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)

54876
I

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

----

, i

i

three sets of mice in the Gass study is
discussed above focused on the question (NCTR) has been performing relatively
appropriate (G-21). In order to make this whether the effect observed with the
large scale carcinogenicity studies with
determination, he carried out a Chi- ,
6.25 ppb group in the Gass study was
DES. (Apparently manufacturing party
squared test on each of the three datareal. Preoccupation with the 6.25 ppb
Hess &Clark is also doing, or has
sets (i.e., the results observed with each 'result threatens, however, to obscure the completed, an animal DES study, whose
of the groups of mice tested). The Chireally important point about that study.
results it has not revealed in the record
squared statistic is based on the sqdared No one, not even among the
(Tr. at 1460, 1469).) Neither the
distances of the observed probit values
manufacturing parties' witnesses,
manufacturing parties nor the Bureaus
from the fitted probit-log dose regression disputed that this study showed that
were able to introduce evidence as to
line. The calculated value is small if the
DES causes mammary cancer in mice in
the final results of the NCTR studies.
observed probits do not deviate greatly
doses at 50 ppb and above. In fact,
Each side, however, had witnesses
from the fitted regression line; an
several manufacturing parties' witnesses testify about preliminary results that
agreed that there is a dose response
absence of large deviation indicates
seemed to be favorable to its position.
relationship observable above that level.
The Bureaus introduced the testimony
with a high probability that the probitIf a substance causes cancer at the
of Dr. Benjamin Highman of NCTR (Glog dose model is adequate.
higher dosages in an animal assay and
Unlike the manufacturing parties'
54). Dr. Highman testified that he had
does not cause cancer at lower dosages, examined tissue slide preparations of
witnesses, who disregard the 6.25 and
a scientifically sound interpretation of
mice from one of the ongoing NCTR
12.5 ppb dosage levels in their
those resdlts is that the test was-not
experiments (id. at 2). He stated that he
computations, Dr. Condon used the data
sensitive enough to detect the lower
from all the dosage levels. He
found DES-related adenocarcinomas of
response that would be expected at
the cervix and endometrium in test
commented on his Chi-squared
lower dosages. Another conceivable.
animals and did not find any such
calculations as follows. (G-21 at 2):
interpretation of such results is, of
tumQrs in the control mice (id.). The
The observed values do not differ
number of such effects he had found as
significantly from those depicted by the fitted course, that the dosages that did not.
cause an observed effect are not
of the-date'of his testimony (March 22,
dose response curve at any of the'doses used
carcinogenic. Nothing in this record
1977) was not large enough to be
in this experiment. This means that none of
convinces me that the latter
the observed response values (mammary
-characterized as statistically significant'
interpretation is the correct one and I
tumor incidence) should be dismissed as
(id.).
aberrant values.
cannot presume that it is.
Dr. Highman's testimony was updated
I therefore do not rely upon a showing at the time of cross-examination (May
If the probit-log dose model is accepted
that 6.25 ppb DES did cause cancer in
16, 1977) to include findings of
as correct, then there"is no threshold
the Gass study. Rather, I rely upon the
additional tumors since the time when
level because this model presupposes
fact, discussed above in my evaluation
the direct examination was submitted
that every dosage level induces a
of the manufacturing parties'
(Tr. at 109-117). The additional
response.
contentions, that routine animal
information did not make the figures
Dr. Jerome Cornfield and Dr. Adrian
carcinogenicity tests currently cannot
statistically significant (id. at 136). Dr.
Gross in effect'incorporated in their
show a no-effect level for a carcinogen.
Highman noted, however, that the
testimony (G-25 at 2; G-34 at 1) a 1971
(The support in the record for this
adenocarcinomas are extremely rare
memorandum from Anne Alderman to
and he-stated that the rarity itself made
Dr. R. L. Gillespie (G-23). (Ms. Alderman proposition has been cited in my
-previous discussion of it.)them significant from a pathological
did not testify.] That memorandum also
, It noteworthy that few carcinogens
standpoint (G-54 at 2).
noted that the probit-log dose curve over haveisbeen shown to cause cancer in
A manufacturing parties' witness, Dr.
the entire range of doses used could be.
animal studfes at levels as low as 50
Jukes, testified that the NCTR had just
observed in the C3H female mice in the
-ppb, the level at which the Gass study
completed (as of September 12, 1977) a a
Gass study. The interpretation of the
unambiguously shows DES to cause
confirmatory experiment in which C3H
Gass study data advanced by Drs.
cancer. Yet, the agency has not taken
mice received DES. He stated that the
Condon, Cornfield and Gross is at least
the position that no-effect levels have
mice receiving 10 ppb of DES had a
as persuasive as the manufacturing
been established for carcinogens that
parties witnesses' conclusion (discussed not show effects at levels that low. do lower incidence of tumors than the
control mice, From this information he
above) that the results with the 6.25 ppb
The manufacturing parties, as noted,
drew the conclusion that the 6.25 ppb
group are inconsistent with a dose.
have *argued that DES is different from
result in the Gass study represented
The Alderman memorandum also
'other carcinogens because the
insignificant fluctuation above the
contains the following observation (G,
carcinogenic effects of DES occur only
control value (M-99 at 5). This testimony
23):
at levels at which-it causes physiological was first stricken by the Administrative
When the three lowest dosage groups (6.25, effects (such as ovarian weight
Law Judge and then reinstated (Tr. at
12.5 and 25 ppb) are combined, they show a
depression (see M-62 at 3)) associated
2141).
significantly (P<.025) higherincidence than
with its estrogenic activity. I have
Dr. Jukes seems to have. admitted at
the control group, indicating that there is
discussed in section III(D)(1) above my
the time of cross-examination
evidence of an effect somewhere in this
reasons for rejecting that theory. I must
(November 4, 1977), that his statement
range.
note that in any case the Gass study
that the test had been just completed
Because the Gass study would not be
does not show, as the manufacturing
was not entirely accurate, or at least did
expected to be sufficiently sensitive to*
parties dontend, that there is a no-effect
not mean that the histology and analysis
produce interpretable ktesults at levels in level for DES's estrogenic properties. An had been completed (Tr. at 2140). During
the 6.25 to 25 ppb range, I do not rely
equally plausible interpretation of the
cross-examination, Dr. Jukes also agreed
upon the argument by the Bureaus that
data from that study is that the study
that his statement was referring only to
this study shows DES to be a carcinogen was not sensitive enough to detect
mammary tumors and not to all tumors
at such low levels.
_estrogenic effects below 25 ppb.
in the tes't animals (Tr. at 2206).
(d) ConclusionAs to GassStudy. The
(ii) NCTR Studies. FDA's National'
The question of how to deal with
testimony of the Bureaus' witnesses
Center for Tokicological Research
ongoing studies in an administrative

t

Federal Register / Vol. 44, No. 18 5 I Friday, September 21, 1979
hearing is a difficult one, Until a test is
completed and properly analyzed,
reports of its results can be misleading.
The FDA occasionally has to rely on
preliminary analyses of test results in
regulatory decisionmaking. It does so
reluctantly and only in circumstances in
which it has obtained all the information
available about the study in question.
The restrictions inherent in a regulatory
hearing make full knowledge about the
NCTR study impossible.
I have concluded that I should not rely
upon the preliminary reports of results
of the NCTR study. I have, however,
considered whether my findings would
change in any way if I were to accept as
valid Dr. Jukes' report that the group of
mice receiving 10 ppb DES in the NCTR
study had a lower incidence of
mammary tumors than control mice. For
the reasons that follow, my findings
would not change.
Dr. Jukes did not report that DES did
not cause cancer in mice treated with
higher levels of DES or even that those
results did not show a dose-response
relationship. As I have discussed, I must
presume that the 10 ppb result, if
reported correctly, is attributable to the
insensitivity of the test system. That
result alone, or together with the Gass
data, would form no basis for
determining that a no-effect level for
DES's carcinogenicity had been
identified.
(iii) Conclusion As to Animal
CarcinogenicityData. I find that DES is
a carcinogen and that the results of the
Gassstudy do not demonstrate a noeffect level for the carcinogenicity of
DES. The NCTR data are not complete
and cannot be relied upon, The results
of the NCTR study reported by Dr. Jukes
would not, at any rate, justify a finding
that there is a no-effect level for DES.
These findings warrant a conclusion
that DES has not been shown to be safe
and that it is unsafe.
(b) Human CancerData. It is entirely
appropriate for a regulatory agency such
as the Food and Drug Administration tq
conclude from data showing a substance
to be carcinogenic in animals that that
substance presents a caner risk to
human beings. Indeed, FDA has done so
often. See, e.g., Certified Color
ManufacturersAssociation v. Mathews,
543 F.2d 284 (D.C. Cir, 1976) (Red No. 2);
Bell v. Goddard,supra (DES as a poultry
implant]. The Bureaus have, in any case,
presented expert opinion in this case to
support the association between animal
and human cancer. See, e.g., G-85 at 3
(Dr. Marvin Schneiderman). Thus, the
evidence that DES is a carcinogen in
humans is simply corroborative of the
conceded animal carcinogenicity of
DES, unless the human data can be said

