Responses to ICF Consulting Questions for 

External Expert Review Panel on n-Propyl Bromide

Expert Review Panel Members

Dr. Ulrike Luderer, M.D., Ph.D., M.P.H.

Dr. George Daston, Ph.D. 

Dr. Jodi Flaws, Ph.D. 

Question 1.  Do you consider a statistically significant decrease in the
number of estrous cycles within a given time period and an increase in
estrous cycle length (due to sustained diestrous) to be a
toxicologically significant finding?  Do you agree with the rationale
presented below?

Rationale:  An increase in estrous cycle length, leading to a reduction
in the number of estrous cycles in a given time period, can be
considered evidence of reproductive toxicity.  However, interpretation
of these data is open to question, in part because of normal
intra-individual variation in cycle length. Alterations in the
distribution of estrous cycle length alone have not been shown to be
reliable predictors of reproductive toxicity.   However, ICF considers
the weight-of-evidence for nPB-induced reproductive toxicity in females
to be sufficient based on the estrous cycle length effect as an early
biomarker or indicator of this reproductive toxicity, for the following
reasons:

A constellation of other reproductive effects occurred at higher doses,
with increasing range and severity of effects observed with increasing
dose, resulting in total infertility at the highest dose tested.  These
statistically significant effects included:

A decrease in the absolute and relative weights of the ovaries at the
highest dose, although it is not clear if evaluation of this endpoint
properly controlled for cyclic variation and infertility;

A change in the distribution of follicle subtype at the highest dose
relative to controls (an increase in the number of primordial follicles
and a decrease in the number of corpora lutea); however, this endpoint
was not measured in the other three dose groups, and so a dose-response
assessment was not possible;

A significant decrease in fertility in the two highest dose F0 groups:
no animals in the 750 ppm dose group produced litters and the fertility
index was significantly reduced in the 500 ppm group;

An increase in the number of females that displayed evidence of mating
without delivery in the two highest dose groups; 

An increase in the number of females for whom no mating was observed in
the two highest dose groups;

An increase in the  mean number of days between pairing and coitus in
the two highest dose groups; and

A decrease in the number of implantation sites in the second highest
dose group of 500 ppm (no litters were conceived in the highest dose
group of 750 ppm). 

The results of a study by Yamada et. al. support the selection of
increased estrous cycle length due to extended diestrus as a critical
endpoint for nPB-induced reproductive study effects.  In Yamada et. al.,
female Wistar rats were exposed to nPB at concentrations of 0, 200, 400,
and 800 ppm daily for 12 weeks and selected reproductive parameters were
examined.  Animals in the 800 ppm group became ill as a result of
treatment and were euthanized at week 8; their results were excluded
from statistical analysis by the study authors.  In other dose groups,
the following effects were observed:

No changes in absolute or relative ovary weights.

A statistically significant decrease in the number of estrous cycles at
400 ppm, due to extended diestrus and similar to that observed in the
Stump (2001) study.

A statistically significant decrease in the number of antral follicles
at 200 and 400 ppm, and in the number of growing follicles at 400 ppm.

A change in the number of primordial follicles that was not
statistically significant and did not show a dose-response as compared
with controls. Specifically, the number of these follicles fluctuated
inconsistently with dose, decreasing compared to controls at 200 ppm,
but increasing compared to controls at 400 ppm.

The toxicological significance of the changes in antral and growing
follicles is uncertain.  Changes in both follicle subtypes occurred at
400 ppm, the dose at which decreased estrous cycle length was observed. 
At 200 ppm, the only significant effect was a decrease in antral
follicles.  Although changes in the distribution of ovarian follicle
subtype may be indicative of reproductive toxicity, this is an endpoint
that has rarely been examined in rodent studies, and thus is not yet
well characterized.  Under the conditions of this experiment, the
functional significance of decreased antral follicle numbers, in the
absence of estrous cycle alterations, is not known.  Therefore, ICF
considers the critical effect in this study also to be increased estrous
cycle length, similar to the Stump study. 

