Natural Resources Defense Council, Inc. et al v. United States Food and Drug Administration et al

Filing 44

DECLARATION of Amy A. Barcelo in Support re: 40 MOTION for Summary Judgment.. Document filed by Center for Veterinary Medicine, Bernadette Dunham, Margaret Hamburg, Kathleen Sebelius, United States Department of Health and Human Services, United States Food and Drug Administration. (Attachments: # 1 Exhibit A, # 2 Exhibit B, # 3 Exhibit C, # 4 Exhibit D, # 5 Exhibit E, # 6 Exhibit F, # 7 Exhibit G, # 8 Exhibit H, # 9 Exhibit I, # 10 Exhibit J, # 11 Exhibit K, # 12 Exhibit L, # 13 Exhibit M, # 14 Exhibit N, # 15 Exhibit O, # 16 Exhibit P Part 1 of 5, # 17 Exhibit P Part 2 of 5, # 18 Exhibit P Part 3 of 5, # 19 Exhibit P Part 4 of 5, # 20 Exhibit P Part 5 of 5)(Barcelo, Amy)

54864

m

4 Federal Register / Vol. 44, No. 185 / Friday. September 21, 1979 / Notices
.
i

a residue was even found in muscle
tissues),-with implants smaller than the
size permitted, 120days after
implantation (and also shorter periods
after implantation), the results reported"
show that DES residues will appear at
low levels in the edible tissues of cattle
implanted in accordance with approved
conditions of use. The fact that part of
the DES implants still existed in some
steers at the end of 120 days (and were
thus presumably sending DES into those
animals' skstems at the day of slaughter)
buttresses this conclusion.
(c) OralDosagesin Sheep.-(i) Study.
DES is permitfec in sheep feed at up to 2
mg per head per day, again with a 7-day
withdrawal period. 21 CFR 558.225. One
stuly was done with sheep by Dr.
Aschbacher (G-4). In this study, 6-sheep
were sacrificed 7 days after feeding with
a single dose of 11C-DES. Neither the.
report nor the testimony is clear as to
the amount of the '4C-DES dose. Each of
the sheep had been fed DES for the 7
days prior to the C-DES feeding, 2 at the
rate of 100 mg per day, 2 at 4 mg'per
day, and 2 at 3 mg per day. DES feeding
was continued in the first two groups for
an additional 7 days but was stopped-.
after the date upon which'the 14C-DES
was fed in' the third (3 mg) group. All
sheep were sacrificed on day 15 (7 days
4
following 1C-DES dosing).
No measurable radioactivity was
observed in the tissues of any of these
sheep, with the exception of the adrenal.
glands in three sheep. In his testimony,
Dr. Aschbachdr stated that the design of
this experiment and its analytical procedures would have prevented the
quantitation of radioactivity present at
the level of lqss than 1 ppb of DES
equivdlents in the animals receiving 4 or.
3 mg of unlabbled DES per day (G-4 at
3). (He did not address the sensitivity of
his methods as they apply to animals
receiving 100 mg per day. The rack of
sensitivity would also, hoivever, mean
that residues below I ppb from thoseanimals would not be detected.) "
(ii) Conclusion As to OralDosdges in
Sheep. The most likely reason for the .
failure of this study to show residues in
tissues other thai the adrenal glands is*
the relatively high limits of detection of
the test methods.', (Many of the residues
observed in the cattle studies were
observed at levels below'the ldwest
level of sensitivity (1 ppb) of this test.)
The presence of DES,residues in the
adrenal glands of the sheep tested is
evidence that DES residues remain
within the sheep's bodies rather than
passing totally out of their systein.
Because no rationale has been advanced
to support a theory that all DES.residues
in the sheep's body would be
concentrated in adrenal glands, Imust

T

conclude that DES residues would be
present, at non-observable levels, in the
other tissues of sheep fed DES.'
My conclusion on this subject is
isupported by the results observed in the
cattle studies discussed above. The fact
that both cattle and sheep respond to DES by increased growth warrants, for
present purposes, the assumption that
the two animals deal with ingested DES
in a similar manner. Such an assumption
is biologichlly plausible (and more likely
than the contrary assumption); and -nothing in the record contradicts it. Both
cattle and sheep are reminants and are
good models for ruminant metabolism.
I find that the results of the
radiotracer study in sheep, taken
together with the evidence from the
cattle studies, show that DES used
under approved (or actual) conditions of
use results in DES'residues in edible
tissues of treated sheep.
(d) Implants in Sheep. No radiotracer
study was performed with implanted
sheep. Although the question is not
without difficulty, I conclude that the
conservative approach appropriate for
safety determinations sanctions
extrapolation from the cattle-data,
despite spedies differences, to determine
that DES implants in sheep. result in DES
residues in the edible tissues of sheep. I
have discussed above my reasons for
concluding that sheep are likely to deal
with orally administered DES in a
maner similar to cattle. The same
considerations apply to DES implants.
The results of the radioisotope test of
DES fed to sheep (whick showed that
fed DES does remain in at least some
tissue of these animals) also lend some
"pport to the'conclusion that DES
implantation in sheep leads to tissues
residues. I therefore find that the
radiotracer studies show that use of DES
implants in'sheep in accordance with
approved.conditions of use results in
DES residues in edible tissues of the
treated-animals.
(e) The Pseudo-DESIssue. The Court
ordered a hearing on the withdrawal of
approval of the DES NADA's in part due
to applicants' argument that the residues
identified by the radioisotope procedure
were caused by an impurity in the DES
implants supplied to the government
researchers by Hess & Clark, Hess &
Clark,"Division Rhodia,Inc., v. FDA,
of
supra, 495 F. 2d at992. In particular, the
applicants argued that the implants
were contaminated with "pseudo-DES,"
which is somewhat similar in chemical

i

i|

, , ,m

when it became in Hess &Clark's
interest to do so (Bureaus' Brief at 02:
see also the cross-examination of Dr.
Tennent (Tr. at 1274-76)). The
manufacturers, on the other hand, argue
that they had forewarned the agency
that there were impurities In the
implants (Tr. at 1275). 1need not decide
this issue.
To resolve the pseudo-DES question,
Dr. Williams further tested the liver
samples from Dr. Rumsey's steer
studies. These tests showed that the
radioactivity identified by him as DES
or its conjugates was, in fact, authentic
DES or its conjugates and not the
impurity, pseudo-DES (d-99 at 3-5: G101; G-102).

Two of the manufacturers' witnesses
discussed the pseudo-DES problem. One
of them, Dr. Liebermnn.admitted on
cross-examination that in light of Dr.
Williams' work all of the observed
residue couldnot be pseudo-DES (Tr. at
2110).
The, manufacturing parties' Dr.
Tennent presented, in his direct
testimony, calculations he had made
from Dr. Williams' results. He stated
that he had made corrections for
contamination. He found 0.035 ppb of
DES and its conjugates in the 120-day
steer implanted with I implant and 0.120
ppb'in the 120-day steer implanted with
2 implants (M-132 at 15). A table made
up by Dr. Tennent for the samples taken
from the orally dosed animals showed
0.037 ppb DES and conjugates in one
animal slaughtered after a 7-day
withdrawal period and 0.011 ppb In the
other animal with the same withdrawal
period (id. at 17). Thus, Dr. Tennent's
analysis seems not to dispute the fact
that there was some actual DES and/or
its conjugates in the livers of some of
these animals. Although Dr. Tennent
stated that he considered the data '
marginal, due to inherent counting errors
at low levels of activity (M-132 at 10),
the record shows that Dr. Williams
minimized counting errors by extending
the counting time in his test procedure
(Tr. at 684).
I find that the residues detected
cannot be attributed wholly to pseudoDES or other impurities. Whether or not,
as the manufacturing parties' witnesses
contend, some of the residues detected
as DES and/or its conjugates are
impurities, it is clear that part of the
observed residues are in fact DES
residues. f cannot find, for the reasons
discussed below in the section of this
I
structure to actual DES.
Decision (sections III(C) and (D)) dealing
with the carcinogenicity and other
The Bureaus argue that Hess &Clark
withheld the information that there were, adverse effects of DES and Its
impurities in the implants until the
conjugates, that any amount of DES
radioisotope studies were-completed
residues is safe. Therefore, the fact that
DES residues have been shown to occur
and revealed that information only