to show or to disprove the claim that
very small amounts of DES have no
effect when administered to humans.
The data presented in the hearing do
neither.
(i] Dr. Herbst'sData.Dr. Arthur-L.
Herbst discovered a link between the
use of DES by expectant mothers as a
drug to prevent miscarriage and a
variety of changes in the genital tracts of
female children born to those mothers.
Chief in importance among these
changes, which are manifested in most
cases when the daughters are teenaged
or older, is the finding of
adenocarcinoma in the daughters'
genital tracts. Dr. Herbst refers to these
cancers as "clear cell adenocarcinoma,"
a type of tumor that he regards as rare
(see, e.g., G-38).
After publication of his initial findings
of a relationship between this type of
cancer and maternal ingestion of DES,
Dr.Herbst was instrumental in setting
up and has directed a registry of clear
cell adenocarcinoma in the genital tract
of young females (G-37). Discoveries of
this type of cancer have been reported
to him and he has sought to determine
whether each such cancer is, in fact,
associated with maternal DES use. Dr.
Herbst has reported his findings in a
series of articles in medical journals {C38 through G-43; G-45- M-26).
One hundred fifty-four of the 302
cases of clear cell adenocarcinoma
reported in the most recent article were
in women whose mothers had been
treated with DES; 65 were not; 25 of the
302 reported cases were in women
whose mothers had been treated with
unidentified medication; and the history
of the remaining 58 was uncertain (M-26
at 44). About 50 out of the 302 cases in
his Registry were fatal. (G-37 at 3]. (The
Administrative Law Judge mistakenly
states that 50 percent of the cases were
fatal [I.D. at 25].)
Dr. Herbst referred to the "'now
generally-accepted relationship of these
cancers to maternally ingested DES" (G37 at 1] See also his statement that "the
association of DES with these cancers is
now an accepted fact" (id. at 2).
Dr. Herbst stated in his testimony that
he was unable to calculate a risk figure
to predict what percentage of those
exposed to DES in utero will develop
cancer (C-37 at 4]. fle staled that the
"risk rate" through age 25 may be
approximately 1 cancer per 1,000
exposures to the DES anti-abortion
treatment, a risk that he regarded as
significant (id. at 5). He declined to
predict whether the rate will increase'as
!he exposed individuals grow older {id.).
In a paper submitted by the
manufacturing parties (M-26), Dr.
Herbst and others utilized data obtained

/

Notices

5487'7

through his registry to make calculations
of the risk of cancer from maternal DES
use. These calculations are extremely
questionable: They are based upon a
ratio of the cases reported to him to the
total number of female births during the
various years inquestion. This ratio is
then adjusted by a variety of estimates
of the percentage of the total births in a
given year that involved the
administration of DES to the. mother. It
would appear obvious that the
numerator (number of cases of this kind
of cancer) would not represent all of the
cases of this cancer during the years in
question and that the denominator
(number of births involving DES
treatment of the mother) is based on
speculation. The risk figure computed
(for subjects 24 years old and youngerl
is between 0.14 and 1.4 per 1,000 (M-26
at 47]. The only possible value of these
calculations would be as an illustration
that the number of DES-related vaginal
tract cancers in proportion to the
number of females exposed in utero is
relatively small,
The Administrative Law judge's
decision summarizes the Herbst data,
but does not discuss the manufacturing
parties' attacks upon that evidence..
Those attacks are four.
First, the manufacturing parties argue
that Dr. Herbst has not shown a
carcinogenic effect caused by DES. The
manufacturing parties argue that the
effect observed is teratogenic rather
than carcinogenic (Manufacturing
Parties' Exceptions at 142-144]. Dr
Herbst himself has stated that the effect
may be teratogenic (G-41 at 17 Tr. at
1165-66). This is also the conclusion of
the manufacturing parties' witness, Dr.
Jensen (see M-69 at 12). The teratogenit
effect would be an alteration in the
vaginal tract during the growth of the
embryo that would lead occasionally to
cancer (id.). The importance of this
distinction is that a teratogenic effect
leading to cancer would not be evidence
that DES would cause cancer in those
not exposed in utero. On this record. I
have insufficient basis to determine that
the effect observed cannot be
characterized as carcinogenic. But in
any case, a teratogenic effect would be a
sufficient basis for a finding that DES is
not shown to be safe.
Second, the manufacturing parties
attempt to discredit the association
between maternal use DES and the
effects observed IManufacturing Parties'
Exceptions at 144-146). During crossexamination, Dr. Herbst was asked
whether he knew what proportion of the
mothers of affected daughters were
diabetics or were taking insulin or Were
subject to high blood pressure.-Dr.

54878
--

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

II

I

II

""

Herbst answered that he did not know
Apparently the manufacturing parties
the specific figures, but that few were
are suggesting that all the other cases of
diabetic (Tr. at 1158-59). Dr. Herbst's
DES-related cancers reported by Herbst
involved very large doses of DES. Dr.
'responses to. questions asked on crossexamination demonstrate that he did, in Kliman testified that the usual dose for
-anti-abortion therapy was 5 to 150 mg
fact, look for other potential causitivefactors for the cancers (id. at 1159, 1162). per day (M-110 at 22).
The manufacturing parties' attack on his
The argument about size and dosage
becomes'important in light of the
data on that ground is thus not
manufacturing parties' argument that
persuasive.
any cancer-causing effect of DES is
On cross-examination, Dr. Herbst
associated with its estrogenic
stated that-a high proportion of DESproperties. (See discussion of this
treated mothers had a history of
question in section III(D}(1) above.)
previous abortions or bleeding as a
Thus, they argue that the dosage of DES
complication of pregnancy (Tr. at 1159).
administered as medication would-be
Since DES was used to prevent
much greater than- the amount of
abortions, this fact would be expected
endogenous e.stradiol that humans
(cf. Tr. at 1155, 1158). Manufacturing
normally produce (M-69 at 10). The
parties' witness Dr. Kliman appeared t6
amount of DES that might be consumed
suggest that the cases of vaginal
daily through ingestion of part per
adenocarcinoma found by Herbst might
billion residues of DES in meat, on the
be associated with the saving of
other hand, would not add significantly
otherwise "high risk" pregnancies (Mto the amount of endogenous estradiol
110 at 21-22). Dr.,Kiiman criticized the
(id. at 10-11). As discussed in section
failure of Dr. Herbst to compare his
III(D)(1), however, DES'differs
findings to a control group of
ignificantly from other estrogens.
comparable individuals (id. at 22). Dr.
The fourth manufacturing parties'
Kliman argued that the best control
attack on the Herbst data involves the
group would have been sibllngs of the
charge that those data do,not
treated mothers (apparently those
demonstrate a-distinction between DES
siblings carried by the mothers during a
and natural estrogens (Manufacturing,
timewhen*they were not treated with
Parties' Exceptiois at 148-149). D .
DES (id. at 21-22)). Since, he apparently
Jensen argued that animal data show
theorized; all these children would be
that abnormalities in developing
dead (id.), by his definition such a
reproductive organs, including
comparison could not be made.
production of tumors, can be induced
Dr. Kliman's argument is based upon
with natural"steroidal estrogens as well
speculation that there is some
as with DES (M-69 at 14, 15). As
correlation between the need for DES as discussed in section III(D)(1), however,
an anti-abortion agent and the cancer
DES is in-some ways significantly
observed. He does not suggest a basis
different from natural estrogens. In view
for this theory. Thus, while Dr. Kliman . of these differences, I can not assume
has pointed to another variable'that
that natural estrogens, if used as DES
cannot be controlled in the analysis of
was used in the treatment of pregnant
the Herbst data, his arguments do not
women, would result in the
form a basis for disregarding the
abnormalities in their offspring observed
association that Dr. Herbst has
as the result of usage of DES.
observed.
The attacks on the Herbst data, and
The manufacturing parties' third
on the conclusion that these data show
criticism of Dr. Herbst's data is that is
an association between DES and
shows effects only at "extremely high
cancers in humans, are thus without
d6sages" (Manufacturing Parties'
merit.
Exceptions at 146--148). They argue that
(i) Mayo ClinicData. In a study
the Herbst data do not show a dose
supported in part by FDA and NIH, 1,719
response relationship with DES and thus children born to mothers who had used
do not show that very small doses of
DES during pregnancy at Mayo Clinic
DES cause a response. Dr. Herbst had
obstetric facilities from 1943 through.
identified one case in which the mother
1959 were followed to determine
whether any.haid developed cancers. No
had received as little as 1.5 mg of DES
per day or 135 mg during the entire
cancers of the vaginal or.(for males)
pregnancy (G-39 at 716; G-37at 4), but
urinary tract were found. The authors of
the manufacturing parties argue that this the report of this followup project
concluded that their findings did not
result is consistent with the hypothesis
that low amounts of DES do not cause
show a lack of correlation between DES
the effects Herbst observed.-They would and the vaginal tract cancer observed by Herbst. Rather, they concluded that
lump this low dose case with.cases
their work showed the association to be
reported for which there is no evidence
rare (G-44 at 797). '
that the mother was administered DES.