The mechanism of action of alteration in estrous cycle length is not
known for nPB, but is likely to be hormonally-related, either at the
receptor level or at the level of the neuroendocrine-ovary feedback
loop.  Although the nature and extent of neuroendocrine control of
estrous/menstrual cycling varies significantly between rodents and
nonhuman/human primates, a decrease in regular cycling and/or increase
in cycle length is generally associated with infertility in both orders.
Specifically with nPB, a dose-related decrease in fertility was observed
at higher dose groups.  Although it is not possible to separate out the
contribution of each parent to overall reduced fertility (reproductive
parameters in both sexes were adversely affected by nPB treatment), it
is likely that changes in estrous cycle length affected reproduction.  



Response by Dr. Luderer

“A statistically significant decrease in the number of estrous cycles
within a given period of time and an increase in estrous cycle length
are toxicologically significant findings in this instance for several
reasons. 

First, there is a clear dose-response relationship of decreasing estrous
cycles with increasing dose. 

Second, the alterations in estrous cycle length fall outside the
historic range of estrous cycle duration for the laboratory that
conducted one of the studies (Stump, 2001). 

Third, the alterations in estrous cycle length are not isolated
findings, but occur in the context of other effects on the female
reproductive system. 

Fourth, the effects on estrous cycling were observed in two independent
studies of n-propyl bromide (nPB) exposure (Stump, 2001; Yamada et al,
2003).

Yamada et al (2003) exposed adult, female Wistar rats to 0, 200, 400, or
800 ppm 8h/day for 12 weeks. Animals in the 800 ppm group became ill and
were sacrificed early during the 8th week. This group displayed
lengthening of estrous cycles and constant diestrus prior to sacrifice.
Prolonged or absent estrous cycles were also noted in the 400 ppm group
(statistically significant effect at 400 and 800 ppm). The toxicological
significance of these estrous cycle-related findings is strengthened by
additional findings: dose-dependent decrease in number of antral
follicles (statistically significant at 200 and 400 ppm), dose-dependent
decrease in growing follicles (statistically significant at 400 ppm),
and increase in primordial follicles and decrease in absolute and
relative ovarian weights in the 800 ppm group. The results from the 800
ppm group were not included in the statistical analyses because there
were no concurrently sacrificed controls. It was also noted that there
were no fresh corpora lutea in the 800 ppm ovaries. These findings
relate to each other in that the reduced ovarian weight is likely due to
the decreased corpora lutea, which may indicate a decreased ovulation
rate, which is consistent with the prolonged estrous cycles. The absence
of an effect of nPB on luteinizing hormone (LH) or follicle stimulating
hormone (FSH) levels suggests that the primary effect was not on the
pituitary or hypothalamus; however, these hormones are secreted
episodically. Thus, a single determination might miss more subtle
changes in pulsatile LH or FSH secretion. In addition, a decrease in
ovulation might be due to a lack of normal preovulatory gonadotropin
surges. However, the reproductive hormone measurements, which were done
on diestrus I, were not timed to be able to assess the preovulatory
gonadotropin surges, which occur on the afternoon of proestrus.

Stump (2001) exposed F0 female Sprague-Dawley rats to 0, 100, 250, 500,
or 750 ppm nPB for 6h/day for at least 70 days prior to mating, then
continued through necropsy at PND 21, except for PND 0-4. F1 females
were exposed to 0, 100, 250, or 500 ppm for 6h/day for at least 70 days
prior to mating, also continued until PND 21 and excluding PND 0-4. Dams
were removed from litters during exposures. Dose-dependent increases in
estrous cycle length and decreased numbers of estrous cycles were
observed in F0 and F1 females.  The estrous cycle data were not analyzed
statistically by the (study) authors, but were considered
treatment-related in the 500 and 750 ppm F0 groups and in the 250 and
500 ppm F1 groups. Statistical analyses of the estrous cycle data by ICF
showed that the mean number of estrous cycles was significantly
decreased in the 250, 500, and 750 ppm F0 groups and in the 500 ppm F1
group and that estrous cycle length was significantly longer in the 500
and 750 ppm F0 groups. The toxicological significance of these findings
is underscored by other findings at the same doses in the F0 and F1
generations: decreased fertility (significant in 500 and 750 ppm groups,
though both males and females were treated, so decreased fertility could
be partially due to the observed male reproductive effects), decreased
litter size (significant in both 500 ppm groups), decreased implantation
sites (significant in both 500 ppm groups), decreased ovarian weights
(significant in 750 ppm F0 group), and decreased numbers of corpora
lutea (significant in F0 750 ppm group). Number of primordial follicles
was significantly increased in the F1 500 ppm group compared to the
control group and non-significantly increased in the F0 750 ppm group
compared to the control group. This endpoint was not assessed in the
other treatment groups.