Federal Register / Vol. 44, No. 185 / Friday, September 21, 1979
yet a third time using a different
derivative of DES (id. at 2-3).
A liver found to contain DES by one
or both of the first two gas
chromatography procedures but not by
the third test (whether it be-mass
spectrometry or gas chromatography) is
recorded as a presumptive violation (see
G-58 at 2-3). Dr. Spaulding noted
correctly that the conservative policy of
the Department of Agriculture (USDA)
in requiring confirmation of the first gas
chromatography test by the second may
result in an understatement of the
number of residues that actually occur
(0-94 at 3).
The Bureau submitted the testimony
of Dr. Bert Levy, a statistician from the
Department of Agriculture (G-58). Dr.
Levy stated the number of cattle and
sheep slaughtered during the years 1971
through 1975, the number of cattle and
sheep tested for residues from 1971
through 1976, and the number found to
contain violative levels of DES residues
from 1971 through 1976. (The total
number of cattle and sheep slaughtered
in 1976 was not available at the time the
testimony was submitted.) Or the basis
of these data he calculated, at a 95
percent confidence level, the percentage
range (i.e., the lowest and highest
possible percentage) of the total cattle
and sheep slaughtered in that year that
had violative DES residues. The
numbers for cattle range from a low of
0.2-1.0 percent (reflecting 9 livers
containing residues of an estimated 1780
tested] in 1976 to a high of 1.3L2.5
percent (reflecting 36 livers containing
residues of 2003 tested) in 1972. Dr.
Levy's calculations for sheep ranged
71).
from .09-0.6 percent in 1971 (5 livers
A portion of the liver is shipped to'a
containing residues of 1810 tested) to 0laboratory where it is analyzed by the
3.7 in 1976 (0 livers containing residues
gas chromatography method. This
of an estimated 100 livers tested).
method can detect (but apparently not
Dr. Levy's calculations illustrate the
positively identify) DES at levels as low
fact that the number of DES residues
as 0.5 ppb (Tr. at 492-93). If the gas
found represents a much larger number
chromatography analysis is negative.
of residues in the total treated
the liver is considered to be free of DES
population. If must be noted, however.
residues (-94 at 2). If the analysis is
that the percentages calculated depend
positive, the entire liver is then
more on the sample size than on the
requested and a second analysis is
number of residues found. This fact is
perfo'rmed, again using gas
apparent from the calculations as to
- chromatography procedures (id.). If this
sheep stated above. In 1976, when 100
test does not confirm the first result, the
sheep livers were tested and no
liver is not recorded as having been
violations were found, the computed
shown to contain DES residues (id.).
range of violations was 0-3.7 percent.
If the second gas chromatography
This calculation is not Intended to be
analysis is positive, a third test is run
(id.). If the level observed in the first two evidence that the violation rate was as
high as 3.7 percent. Indeed, as Dr. Levy's
-tests is high enough, this reconfirmation
calculations show, the percentage of
will be performed by mass
actual residues could be as low as zero.
spectrophotometric analysis (id.). This
The gas chromatography method of
procedure can detect levels of
approximately 2 ppb (id.). If the first two analysis was first supplemented by
mass spectrometry in either 1974 or 1975
gas chromatography tests had detected
(compare Tr. at 496 (Dr. Spaulding) with
DES at a level lower than 2 ppb, the gas
Tr. at 725 (Dr. Levy)). Therefore, Dr.
-chromatography procedure is performed
at low levels in the edible tissues of DES
treated animals (together with the
evidence on toxicity discussed in
section I(D) of this Decision below] is
cause for concluding that the approved
uses of DES have not been shown to be
safe and have been shown not to be
safe.
(f) ConclusionAs to Radiotracer
Studies. I recognize that application of
the results of the radiotracer studies to
approved (and actual) conditions of use
involves, in some cases, extrapolation.
Such extrapolation is commonplace in
science and is valid here. For the
reasons stated above, I find that the
radioisotope evidence discussed above
demonstrates that approved (and actual)
animal drug uses of DES, in sheep as
well as cattle, will result in DES
residues in edible tissues.
(3) Findingsby Departmentof
AgricultureMonitoringProgram.-(a)
Evidence of Residues. The Bureaus rely
upon evidence that DES residues have
been discovered in animal tissue by the
Department of Agriculture as part of its
monitoring program.
Dr. John Spaulding, Chief of the
Residue Evaluation and Planning Staff
of the Department of Agriculture,
testified concerning that Department's
residue monitoring program. He stated
that steer and heifer livers are selected
at slaughterhouses by inspectors in
accordance with (1) a random sampling
technique (described in some detail by
Dr. Levy JG-58)] and (2) a number of
sampling procedures designed to follow
up on evidence of potential violations
with particular lots of meat (Tr. at 470-

1 Notices

54865

Levy's 1975 and 1976 figures (29 livers in
1975 equalling 0.8-2.7 percent violations
and 9 livers in 1976 equalling 0.2-1.0
percent violations) were confirmed by
mass spectrometry. The manufacturing
parties emphasize that the gas
chromatography method alone is not
sufficient to identify residues positively
as DES. This position is consistent with
USDA's requirement of confirmation of
gas chromatography findings by mass
spectrometry in 1975 and 1976. In the
table at the end of Dr. Levy's testimony
(G-58), a number of apparent DES
residues (15 in 1975 and 29 in 1976) are
reported as presumptive violations, that
is, violations that were found by at least
one gas chromatography test but were
not confirmed by "mass spectroscopy."
(The term "mass spectroscopy" used in
Dr. Levy's testimony is a synonym for
the term "mass spectrometry" used
elsewhere.) Because these residues have
not been positively identified as DES, I
place less weight on them than on the
residues (stated above) that were
confirmed by mass spectrometry.
As discussed above in the section on
analytical methods, the gas
chromatography/mass spectrometry
method has not been shown to be
sufficiently specific to serve as an
analytical method for DES. Though this
lack of specificity might make absolute
confirmation of the residues as DES
impossible, the USDA results are
nevertheless probative evidence that
DES residues exist in the tissues
identified as containing DES residues by
this method. An analytical method that
does not meet all the requirements for
routine regulatory use may nonetheless
provide credible data for use in an
evidentiary hearing.
The manufacturing parties argue that
the Department of Agriculture findings
of DES residues must be discounted due
to three documents (M-18; M-19; G-281,
which, they allege, show "apparent
failures by Department of Agriculture
employees to follow procedures that had
been agreed upqn with the FDA for the
handling of samples of livers to be
analyzed for residues of DES"
(Manufacturing Parties' Exceptions at
49). The inference that the
manufacturing parties seek to draw from
these exhibits. i.e., that there was
something wrong with the procedures
utilized by USDA, is not warranted.
Two of the three memoranda reflect
an agreement, reached in early 1974, to
have USDA and FDA use the same
method of gas chromatography analysis.
An October 22,1974, memorandum
suggests that USDA had not made much
progress in utilizing the agreed method
(M-18 at 3). On April 18,1975 (M-19).

54866,

Federal Register / Vol. 44, No.- 185 / Friday, September 21,- 1979 / Notices

the Bureau of Foods reported to the
FDA's Associate Commissioner for
Compliance thiat a new agreement had
been worked out in accordance with
which USDA would utilize the FDA
method exclusively and then confirm by
mass spectrometry the identity of any
residues'found. USDA would then report
as positive any reading so confirmed.
Neither of these memoranda shows that
the procedures previously hsed by
USDA were invalid, and thus neither
provides a basis for disregarding the
USDA residue findings.
The manufacturing parties take a
sentence out of context from the third
document referred to, a December 17,
1975, memorandum to the FDA's Chief
Counsel from the Bureau of Foods [G28), to imply thai the FDA'was not
satisfied with the sampling technique
utilized by USDA. In fact, the question
raised there was whether USDA was
cooperating correctly in a multilaboratory test of the FDA's gas
chromatography method. This
document, also, provides no basis for
disregarding the USDA residue findings.
It is apparent, therefore, that DES
residues have been fbund in the past
few years in the livers of cattle by the
methods utilized by the Department of
Agriculture's sampling program.
(Although the majority of these residues
appear to result from use of DES in feed,
some result from the use of DES
implants (Tr. at 769).) These residues
havebeen identified in only a relatively
small percentage'of the animals tested,
but it must be recalled that (1] not allresidues will be caught by this system
because the lowest level of
measurement claimed is 0.5'ppb and (2)
the residues found represent a
significant amount of meat (1 percent of
25 million steers is 250,000).
(b) The Question of Misuse.---(i).
Necessity of Determining Whether
Residues in Edible Tissues Result From
Misuse. The manufacturing parties

Proceeding as well. The section in
question requires the Commisiioner, in
determining whether a drug is safe, to
consider "among other relevant factors"
four specified factors. One of these is
"whether the conditions of use
prescribed, recommended, or suggested
in th proposed labeling are reasonably
certain to be followed in practice," 21
U.S.C. 360b(d)(2)(D).
The manufacturing parties seem to'
argue that at'some arbitrarily selected
percentage of misuse of all animal drugs,
"reasonable" misuse (to be tolerated) is
divided from "unreasonable" misuse (to
be the basis for a withdrawal). Then,
they seem to argue, if residues are not
found that prove that that percentage of
misuse had. been exceeded, the drug
must be declared safe no matter how
harmful the residues found may be to
the consuming public,
This interpretation is inconsistent
with the statute's terms. Whether
conditions of use are reasonably certain
4o be followed is only one of severdl
factors to be considered, and the
ultimate issue is whether the animal
drug is safe.
The term "reasonably certain to be
followed in practice" must, in any'case,
be interpreted in the context in which it
appears i.e.,.as a consideration irr
deciding whether the use of a drug is
safe, Thus, the amount of certainty that
is reasonable necessarily i'aries with the
danger posed by the drug. One degree of
certainty would be required (i.e.,
reasonable) for a drug whose residue
would kill a human consumer on the
spot, whereas another degree bf
certainty would be required for a drug
whose residue represented only a
relatively remote danger to the ultimate
human consumer. The failure to show
the extent of'the danger associated with
residues of DES above 0.5"ppb (or above
any level of residues-see section II(a)
(2 and 3) of this Decision) prevents a "
determination that the reported residues
are consistent with, "reasonable"
argue: '"he question * * * is whether
the number of violations is so great as to certainty that approved conditions of
use will be followed in practice.
show that the approved conditions of
The manufacturing parties sought to
use are not 'reasonably certain to be
introduce into evidence a document
followed in practice"' (Manufacturing
Parties' Exceptions at 59), The question-, -showing the extent of detected residues
tolerated by the FDA for other animal
however, is whether DES causes
drugs (M-148a). This document was
residues that have not been "shown to
properly excluded from the evidentiary
be safe," 21 U.S.C. 360b(e)(1)(B).
record (see discussion of evidentiary
The manufacturing parties refer to 21
rulings (section VI of this Decision)
U.S.C. 360b(dJ(2), which~sets out factors
below). In any case, the.argument that,
that the Commissioner must consider in
the percentage of residues detected for
determining an animal drug's,safety in
DES is no greater than the percentage of
the context of a refusalto approve an..
residues detected for other animal drugs
NADA. Because that section provides
is irrelevant. Because no safe-dose for
evidence of congressional intent with
DES may be computed, DES cannot be
respect to the meaning of the term
"safe," as used in the statute, it is
compared to other animal drugsfor.
appropriate-to refer to it inca withdrdwal which a safe dose can be computed.

Agency policy requires that the level
of detection of the analytical method for
an animal drug be set to pick up any
residues above the safe dose for that
drug.-For carcinogens, a "virtually safe"
dose or "no residue" level is utilized (C24, see also 44 FR 17070; March 20,1970).
The percentage of detected residues for
other animal drugs should, therefore, be
the percentage above the safe dose. The
percentage of residues computed for
DES represents, at best, only the
percentage of residues above o.5 ppb,
the lowest limit 6f detection of the gas
chromatography method of analysis. We
do not know how many residues occur
above the "safe dose" of DES because
no "safe dose" has been identified. Even
if one accepts the Bureaus' witnesses'
calculations ofi ppt as a "virtually
safe" dose, aS I do not, no one knowi
how many residues occur above that
level.