II

__

I'.ll

The researchers calculated the upper
limits of the risk from the use of DES as
a carcinogen that would be consistent
with their findings of no such cancer In
803 live born females. They calculated
an upper risk limit with a 95 percent
confidence level of 4 cancers per 1,000
exposed subjects (G-44 at 798). The
researchers also considered 'the
potential risk if their study were limited
to those children of mothers exposed to
DES during the first trimester of
pregnancy. (The cases observed by
Herbst had involved such exposure.)
Using this group, and adjusting for the
age of the patients at the time of
followup, the researchers calculated an
upper limit risk of 7 per 1,000 of
developing this kind of cancer by age 13
and a risk of 13 per 1,000 of developing'
the cancer by age 22 (id.). It must be
remembered that the results of this
followup study are also consistent with
a risk of zero per 1,000 for any of the
groups considered. The upper risk limit
ismerely a function of the number of
persons included in the followup group.
I accept the researchers' conclusion
about this study-that it does not show
that there is no association between
maternal DES therapy and vaginal tract
cancer in offspring but that it does show
that that association is relatively rare,"
(iii) Chicago Study. The University of
Chicago sponsored a followup.study of a
controlled efficacy trial for DES use in
pregnancy that had been conducted
during 1951 and 1952, A report of early
findings in this study was published in
January 1977 and included in the record
(G-10. A later, unpublished report of
further progress'of the study, submitted
by the researchers to, their contract
monitor on August 31,1977, was added
to the 'record later (G-192). Each report
states that no statistically significant
correlations between cancer and DES
treatment had been observed, either in
the mothers treated or in the children
exposed in utero. (There were other,
noncancerous, effects of treatment. See
section III(D)(3) of this Decision below.)
There were differences in the cancer
incidences between the DES mothers
and their control counterparts. 4.9
percent of the DES exposed women
contracted breast cancer while 3.1
percent of the control women were
similarly afflicted;'5.9 percent of the DES
exposed women had cancer in
"endocrine related sites" (breast,
endometrium, ovary, and colon), while
4.2 percent of the control women had
such cancers; 3.6 percent of the DES
exposed women had cancers at other
sites, while 2.7 percent of the control
women hadsuch cancers (C-192,
Appendix 4-15a). None of these

Federal Register / Vol. 44, No. 185 / Friday, September 21, 1979 / Notices
increases was statistically significant.
however.
The Bureaus, in their brief to the
Administrative Law Judge, argued that
G-192 illustrates that the increased risk
of cancer in the DES exposed mothers
was significant over time, i.e., that the
DES mothers contracted cancer earlier
than the control mothers (Bureaus' Brief
at 34). Though the results reported do
show that DES-treated mothers
developed breast cancer earlier than
women in the treated group (G-192,
Appendix 4-14a), there is no showing
that this effect is statistically significant.
The manufacturing parties submitted
a statement of Dr. Herbst on this issue
(M-209). In this proffered testimony,
which was not received in evidence by
the Administrative Law Judge, Dr.
Herbst stated that he was now the
principal investigator on the University
of Chicago followup study, replacing Dr.
Bibbo, who had authored G-192. Dr.
Herbst stated that he, Dr. Bibbo, and the
biostatistician involved in the research
project agreed that G-192 did not
establish that DES ingestion by mothers
during pregnancy had caused an
increased risk of breast cancer or that
the report otherwise was evidence of
carcinogenicity of DES in humans.
I concur with Dr. Herbst's analysis of
this study. The data from the Chicago
study taken alone would not be a basis
for a finding that DES is a human
carcinogen. Those data are, however,
not inconsistent with that proposition.
The'results referred to by the Bureaus
do raise questions about whether a
larger, and thus more sensitive, study
might show the effects the Bureaus
contend exist.
(iv) Conclusion as to Human Cancer
Risk I find that evidence in the record
concerning the incidence of clear cell
adenocarcinoma in daughters of mothers
treated with DES (the Herbst data)
supports the conclusion (which may also
be drawn from animal carcinogenicity
data) that DES presents a human cancer
risk. The evidence from the treatment of
women with DES provides no basis for
concluding'that there is a no-effect level
for DES with respect to cancer. These
findings warrant the conclusions that
DES has not been shown to be safe and
that it is unsafe.
(3) Adverse Effects of DES Other
Than Cancer.As noted above, the
"safety clause".must be invoked if
serious questions about the safety of
observed residues are raised by the
Bureaus, and the manufacturing parties
fail to show that the DES residues are
safe. I find that safety questions about
DES have been raised not only by the
substance's carcinogenic effects but also
by other adverse effects with which it is

associated. These questions have not
been resolved.
(a)
TetatogenicEffects. Dr. Thomas
Collins of the Bureau of Foods testified
about his review of articles suggesting a
teratogenic effect associated with DES
(G-12). He defined "teratology" as the
science concerned with the generation
of structural or functional alterations or
malformations in organisms during their
development, both prior to and
subsequefit to birth (id. at 1].
Based upon his review, Dr. Collins
gave his opinion that DES is a teratogen
in mice and humans and that it has
specific effects on male and female
reproductive systems and on the
cardiovascular system (G-12 at 6; see
also G-57 at 5; G-72 at 7-8). He based
his conclusion on the following: (1]
observations of anomalies of cervix
development in females after prenatal
exposure to DES (see discussion in this
section below); (2) reproductive tract
lesions in male mice exposed prenatally
to DES in a study by McLachlan. et al.
(see discussion in this section below);
(3) observed effects on male genital
tracts associated with the
administration to the subjects' mothers
of DES prior to the subjects' births (also
discussed in this section below): (4)a
letter to Lancet (the British Medical
Journal) reporting one case of functional
incompetence of male gonads,
apparently associated with human
prenatal DES exposure; (5) a report that
four female human infants and children
exposed to DES in utero exhibited a
degree of masculinization, in a report
that also stated that the offspring of 700
DES-treated women were shown to be
normal; (6) three studies demonstrating
teratogenicity of DES and DES
dipropionate in mice; (7) a report that
cardiovascular malformations were
found at birth in children exposed
prenatally to oral contraceptives during
the first month of pregnancy at the rate
of 18.2 per 1,000 versus 7.8 per 1,000
among children not so exposed. The
reports relied upon by Dr. Collins are
found in the administrative record at G13 through G-20.
The manufacturing parties' Dr.
Bernard Kliman contended that Dr.
Collins' summary of published articles
on the teratogenicity of DES is worthless
"because he has failed to provide any
analysis of these reports" (M-110 at 19).
Dr. Kliman contended that these reports
do not support Dr. Collins' statement
that DES is a teratogen (id.). His own
review of these reports was rather
sketchy, and the -criticisms he makes of
them are not persuasive. Dr. Kliman
discounted, for example, the three
studies of Gabriel-Robez and colleagues

54879

(G-17, G-18, G-19) (the, sixth basis for
Dr. Collins' opinion as cited above]
because DES dipropionate was
administered instead of free DES (M-110
at 19)). DES dipropionate is, ho;ever.
hydrolized by esterases (enzymes which
catalyze the hydrolysis of esters into
their alcohols and acids) to yield DES
and propionate. Due to the abundance
and ubiquity of these esterases. the
proposition that DES was the underlying
cause of the observed teratogenic effects
cannot be disregarded.
One of the articles by Dr. Herbst
details the benign abnormalities of the
vaginal tract found in a study of 110 DES
daughters and a control group of 82
unexposed females (0-40]. He found an
association with very high statistical
significance (p = <.0001) between DES
and the following abnormalities: vaginal
or cervial fibrous ridges;.cerical
erosion identified in biopsy specimens:
failure of part of the cervix to stain with
iodine; vaginal adenosis identified in
biopsy specimens; failure of part of the
vagina to stain with iodine (id., Table 3).
These and other noncarcinogenic
abnormalities observed in the daughters
of DES-treated mothers may be
characterized as "benign" (0-40 at 338).
Any change in the human body causedby the administration of a foreign
substance is, however, reason for
concern. Although there is apparently no
evidence of the direct transition from
adenosis (the presence of glandular
epithelium or its mucinous productsJ.
one of the observed abnormalities, to
adenocarcinoina. it is noteworthy that
adenosis is present in nearly all of the
adenocarcinoma victims (id. at 339; see
also G-42 at 10; cf. G-138 at 3).
Dr. Gill. in reporting his followup
study of a controlled test of the effectiveness of DES in pregnant women
(the Chicago study (G-10)]), also
observed statistically significant
associations of maternal ingestion of
DES with circumferential ridges of the
vagina and cervix and dysplastic lesions
in these tissues in female offspring. This
study also demonstrated with statistical
significance (p<.01 and p<.&05) that
DES is related to observations of
epididymal cysts (the epididymis is the
cordlike structure, near the testis, whose
ducts store the spermatazoa), and
hypotrophic (underdeveloped) testes in
the male offspring. In addition, a
substantial percentage (28) of the group
of males exposed to DES in utero had
severely pathologic decreases in sperm
production: no such effect was found in
the control males. Dr. Gill reported
adenosis in 66.8 percent of the DESexposed females compared to 3.6
percent in the control group. A hiter