I agree with ICF’s conclusion that the critical effect in the Stump
(2001) study is the estrous cycle effect observed at 250 ppm. However, I
disagree that the estrous cycle effect observed at 400 ppm is the
critical effect in the Yamada et. al. (2003) study. The decrease in
antral follicles in the 200 ppm group should be considered a significant
effect because a decrease in antral follicle numbers may result in
decreased numbers of ovulated oocytes and reduced fertility. The
question of whether the decreased antral follicle number was associated
with decreased numbers of ovulated follicles could be answered by going
back to the serially sectioned ovaries and counting the number of
corpora lutea, as was done in the Stump study for the highest dose
groups. It might be worth asking the authors of the Yamada et al study
if they would be willing to do this or would permit others to do it.
Even in the absence of any additional data, if the data on antral
follicle numbers in the Yamada et al study are sufficient to perform a
benchmark dose analysis, this should be done.”

To summarize, the decrease in estrous cycle number and increase in
estrous cycle length observed in the Yamada et al (2003) and Stump
(2001) studies are toxicologically significant for the reasons outlined
above.

ICF Response to Dr. Luderer

Dr. Luderer suggested that the decrease in estrous cycle number and
increase in estrous cycle length observed in the Yamada et. al. (2003)
and Stump (2001) studies are toxicologically significant.  However, she
suggested that the decrease in antral follicles at 200 ppm, in the
absence of changes in estrous cycle length, was the critical effect.  To
explore this endpoint further, given that there are not additional data
on corpora lutea numbers, a teleconference between ICF scientists and
Dr. Luderer was held on October 27, 2004.  Dr. Bonnie Ransom Stern, the
senior ICF toxicologist on this project, summarized the discussion.   

Dr. Stern noted that there was agreement between ICF and Dr. Luderer on
the statistical and toxicological significance of changes in estrous
cycling, as measured by increased estrous cycle length and decreased
number of estrous cycles within a specified time period. Dr. Luderer’s
supporting rationale is strong, and consistent with ICF’s
interpretation.

Dr. Luderer considered the decreased percentage of antral follicles in
the 200 ppm group to be a possible indicator of reduced ovulated
oocytes, associated with reduced fertility.  However, she agreed that
changes in distribution of follicular subtype is an endpoint in the
female reproductive physiology that has rarely been measured in
toxicological studies, and that its functional significance in terms of
fertility effects is not well understood in the absence of additional
data. Antral follicles represent the follicle subtype/stage that
precedes oocyte ovulation (earlier subtypes in the progression of
follicular development are primordial and growing follicles,
respectively).  However, as with statistically significant measures of
changes in sperm count and/or sperm motility, the magnitude of the
percent decrease of antral follicles would be the determining
characteristic in assessment of toxicological significance.  That is,
there is likely to be a threshold effect – decreased production of
antral follicles below a certain threshold level (in terms of magnitude
of effect) may have no impact on fertility, whereas a clear effect on
fertility may be observed above a threshold.

 

Dr. Luderer agreed with this interpretation and noted data on corpora
lutea numbers in conjunction with those on antral follices would provide
a better indication of toxicological effect.  Statistically significant
decreases in both antral follicles and corpora lutea would provide
strong evidence of a significant toxicological effect, and in her
judgment, would warrant the use of this end point as the critical effect
for development of an AEL for nPB.  In the absence of data on corpora
lutea, it is not possible to interpret the antral follicle findings;
therefore, the use of estrous cycle changes as the critical effect is
scientifically justifiable and defensible.