'It is true that some animal drugs have
been approved by the FDA using
analytical methods that do not have a
lowest limit of reliable measurement
corresponding to a safe or "no residue"
level by today's standards. Conceivably,
some such NADA's may have been
approved by mistake. Some are under
review-by the FDA now (see, e.g., 42 FR
43770; Aug. 30,1977 (penicillin) and 42
FR 56254; Oct. 21, 1977 (chlorteracycline
and oxytetracycline). (The cited
documents are notices of opportunity for
hearing in which one of the issues raised
is whether the tolerance levels approved
for those drugs are in fact, "safe levels.")
The approval of other NADA's will be
reviewed in an orderly manner in
accordance with agency priorities
pursuant to its ongoing "cyclical revier"
program (see 42 FR 64369; Dec. 23,1977),
It may be that the FDA will find, after
careful review, that it cannot determine
the percentage of residues above a "safe
level" or "no residue" level for these
other animal drugs. If it makes that
determination it will find, as I have done
with respect to DES, that the existence
of any amount of residues in edible
tissues means that the approved
conditions of use can not be found safe
as "reasonably certain to be followed Id
practice." The comparison of the
number of DES residues detected above
0.5 ppb with the number of residues

detected for these other drugs is
meaningless at this point.
I need not decide whether or not the
residues found r6sult from approved
conditions of use. The residues present a
safety question and (1) If they result
from approved conditions of use, those
conditions have not been shown to be
safe or (2),if they result from misuse,
then I can not find that the approved

Federal Register / Vol. 44, No. 185 / Friday, September 21, 1979 / Notices

5M7

conditions of use are reasonably certain
D at I n. *). Although my review of these above, the Bureaus have the burden of
to be followed, for the reasons discussed EIR's reveals a somewhat larger number showing that residues are occurring
above. In either Case, residues that have of EIR's lacking a showing of misuse, I
under the approved conditions of use if
not been shown to be safe are entering
cannot find that these reports
a decision on that issue must be made at
the food supply in amounts that must be
demonstrate that DES residues occur
all. The Bureaus have failed in their
considered to pose a significant risk to
when the approved conditions of use are burden, and the residues are therefore
the health of consumers.
followed.
attributed to misuse.
(ii)Evidence As to Causes of
Acceptance of the investigator's
(2) In light of the misuse
Residues. I have, in any case,
findings as evidence that residues will
demonstrated. I find that the approved
considered whether the record shows
occur when the DES is used under
conditions of use are not "reasonably"
that the DES residues detected by USDA
certain to be followed in practice.
approved conditions of use would
result either from following approved
reflect an unjustified confidence that
(4) GLCResidue Study. Dr. Rumsey et
conditions of use or from misuse of the
where FDA inspectors had not found
al. performed one study of the fate of
drug. The only evidence of potential
evidence of misuse there was no misuse. implanted DES in which radio-isotopes
value in resolving the issue are reports
As misuse is a violation of the law, there were not used (G-78).
by FDA investigators. The Food and
would, of course, be incentive for feed
Four lots of 16 steers were implanted
Drug Administration follows up on
with two 30 mg-DES implants each.
lot operators to clean up before the FDA
reports of DES residues 'madeto it by
inspectors got to them. It would thus be
Steers were sacrificed at14 days. 28
USDA. in most cases by visiting the
surprising if FDA inspections caught the days. 56 days. 84 days and 119 days.
facility at which the animal was treated
misuse in every instance. Therefore. I
Animal tissues were analyzed using
with DES. The Bureaus presented a set
can not rely upon the relatively small
identical gas chromatography
of approximately 140 establishment
techniques in two different laboratories.
percentage of investigations of residues
investigation reports ("EIR's") prepared.
that do not show misuse as proof that
This test did not show the presence of
by FDA inspectors who were seeking
residues result when there is no misuse. DES in the tissues of animals
the'cause of reported residues.This set
I conclude that the record does not
slaughtered after more than 28 days.
of papers has been marked as Exhibit
permit resolution of the question
One of the two analytical laboratories
G-89. The Bureaus also presented a
found measurable DES in two of the
whether the residues found by USDA
summary of EIR's from investigations of
are or are not the result of misuse of
animals slaughtered after 28 days but
the causes of reported residues. That
the other laboratory did not make that
DES.
finding (G-76 at 2). The report of this
(c) ConclusionAs to Findiogsby
summary was marked as G-137.
USDA MonitoringPrograim The USDA
The manufacturing parties note
study.-and Dr. Rumsey, stated that the
level of sensitivity of the gas
(Manufacturing Parties' Exceptions at
reports demonstrate that residues in
56) the discrepancy between the listing
edible tissues do occur as the result of
chromatography-method is 0.5 ppb (G-76
of the DES findings in Dr. Levy's
'the use of DES pursuant to its approved
at 2 G--78 at1]. This study, asDr.
testimony (G-54)-and the DES residues
(or actual) conditions of use. both In
Rumsey stated (G-76 at 2). neither
noted in the summary of FDA food and in implants, as an animal drug
proves nor disproves that DES residues
investigations (G-1371. For some years.
in cattle. The reports do not. due to the
appear in tissues at leels below 0.5 ppb
1
-- 37 lists more residues than does
small number of tissues sampled in
when DES implants are used in
recent years, show whether or not use of accordance with their approved
Levy; in other years, it lists fewer. The
Bureaus have, however, explained this
DES as an animal drug results in DES
conditions of use.
Part of the DES implants (about 20
discrepancy. the FDA inspection figures
residues in the edible tissues of sheep.
I conclude that it is not necessary to
are based upon not only the "objective"
percent of the initial weight) remained in
decide whether the residues found result the steers 119 days after implantation
(i.e.. random) sampling program
from the approved conditions of use or
(G-76 at 2-3). This fact suggests that at
described by Levy [see 4iscussion
above) but also the "for cause" program. from misuse of the drug. Whether or not
least some DES implants remain in
animals, releasing DES to their systems.
which involves followup sampling of the the residues result from approved uses,
products of previous offenders (Bureaus' the record demonstrates, as discussed in 120 days after implantation. This finding
the sections on safety below, that these
supports my conclusion that approved
Reply to Exceptions at 6). Thus. in those
conditions of use of DES implants result
residues are potentially hazardous and
years when Levy reports more residues
have not been shown to be safe. To the
in residues in the tissues of the animals
than the FDA, the FDA did not
extent that the possibility that DES will
investigate each residue reported.
at slaughter.
Where the summary shows more
be misused is a factor in this safety
(C) The DES ConjugatesIssue
residues, the FDA has investigated some decision, that factor does not support
The Court inHess & Clark stated as
the safety of DES. The record provides
residues found in the "for cause"
one issue to be considered in the DES
no basis for a conclusion that the
program.
hearing: "[Wihether the detected
approved conditons of use are
The manufacturing parties object to
"reasonably" certain to be followed.
residues are composed solely of DES
any reliance upon G-137 since the
I have also made an alternative
conjugates, and whether that substance
person who male up this summary was
isharmful;
finding to obviate any need for remand
* "",495F.2dat994.The
not presented for cross-examnation.
in case a reviewing court should decide
context indicates that the adverb
Some but not all of the EIR's
"solely" refers to the manufacturers"
summarized in G-137 were made part of that I am obliged to determine whether
or not observed residues result from
arguments thatthe residues detected are
the record as part of G-89. In reaching
my decision I have relied exclusively on
misuse. That alternative finding is as
solely DES conjugates as opposed to
follows:
DES itself, and that the harmfulness of
the EIR's actually made a part of the
(1) The observed residues result from
DES conjugates had not been put in
record in G-89.
misuse. Where the record does not
The manufacturing parties suggest
issue.
contain sufficient evidence to decide a
Conjugates of DES are. according to
that only 12 of the 140 EIR's in G-89 do
question, it is decided against the party
not show evidence of misuse
the Bureaus' Dr. Kenneth Williams.'
"compounds composed of DES,
{Manufacturing Parties' Brief, Appendix
with the burden of proof. As discussed

54868,

. Federal Register

/

Vol. 44, No. 185 / Friday, September 21, 1979 / Notices
I

chemically linked to another molecule oi shows that use of the drug results in
molecules through one of its hydroxyl
residues 'of unidentified substdnces, the
groups in such a fashion that hydrolytic
Commissioner must decide whether,
[chemical or enzymatic] procedures may despite his lack of knowledge of these
regenerate the parent compound" (C-99
substances, the drug may be considered
at 2). Dr. Williams stated further. "In
to be "shown to be safe."
DES conjugates, the DES molecule is
I reject the contention that the Court
attached to another molecule but is
in Hess &Clark-was demanding that the
otherwise structurally unaltered" (id.).
FDA identify the DES residues foimnd
The mantifacturers' Dr. Sieck stated
and demonstrate that those residues are
under cross-examination that a test on
not safe. Such a requirement would
which he was working had identified as
place the public in danger during the
conjugates of DES, the following: sulfate period (perhaps years) necessary to
of DES, the monoglucuroi'iide of DES, thi characterize and test suspect residues of
e
monoglucuronide of methoxy DES and
approved drugs. It Would also put the.
two other uncharacteized glucuronide
FDA in the bbsiness of drug t6sting, a
conjugates (Tr. at 1370). Dr. Kaltenbach,
task that Congress intended to be the
another expert'supporting the
responsibility of the manufacturers of
manufacturers' interests, stated that not
regulated proiducts (see,-e.g., H. R; Rept.
all residues had been identified (Tr. at
No. 2284, 85th Cong., 2d Sess. 1 (1958)).
2087).
As noted, the Court in Hess &Clark
(1) Burden of Proofon Residue Issue.
did require "some evidence" of a link
The Court did not state who would havE between the iesidue and safety bef6re
the burden of showing whether residues
any burden is placed upon the applicant
found are solely DES conjugates and
to identify observedxesidues and show
whether thcse conjugates are,.harmful. I
theii (and, thus, the approved drug's)
did make clear Its rejection of-the FDA,
safety Tlifsiieqfiirement, not evident
argument that a new discovery of
from the statutef is nevertheless met.
unidentified residues is itself sufficient
here. Thos6 residues resulting from thb
to show that an animal drug is'no longeiL use of DES that hive been identified
shown to be safe. The Court stated ihat
-hav bedn'diitii d as DES and/or its
the agency "has an initial burden of
conjugate6'j(s&,e.g., G-99 at 5-6; see
coming forward with some evidence of
als6 disd6sidan6ove in section H(B)
the relationship between the residue
-anddiscugsion below). It is elementary
and safety to warrant shifting to the
biochemi 6ty iaht.the conjugation of a
manufacturer'the burden of showing
ifolecule, althdugh it may change that
safety. This is at least the case where,
moledule' 'activity quantitatively, rarely
as here, the residues are of unknown
composition" 495 F.2d at 993 (emphasis -"elimixzites it.
This change in but failure to eliminate
added); see also Chemetron,supra,495
.the activitylof DES has been shown to
F.2d at 1000.
I The question of what happens when
_occur ithriesp'ct to the estrogenic
activ
f
6o 66njugate of DES (see Mnew evidence shows that an approved
animal drug adds unidentified iesidues -110 at 3; G-102: Comments on Vineland
to the human food supply"ione of great Laboratories Submission at 1; see also
discussion in section III (c)(2) of this
importance to the FDA's ability to deal
-Decision below). Also, as discussed in
not only with DES, but also with other
detail below, DES conjugates would be
animal drugs, Chemicals-such as animal
expected to hydrolyze (break down) in
drugs invariably are metabolized, at
the human body to form free DES, tlus
least in part, into other substances in an
making DES conjugates as dangerous as
animal (or human) body. It is for this
DES-itself. Therefore, there is
reason that the FDA requires
substantial evidence in.the record that.
,identification of the principal
warrants an inference that the DES
metabolites of an animal drug, and
conjugates are active in a manner - '.
demands toxicity testing and analytical
similar to that of DES itself. Due to -the'
methods for those metabolites, before it
recognized dangers associated with DES
will approve an NADA-(cf.:,G-24; 44 FR
(see the-discussion of the safety data
17081 et seq. (March 20, 1979)), The
with respect to DES below), there is,
agency's concern about these
therof6re, without question "some,
substances -"formed in or on food
because of the use of" the animal drug i4 evidence"_iat residues identified as
DES and/or its conjugates are unsafe.
in accord with -the statute's
-Thus, if some evidence of a
,requirements, 21 U.S.C. 360b(d)(2)(A);
relationship between the residues found
Once an NADA is approved, as
and safetjis necessary, that evidence is
discussed previously, the agency can
withdraw approval if "new,. ,
presenthere, The manufacturing parties
therefore have- the burden of identifying
evidence t * * shows that such drug is
the residues and showing .them'to be -.
not shown-to be safe," 21 U.S.C. . safe.., ....-- ,,.
_
, . ,
.
360b(e)(1)(B). Where new evidence. -