54880

Federal Register / Vol. 44, No. 185

/

Fridayi, September 21, 1979

/

Notices

X"M

report of this study,(G-192). which

includes more data, also found
significance in these tireas.
Dr. John McLacllan was an author of
a book chapter (G--61) dealing with the
transplacental toxicity of DES. It details
a number of animal and human reports
that have shown problems With the in
utero exposure of animals and humans
to DES. Some of the articles cited have
been included in the administrative
record. For example, G-60 is a report of
a test of male mice exposed in utero to
DES (100 mg per kg of maternal body
weight administered daily from day 9
through day 16 ofgestation). Six of i0
males born of DESr.reated mothers were
sterile, while none of a similar size
group of males born to control mothers
was sterile. Uponsacrifice, 15 of 24 of
the males born to DES-treated mothers'
and none-of the 15 males born. to control
mothers was found to have testicular
changes (id. at991).
Dr. McLachlan testified that
carcinogens that require long term, high
dose administration to induce
'detectable cancer in adult testanimals
have been shown to be capable of
producing cancer-in offspring of treated
mothers at much lower doses
administered for shorter periods of tine
(G-59 at 2). He identified this
phenomenon as transplacental
carcinogenicity and suggested that the
human carcinogenicity data discussed
above show DES to be a transplacental
carcinogen (id.).
Dr. McLachlan is performing a series
of studies ion the teratogenic effect s of DES (id. at 5). He described one such
study, in which he observeda
statistically significant dose response
relationship between DES
administration and loss of fertility of
female progency of DES-treated mothers
in a mouse study (id. at 4). The dosages
range from 0.01 to 100 micrograms (p)
:per kilogram (kg) of animal body weight.
Although there was no statistically
significant difference between the
lowest dosage and the control animals,
the dose-response relationship.observed
and the fact ihat higher levels caued an
effect is significant
Dr. Kliman objected to Dr.
McLachlan's studies because "no
control experiments were conducted
with any natural estrogen (M-110 at 14),
so that it is impossible to determine
whether the observed-effects would also
have been caused by natural estrogens.
However, Dr. McLachlan's objective
was to ascertain the transplacental
toxicity of DES, in which he succeeded;
and not to establish that DES is the only
estrogen that exhibits fransplacental
: .
toxicity. - .

Dr. McLachlan described a theory that
would diffeientiate DES from other
estrogens with respect .tolransplacental
toxicity
In the normal pregnant female, the
presence of high levels of the endogenous
estrogens may be less of a threat to the
'developing fetus because of the presence of
alpha-fetoprotein. a substance that -acts as a
high affinity binder of natural estrogens and
so renders them relatively nontoxic to the
fetus. It has been demonstrated that DES
does not bind to alphq-fetoprotein with the
same high affinity (id. at 5).
'In addition, he cited the same type of
relationship'in mammals for TeBG
(discussed in.section II(D)(1) above).
For this and other reasons (id. at 6), it
was his opinion' that DES plays a more
critical role than the endogenous
estrogens in transplacental toxicity.
According to the manuffacturing
parties' Dr. Bemard'Kliman, Dr.
McLachlan misinterpreted the data of
Uriel, et al. (G-63) in.deVeloping his
theory . Dr. Kliman'stated that DES has
40 percent of the activity ofestradiol
and nearly the same activity as estradiol
in binding to -these proteins. Also, the
lower binding activity o[DES only
allows DES to be metabolized more
quickly by the liver {M-=I0 at 15-16).
As discussed in sectionlII[D)(1) of
this Decision, data sufficient to resolve'
the arguments presented by Dr.
McLachlan and those presented by Dr.
Kliman on this issue is lacking. It is
therefore not possible to determine with
assurance that the teratogenic (or
mutagenic) effects of DES either differ
from or are the same as the effects
-associated with endogenous estrogens.
I must conclude, on the basis of the
evidence discussed in this section, that
DES is a teratogen in animals and in
humans. "
(b) Mutagenic Effects Dr. Sydney
Green, who, at the time of his testimony,
headed the Genetics Toxicology Branch
of the Division of Toxicology, Bureau of
Foods, reviewed two published reports
(G-32 and G-33) that establish the
mutagenicity of DES dipbosphale. The
first study-revealed that DES
diphosphate resulted in monosomies
(cells with one chromosome less than
normal) and trisomies (cells with one
chromosorihe more" than normal) in the
bone marrow of mice G-32]. Dr. Green
classified this as a mutageniq effect (G31 at 2):
The znonosomies are notsignificant
contributors to hereditary diseases or

disorders because cells possessing such

"chromosomal abnormalitiesJ rarely survive.

However, the presence of trisomies can be said to be a true -mutagenictffect. Such cells
usually survive and pass on their abnormal
,.characterisitcs io.futuregenerations-1f these,

effects are seen ingerminal (sex) cells they
can lead to mongolism and other hereditary
disorders.,

The second study also uncovered the
production of trisomies in offspring of
mice whose mothers were treated with
DES diphbsphate (G-33).
During cross-exanination, Qr. Green
stated his opinion that DES could be
considered as the underlying cause of
this mutagenic effect (Tr. at 578-79). He
noted, among the bases for his opinion
on this issue, the fact that when DES
diphosphate is hydrolized it yields DES:
"So, in essence, one would be testing
diethylstilbestrol within the cells, as
opposed to diethylstilbestrol
diphosphate"(Tr at 579).
Dr Kliman criticized the testimony of
Dr. Green, particularly because Dr.
Green failed to mention that similar
mutational aberrations are also
associated with the natural estrogens
(M-110 at 20). But, during cross-

Lxamination, Dr. Green acknowledged
that studies have shown estrogens to be
mutagenic (Tr. at 578-79). Like Dr.

McLachlan, he did not claim DES is the
only estrogen that produces adverse
effects.
I find, on the basis of the evidence In
this record, that DES does cause
mutagenic effects in some
circumstances.
(c) OtherEffects; Dr. Roy Hertz
reported on the extreme potency of DES
evidenced by accidental absorption In
industry and the home, such as "the
occurrence of breast development in'
children ingesting accidentally DES
contaminated vitamin capsules," and t
"the precocious development of the
breasts and external genitalia when the
prepubertal daughter of a worker (in the
animal drug industry) used her father's
bed while he was at work" (G-40 at 7).
These reports add marginally to the
impression that DES poses genuine and
serious risks to humans and that its
activity produces toxic effects that are
not now totally understood.
1d) No-Effect Level. The Bureaus'
witnesses testified that no-effect levels
for the adverse effects of DES could not
be established. With regard to'the
teratogenic effects of DES,'Dr. Collins
testified that "none fof the reports he
evaluated]'are sufficiently complete to
allow us to establish safe tolerance
levels for DES" (G-12'at 6). Dr.'Gill, who
reported on the'effebts of DES
discovered in the Chicbgo study, stated
that "it is not possible to calculate a safe
tolerance level for such exposure for the
data reported" (G-138 at 3; see al96 G37 at 5).
Dr. Kilnan testified that Dr.
McLachlan's work points toa no-effect
level of DES because "fenale mouse

Federal Register / Vol. 44. No. 185 / Friday, September 21, 1979 / Notices
fertility was not significantly altered by
the lowest dose, 0.01 jig per kg per day
on days 9 to 16 of pregnancy" (M-110at
14-15). The failure of this test to
demonstrate a response at its lowest
dosage could, however, be the result of
the relative lack of sensitivity of the test
system. Dr. McLachlan himself testified
that it is not possible to determine a noeffect level from his studies (Tr. at 92).
Dr. Green testified that the mutagenic
studies he reviewed also did not support
the existence of a no-effect level for
DES:

Early in 1977, the manufacturing
parties filed a motion in the United
Stat6s Court of Appeals for the District
of Columbia Circuit to compel the
agency to consider the societal benefits
from DES, including any adverse
environmental consequentes from
withdrawal of the NADA's, as part of
the hearing. In a memorandum of March
22. 1977. Acting Commissioner Gardner
mooted that question by directing the
Administrative Law Judge to consider
the benefits issues. He did so by means
of a memorandum to the Administrative
Law Judge in which he noted that he
These studies, however, do not provide
was taking no position on the relevance
quantitative data which would allow a
of these issues to the proceeding. He did
calculation of a no-effect level for these
state in that memorandum: "In making
effects and the subsequent estimation of safe
this safety determination [under the
tolerance levels for humans (G-31 at 3).
"safety clause"] societal benefits and
(e) Conclusion As to Adverse Effects
environmental effects have historically
of DES Other Than Cancer.I find that
not been considered to be legally
the evidence presented by the Bureaus
relevant" (Record No. 110 at 2).
demonstrates that DES is a teratogen
(a) Legislative History.- (i) New
and a mutagen. It is not possible from
Animal DrugProvisions.The new
the evidence in this record to establish
animal drug applications that are the
the existence of a no-effect level for DES subject of this hearing are creatures of
for these effect. Thus, the fact that DES
the animal drug amendment of 196,. As
causes teratogenic and mutagenic
noted in section 1. that amendment was
effects is an in.dependent basis for my
intended to consolidate the agency's
conclusion that DES has been shown not review of animal drugs, which was at
to be "shown to be safe," and that it is
that time being conducted under the
unsafe, for its approved uses.
food additive Amendments and the new
(E) The Risk-Benefit Issue (1)
drug provisions. Congress did not, in
Proprietyof Risk-Benefit Analysis. The
writing the new animal drug provisions.
Administrative Law Judge held that
include language authorizing the FDA to
under 21 U.S.C. 360b consideration of
consider the benefits of an animal drug
the alleged societal benefits of the use of in determiningwhether it is safe.
DES is not an appropriate part of the
(Compare 21 U.S.C. 346 and 346a, in
decision whether approval of the new
which, Congress required the
animal drug application should be
Commissioner and (now) the
withdrawn (I.D. at 15). This
Administrator of the Environmental
interpretation of the statute is supported Protection Agency to consider the
by the legislative history of the statute,
benefits of the products regulated under
is consistent with positions the agency
those sections in the setting of
has taken previously on this issue, and
tolerances.) Nothing in the legislative
reflects sound public policy. In Hess &
history of the 1968 amendments supports
Clark,Division of Rhodia,Inc. v. FDA,
the proposition that the FDA should
495 F.2d 975 (D.C. Cir. 1974), however,
copsider the socio-economic benefits of
the Court stated in dictum that the FDA
a drug in deciding whether it is safe. The
should consider the benefits of the use
manufacturing parties have. however,
of DES should it proceed under the
relied upon legislative hist-ory of the new
"safety clause" (495 F.2d at 993-94):
drug and food additive provisions for
support of their position that benefits
Outside of the per se-rule of the Delaney
should be considered. Little support
Clause, the typical issue for the FDA is not
exists.
the absolute safety of a drug. Most drugs are
unsafe in some degree. Rather. the issue for
(ii) New DrugProvisions.DES was
the FDA is whether to allow sale of the drug,
approved for animal use in the 1950's
usually under specific restrictions. Resolution
under the then-existing new drug
of this issue inevitably means calculating
provision. Prior to 1902. when
whether the benefits which the drug produces effectiveness was made
an additional
outweigh the costs of-its restricted use. In the
consideration, new drug applications
present case, DES is asserted to be of
were approved if use bf the drug was
substantial benefit in enhancing meat
shown to be safe. At that time, the
production, and this is not gainsaid by FDA.
agency took the position that
The FDA must cqnsider, after hearing.
effectiveness was an element in the
whether DES pellets would be safe in terms
consideration of the safety of a drug
of the amounts of residue consumed.
when that drug was to be used in
(Footnotes omitted.)

54 F31

"treatment of a life-threatening disease or
where there was an indication that that
drug would occasionally produce
serious toxic or even lethal effects. The
manufacturing parties argue that. in
taking this position, the FDA was stating
its understanding that a safety
determination necessarily involved a
risk-benefit analysis.
The evident flaw in the application of
the manufacturing parties' argument to
the instant proceedings is that
effectiveness in a human drug context is
different from effectiveness in an animal
drug context. The risk to the patient
from a human dru- may be justified by a
therapeutic benefit to thatpatient frornthe drug. It is an entirely different
question, however, whether the risk to
human consumers of the products of
animals are justified by an economic
benefit to animal drug manufacturers.
animal producers, or meat consumers
generally.
(The only time where the theory that
effectiveness is part of safety would be
applicable to an animal drug would be
circumstances in which the risk was to-.
the animal itself, as opposed to any
human consumer. FDA considers that
type of benefit relevant to a
determination of safety; but that type of
benefit is not at issue in this proceeding.
The types of benefits urged by the
manufacturing parties are alleged health
benefits to humans and economic and
environmental benefits.)
There is. of course, an obvious
difference between the therapeutic
benefits of a drug. which often alleviae
a risk to the person to whom the drug is
administered, and so-called "socioeconomic" benefits associated with the
use of a drug. The former are the only
type of benefits that the FDA considers
in determining w'hether a human drug is
safe. The agency never considers socioeconomic beneftis in making that
decision.
Moreover, the consideration of risks
and benefits with respect to human
drugs is always based on the premise
that before being exposed to the risk. an
individual patient will have the
protection of either a physician's
evaluation (in the case ofa prescription
drug] or adequate directions for use
enabling the patient himself to decide
whether to run the risk (in the case of an
over-the-counter drug). No such
protection is available to those exposed
to the risk from residues of DES in moat.
The asserted similarity between the
treatment of human drugs and animal
drugs is, course, critical to the
of
manufacturing parties' argument on this
subjecL Apparently the distinction
between the two systems of regulation
was not adequately pointed out to the-

54882
II

Federal Register / Vol. 44, No. 185 / Friday, September 21, 1979
m.=..a

Hess & Clark Court, however. In a
footnote, the Court quoted extensively

from an article by Richard Merrill on
prescriptiondrug injuries as support for
the proposition that effectivenss
considerations are relevant to safety
determinations (495 F. 2d at 994 n. 59).
* (iii) Food Additive Provisions.The
manufacturing parties rely upon the fact
.that one impetus for passage of the food
additives amendment was a desire by
the FDA and the regulated industry to
allow FDA to set tolerances for products
that were-hazardous at some levels and
not at others. Congress, in
accommodating this desire by allowing
the setting of-tolerances, allowed the
agency to consider the level of the
ingredient that would be required to
serve its functional'purpose. Where a
tolerance limitationis required for a
product, the tolerance may not be
greater than the.amount necessary to
accomplish the additive's intended
purpose see 21 U.S.C. 348(c)f4)(A)-

Similarly, where a iolerance is required,
no food additive petition may be
approved unless it contains evidence
thatestablishes that the additive will
accomplish its intended physical or
other technical effect: see 21 U.S.C.
348(c](4)(B).

Thus, where an additive is shown to
be safe at some level, the FDA.is
authorized to consider whether it does
what it is intended to do. The FDA. is
not, however, authorized to consider
whether whatthe additive is supposed
to do provides'any ,benefit to society.
Congress was explicit in its reports on
this bill that the FDA would not be
allowed to consider the societal benefits
to be derived from use of the fQod
additive in question. See, e.g., S. Rept
No,2422, 85th Cong., 2d Sess. 7 (1958):
Determination of a proper tolerance
level "does not involve any judgment on
the part of the Secretary'of whether [the
food additive's) effect results in any
added 'value' to the consumer of such
food or enhances the marketability from
a merchandising point of view.".Accord,
H.R..Rept. No. 2284, 85th Cong., 2d Sess.
(1958). (Congress thus rejected the
position apparently advanced by
Commissioner of Food and Drugs
Larrick in 1956 that some consideration
of the "benefit to the producer or
consumer" should be permitted in the
evaluation of food additives. Hearings
Before Subcommittee of the House
Interstate and Foreign Commerce
Committee onHRL4475, etc., 84th Cong.,
2d Sess. 194-95 (1956)-) Therefore, under
the food additive amendment, were a
tolerance applicable for a substance
such as DES, PDA would be barred from

m

l

considering societal benefits in setting
that tolerance.
(ii) Conclusion.As to Legislative
History.Congress thus did not authorize
or require consideration of the socioeconomic benefits of-an animal drug in
determining its safety. Indeed, the
language adopted by Congress, having
its roots in the human drug and food
additive provisions of the law, clearly
reflects an intention that FDA definitely
not consider socio-economic benefits in
making decisions on the safety of
animal drugs. I thus conclude that
Congress has made the determination
that an animal drug that poses a risk to
humans can never be considered "safe"
because it provides an economic or
other social'benefit to society.
(b) The Agency's Position.The FDA
has never considered the benefits of an
animal drug that po~ed a risk to ultimate
human consumers when deciding
whether that drug is safe. The
'manufacturing parties do not contend
that the agency has done so. Ihdeed,"
Bell v. Goddard,supra,which also dealt
with DES--there as a drug for poultrydescribes an-FDA action with respectto
an animal drug in which not even the
proponents of the drug-contended that
benefits should be considered.
The manufacturing parties do quote
from the preamble to regulations issued
by the FDA in 1976 that deal not with
animal di-ugs but rather with food
additives. As first proposed in
September of 1974. thdeL-regulations ,
,would have defined "safe" and*"safety"
to include consideration of. among other
factors, "[tjhe benefit contributed by the
substance" (39 FR 34194 (September 23,
1974)). When the final regulations were
issued, this consideration was deleted.
In an apparent attempt to explain the
agency's rationale for the original
proposal, however, the preamble to the
final regulation made the following "
statement (41 FR 53601: December 7,
1976):
The Commissioner concludes that it is
appropriate to recognize that the benefit
contributed by a substance is inev.itably a
factorlo be considered in d~termining
whether a particular substance is "safe" (or
generally recognized as "safe") for its. intended use. The term "safe" is to be given
its ordinary meaning, and in its common
usage the term is understood to carry an
assessment of benefits and risks. It is true, as
the comment states, that minor food additives
are not approved at levels that may present a
hazard to the normal consumer. This result is
required by the act because the benefit of a
minor food-additive is too small to justify the
imposition ofa known risk to normal
consumers; use of such ingredient at levels
that may present a hazard'to the normal
consumer would nat be "safe." However, this
re.ult does not necessarily follow in the case

1 Notices

I I'