Dr. Luderer suggested that one possible solution would be to contact
Drs. Yamada and Ichihara to see if serially sectioned ovaries
(presumably stored in their laboratory) from this study could be
obtained from all dose groups and the number of corpora lutea counted. 
The same procedure could be done with the Stump study (2001) if ovarian
tissue samples were still available; in this study, both antral
follicles and corpora lutea were significantly decreased at the high
dose relative to controls; however, the authors did not analyze these
end points in the low and mid-dose groups.  It was agreed, however, that
these activities would be both cost-prohibitive and time-consuming. 
Therefore, it is not practical to obtain additional data from
histopathologic examination.  Based on these limitations, ICF and Dr.
Luderer agreed that the selection of increased estrous cycle length as
the critical effect for nPB reproductive toxicity is both scientifically
justified and scientifically defensible.

Response by Dr. Daston

“It is my opinion that a significant increase in estrous cycle length
(or decreased number of cycles per unit time) should be considered to be
an adverse effect when accompanied by other endocrine or reproductive
endpoints more directly tied to fertility.  

I agree with the rationale provided by ICF for categorizing increased
cycle length as a relevant finding.  The weight of evidence presented in
the rationale supports the conclusion that the increased cycle duration
is real and that it is one of a number of effects on female reproductive
function in rodents.  

That said, I believe the presentation of the weight of evidence argument
needs to acknowledge that at least some of the points listed are estrous
cycle stage-dependent; i.e., they are not completely independent
variables and as such may not be as weighty as if they were.  This is
appropriately acknowledged for ovarian weight (point “a” in the
first full paragraph of p.3), but should also be acknowledged for
follicle subtype (point “b”) and mean number of days-to-coitus
(point “f”).  Even with these changes, the weight of evidence is
still strong enough to support the conclusion that 1-bromopropane has
adverse effects on reproduction.  Furthermore, linking the mean number
of days to mate with the estrous cycle duration provides support that
the estrous cycle effect is real and biologically important.

Although I agree that increased cycle duration is an adverse effect, I
do not believe that the estrous cycle data have been analyzed
appropriately.  There is considerable intra- and inter-individual
variability in estrous cycle length.  Although the textbook cycle length
of 4-5 days is the most commonly observed pattern, cycle duration can be
irregular, especially (but not exclusively) when the animals are
maintained in all-female environments.   The problem of variability is
compounded by the fact that assessment of estrous cycle stage is
typically done only on a daily basis during the course of rodent
subchronic or reproductive toxicity studies; therefore, if the
assessment is done at a time when the animals are transitioning from one
stage to the next it could lead to the erroneous conclusion that some of
the animals had a cycle that was one day longer (or shorter) than
normal.  

These issues should not preclude the use of increased cycle length as
the critical effect for risk assessment, but the data need to be
analyzed differently.  My first criticism is that the data for cycle
length are expressed in the tenth or hundredths of days (!!), values
that suggest a degree of precision that simply doesn’t exist.   Would
the difference in cycle length at the LOEL still be significant if
estrous cycle data were rounded to the nearest day?  

My second concern is with the use of a 10% change in mean as the
benchmark response level.  Setting a BMR for a continuous variable takes
a lot more thought, and input of toxicological expertise, than setting a
BMR for a quantal variable.  Importantly, a 10% increase in the mean for
a continuous variable (especially one that displays a fair amount of
variability) is in no way equivalent to a BMR for a quantal variable in
which there is a 10% increase in number of affected individuals.  The
toxicological and clinical literature are replete with examples in which
the range of normal values extends well beyond the +/- 10% range. 
(E.g., a systolic blood pressure of 108 or 132 would not be considered
adverse, even though the “right” number is 120.)  In the specific
case of estrous cycle length, a 10% increase over the mean F0 control
value in the study being modeled (4.2 days) is 4.6 days.  This is still
within the range of what is considered normal by the experts in the
field.  The ideal solution to this problem would be to have a consensus
among experts as to what they consider to be the normal range.  In this
case, it would not only be the acceptable range in days, but also the
number of cycles that would need to be outside the normal range in order
for the individual to be categorized as an abnormal cycler.  I know of
no such consensus in the literature.  A simple approach would be to
consider any individual with cycle lengths 1 day or greater than the
controls to be abnormal, thereby converting the continuous variable to a
quantal one (i.e., number of abnormal individuals), and use a BMR of 10%
added risk.