(2) Failureof ManufacturingPartiesto
Satisfy Burden of Proof.It Is clear that
the manufacturing parties have shown
neither that the residues found are
solely DES conjugates (rather than
totally or partially DES Itself), nor that
DES conjugates are safe.
The manufacturing parties presented
no data to show that all DES residues
found would be in the conjugate form,
They have not even advanced a
theoretical basis that justifies an
expectation that all residues would be
conjugated.
The only investigation made of any of
the residues detected to determine
whether or not they contained free DES
showed that in fact free DES residues
were present, see G-103 at Tables V,
VII, IX, X, XII, and handwritten tables.
The Bureaus' expert witnesses did not
rely upon this finding, however, and, as
discussed below, the analyst who
detected free DES noted that it can not
be proven that the free DES he observed
did not arise from hydrolysis of a DES
conjugate during analysis (G-212:
Comm6nts on the Vineland Laboratories
Submission at 1). 1 am thus left with a
record devoid of support either for the
proposition that the'rbiidues found are
"solely" DES conjugates or for the
converse of that pioposition. Tie
manufacturing parties have thus failed
in their burden of proof on this Issue.
Even a'ssuming that all the residues
discover~d were DES conjugates, the
manufacturing parties have failed to
show that DES conjugates are safe. The
only evidence in the record on this
question is Dr. Kilman's testimony that
DES-monoglucuronide had not caused
renal (kidney) tumors in hamsters after
15 montha (M-10 at 4 M-25) though it
apparently did cause dysplastic changes
in th'dse inimals (Tr. at 1827-208. (Cf. M-113 at 764 in which researchers suggest
that'it is a conjugated form of DES that
is responsible for kidney tumors in
hamsters.) The test cited by Dr. Kilman,
of one animal species,'for less than the
animals' lifetime, in which the
investigators looked only for one type of
tumor, can hardly be accepted as
evidence that DES conjugates are shown
to be safe in man. It is perhaps
noteworthy that the DESmonoglucuronide was adiinistered
-subcutaneously in the hamster
experiment {M-25 at 1252), a route that,
would be expected to prevent the
metabolism of-the glucuronide to DES
itself (id. at 1255; M-110 at 3). As
discussed below, the record provides
evidence that DES conjugates are unsafe
because they hydrolyze in the human
body to DES itself.-Dr. Kliman also testified (M-110 at 3)
that DES-monoglucuronide, when

Federal Register / Vol. 44, No. 185
administered by the subcutaneous route,
had been shown in one test (M-111) to
have 6 percent of the estrogenic potency
(measured by effects on the cells of the
vagina] of DES itself in rats and in
another study to have 9 percent of the
estrogenic potency (measured by effect
on the weights of uteri) of DES in rats
and 16 percent in mice (M-24). Dr.
Kliman neglected to mention that the
latter test showed that, when
administered orally, DESmonoglucuronide had 40 percent of the
estrogenic activity of DES in rats and 28
percent in mice id. af 651). If one were
to accept the manufacturing parties'
argument that estrogenic activity is
associated with carcinogenicity and
toxicity, the evidence cited by Dr.
Kliman in fact might be taken as some
evidence that DES-monoglucuronide is
unsafe. In any case, these data do not
show the safety of DES conjugates.
Thus, I find (1) that the Bureaus have
presented enough evidence (see
subsection 1 of this section above) to
raise substantial questions about the
safety of the residues of DES; (2) that
these residues consist of free DES or its
conjugates or combinations of free DES
and its conjugates; (3) that the
manufacturing parties have not shown
that the residues detected are solely
DES conjugates; (4) that the
manufacturing parties have not shown
that DES conjugates are safe; and (5)
that therefore the safety questions
raised by the Bureaus remain
unresolved. These findings, together
with my finding (discussed above) that
new evidence has shown that use of
DES as an animal drug produces
residues in edible tissues of treated
.animals, constitute a sufficient basis for
withdrawal of approval of the DES
NADA's.

(3) FindingsAssuming That Bureaus
Have Burden of Proof.The
manufacturing parties read the Court in
Hess &Clark and Chemetro as
assigning to the Bureaus "the burden of
coming forward with evidence sufficient
to resolve * * * in their favor" the

issues of the identity of the residues
found and whether those residues are
harmful (Manufacturi0g Parties'
Exceptions at 70-71). I now consider the
evidence in the record under this standard.
(a) Evidence That Residues Contain
FreeDES. Dr. Williams analyzed the
livers of steers implanted by Dr. Rumsey
et al. with radioactive DES (see,
generally. C-99). (These radio-isotope
studies are discussed in detail in section
III(B)[2) of this decision.] Dr. Williams
sought to determine whether any of the
radioaclive-residues that were found in

/

Friday, September 21, 1979 / Notices

the livers of the treated steers were in
fact free DES. He found free DES. (GX,
103 at Tables V. VII, LX. XII, and
handwritten tables G-102: Comments on
the Vineland Laboratories Submission
at 1).
The manufacturing parties take the
position that no free DES was actually
found by Dr. Williams (Manufacturing
Parties' Exceptions at 75-76). They focus
on Dr. Williams' analyses of residues
found in the liver samples from the two
steers implanted with radioactive DES
that were slaughtered after 120 days.
The attack on the findings in the first
of these two liver samples is premised
upon a mischaracterization of Dr.
- Williams' testimony on crossexamination. The manufacturing parties
state, incorrectly, that Dr. Williams
conceded that the amount of
radioactivity detected in the "free
fraction" of this first sample was so
close to background radiation as to
make his finding of free DES
meaningless (id.). It is important to note,
however, that Dr. Williams analyzed for
free DES three separate subsamples of
each sample of liver provided by Dr.
Rumsey (see, e.g., G-103 at Table VII).
At the hearing, Dr. Williams was asked
about the subsample in which the
radioactivity of the fraction of the
residue identified as free DES was the
lowest. He stated that the accorded no._
particular significance to the results for
that subsample because they were so
close to background (Tr. at 702]. The
manufacturing parties rely on this
comment by Dr. Williams. The comment
applies only to one of the three
subsamples analyzed from the liver
samples from the first 120-day steer. The
fact that each of the three subsamples of
the first liver sample produced a result
above background provides'more
assurance that the result was a true one
than would a single subsample standing
alone. In addition, each of the other two
subsamples of this first liver sample
produced a result higher than the one
-about which Dr. Williams was
questioned. Dr. Williams stated that he
thought his findings for this whole
sample (and the sample from the second
120-day steer) were significant (G-102:
Comments on Vineland Laboratories
Submission at 1).
The liver sample from the second 120day steer produced slightly higher
findings of free DES than the sample
from the first steer. The manufacturing
parties also attack Dr. Williams'
findings with respect to the sample from
the second 120-day steer, in part by
taking out of context statements made
by Dr. Williams.
"Counts per minute" are the units of
-measurement of the method by which

54869

Dr. Williams analyzed the residue. In
the liver sample from the second 120day steer, Dr. Williams observed free
DES that provided a response of about 2
counts per minute above the background
rate (Tr. at 702]. The manufacturing
parties rely upon statements by Dr.
Williams dealing with his analysis of a
different part of the residue (the hexane
fraction] found in the livers
(Manufacturing Parties' Exceptions at
76). He stated that "for these particular
samples" (i.e., the samples tested in the
haxane fraction analyses] 2 or 3 counts
per minute would be "on a shaky line"
(Tr. at 684) and elsewhere stated that 2.1
cpm would be "marginal above
background" in the hexane analyses (Tr.
at 091]. While these statements were
equivocal, I take them to mean that. for
the analysis of the hexanefraction,2-3
counts per minute was too low to
produce a reliable result. Dr. Williams
does not seem to have admitted, as the
manufacturing parties suggest, that his
findings in his analysisforfree DES
with the second 120-day steer were
insignificant. In fact, he stated
unequivocally that these results were
not as was suggested to him during
cross-examination. "meaningless" (Tr.
at 702).
The manufacturing parties state that
the Bureaus' Dr. Aschbacher testified
that it was necessary to detect counts
per minute of more than twice the
background rate (not found for the two
120-day steers) in order to have
meaningful results (Manufacturing
Parties' Exceptions at 77). Yet the
transcript reference cited makes it clear
that Dr. Aschbacher's conclusion was
applicable only to his own study,
because of that study's design (Tr. at
597-98).
The manufacturing parties' witness
Dr. Tennant stated his opinion that the
low number of counts per minute
observed in the residues found in the
livers of the two 120-day steers were
"marginal" (M--132 at 16].
(Manufacturing parties' Drs. Lieberman
and Kliman also made conclusory
statements about the validity of the
results observed with the 120-day steer
livers (M-122 at 2 M-110 at23.) The
record shows, however, that Dr.
Williams minimized the likelihood of
error in his analysis by utilizing a
relatively long counting time (Tr. at 684].
I accept Dr. Williams' analysis of his
own results.
The manufacturing parties argue that
it has not been proven that an
unidentified impurity was not
responsible for the free DES observed
Manufacturing Parties' Exceptions at 7778). My conclusion that Dr. Williams'