=m

of important food additives. For example, If it
were found that a major food source such as
meat or grain was associated with the
development of chronic diseases In normal
individuals, it would not necessarily follow
that the food was unsafe within the meaning
of the act. The ordinary understanding of the
term "safe" would require some benefit-to.
risk analysis in such circumstances,
Another example relates to the incidence
of allergic reactions to particular food
-ingredients. Adverse reactions caused by
allergy are clearly a consideration in
detennining whether a food ingredient Issafe.
Ordinarily, the incidence of allergic reactions
from a food additive cannot be considered
because data and test protocols do not qxlst.
When data exist, however, they may be
considered, andan assessment of benefits
and risks becomes relevant. For example, if It
were determined that both a particular
emulsifier and a particular fruit resulted In
the same unusually high Incidence of allergic
reactions, one might reasonably conclude
that the emulsifier was not safe but that the
fruit was safe. Such conclusions would
simply represent common understanding of
the safety * *
The Commissioner has, however, deleted
from the regulations the reference to
consideration of benefits on the ground that
this separate consideration is legitimately
included within the concept of safety as used
in'the act. Furthermore, explicitly retaining
the criterion of benefit in the regulations
might be construed as requiring routine
formal analysis of a factor that the agency
will only occasionally need to take Into
account, because the agency's general
guidelines will result In disapproval of food
additives that may cause toxic effects in
normal individuals.
This language is quoted in full
because I am, on behalf of the FDA.
disavowing it. It has never been the
basis for an agency decision. As
discussed, there is no justification for
such a statement either in the statute
itself or in its legislative history,
The manufacturing parties argue that
this statement in the preamble to a
regulation is an advisory opinion
binding upon the agency. They citefor
this proposition 21 CFR 10.85(d) (1) and
(e). Subsection (d)(1), In fact, does
identify the preamble to a final
regulation as an advisory opinion.
Subsection (e) states that an advisory
opinion "obligates the agency to follow
it until it is amended or revoked." An
advisory opinion may, however, be
amended or revoked in the Federal
Register at any time after it has been
issued (21 CFR 10.85(g)). To the extent
that the language quoted above may be,
considered such an "advisory opinion,"
that opinion has been superseded (and,
by virtue of 21 CFR 10.85(g), revoked) by
at least one subsequent Federal Register
statement that directly contradicts it.
See 42 FR 1.9996 (April 15,1977)
(Saccharin and Its Salts): "The

Federal Register
Commissioner/' * * notes that under

/

the provisions of the law relating to food
additives, FDA is not empowered to
take into account the asserted benefits
of any food additive in applying the
basic safety standard of the act." In any
case, the language cited by the
manufacturing parties deals with safety
in the context of GRAS substances and
food additives, not in the context of new
animal drugs. It thus would in no case
be binding in this proceeding, Nor could
it be said that the manufacturing parties
have relied on the cited language or that
may disavowal of that language is in
any respect unfair, to them.
1c) PolicyArguments. There are
persuasive policy arguments against
having an administrative agency such as
the FDA make the kind of risk-benefit
analysis sought by the manufacturing
parties here. It may be that preliminary
issues in this analysis are of the type
that the FDA is qualified by experience'
and expertise to resolve. The agency is equipped, for instance, to evaluate
calculations of the risk from a drug such
as DES if the necessary data are
available {they are not here). Once the
risk and the benefits of an animal drug
are determined. however, the ultimate
issues require pure value judgments. {On
the difficulty such judgments present for
the administrative process and for
judicial review, see Cooper, "The Role
of Regulatory Agencies in Risk-Benefit
Decision-Making." 33 Food. Drug.
Cosmetic L. 1. 11978).)
755
It may be suggested that the agency
makes risk/benefit analyses often with
respact to such products as human drugs
and medical devices. This suggestion is.
however, incorrect. Properly understood,
the agency's evaluation of, for example.
a human drug -s a comparison of risk to
risk. The risk f using the drug is
weighed against the Tisk of not using it.
Moreover, the risks and benefits {or
avoidance of other risks) are of the same
type (relative to health], accrue to the
same persons [patients). and are subject
to a well-established scientific and
professional discipline Imedicine). Even
so, this type of evaluation is rarely easy.
Often a calculated risk of one harm must
be weighed against a significantly
smaller fisk of a much greater harm.as
with a useful drug that occasionally
produces severe side effects. The factors
considered are all detriments to the
public's health, however, and the
decision may be appropriately
considered to be a medical one.
Here. however, the manufacturing
parties ask the FDA to weigh a risk of
cancer and other serious adverse effects
against an economic benefit. Arguably.
the persons at risk also receive part of

VoL 44. No. 185

/

Friday. September 21, 1979 / Notices

the economic benefit because the meat
they purchase may be available at a
lower price because of the use of DES.
But much of the economic benefit, as
evidenced by the tenacity with which
the withdrawal of the DES NADA'shas
been fought. goes to parties other than
the consumers of the meat products or
DES-treated animals.
Perhaps society is wilLtig to expose
all of its meat-consuming members to a
relatively small risk of cancer and other
adverse effects in order to provide a
small economic benefit to those
.consumers and a larger economic
benefit to DES producers and.
potentially, users. The FDA is not.
however, qualified in any particular way
to make that value judgment for society.
The value judgment could not be
supported by a record. a record could
support only factual firdLigs, not value
judgments. Nor could the value
judgment be effectively reviewed by a
court, which in general is limited to
consideration of facts, law. and
procedures. In a democratic system, the
appropriate place for value judgments to,
be made is the !egislature. Here, as
discussed above, it is apparent that
Congress has shouldered the
responsibility for resolving this issue. It
has decided that no economic benefit
justifies use of an animal drug that
presents an identifiable risk to the
health of consumers.
The manufacturin- part-es also ask
FDA to consider general nontherapeutic
health benefits from the use of DES.
Nothing in the language er leg*slative
history of the statute or in FDAs prior
interpretation or application of the
statute suggests that consideration of
such benefits is eith.er requre" ,or
permissible. There is nothm.g to saz-est
that Congress or FDA has ever thou;ht
that such benefits might fl v: from the
administration of drugs to animals.
Thus, it is understandable that Congress
did not contemplate-that FDA would
consider such benefits in determining
the safety of animal drugs, and that FDA
has not done so.
The argument that FDA chould
consider such benefits appeals to some
as a public policy but that appeal can
hardly outweigh the combined, force of
language. legislative history. and agency
practice that weighs ogbinst
consideration of such benefits. In view
of the importance of the question. I
believe it should be resolved only on a
record that squarelypresents it. Here. as
discussed in sections IIJfE)(2) Ic) and (d).
the manufacturing parties have not
shown that DES presents health benefits
that could outweigh its risks. The quality
of the evidence in this record on health

54833

benefits is so unsatisfactory that it does
not provide a sufficiently powerful
policy argument for raising the legal
issue. Therefore, I would rather leave
the legal question open, while
recognizing that it would require a very
powerful showing indeed to outeigh
the strong legal arguments against
consideration of such benefits .
(d) ConcJusionAs to PrrprL-earyof
Risk-Benefit Analysis. The law is clear
that the FDA may not consider socioeconomic benfits in the determination of
the safety to human beings of a new
animal drug. and lam not prepared to
conclude that it permits consideration of
human health benefits. In order to
provide as complete a record for judicial
review as possible, however. I will
discuss, as did the Administrative Law
Judge. the evidence presented at the
hearing with respect to both tyTes of
benefits.
(2) Risk.Beaefit Analysis. It is clear
that the applicant has the burden of
showing that an animal drug is "safe!' If
a riskibenefit analysis were'
appropriately a part of an animal drug
safety decision, the applicant would.
therefore, have the burden of showing
that the benefits of the drug outweigh its
risks. The allocation of the burden of proof is important because the record of
this proceeding is totally inadequate
even to determine vwhat the risks and
benefits of DES are [or how geat the
risk of DES use is). much less to provide
any guidance on how the weighing of
risks against benefits should he
accomplished.(a) QuantitativeRisA As, .- meL
Some manufacturing parties' witnesses
extrapolated from the Herbst data to
calculate extremely low levels of risk of
human cancer in females from the
ingestion of DES-contaminated meat
(see M-83: M-99; M-104)] flhese risk
calculations do not address the question
of how great a risk DES poses to human
males.) Other manufacturing parties'
witnesses argued that there is, in effect.
no risk from the present uses of DES [M69. M-40). For the reasons discissed
below, I do not regard either of thes2
contentions as valid. In additin. I find
that the data on DES are toi meager to
allow any risk calculation zcceptab!e fo
the purpose of supporting ceaticced
approval of DES as an animal drug.
(i) Calculatio s FromHerbstDasa Or.
Uerbst testified that he regards risk
estimates based on his data as highly
suspect (G-37 at 51:
I am informed that cthers have attempteto cal date and etrapolate risk estima:es
and "no-effect levels" in the -- oWe United
SWates population foe DES Zn
lodusiagdata-

from our Registry, but ! do not believe these
calculations can properly be made from our

54884
m

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

Em

data, nor that "no-effect levels" can be extrapiola ted from our epidemiological
observations of effect levels.