A different approach, but one which I believe is reasonable in this
circumstance, is to use the variability of the data as the basis for
setting a BMR.  There are examples in the literature (Kavlock et al.
(1995) Fund Appl. Toxicol, 26: 211-22) in which the following cutoffs
were tried for continuous variables:  two standard errors away from the
mean; 0.5 standard deviations from the mean; being less than the 25th %I
le (or in the case of an increase, greater than the 75th) of control
values; or a 10% added risk of being in the top decile.  In any of these
cases, the choice will need to be justified, but for any the
justification is likely to be much more reasonable than the BMR of a 10%
increase from control mean.”

ICF Response to Dr. Daston

Dr. Daston indicates that the weight of evidence is strong enough to
support the conclusion that 1-bromopropane has adverse effects on
reproduction and that the change in estrous cycle length is the critical
effect in the cascade of reproductively toxic effects occurring with
increasing dose.  

Dr. Daston noted that at least some of the reproductive findings listed
by ICF to support the choice of critical effect are estrous cycle
stage-dependent and that this dependency should be acknowledged in
ICF’s weight-of-evidence (WOE) presentation.  Ovarian weight is noted
as being estrous cycle stage-dependent: follicle subtype distribution
and mean number of days-to-coitus should be similarly attributed.  ICF
agrees with this recommendation.  

Dr. Daston also had some concerns about the design and methodology used
in the Stump (2001) study.  First, estrous cycle data were not
appropriately analyzed by Stump (2001) because intra- and
inter-individual variability in estrous cycle length were not taken into
consideration to the extent needed to describe this variability. 
Second, he noted that the data for estrous cycle length are expressed in
the tenth or hundredths of days and this precision is misleading because
it does not actually exist.  ICF has acknowledged the limitations in the
way estrous cycle data were analyzed and has reanalyzed the estrous
cycle data, using available individual animal data, thereby eliminating
the need for averaging. 

Dr. Daston’s second concern was the use of benchmark dose modeling
based upon a 10% change in benchmark response (BMR) for mean number of
estrous cycles occurring during the 3-week period prior to mating.  He
suggested that benchmark dose modeling be redone, using as the BMR a
change representing 50% of the standard deviation (0.5 SD) of the
control data. The use of any multiple of a standard deviation implies
that the data are normally distributed. This assumption is clearly not
true in the case of data from our analysis of estrous cycles during the
3-week period data. Without normally distributed data, use of the
standard deviation is not advisable. Therefore, a 10% change in the mean
number of estrous cycles in a 3-week period, a change that ICF
reproductive/developmental experts believe is biologically significant,
is the selected BMR.  ICF therefore disagrees with Dr. Daston’s
recommendation to re-do the benchmark dose modeling for mean number of
estrous cycles in a 3 week period.

 

Response by Dr. Flaws

“I do consider a statistically significant decrease in the number of
estrous cycles to be a significant finding.  In general, the cycles in
rats are fairly regular (4-5 days) and when they become abnormal, its
often an indicator of toxicity at the level of the vagina, ovary, or HPG
(hypothalamic-pituitary-gonad) axis.  Since some of the data indicate
that there are more primordial follicles and fewer corpora lutea, this
suggests that the chemical may be inhibiting the ability of primordial
follicles to grow to the ovulatory stage.  If this is the case, the
animals may have lower levels of estrogen and this could lead to
abnormal cyclicity.  

ICF Response to Dr. Flaws

Dr. Flaws’ comments support ICF’s conclusions and no follow-up is
necessary.

Question 2.  Do you consider the changes in estrous cycle number and
length in Sprague-Dawley rats, in the absence of knowledge about
mechanism of action, to have human relevance?

Rationale:  The relevance of these changes in humans can be questioned
for the following reasons:

Human reproductive function differs from non-primates and rodents. The
toxicological data are garnered from rodent or rabbits studies, in which
the pattern, timing, and control of factors affecting estrous cycling is
not concordant with fertility variability in women. 

