Comments	Action/Response	2nd comments 	Action / Response 2	Page #

EPA





Analyses of fetal examination data should combine incidences of related
findings to better show trends that may be treatment related. Examples
of related findings that could be combined include:

vertebral centra variations (including thoracic and lumbar; including
unossified, bifid, and dumbbell-shaped)

supernumlerary thoracic/lumbar ribs (including rudimentary and
well-formed ribs on Ll and T 14)	Observations combined as necessary for
the reanalyses

	OK, with one exception:  that the combination for observations that may
reflect delayed sternabrae ossification may not be the best.  Based upon
page 234 of the report (K-4), I think that the most appropriate
combination of observations is 24, 25, 26, and 28.  This differs from
what the lab combined (25, 26, 27, 29).  This may be a typographical
error, since 29 doesn’t deal with sternabrae but deals with vertebrae.
See final draft report	Appendix K

The statistically significant increased fetal incidence of unossified
sternebra #5 was not regarded as evidence of developmental toxicity by
the study authors because the litter incidence was not statistically
significant.

True, the litter is the appropriate unit of analysis; however, as stated
above, the litter analyses (chi-square and Fisher' exact) presented in
the report do not consider withinlitter incidences and can lead to false
negative results. Thus, pending re-analysis, we tentatively regard the
increased incidence of unossified sternebra #5 to reflect developmental
toxicity at 20,000 mg/m3.

The study authors point out that sternebra #5 is the last to ossify and
that it's appearance is the most variable. This is hardly a reason to
dismiss the finding; to the contrary, it is this normal variability that
makes it conducive to detecting treatment-induced delays in
ossification.

Although sternebral ossification provides an index of developmental
delay, other ossification sites (e.g., caudal vertebrae, metacarpals,
metatarsals, phalanges) also show normal variability and, when
quantified, often are more sensitive in detecting developm.enta1 delays.
Apparently, however, the skeletal examinations did not attempt to
quantify the extent of ossification at such sites. In view of the data
suggestive of a treatment,-induced delay in sternebral ossification, we
recommend that the study authors evaluate the number of ossification
sites at one or more of these sites (e.g., the number of ossified caudal
vertebrae).

	The revised statistical analyses did not detect any statistically
significant differences between the control group and the groups exposed
to the test substance.  Therefore, these statement have been removed
from the report.	OK – see earlier response regarding combined
observations noted in Appendix K.  Re-analysis may alter the statistical
findings	See final draft report	i, 1-2, 4-2, 4-3, and 4-5







Goldsworthy





The only potential treatment related statistical difference between
control and treated groups was an increase in unossified sternebrae in
the high dose group on a fetal basis. This finding was not statistically
increase on a litter basis, the basic unit of analysis. Additionally,
increased abnormality was noted in the last forming sternebrae that
exhibits the most variable appearance in normal specimens. Collectively,
it is appropriate to not consider this finding in the NOAEL setting for
developmental toxicity. However, text should (1) provide a table with
the unossified sternebrae data in the results section for both fetus and
litter basis (Section 4, pages 4-1 to 4-3), (2) clearly state on 4-3,
skeletal observations, that the statistically significant increase was
noted on a fetus basis, and (3) consider noting this finding in the
abstract (as it was noted in the Summary and Discussion sections).	All
references to the unossified sternebrae removed from the text.  The
revised statistical analyses did not identify this as a statistically
significant observation.

.	Please see the earlier comment regarding re-analysis of sternabral
findings as a different combination than reported in the report	See
final draft report	i, 1-2, 4-2, 4-3, and 4-5

211 (b) Research Group Reviewer





- The one area of statistically elevated fetal findings in the study,
unossified sternabrae, should be compared to the historical control data
range as part of the discussion of interpretation.	This incidence of
this observation in the treated groups was not statistically
significantly different from the control incidence in the revised
statistical analyses.