I

reasonable assumptionsfor the six just
listed the risk of human cancer in
females from the ingestion of DEStreated meat is 2 in 100 trillion (1
trillion=1019, or equivalent to one
cancer every 10 million years in the
United States. (Id.)
Dr. Thomas H. Jukes, the author of
"Diethylstilbestrol in Beef Production:
What Is the Rk to Consumers?" (M104), calculated his risk estimate in a
manner similar to that of Dr. Jones and
Dr. Grendon. Dr. Jukes assumed a lower
daily intake of DES-1.9 nanograms (1
nanogram (ng)=10 - 9grams =1 billionth
of a gram) DES/day as compared to 100
ng DES/day resulting from assumptions
(1) and (2) above. Also, in his linear
extrapolation from the 1.5 mg DES/day
dose level, he assumed that the risk of
human cancer pregnant women
receiving DES therapy was 4 in 1,000, an
upper limit estimate of risk comp*uted by
Lanier, et al. (C-44 at 798). Dr. Jukes
then arrived at a risk estimate of less
than 5 in I billion from consumption of
DES-treated meat, or approximately 1
cancer every 133 years in the United
States (he assumes 1.5 million female
births per year in the United States). In
his written testimony, Dr. Jukes revises
his estimate to I case of cancer every ,380 to 3,800 years (M-99 at 10), because
he substituted for the 4 in 1,000 risk
estimate (of cancer to females exposed
in utero to DES) the 0.14 to 1.4 in 1,000
estimate proposed by Dr. Herbst (M-26

I agree with Dr. Herbst's opinion on this
issue. The manufacturing parties' .risk
assessments from the'Herbst data
merely demonstrate that the results of a
risk calculation are dependent on the
assumptions on which it is based.
The following assertions about the
risk of DES should be read in light of the
unsupported assumptions upon which
they rely, i.e., that (1) the only cancer
DES causes in women is vaginal
adenocarcinoma in'the daughters of DES
exposed mothers; (2) there is a straight
line dose-response for DES from the
lowest DES dose that has been
associated witft vaginal
adenocarcinoma; (3) the risk of lifetime
exposure to DES is identical to the risk
of exposure of the child to DES during
the mother's pregnancy; ai~d (4) we
know the incidence of vaginal
adenocarcinoma ass6ciated with DES
exposure in utero. I will first describe
the calculations mad&from the Herbst
data and then elaborate upon my
reasons for not accepting the
assumptions upon which those
calculations are based.
In an article entitled "Environmental
Factors in.the Origin of Cancer and
Estimation of the Possible Hazard to
Man" (M-63), the authors, Dr. H. B.
Jones and Dr.'A. Grendon, calculated
that under "very'conservative"
at 47).
assumptions the risk of DES-related
As Bureaus' witness Dr. Hoel stated,
cancer from meat consumption to the
female population of the United States is "the central assumptions upon which
these authors based-their calculations
3 in 100 million. The authors then
have not been validated" (G-55 at 2).
assumed that there are 4 million births
My discussion. of the four unsupported
per year in the United States, so that
assumptions made by the manufacturing
this risk is equivalent to one cancer
parties' witnesses that I regard as most
every 8 years. Their "conservative"
important follows:
assumptions are as follows:
I First, as the Bureaus' Dr. Condon
(1) A pregnant woman eats 10 oz. of
stated, one reason for rejecting these
beef muscle every day, except 1 day per
risk calculations is the fact that "they
week, in which 6 oz. of beef liver are
assume that the only type of cancer risk
substituted.
due to DES is vaginal carcinoma
(2] Beef liver contains DES at a
because it was the only human cancer
concentration of 2 ppb, and the
.on which they based their calculations"
concentration in beef muscle is 0.2 ppb.
(G-21 at 4). See also Dr. Cornfield's
(3) 100 DES-related cancers resulted
statement that "[blecause of the lack of
from pregnant women receiving DES
studies of other forms of cancer in the
treatment who gave birth during the
women exposed, the human evidence
period 1951-1955.
[upon which the manufacturing parties'
(4) DES was prescribed for only 1
- witnesses rely) cannot be used to
percent of the 10 million, pregnant
women during the period 1951-1955.
estimate a safe dose of DES in food" (G25 at 2). Particularly in light-of the fact
(51 The dose that elicited the response
ineach of the 100 cancer victims was 1.5 that animal carcinogenicity studies
show that DES causes cancer in a
mg DES/day, the lowest dose
variety of organs (see, e.g., G-47; G-84),
administered to pregnant women, as
I see no basis for the assumption that..
reported to'Dr. Herbst in his Registry.
(6) The dose-response relationship to - DES is associated with only this one
rare type of cancer.
DES is linear in the 0 to 1.5 mg DES/day
A second reason why the risk'
range. Dr. Jones and Dr., Grendon claim
estimates presented by the
that when they substitute more
-

manufacturing parties are extremely
small is that they have assumed that
there exists a dose-response
relationship between the incidence of
vaginal cancer in females and the
dosage of DES administered to their
mothers. Put in its simplest terms, a dose
response relationship in this context
means that an increase in the dosage of
DES administred results in an increase
in the percentage of persons who are
afflicted with cancer. Thus, again to
simplify the matter, if cancer were found
in I in 1,000 persons treated with 1.5 mg
of substance X, 1 in 100 persons treated
with 15 mg X, and I in 10 persons
treated With 150 mg X, a dose response
would be shovn. If that effect were
observed, it might be valid to estimate
that 0.15 mg X would cause cancer In 1
in 10,000 persons.
Another possibility, however, Is that
0.015 mg (or some even lower amount) of
substance X causes cancer in I in 1,000
persons, and that increases in dosage
above that do not add appreciably to
that risk. Thus, persons administered
0.015 mg X would be at the same risk as
those administered 1.5 mg X or 150 ing
X. The assumption that because 1.5 mg
X-caused one cancer in 1,000, 0.15 nig X
would cause I cancer in 10,000, would
then be incorrect.
The above example oversimplifies this
question, but does illustrate the problem
with the assumption utilized by the
manufacturing parties' witnesses. As is
often true with retrospective
epidemiological studies, it-can not be
determined from the Herbst data
whether there is in fact a dose-response
relationship between DES dosage and
the cancers observed. As Dr. Condon
noted, "no such relAtionship [between
dose and response] has been
established" (G-21 at 4). Dr. Heel
reiterated this fact (-55 at 3):
There is no scientific support for this

assumption. The reported studies of A,L,
Herbst. et a]. provide no basis for
constructing a dose-response relationship for
the observed carcinogenic effects. Without
such an established relationship, it is not
valid to extrapolate these data to low levels
of risk.
In general,'if no dose-response
relationship has been established in the
observable dose range, there is no
justification for extrapolating to the low
dose range via a dose-response curve.
The third unsupported assumption
made by the manufacturing partibs'
witnesses is equally likely to produce a
misleading risk assessment. A proper
analysis of the risks associated with
DES as an animal drug should deal with
low-dose, long-term (lifetime) exposure,
whereas the women in Herbst's Registry
faced high-dose, short-term (during

Federal Register

I

pregnancy only] exposure to DES. This
fact alone invalidates any risk

assessment, based on the Herbst data.
of human carcinogenesis from
consumption of DES-treated meat (see
G-21 at 4; C-55 at 3).

A fourth, though less important,

unsupported assumption by the
witnesses seeking to'calculate the risks
of DES use is the assumption that the
incidence rate of vaginal carcinoma in
women exposed to DES in utero is
known. As Dr. Condon noted, however.

there is a long latent period for vaginal
adenocarcinoma; consequently, more
cases may occur as the women exposed
age tG-21 at 4-5). In addition, there is no
certainty that all cases of this type of

cancer that resulted from use of DES
have been diagnosed and reported to Dlr.
Herbst.
For the reasons I ha.ve discussed. I
regard the risk assessments provided by
4he snanufacturing parties to prove the
safety of-DES in meat as unsupported
and unreliable. {Note that, in any case.
these estimates say nothing about the
risk of cancer posed by DES to the
approximately half of the population
that is male. I cannot assume, on the
basis of the evidence in this record, that
DES does not cause cancer in males.)
(ii) Argument That Approved Uses of
DES PresentNo Risk. Some witnesses
for the manufacturing parties attempted
to downplay the risk of DES to humans
either because it contributes very little
to the lotal amount of endogenous
estrogens or because the amount of DES
ingested from meat is well below what
are alleged to be no-effect levels.
Dr. Elwood V. Jensen argued that the,.
daily consumption of DES in meat is at
most 40 ng and that this amount is
insignificant:
It is my considered opinion that ingestion
of 40 oreves 400 ng of diethylstilbestrol per
day would have no physiological significance
in comparison with the 20,000 to 400.000 ng of
endogenous estradiol that humans normally
produce (in addition to estrone which also
makes a contribution to the total estrogen
level).
(M-69 at 10). As I have discussed
above {section tII[D)(1). however, DES
is not an endogenous estrogen, and !
-cannot find that its carcinogenic and
other adverse effects result only from its
estrogenic properties.
Dr. Nicholas H. Booth apparently
assumed that DES can have no
carcinogenic or other adverse effect at a
level at which it does not induce a
uterine response. He claimed that the
no-effect level from the parenteral
administration of DES is 0.29 1Lg/kg
body weight tM-40 at 2-4) because (1)'
the no-effect level from estradiol is 0.166
fg/kg body weight in rats (M-49), and