54870

Federal Register / Vol. 44, No. 185 / Friday, September 21, 1979 -/ Notices

results are not totally attributable, to the
impurity called pseudo-DES is discussed,
above in section III(B)(2)(c) of this
Decision. There is no reason to believe
that significant impurities other than
pseudo-DES existed in the radib-labeled
DES or that, if they existed, they would
have caused the tests to reveal free DES
erroneously. Thus, this speculation does
not provide a basis for discounting Dr.
Williams's observations.
The manufacturing parties do not
attack Dr. Williams' finding of free DES
at much higher levels in the i4 C-DES
residues found in the tissues of steers
slaughtered less than 120 days after
implantation with DES (see G-103,
Table VII.) They provide no
explanation-and I am aware of none--rfor why free DES would be part of-the
i4 C-DES residues in. animals
slaughtered at less than 120 days but
would not be part 6f residues found at
120 days (cf. Tr. at 2122). The results
found with the sub-120-day samples thus
confirm the'results seen by Dr. Williams
with the 120-day samples.
Although the results of Dr. Williams'
analysis oflivers from'animals fed DES
(as opposed to those implanted with
DES) were not discussed; Dr. Williams'
tables reveal that he also found free'
DgS in the.livers from the steers fed
radio-labeled DES (G-103: handwritten:
tables). The manufacturing parties have'
suggested no reason why, id any case,
the evidence on this subject from DES
- implants would not be applicable to
DES used in feed.
I find therefore that Dr. Williams'
analysis revealed free DES.-This finding;
however,' does not necessarily mean
that it has been demonstrated that use
of DES as an animal drug results in
residues that contain free DES.-.
-According to the analyst, Dr.
Williams; it can not "be proven that the
free DES did not arise from hydrolysis of
.
'
some conjugate (other than
monoglucuronide) during the work-up of
the samples" (G-102: Comments on the
Vineland Laboratories Submission at 1).
(Dr. Williams added tritium-labeled
DES-monoglucuronide to some of the'
DES tested. His parenthetical exclusion
apparently was meant to-make clear.
that the free DES did not come from
hydrolysis of the added product.) Dr.
Williams' analysis thus shows that the
residues contain either free DES or a
conjugate hydrolyzable 'to free DES.
As I found in section III(B) of this
Decision dealing with the detection of
DES residues, the 'recordshows that use of DES as an animal drug results in.
residues, in the edible tisues'of treatedanimals, of DES and-or.its conjugates.
As discussed above, there 'is no reason,
to believe that these-residues -would be'.

with radioactive tritium.'Their excretory
"solely" DES conjugates as opposed to
products were then analyzed. The
DES itself. Based on the evidence in-the
researchers found that the DES.
record, however,'I cannot exclude that
monoglucuronide and the DES itself
possibility.1 thus consider the question
resulted in simular metabolic products
whether DES conjugates have been
in the urine of the volunteers. (The
shown to be unsafe. (b) Evidence of Lack 6f Safety of DES
different radioactive labeling of the DES
and the conjugate made it possible to
Residues. I find, on the basis of trace the metabolites to their parent
evidence in the record,.that if the DES
compound.) This finding, together with
residues in the edible tissues of treated
animals are, conjugates of DES, those
other indirect evidence, showed that the
conjugate was hydrolyzed to DES In the
conjugates would be expected to break
intestinal tract prior to absorption into
down (hydrolyze) in the human body to
DES itself. Evidence in the record that
the bloodstream (see, generally, C-97.)
Ms. Weissinger concluded that the rat
DES is unsafe, therefore, is equally
and human studies showed that
applicable to residues of DES
diethylstilbestrol glucuronide is
conjugates.
hydrolized in the intestine to produce
The finding that the residues found, if
they consist of DES conjugates to the
free DES (G-95 at 2). Ms. Weissinger
exclusion of free DES,; would
stated her opinion that the conversion of
nevertheless hydrolyz6 in the human
the conjugate to DES in the intestine Is
catalyzed by an enzyme known as Betabody to free DES is supported by the
testimony of expert witnesses. Bureaus'
glucuronidase, which is present in
microorganisms normally found in
witness'Dr Williams stated: "I feel that
animal and human intestines (id.).
it is most probable that conjugated DES,
Manufacturing parties' witness Dr.
occurring in animal tissues, will give rise
"Kliman attached Ms. Weissinger's
to free DES-after ingestion by humans"
conclusions on several grounds. Chief
(G-102: Com
yments on the Vineland
among them is that the upper part of the
Laboratories Submission at 2).'
human small intestine does not contain
Manufacturing parties' witness Dr.
bacteriai gluc'uronidise,'whlch Dr.
Liberman-made clear his opinion that
Kliman stated is essential to the
whatever DES conjugates were found in
hydrolysis of the conjugate (M-110 at 10,
the radio-tracer studies would be
cf. Tr. at 850 (Weissinger cross-hydrolyzable by enzymes to free DES
examination)). Dr. Kliman stated that
-(TY. at 2123-24).'
' Evidence in the'record that supports
absorption takes place in the upper part
of the human small intestine (M-110 at
-these opinions includes (1) studies
18). Therefore, he seems to argue,
(discussed in the-following paragraphs)
hydrolysis of'the conjugated DES would
showing that one conjugate, DESmonoglucuronide, hydrolyzes to DES,
not take'place at a point in the digestive
tract at which absorption of the freed
(apparently in the digestive tracts) in
human and animal, bodies (G-96-98) and DES could follow. The test showed,
however, that DES metabolites
(2) the discovery of free DES, discussed'
traceable to hydrolysis of DES,above, in the radioisotope tests of DES.
(Evidence in the record shows that the
monoglucuronide did appear In the urine
of the'human volunteers (G-97 at 01,
.free DES found by Dr. Williams either
was an actual free DE5 residue or was
602). They could not have done so had
the result of hydrolysis of a conjugate of
there been no absorption.,
-Dr. Kliman also argued that the
DES. My reliance on the-Williams'sdata
studies 'referred to by Ms. Weissingor
-here assumes the latter explanation to
nfust be discounted because the subjects
be correct:The William's study may be
(both humans and rats) were fasting,
taken as showing that DES conjugates
and introduction of tile DES with food
-ate hydrolyzed to free-DES. It does not,
might affect the absorption or hydrolysis
'however, prove, that the conditions
being considered (M-110 at 17-18). In
necessary for that hydrolysis occur in
-the absence of data showing that the'
the human, body.), results of such a study would have 6een
Studies showing that a conjugated,
different under nonfasting conditions,
* form of DES, DES-monoglucuronide,.
will be transformed back to DES itself in. however, this criticism provides no
basis for discounting the results.
human consumers were introduced.by
Dr. Kliman further criticized Ms.
the Bureaus' witness Ms, Wefssinger (GWeissinger's testimony concerning the
95). These studies were done with rats
study on two human volunteers (M-110
in various stages of early development'
at 18). Dr. Kliman argued that there Is no
(G-96-97) and, in one case, with two
evidence to show whether the conjugate
human volunteers(G-97).
I In the human study,, two.men were
of DES was absorbed in the presence or,
absence of its glucuronide component
each administered simultaneously DES(id.). He then qtated that there was no
monoglucuronide.labeled with
demonstration.of conversion of the
radioactive carbon-and DES labeled

Federal Register / Vol. 44. No. 185 1 Friday. September 21. 1979 / Notices

54871

I
L

conjugate to DES in the intestinal tract
(id.). Neither of these points addresses
the issue, however, because the study
did show. according to its authors, that
DES and the conjugate of DES
administered simultaneously resulted in
the same metabolic products in the body
(G-97). The report of the study states
further.
Since the ingested glucuronide conjugate
was excreted as products other than DESG
[the DES conjugatel. it appears that conjugate
hydrolysis occurs in the body. Hydrolysis of
DESG to DES may be nearly complete, since
similar amounts ofsulfate conjugates and
polar non-hydrolyzable metabolites were
excreted in the urine after ingestion of DES
and its glucuronide conjugate'*
(G-97 at 601). Thus, there is no neea to
determine whether the glucuronide
portion of the conjugate was present
during absorption from the intestinal
tract and subsequently removed or was,
split from the DES molecule before
absorption. The material fact is that the
conjugate was hydrolyzed to DES within
the human system.
Steers have been shown to conjugate
DES to DES-monoglucuronide (as shown
by the presence, in the urine of steers
treated with 14
C-DES, of DESmonoglucuronide attributable to that
11 C-DES( (G-3 at 47-48]. This evidence
supports a finding that DES conjugates
found in edible tissues of cattle and
sheep include DES-monoglucuronide.
In any case, as discussed above in
subsection (a) of this section, however,
analysis of residues actually observed in
the radiotracer studies revealed that
those residues contain, if not free DES
itself, then DES conjugates that
hydrolyze to DES. That evidence
suggests the likelihood that whatever
conjugates do occur in animal tissues
will be hydrolyzed to DES in the human
body.
-{c) ConclusionAs to Conjugates
Issues Assuming Bureaus Have Burden
of Proof.For the reasons stated, I find
that. if the Bureaus have the burden of
showing that the residues found are
harmful, they have carried that burden.
The residues contain either free DES or
DES conjugates that would hydrolyze to
DES. Because DES conjugates hydrolyze
in the human body to free DES. the
questions raised about the safety of DES
'apply equally to the conjugates of DES.
(D)
Evidence That DES Is Not Shown To
Be Safe
(1) Relationship ofDES to
EndogenousEstrogens.-(a)The Issues.
As discussed below. DES is not a
natural estrogen. Yet, because DES has
estrogenic effects, the manufacturing
parties contend that it should be judged
as if were in fact a natural estrogen
it

(Manufacturing Parties' Exceptions at 94

ff).