Human data are primarily obtained from individuals presented as
infertile; we generally do not have base cases or reliable control
groups for comparison with these individuals.

Menstrual cycle length varies significantly among women and is also
affected by a number of endogenous and exogenous factors, including
psychological variables including stress, as well as hormonal
interactions.

Non-clinical menstrual variation is prevalent and non-pathological. 

There is some evidence that many human couples experiencing reproductive
problems are able, without intervention, to reproduce successfully at a
later time. and

Reproductive outcome is dependent on performance of both sexual
partners.

Nonetheless, according to EPA policy, in the absence of mechanistic or
mode of action data, the Agency considers adverse findings in rodent
reproductive parameters to be potential indicators of similar adverse
effects in humans.   

Response by Dr. Luderer

“The rationale for this question correctly notes that there are
differences between humans and rodents in terms of the regulation of
menstrual cycles and estrous cycles. However, there are also many
similarities, and indeed much that we know about the human
hypothalamic-pituitary-ovarian axis was originally discovered in
rodents. Many modes of action by which chemicals can disrupt estrous
cycles in rodents are also relevant in humans. For example, suppression
of hypothalamic gonadotropin releasing hormone and/or pituitary
luteinizing hormone and follicle stimulating hormone secretion,
disruption of follicular steroid synthesis, or direct toxicity to
components of ovarian follicles can all adversely affect fertility and
menstrual cycle length in humans, as well as fertility and estrous cycle
length in rodents.  

The rationale also points out that there is significant inter- and
intra-individual variability in cycle length in women. Such variability
also exists in rodents, as illustrated by the range of estrous cycle
lengths cited in the Stump study for historical controls from their
laboratory. However, in both humans and rodents cycle lengths that fall
well outside the range of normal are associated with reduced fertility.
In humans oligomenorrhea is strongly associated with anovulation and
with decreased fertility (Haddad, 1984). Finally, the relevance of the
changes in estrous cycle number and length is strengthened because these
changes were observed in two separate studies using two different
strains of rats (Yamada et al, 2003; Stump et al, 2001), and they were
not isolated findings in either study, as discussed under my response to
Question 1. Thus, in my opinion the changes in estrous cycle length and
number observed in rats exposed to nPB have relevance to humans.”

ICF Response to Dr. Luderer

ICF finds that Dr. Luderer’s assessment of human relevance supports
ICF’s conclusions.  During the teleconference call between ICF staff
scientists and Dr. Luderer on October 27, 2004, Dr. Bonnie Ransom Stern
noted that ICF agreed with Dr. Luderer’s interpretation of human
relevance with regard to changes in estrous cycle length, and the lack
of utility of single measures of LH and FSH levels which are highly
variable, correlated with stage-of-estrus, and cannot be meaningfully
interpreted unless the precise stage of the estrous cycle during which
they are taken is known.  Further, there are numerous differences
between nonhuman/human primates and rodents with regard to the role of
the hypothalamus-pituitary-gonadal (HPG) axis in regulation of
ovulation.   Human relevance thus continues to be a concern, and at
present, there are no guidelines on these issues.   Nonetheless, in the
absence of guidance, EPA’s policy is to assume human relevance.

Response by Dr. Daston

“The short answer is, without mechanistic information, I don’t know;
however, there are a number of circumstances under which it could have
human relevance.  The rodent estrous cycle (and the vaginal cytology,
which was the actual parameter evaluated in the reproductive toxicity
study) is under the control of estrogens and progesterone, which in turn
are under the control of gonadotropins from the pituitary. These same
hormones control the human menstrual cycle.  While there are examples of
rodent-specific mechanisms of toxicity that perturb this endocrine axis,
there are also examples in which the mechanisms have human relevance. 
In the absence of mechanistic data, it is prudent to assume that the
rodent estrous cycle data have human relevance.  

Points d, e, and f on page 5 of the ICF letter, while all true, are
immaterial to the question of human relevance.  These all have to do
with the question of whether the effect is adverse, not whether it is
relevant to humans.”

