented within the structure of the
charge by the Agency.
41

-------