97) that natural estrogens have been
shown to cause cancer. See also Tr. at
1890. 2166-67. Estrogens have, in
addition, been hassociated with other
adverse effects (see. generally. 42 FR

The manufacturing parties' theory is
that the cancer and other adverse
effects that natural estrogens cause
occur only when those estrogens exceed
37636. 37642 (July 22.1977)). The fact
the level at which they normally appear
that a dangerous substance occurs as a
in the body (id. at 105-06). They argue
component of human tissues, cells. etc..
further that the relatively small amount
(or is identical to a substance that so
of DES added to the body through the
occurs) does not of itself justify
ingestion (eating) of meat containing
approval of the addition of more of that
DES residues would not make the total
substance to the human system by
level of estrogens in the body exceed
artificial means. Cf. I.D. at 35: Bell v.
normal levels (id. at 98-102). and that for
Coddard.
supra. 366 F.2d at 182. Because
that reason DES does not present a
DES can not legitimately be equated to
human cancer risk. It thus follows, they
endogenous estrogens, I do not reach the
argue. that there is no danger in adding
difficult question of how much (if any) of
small amounts of DES to the human
a substance chemically
system (id. at 102 ffl.
indistinguishable from endogenous
An assumption essential to the
estrogen could be added to the human
manufacturing parties' theory on this
body safely.
issue is that DES is simply another
In discussing endogenous estrogens.
estrogen and that it has no carcinogenic
the manufacturing parties refer most
or other adverse effects not associated
often to estradiol. Estradiol is a steroid
with its estrogenic effects. The Bureaus
(cf. G-189 at 2) that is produced by
dispute this assumption. They argue that
animals and man and is required for
there are significant differences between
their proper functioning (cf. M-110 at 7].
DES and natural estrogens and that DES
It influences biochemical physiological
may cause cancer and other adverse
events associated with conception,
effects that would not result from
birth, growth and development, and the
natural estrogens at comparable
proper functioning of adult individuals
"of the different species of mammals. The
dosages (Bureaus' Brief at 120 ff3.
Manufacturing parties' witnesses.
chemical structure of beta-estradiol (the
seem to assume at the outset the
most common form of estradiol) is as
proposition that they wish to support.
follows:
- i.e., that DES, which is not an
endogenous estrogen, must be
considered to be no different from an
endogenous estrogen unless proven
otherwise. They conclude, in effect, that
because it has not been shown that all
the adverse effects of DES are not
associated with its estrogenic activity, it
must be concluded that an association
between DES estrogenicity and all of its
adverse effects exists (see M-69 at 6
("no compelling evidence" that tumorenhancing properties not linked to
White et al.. Principlesof Biochemistiy
estrogenic activity: M-110 at 6; M-62 at
(5th Ed., 1973) at 1062.
5). Bureaus' witnesses, on the other
DES is a stilbene (G-189 at 2: Tr. at
hand, expressed the opinion that the
228). It is not produced by any species of
lack of evidence that the adverse effects* animals, mammalian or otherwise, and
of DES are associated with its
is not required for the proper functioning
estrogenic aclivity prevents acceptance
of living organisms. It is produced
of that conclusion (see. e.g.. G-80 at 8
synthetically. DES does, however, cause
Tr. at 164; G-90 at 6]. Particularly in light
in mammals an array of physiological
of the demonstrated differences
and toxicological effects that are
between DES and endogenous estrogens
remarkably similar to the effecfs
and the theoretically different ways in
produced by endogenous estrogens such
which the body deals with these
as estradiol (and its metabolites, estriol
substances (discussed below), I
and estone). DFS has the following
conclude that the record shows that DES
chemical structure (G-47 at 419):
cannot be considered as simply another
estrogen.
CIHS
Even were DES "just another
estrogen," it is by no means clear that it
OH
HO _Q&C = C -- Q
would be judged safe on that ground.
The manufacturing parties agree
Ial
(Manufacturing Parties' Fxceptions at

I

548172

5

Federal Register

• It

/ Vol. 44, No. 185 / Friday, September 21, 1979 / Notices

I

The manufacturing parties,. while not
disputing the validity of this rendition of
the structurej proffer the following,
which they apparently believe looks
more like the structure of estrone given
by the Adininistrative Law.Judge (I.D. at
37 n23):

>CH.
J

CH 2
H:3C
Manufacturing Parties Exemptions at
114, citing to Heftman &Mosettig, " .
Biochemistry of Steroids (1960) at 167.
(b) Differences Between DES and
NaturalEstrogens,All parties agree that
there are significant similarities
between DES and endogenous
estrqgens. The hearing record
establishes, however, that there are also
incontrovertible differencEs in the
chemical properties and in the
biochemical and physiological effects of
DES on the one hand, and estradiol (and
other endogenous estrogensi on the
other. For the reasons stated in the
following discussion of these
differences, I find, as did the
Administrative Law Judge, that the
observed differences bear on the
toxicological significance of trace
amounts of DES in meat from food,
animals.
(i) Chemicalandfliochemical
"
Differences. The Bureau's witnesses
pointed to two areas in which the
structural differences between DES and
endogenous estrogens may lead to
differences in effects. Each deals with
the fate of DES and endogenous
estrogens (specifically estradiol).within
the body and raises unanswered "
questions about the claiimed equivalence
between DES and estradioLFirst, Bureaus' witnesses testified thuir
there are differences in the W'ay that the
two substances bind to macromolecules
in the body. These macromolecules,
plasma proteins, attach themselves to
smaller chemical molecules, such as
those of estradiol and DES [G-191 at 2.
Once bound, the-molecules are hindered'by the size of the macromolecule from
leaving the circulation and reaching a
target organ [id.) and, once there,
entering the cell itself to do damage (Tr.
at 73-74).

I

Although both estradiol and DES bind 'hydroxyphenyl)2. 4-hexadiene), onlega
to the macromolecule albumin, estradiol, hydroxy dienestrol (3,4 bis (p.
but not DES, binds to the much stronger
hydroxyphenyl)2-4-hexadene1-oIJ (Cbinder, testosterone-estradiolbinding
189 at 2-3: G-187 at 443) and ouregaglobulin (TeBG) (G-191 at 2). There is
hydroxy DES (G-187 at 443; cf, C-109 at
less TeBG than albumin in the body but
3). Other substances, such asparaTeBG binds so much more strongly to
hydroxy-prbpiophenone, have been
es~radiol that its failure to bind DES
lentatively identified as metabolites (Gmust be considered'significant. This is
189 ht 3]. Bureaus' witness Dr. Helton
particularly the case because all active
testified that dienestrol and onmegaPestrogens cause an increase in TeBG,
hydroxy dienestrol are neither known
i.e.. the body protects itself from natural
nor expected to be metabolic products
6s'trogens in a manner not available to'
of any endogenous estrogen (G-189 at 3counteract DES (id. at 3; G-90 at 6).
4). No known metabolites of enlogenous
Bureaus' witnesses point out that if
estrogens are-similar to these
significantly less DES than estradiol is
substanobs in terms of structure or
prevented from reaching target cells.
anticipated reactivity (cf. id.) This
DES would be more dangerous than
record does not provide a basis for
estradiol even if both had identical
determining whether the metabolic
effects on the cell once they reached it
products unique to DES are the causes
(G-19"1 at3; G-159 at 7).
of some or all of the toxicity and
It is noteworthy that this difference in carcinogenicity associated with DVS (cfr
binding resembles the effects observed
M-203 at 5]. I cannot discount the
in rats, though there it is alphapossibility that DES's metabolites exert
fetoprotein rather than TeBG that-"
effects that would not be associated
cauqes the differential (G-159 alj 2-7).
with estrogens and their metabolites.
Human alpha-fetoprotein binds well toAs the Administrative Law Judge
neither estradiol nor DES (Tr. at 2309;
noted, there is some evidence in the
M-203 at,5). Nevertheless, the analogy
recoid that DES binds covalently to
between rat experience Vvith alphaDNA (G-64 at 644) and is capable of
fetoprotein and human experience with
damaging DNA (id. at 646]. See also GTeBG, postulated by Dr. Sheehan (G-159 59 at 6. According to the manufacturing
at 7J, supports the question raised about
parties' own Dr. Jensen, such reactions
differences in the human body's
are typical of chemical carcinogens
reactionis to DES and estradiol.
foreign to ihe body or radiation, but are
The manufacturing parties' Dr. Jensen- not typical of estrogenic hormones (Mexplained in proffered surrebuttal
69 at 6-7; see also Tr. at 2198, cf. G-59 at
testimony his reasons for rejecting this
6). Thus, the fact that DES and/or Its
theory. He-stated that estradiol binding
metabolites is capable of binding with
to TeBG is freely reversible, that
and damaging DNA is some evidence
albumin binds most estradiol, and that,
that DE may cause its carcinogenic
even in pregnancy. TeBG binds only a
effects (and other adverse effects such •
relatively small fraction of the estradiol
as teratogenicity and mutagenicity) by a
available (M-203 at 1-4). I explain in
mechanism that would not be expected
Part VI of this Decision dealing with
of endogenous estrogens.
evidentiaro questions my reasons for
In their exceptions, the manufacturing
agreeing with the Administrative Law
parties attack the study that shows DES
Judge that Dr. Jehsen's "surrebuttal"'
reactions with DNA. They argue that. of
testimony was n6t proper surrebuttal,
the two tests reported, one presented at
and shduld not have been admitted. I
artificial environment and the other
have, nevertheless, considered his
produced only a relatively small effect
comments.(Manufacturing Parties' Exceptions at
The record does not contain
122-23). The study that they contend
-quantitative analysis of available data
to support or reject either the theory that involved an artificial environment does
show that appropriately activated DES
there are differences in the way DES
can reactwith DNA to modify it (0-A
and endogenous estrogens bind to
at 644). The second study shows that
macromolecules in the human body or
Dr. Jensdft criticism of that theory. This this reaction does occur to some extent
under more natural circumstances (id, at
potential difference between DES and
646). These two studies do not provide
estradiol, however, does raise an
unambiguous evideice that DES does
important question about the claim that
indeed bind to and modify DNA. Yet the
the two substances are identical in their
production by DES of reactions not
effecti.
A second, less theoretical, area in
expected to result from natural
estrogens,'like the production of
which DESand estradiol are different is
in the iiietabolites they produce. DES " metabolites not associated with natural
has been shown to yield, among other
estrogens,,raises yet another unresolved
substances, dienestrol (3,4 bis fpquestion about the manufacturing