ICF Response to Dr. Daston

Dr. Daston’s comments support ICF’s interpretation of the human
relevance of estrous cycle changes in rodents, in the absence of
mechanistic data.   

Response by Dr. Flaws

“I do consider the changes in estrous cycles in rats to potentially
have human relevance.  While I agree that the human and rodent cycles do
differ in many aspects, I also think that there are many similarities
between rats and humans.  The process of follicular development and
ovulation is controlled by the same hormones and receptors in both
species (the major difference in the timing of events-a 28-day cycle in
women and 4-5 day cycle in rodents).  Chemicals that disrupt cyclicity
in rats often have toxic effects in humans.  I agree that more
information about mechanism of action would be helpful- to absolutely
defining human relevance.  For example, I would like to know if the
chemical binds to estrogen receptors and whether it affects follicle
growth or numbers for the reasons mentioned above.”

ICF Response to Dr. Flaws

Dr. Flaw’s comments generally support ICF’s interpretation of the
human relevance of the observed estrous cycle  changes in rodents, and
agreed with Dr. Daston’s comment that additional information on
mechanism/mode of action of nPB-induced reproductive toxicity would be
useful in confirming (or disproving) human relevance.   

Question 3.  In summary, do you agree that the weight-of-evidence for
estrous cycle effects on reproductive data from animals is sufficient to
assume human relevance, in the absence of mechanistic information?

Response by Luderer

“Yes, the weight of the evidence is sufficient to assume human
relevance for all of the reasons enumerated under my answers to
Questions 1 and 2.”

Response by Daston

“Yes, as noted above, the data are sufficient to assume human
relevance.”

Response by Flaws

“I think that before I would make policy statements regarding humans
that I would like enough mechanistic information to know the site of
action of the chemical. This would help determine whether the potential
exists for the chemical to act in humans.”

ICF Response to All Experts

All experts agree that ICF’s weight-of-evidence presentation is
sufficient to assume human relevance of animal data regarding nPB’s
reproductive toxicity.  Both Dr. Luderer and Dr. Daston agreed that the
weight of evidence is sufficient even in the absence of mechanistic
information.   Dr. Flaws, as noted in response to Question 2, suggests
that additional information on mode and/or site of action of nPB’s
reproductive toxicity would be useful to define the degree of human
relevance of these animal data.  

Conclusion

Based on a critical review of external peer reviewers’ comments, with
follow up as needed, ICF concludes that all reviewers concur with
ICF’s rationale for selecting estrous cycle changes as the critical
end point for derivation of the industrial AEL for nPB.  All reviewers
agreed that the statistically significant decrease in the number of
estrous cycles within a given time period and the significant increase
in estrous cycle length (due to sustained diestrous) observed in the
Stump (2001) study are toxicologically significant findings.  Further,
all external peer reviewers agreed that the reproductive findings in
Sprague-Dawley rats, supported by similar findings in Wistar rats, can
be considered to have human relevance.  Although it would be useful to
have knowledge of mechanism(s) of action of nPB-induced toxicity to
evaluate the degree of human relevance of animal findings more
definitively, it is both reasonable and prudent to assume human
relevance, based on available information and EPA policy.  

References

Haddad, PF. 1984. Anovulatory Infertility. J Obstet Gynaecol. 4(Suppl
1):S44-48. 

NRC, 2001, Evaluating Chemical and Other Agent Exposures for
Reproductive and Developmental Toxicity (2001), Commission on Life
Science, Part I: Assessing Available Data, National Academy Press.

Stump D.G.  2001.  An Inhalation Two-Generation Reproductive Toxicity
Study of 1-Bromopropane in Rats.  Conducted by WIL Research
Laboratories, Inc., Sponsored by Brominated Solvents Consortium.  May
24, 2001.

Yamada T, Ichihara G, Wang H, et al. 2003.  Exposure to 1-bromopropane
causes ovarian dysfunction in rats.  Toxicol Sci 71:96-103.

 A summary of each expert’s background can be found in Appendix A of
this document. 

  DATE \@ "M/d/yyyy"  12/13/2004 	ICF Expert Panel Responses	  PAGE  1 