	Please see earlier comment recommending combined analysis for
observations that can reflect delayed development of the sternabrae.	See
final draft report

	Page 4-5:

Discussion, first paragraph: The report needs to be consistent in the
interpretation of maternal results. In this first paragraph of the
discussion, the report has reiterated that there were statistically
significant decreases in body weight gain, but goes on to indicate that
these decreases were not biologically significant. If they are not
biologically significant, then the maternal NOAEL should be 20,000
mg/m3, not 10,000 mg/m3. The study director needs to make a decision as
to whether the effect upon body weight gain is an indication of adverse
effect or not, and present this position consistently and clearly.
Second paragraph: The report provides a justification for why the
increased incidence of unossified sternabrae is not deemed to be
evidence of developmental toxicity. The report should further back up
this position with regard to the data within the historical control
base. The first sentence conveys the perception that this observation
may be related to treatment. If that is what the director believes, this
is an appropriate way to word this sentence. If, as suggested by the
subsequent dismissal of this finding, the study director doesn't believe
this finding to be due to exposure, the sentence should objectively
(rather than subjectively) state that the only statistically significant
fetal observation was an increased incidence of unossified sternabrae.
Discussion revised to clarify that the decreased body weight change was
an adverse event and to reflect the revised statistical analyses of the
skeletal findings.

	OK, pending possible revision with an additional statistical analysis
of combined sternabral observations.	See final draft report	Page 4-5

211 (b) Research Group QA/QC Reviewer





Page vi, Compliance Statement:

 Since it was the sponsor’s responsibility to maintain the method of
synthesis, fabrication, or derivation of the test fuel, and this had not
been completed, it should be included in the sponsor’s compliance
statement.	We do not consider this a GLP deviation.  We have added this
to the Test Substance section.  	This needs to be the same in all
reports.  	See final draft report	3-1

Appendix H, Fetal Observations, Individual Fetal Observations, Page
H-17:

 The “a” finding is indicated in the data, but is qualified by
“other” in the key. It appears that additional information is
needed. This finding is not included in the “Incidence of Fetal
Observations”. Please clarify.	No change.  Page H-17 appears fine 	 
The finding now appears on page H-16, #386, the ‘a’ is qualified by
‘Other’ .   More information is needed and this finding is not found
in the summary of findings.  	See final draft report	H-17

Appendix H, Fetal Observations, Individual Fetal Observations, Animal
number 353:

 I could not find the skeletal observations designated as footnotes W,
DD, m, and R in the raw data. Please verify these findings.	M CORRECTED
IN FETUS 2  	Not sure what was corrected here.  Page H-51 is not in
question here.  I could not find the noted findings in the raw data for
this fetus.   If these findings are indeed in the raw data for this
fetus then the report is fine.

H-51



Appendix H:   The following findings were not found on the Appendix H
Summary but were found in the Statistical report:  

Head – anophthalmia, dilated lateral ventricle.  

External  – acaudate, anal atresia, anasarca, brachydactyly,
ectrodactyly, hyphosis.

Abd/Thor (Visceral) – malpositioned kidneys, ovaries and uterus, no
cervical spinal column.

	See final draft report	Appendix H



Appendix H and Statistical Report, Appendix K:  There seem to be a few
discrepancies in incidence between these two appendices.  They follow:

Total fetuses and litters with visceral malformations, group 4, 2 in
App. H, 1 in App. K

Visceral Variations in Group 1, App. H is 5, in App. K combined visceral
variations in Group 1 is only 4?  

	See final draft report	Appendix H & K



Appendix K, Table 1, page K-4:  At the bottom of this table, where
combined findings are noted, it appears that the numbers that are
combined are incorrect.  #17 is a skull finding that should not be
combined with ribs.  # 29 is a vertebrae finding that should not be
combined with sternebrae.  Please verify that these combinations are all
correct	See final draft report	K-4



Appendix K, Page K-5, first paragraph.  Should the word in the last
sentence be irrespective?  

	See draft final report	K-5



211(b) Toxicology Research Group

ETBE Rat Developmental Toxicity Study Report 

Reviewer Checklist 2 – Additional Comments

				  PAGE  1 	                                                         
                                