Federal Register / Vol. 44, No. 185 / Friday, September 21. 1979 / Notices
parties' assumption that DES is no
different in its effects from endogenous
estrogens.
Differences. The
(ii) Physiological
record establishes' differences in the
physiological (in this case, hormonal)
effects of DES and those of estradiol.
They are differences in the degree rather
than the nature of the observed effects.
For instance, the record shows the
following: (1) Via the oral route, DES has
about 10 times the estrogenic potency of
estradiol (or of its metabolites estriol
and estrone) (Tr. at 1784-5; cf. M-51 at
21, Table 3; cf. M-118 at 672 (20 times
more effective in spayed mice)).
(Estrogens cause cell proliferation and
thus observable changes in the walls of
the vagina. The potency of an estrogen
is measured by, among other means, the
extent of these changes.) (2)
Intravenously administered estradiol is
a more potent estrogen than DES
administered via the same route in some
species but not in others (M-110 at 9; see
also M-115). (DES may be more potent
relative to estradiol via the oral route
than the intravenous route because by
the oral route it is not oxidized (and thus
neutralized] in the liver as estradiol is
'(cf. M-69 at 3).) (3) DES produces
smaller changes in the vaginal mitotic
index (changes in the rate of the
multiplication of cells in the skin of the
vagina) than does estradiol (M-40 at 4).
The differences in physiological
effects between estradiol and DES
shown by the record are of degree and
not of nature. Endogenous estrogens
may themselves differ in the strength of
their physiological effects. Thus, the
differences in physiological effects
between DES and estradiol noted above
would not be sufficient to reject the
proposition that DES is no different from
other-estrogens.
Two points should be made about
these data, however. First, the
information in the record on the
derivation of the comparisons noted
above (see M-118) shows that they are
based on effects observed at relatively
high levels of DES and estradiol. These
comparisons thus provide little usable
information about the physiological
.
-. effects, if any, of relatively small
residues of DES in the edible products of
animals treated with DES. Second,
because of the differences in
biochemical effects between estradiol
and DES, I must reject the argument that
these physiological effects of DES are
necessarily related to its carcinogenic
and other adverse effects.
I thus find that a comparison of the
physiological effects of DES with those
of estradiol (or other endogenous
estrogens] neither supports nor detracts
from the manufacturing parties'

assumption that DES is equivalent to
endogenous estrogens.
(c) ConclusionAs to Relationshipof
DES to Endogenous Estrogeni.In
summary, the manufacturing parties
have failed to demonstrate that DES is
identical to estradiol (or any other
endogenous estrogen) either in chemical
structure or in biochemical or
physiological (or toxicological) effects
(cf. Tr. at 164-65; Tr. at 228-29). As Dr.
Rosner stated, "There are differences
[between DES and estradiol or other
estrogens]. This is not the same
compound" (Tr. at 2282; see also G-80 at
8; G-90 at 6). There are simply too many
variables (and too many unknowns)
inherent in the metabolic process and
the processes leading to physiologic and
toxicologic effects to conclude that DES
is safe upon. the basis of similarities to
endogenous estrogens. In particular, the
manufacturing parties have failed to
establish that because the small
amounts of DES introduced to the
human body through residues in meat do
not increase the body's level of
estrogens DES presents no human
cancer risk. On this record, I have no
basis for concluding that the
6arcinogenicity of DES results entirely
from its estrogenic activity.
Animal
(2) CancerData.-a)
Data. DES is a
Carcinogenicity
carcinogen (0-22; G-34 at 1; G-37 at 2;
G-46 at 2; G-47; G-59 at 2; G-70 at 2; G80 at 7-8; G-84; G-85 at 6). This fact was
stated unequivocally by one of the
manufacturing parties' witnesses in a
1974 article that is part of this record
(M-101 at 1920). This fact is also implicit
in the analysis by the manufacturing
parties of the results of the animal
carcinogencity study conducted by Gass
et al. (discussed below). (The
manufacturing parties argue that. in that
study, a carcinogenic response is
observable in mice receiving 50 ppb DES
and that that response increases with
increasing dosage.) See also section I
above.
Although the Bureaus submitted
testimony to the effect that DES is a
carcinogen in a variety of animals and
NCI and IARC summaries of the studies
showing that fact (G-47 and G-84). the
only reports of animal carcinogenicity
studies included in the record are the
report of the Gass study and incomplete
reports of an NCTR study.
(i) The Goss Study (a) BacAground.
The Gass study, entitled "Carcinogenic
Dose-Response Curve to Oral
Diethylstilbestrol" (G-22), appeared in
the Journalof the NationalCancer
Institute in December of 1964. In this
animal test, C3H female. C311 male and
Strain A castrate male mice were
divided into test groups that were given
A

54873

feed containing DES at the following
levels: 0 ppb, 6.25 ppb, 12.5 ppb. 25 ppb,
So ppb, 100 ppb. 500 ppb, and 1000 ppb.
The test groups ranged from 50 to 78
mice. The three control groups ranged
from 115 to 136 mice. The experiment
was terminated after 85 weeks when the
then surviving animals were destroyed
in a fire.
A statistically significant incidence of
mammary carcinoma was observed in
the group of C3H female mice receiving
the lowest dosage (6.25 ppb) of DES
administered. The groups of C3H female
mice receiving 12.5 ppb and 25 ppb did
not show a statistically significant
increase in tumors over controls. (Both
of these treated groups showed tumors
in 43.3 percent of the mice as opposed to
33 percent in the controls and 48.2
percent in the 6.25 ppb group.) There is
no question that the C3H female mice
fed 50, 500 and 1000 ppb DES developed
mammary gland cancer and that the
evidence of cancer in the treated groups
increased with increasing levels of
exposure.
The test groups of C3H male and

Strain A castrate male mice were less
sensitive. In each, some tumors
developed in animals fed 12.5 ppb but
statistical significance was not clearly
apparent below the higher levels of
exposure.
(b)ManufacturingParties'
Contentions.The manufacturing parties
agree that this study (1) does not show
that low levels of DES cause cancer and
(2) does show that low levels of DES do
not cause cancer, i.e., that there is a noeffect level (Manufacturing Parties'
Exceptions at 126-27).

The first argument appears to assume
that, if the only evidence that DES is
carcinogenic was seen-at dosages
substantially above the levels of DES
observed as residues, the FDA could not
find that the levels observed as residues
are unsafe or not shown to be safe. As
discussed in the introduction to this
Decision. however, the FDA must of
necessity rely on tests showing effects
of relatively high levels of a substance
in test animals as a basis for the
decision that lower levels of that
substance present a carcinogenic risk to
man. I have previously explained (in
section II1(D)(1) above) my reasons for
rejecting the manufacturing parties'
theory that the carcinogenicity of DES is
related solely to its estrogenic activity.
(If that theory were accepted.
extrapolation from results of the
ingestion of relatively high levels of DES
in animals to predict the results of
ingestion of lower levels of DES in
humans might, of course, not be
appropriate.]

54874

Federal Register / Vol. 44, No. 185

In light of my rejection of the

"carcinogenicity is a function of
estrogenic activity" theory of the
manufacturing parties, their second
contention, that the animal studies show
a no-effect level for DES, must also be
rejected. Routine bioassays are not
capable of establishing a no-effect level
for a carcinogen. This proposition is
well-supported by the opinions of noted
cancer experts who testified at.the
hearing (G-46 at 8 (Dr. Hertz); Tr. at 172
(Dr. Saffiotti); Tr. at 1128 (Dr.
Schneiderman); Tr. a 283 (Dr:-Shimrkin);
cf.Tr. at 1176 (Dr. Herbst)). (The
conflicting testimony of some
manufacturing parties' witnesses is
discussed below.) Thus, I can not find
that the studies discussed in this section
showed a no-effect level for DE$'s carcinogenic-effect. This conclusion.
would stand even if the Tesults of testing
of DES at low levels were
unambiguously negative. In fact,
although the relative lack of sensitivity
of the Gass study (G-22) makes
:
interpretation of its results at low dose
levels difficult, an apparent carcinogenic
result was, as noted above, reported in "
that study at the lowest level tested
(6.25 ppb). '
Witnesses presented by the
manufacturing pairties supported those
parties' contentions concerning the Gass
study as follows: [1) Some witnesses
gave their opinion that the lowest level
of DES that cause a carcinogenic effect
in the Gass study was a level (estimates
varied'as to what that level should be)
above the lowest level of 6.25 ppb. (See,
e.g., M-110 at 5; M-63.) (2) One witness
testified tharthe results observed at the
three lowest dosage levels of-this study
sijould be discarded because of the
confounding effects of the fire that
terminated the experiment (Tr. at1948-51, 1969-70). (3) One witness ,testified
that no valid statistical conclusions
could be drawn from the study (M-139
at 8). My discussion of and evaluation of
this testimony follows.
Neither the Bureaus nor the
manufacturing parties called Dr. Gass as
a witness. The manufacturing parties
introduced an article authored by Gass
and published in the Food, Drug and
Cosmetic Law Journal (not a refereed
scientific journal) in February oF1975.
That article attacks the Delaney Clause.
It comments upon Dr. Gass' own study
as follows: "The lowest dose of DES that
produces mammary cancer in the most
susceptible animal species--the C3H
mouse-required a minimum of 6.25
ppb--and probably four times that
amount" (M-13 at 112). Elsewhere in the
article Dr,Gass'referred to the
requirement of "at least" 6.25 ppb DES

1

Friday, September 21, 1979'/ Notiqes "

in a mouse diet to cause a carcinogenic
effect and referred to the "probable
carcinogenic doselevel" bf 25 ppb in the
C3H mouse strain (id.)'
Another manufacturing parties'
exhibit (M-178) is a memorandum of
conference between a Mr. Thomas
Tomizawa and a Dr. R. L. Gillespie of
the Bureau of Foods' Division of
Toxicology. Dr. Gillespie, who
apliarently authored but did not sign the
memorandum (dated Marbh 23,1976),
quoles himself as having told Tomizawa
"that currently Dr. Gass believed that
6.25 figure to be a biological fluke and
,that he believed the probability was that
the true figure Was somewhere between
25 and 50 ppb" (id.). The memorandum
does not explain how Gillespie would
know What Gass' then current beliefs
were, and Dr.,Gillespie was not called
as a witnesi. Therefore the statement in
the memorandum cannot be relied on.
No explanation is given by anrone as
to why Qr. Gass was not called as a
witness. Becuse the record reveals
neither Dr. Gass' current views nor the
basis for those views, and anyone
disagreeing with them has not been
given a chance to cross-examine him. I
-have accorded statements of his
opinions less weight than those of
witnesses who testified at the hearing. I
cannot accept, without explanation, his
apparent conclusion that some of the
reported results of his study shoUld be

amounts required to produce a
physiological response" (M-'110 at 5).
His only citation for this proposition
was an article whose authors Included
Dr. Gass. This article contains basically
that statement but provides no specific
support for it. The article does state:
"We should like to emphasize, however,
that to the best of our knowledge, the
relationship between the minimal
physiological and minimal tumorigenic
doses has not been determined for any
of the estrogens" (M-64 at 23). (This
article also contradicts the
manufacturing parties' position on
another point. In discussing the Gass
study, it states: "As no levels bolow 6.25
ppb were fed, this study does not
provide convincing evidence of a
noncarcinogenic level in the C31-f
females," M-64 at 21.)
'

As discussed aboverl have found thai

there is no basis for concluding that
there is a direct relationship between
the carcihogenicity of DES and its
estrogenic effects. Thus, Dr. Kliman's
exclusion of the results at 6.25 at 12.5
from his calculations makes his
conclusions invalid.
The lead author of M-64, Dr. 1-1.
11.
Cole, also testified for the manufacturing
parties. Dr. Cole stated that
physiological effects in the Gass study,
i.e., ovarian weight depression, were
noted at or about 13 ppb (M-42 at 3). (It
is unclear where he got this figure.) IIa
disreghrded. ,
stated that 13 ppb would thus be th
Manufacturng parties' witness Dr.
minimum level of DES required t6 cause
Bernard Kliman explained his reasons
a carcinogenic response (id.). although
for believing that the Gass study show
during cross-examination [Tr, at 1640)
that DES does not cause a carcinogenic
Dr. Cole admitted that at lower dosages
-effect at low levels (M-110 at 5),
there may have beeii physiological
effects other than ovarian weight
The log dose-response curve was linear
depression that went unnoticed. Dr. cole
only between 25 and 500 ppb. My further
did not state a clear factual basis for his
analysis of this data by extrapolation of'this
linear curve to intercept with the cancer
hypothesis of a link between observed
incidence of the control animal-group
physiological effects and carcinogenesis,
indicates no effect of DES on tumor incidence I cannot, therefore, accept that
at or below 12.5 ppb.
hypothesis.
Dr. Kliman disregarded the data
Dr. Cole cited a paper by Jones and
points at the 6.25 and 12.5 ppb levels
Grendon (M43) for the'proposition that
when fitting the probit-log dose line, and the-Gass study showed that the
then noted that the observed responses
minimum carcinogenic level for DES is
at these two lower levels did not fall
greater than 27 ppb, A review of M-63
within the 95 percent confidence bounds "reveals n such conclusion. The authors
of his extrapolated probit-og dose line
of M-63 do state that Gass reported that
(Ti at 1832). It is not, of course, proper
"DES induces mammary cancer in mice
to exclude data from statistical analysis
,only at levels causing physiological
without evidence'that those data are
distrubances, not lower levels," (id, at
invalid.
264). M-63 then refers to tables in the
-Dr. Kliman. in dismissing the results at Gass study without commenting upon
6.25 ppb and 12.5 ppb relied upon the
the finding of a statistically significant
fact that in the Gass study the lowest
effect at 6.25 ppb in the female test
feeding concentration at which the
animals.
weight of the.ovaries was found to have
Dr. Hardin B.Jonestestified for the
decreased was 25 ppb.'He stated: "It is
manufacturing parties (M-97). During
reasonable to conclude that estrogens
cross-examination, he stated a new
are associated with carcinogenesis only
theory to explain the finding of'a
when given in amounts greater than the
statistically significant carcinogenic
-

Federal Register

I

effect in the 6.25 ppb group in the Gass
study [Tr. at 1948-51,1969-70). Because
this testimony was introduced only on
cross-examination, the Bureaus were
denied a chance to prepare detailed
cross-examination of it. I have, however,
considered Dr. Jones' theory on its
merits.
Dr. lones relies, in this theory, upon
the fact that the Gass study was
terminated when a laboratory fire
destroyed the remaining test animals
(G-22 at 973). The study called for
sacrifice of any animal in which a
palpable, one centimeter, subcutaneous
mass was found. After sacrifice, the
mass was examined histologically.
Those masses diagnosed as "mammary
carcinoma" were designated as tumors
in the results (id.-at 972). Those animals
destroyed in the fire were, of course, not
examined for tumors. The Gass results
consider these latter animals as having
no tumors.

Dr. Jones argues that one should
exclude from analysis all animals lost in
the fire. Having done that, he finds that
the results in the 6.25,12.25, and 25 ppb
groups are not different at a statistically
significant level from each other.
I This lack of statistical significance.
how4ever, could be due to the reduction
in group numbers and the consequent
reduction in statistical power to detect
differences. Moreover, the results of Dr.
Jones' analysis are, in any case.
dependent upon the number of animals
per group that exhibit non-cancerous
sdbcutaneous masses. If a group had a
relatively small number of animals with

such masses, then the percentage of
animals with mammary carcinoma

would increase, and vice versa. (The
report of this study does not provide
information about how many, if any.
-

mice died of natural causes before the
fire.) Because it is not clear that
noncancerous subcutaneous masses
were a controlled variable in these

groups (and no adjustments can be
made for this fact), it is not appropriate
to utilize the method that Dr. Jones has
suggested to analyze the results of this
test. if it were, as Dr. Jones suggested.

improper to count all of the animals
destroyed in the fire as not having
tumors, then I probably would be best
advised to disregard this study altogether. The weight of th6 expert
evidence, however, including testimony
for both sides in this hearing, suggests
that the test results can be relied upon
when properly analyzed. {See, e.g.. M10; M-62; C-21: G-25.)
Dr. Thomas Jukes testified that the
Gass study showed a dose-response
relationship starting at 25 ppb and that
this relationship" with an absence of
significantly larger numbers of tumors

Vol. 44, No. 185 / Friday, September 21. 1979 / Notices
above controls below this level" showed
a threshold [M-99 at 4). this comment, of
course, ignores the result observed in
the 6.25 ppb group. Dr. Jukes then stated
that any reliance upon the results
observed in the 6.25 ppb group
separately from the results observed in
the groups fed 12.5 and 25 ppb DES
"defies biological common sense" (id. at
5). The Bureaus do not, however, ignore
the 12.5 ppb and 25 ppb results (see
discussion below). Relyng on any of
"thesethree results "separately" would.
of course, be improper.
Dr. Jukes also stated that the
"threshold" for tumor induction of DES
in C3H mice "extends at least as far as
12.5 ppb and perhaps to 25 ppb" (id. at
6). This conclusion is based upon his
report that the NCTR study, discussed
below, showed fewer tumors in mice fed
10 ppb than in control mice. I explain
below my reasons for not relying on
preliminary reports of the NCTR data.
Another, and more persuasive, analysis
of the combined low dose results from
the Gass and NCTR studies would be.
however, that these studies are not
sufficiently sensitive to show clearly
any effect that might be associatedwith
very low dosages. This interpretation is
the conservative one and Iadopt it.
Therefore, these data do net provide a
basis for the conclusion that a threshold
has been shown for DES.
The manufacturing parties suggest
that, because C3H female mice are
highly susceptible to mammary tumor.
(in part because of the presence of a
mammary tumor virus in that strain of
mice), the results of test with this kind of
mouse are not properly appjicable to
man (Manufacturing Parties' Exceptions
at 135-138). The particulal' sensitivity of
these mice. however. only makes tests
with them more sensitive indicators of
the carcinogenic effect of a substance
such as DES. I cannot find that this
enhanced sensitivity is reason for
discarding test results achieved in
female C3H mice.
The manufacturing parties also
contend that this animal test is not
equivalent to human exposure because
in the animal tests the feed containing
DES constituted the entire diet of the
mice and that mice consume more food
per unit of body than humans do
(Manufacturing Parties' Exceptions at
137-38). These factors only mak-e this
test more sensitive to carcinogenic
reactions. For the reasons discussed in
the introduction to this Decision (section
I(D)). it is necessary to use the most
sensitive animal test system available in
seeking information about the potential
carcinogenic effects of substances such
as DES.

54875

The manufacturing parties! statistical
expert. Dr. C. R. Weaver, raised
questions about whether the
environmental effects and the diet
effects were completely separated in the
Gass study (M-139 aL--10). It is true
that, if thereexists "confounding" of
effects, it is nearly impossible to
distinguish statistically between them.
Dr. Weaver's concern is that in the Gass
study all the cages of animals receiving
a particular diet may have been together
(but separated from the cages of animals
receiving other diets). and that therefore
the different diet groups Were subject to
different environmental conditions (M139 at 9). Dr. Weaver relied upon
secondhand hearsay for some of his
assertions (Tr. at 1518). 1 have et aluated
his statements in that light and do not
consider his testimony a proper basis for
a finding that the Gass study did not
have a satisfactory experimental design
to avoid the confounding of the effects
observed.
Dr. Weaver stated that all
interpretations of the Gass study should
be disregarded until further evidence is
available (M-139 at 8):
In view of the inadequate nature of the
Cars data. the anomalous results obtained.
and the suspect nature of the data at the
lower end of the dose range. it is my opLuion
that statistical conclusions cannot propely
le drawn form this study.' * '
Dr. Weaver's position. if accepted.
would mean that the Gass study could
not be used to establish a no-effect level
for DES. He thus directly contradicts the
testimony previously discussed.
(c) Bureaus' Contentions.The
Bureaus' contentions with respect to the
Gass study are straightforward. They
argue that the study shows (1] that DES
causes cancer in test animals and (2)
that 625 ppb DES caused cancer in mice
in that study (Bureaus' Brief at 30,41).
As discussed above, even some
manufacturing parties' witnesses based
their testimony on the conclusion that
the higher levels of DES fed in this study
produced cancer (see. e.g., M-110 at 5).
That proposition is not fairly open to
dispute. and I agree with the Bureaus
that DES at least at the 50.100. 500. and
1000 ppb levels was shown to cause
cancer in animals in the Gass study.
Testimony in support of the Bureaus'
second argument emphasizes that the
6.25 ppb result in logically consistent
with the results observed at 12.5 and 25
ppb and. in turn. consistent with the
hypothesis that any amount of DES
would cause some carcinogenic effect.
Dr. Robert]. Condon testified that he
had investigated whether or not the
probit-log dose model for the incidence
rate of mammary cancer among the