Vol. 44, No. 185

1

Friday. September 21. 1979 / Notices

(2) estradiol is 1.72 times more potent
than DES in mice when the effect is
taken to be artalteration in the vaginal
mitotic count (M-4B}. (The mitotic count
is the proportion of cells that are in the
process of cell division.) This dosage of
DES is 1.5 to 3 times smaller than the
6.25 ppb dosage administered to some of

the mice in the Cass study (discussed in
section III (D)(2(a) of this Decision) (M40 at 40). If beef liver contains DES at 2
ppb. he calculated that the average daily
intake of DES from meat is 3.8 ng (twice
the amount of Dr. Jukes' estimate),
which for a woman weighing 60 kg
'yields 0.063 rig DES/kg body weight fid.
at 6].
Dr. Booth states that if all the DES is
absorbed from the gastrointestinal [GI)
tract, this amount is 4,523 times below
the no-effect level of the rat that he
computed fid.). Whereas if only 3
percent of the DES is absorbed from the

GI tract, which he regarded as the more
realistic situation, this amount (17 x
10-2g DES/kg body weight) is 167.000
times below Dr. Booth's rat no-effect
level id.).
Dr. Booth also compared id. at 4-5)
the 0.063 mg DES/kg body weight to a
no-effect level in humans, which he
calculated from a study of the treatment
of senile vaginitis with DES (M-50). He
estimated the no-effect level for oral
administration in humans to be 0.476 jig
DES/kg body weight, approximately 1.5
times higher than the parenteral noeffect level in rats IM-40 at 4-5).
It must be remembered what Dr.
Booth considered as effects: a uterine
response in the rat, a change in the
vaginal mitotic count in the rat, and a
favorable reaction to the treatment of
senile vaginitis in humans. Dr. Booth. in
his testimony, did not even discuss his
reasons for assuming that these effects
correlate with either carcinogenesis or
any other adverse effect associated wvith
DES. No evidence in this record
demonstrates such a correlation. See the
discussion of my reasons for rejecting
the argument that DES is no different
from endogenous estrogens (section
IlI(D)[1) of this Decision). Finally. the
manner in which Dr. Booth combined
the results from studies with different
species and different methods of
administration in order to calculate noeffect levels has not been justified.
I can not agree that any amount of
DES, no -matter how small, has been
shown to be safe. On this point, my
conclusion is supported by the opinion
of Dr. Rauschen "Because of the lack'of
data concerning the exact levels of DES
which may elicit cancer in humans, we
cannot say how small an amount may
cause cancer nor how long that cancer
will take to appear" (G-70 at 4). See also

54M5

Dr. Saffiotti's testimony that "exposure
to any amount ofa carcinogen. however
small, will contribute to the total
carcinogenic effect in the population
." (G-80 at 6-7).
(iii) Risk Calcuationsfrom
Anima
Data.Having found that the risk
calculations proffered by the
manufacturing parties are invalid, I have
considered whether the available data
permit any reliable estimate of the risk
of DES use. Dr. Hoel noted what he
considered to be the only plausible

alternative method for conducting a risk
assessment of DES in meat (G-55at3]:
Entimation of cancer risks due to longera
(lifetime). low-level exposure to DES is,
for
the present, made only by extrapolatioa from
Vfirtime toxicity sttalies in expedment_.
animals.
Even though such estiziations require
extrapolation from animals to humans, t1e
general absence of risk data on lifetime
human exposure to DES makes it necessary
to use animal data.
None of the manufacturing parties'
witnesses attempted such an
extrapolation from a-nal data.
Some Bureaus! witnesses calculated
from the results of the Gass study that I
ppt DES would present a risk of less
than I cancer in I million exposed (see,
e.g., G-34 at 2). This calculation, even if accepted as valid. is hardly relevant to
present use of DES which, the record
shows, results in DES residues in edible
tissues above I ppt. (See, generally,
section III(B) of this Decision.)
As noted in the section dealing with
the analytical methods for DES
(II(A)[2)), this calculation is, any case,
in
unreliable. As discussed in that section,
substances metabolize in the body, and
the metabolites of a substance may be
more toxic than the parent compound.
Because different metabolites may be
formed by different species (see, e.g, C-24 at 10416), testing of the parent
substance in one species can not
provide definitive information about the
toxicity or carcinogenicity of that
substance in other species. Ilf.
for
example. DES metabolism in the body of
a steer produces a carcinogenic
metabolite that is not produced byDES
metabolism in the mouse, the results of
the Gass mouse study would not reflect
that metabolite. Thus, extrapolation
from the Cass study of DES could show
DES to be less carcinogenic to humans
than it actually is. Because the required
metabolism studies of DES do not
appear in the record, there is no basis
either for the calculation made by the
Bureaus experts or for any calculation
of the risks of present uses of DES.
(iv) Conclusion as to Quantitatire
Risk Assessment. I find that each of the
risk calculations for DES proffered by

54886
I

Federal Register
----

/

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

I

.

I

_

I

_

the manufacturing parties rests on
and thus would not be the subject of
breast and colon cancer in various
unwarranted assumptions and must be
published regulations. Even where
countries. Mr. Wynder did not testify at
rejected. The record does not provide
regulations are published, they show
the hearing and was thus not subjected
data that make possible a reasonably
only that a drug is approved. They say
to cross-examination on his conclusions.
well grounded calculation of the risk
.nothing about its comparative
I do not disagree with the general
from the presently approved uses of
effectiveness, cost, or availability. The
proposition that It would be a good idea
DES.
components of th6 FDA that have firstfor Americans to eat leaner meat,
hand knowledge about animal drugs are, though the record provides little support
(b) Introduction to Discussion of
Benefits. The discussion that follows
of course, not available to-me in making
for that proposition. Nothing in the
this decision.
deals first with the contention that DES
record, however, provides a basis for
The FDA has proposed to withdraw
use provides."health benefits" to society
determining how much of a fat reduction
approval of two potential substitutes for
by (1) decreasing the amount of fat in
would make a meaningful difference In
DES, Synovex-S and Synovex-H
the humandiet and (2)
saving food. I
the health of consumers. Without some
implants, 44 FR 1463 Uanuary 5, 1979).
then discuss the evidence in the record
basis in the record for a finding on the
Those products will, of course, be
that DES use provides an economic
amount of fat reduction needed to
available for some time until withdrawal achieve a positive effect on health, I
benefit to society. Because the argument
of their approval is accomplished.
that one should consider the "health
cannot reach any conclusion about the
Nevertheless, because the FDA is
benefits" of an animal drug in
benefit to health from fat reductions
determining its safety has some appeal, I sbeking to remove these growth
attributable to use of DES.
promotants from the market, they will
have considered the evidence in the
(ii)Effect of Withdrawal of Approval
not be considered a factor in the DES
record regarding claimed "health
of the DES NADA 's on Fat
benefits determination.
benefits" with especially great care.
Consumption.This question itself
(c) Health Benefits: Reduction in Fat. involves a large number of subquestions,
(The manufacturing parties make
The manufacturing parties and the propassing reference to a claimed health
Logically, the difference in theo amount
DES intervenors argued that the ban of
benefit from reduction to animal waste
equal the amount
DES would actually have adverse health of fat cohsumed would between the moat
(Manufacturing Parties' Exceptions at
of the difference in fat
consequences because the edible tissues of DES-treated animals and the meat of
178 n. *]. Preston's statement that
Dr.
of animals not fed DES'contain more fat
"there is potentially less animal waste!,
animals that would be iaarketed after
associated with DES use (M-124 at 4) is. than the tissues of DES-treated animals'
the ban of DES times the amount of beef
(see Manufacturing Parties' Exceptions
all the evidence to which I have been
at 175-77). As the following discussion - that would be consumed by human
consumers after a ban of DES plus or
cited on this question and I 6annot find,
illustrates, the manufacturing parties
on the basis of that single unsupported
minus the amount of fat that would be
have not supplied to this record
statement, that reduction irf animal
consumed by humans from alternatives
sufficient data to make'possible any
waste is a health benefit associated
to beef or lamb should the ban of DES
conclusions on this point.
with the use of DES.)
alter the consumption of those products
The question whether the ban of DES
One factor that the manufacturing
to any significant degree.
would result in significant adverse
parties seem to ignore is-the availability
(a) Amount of FatSaving in Meat,
health effects to the public because of
of alternatives to DES. If a claimed
Each of the factors mentioned itself
an increase in fat in the diet logically
benefit from the use of DES is also
depends on analysis of subfactors. Thus,
mustbe divided into.two
(1)
available from a potential substitute, it- How much of a difference questions:the
the amount of the difference in fat
in fat in
is appropriate as a matter of common
human diet will cause a difference in the between the meat of DES-treated
sense and logic to discount that benefit
animals and that of animals available to
healih of consumers? (2) How much
in determining whether the benefits of
the public after a ban of DES depends
difference in the fat consumed, by
DES outweigh its risks. (This practice is
on what alternatives there will be to the
human beings will result from the
followed by the Environmental
use of DES. It is, as a practical matter,
withdrawal of aproval of the DES
Protection Agency in the risk/benefit
meaningless to compare the use of DES
NADA's?
decisions it must make, see, e.g., 44 FR
siniply to the production of cattle and
(i) Relationship Between Fat Intake
15874, 15876 (March 15, 1979) (2, 4, 5-T);
sheep without DES. Producers
andHealth.The manufacturing parties'
43 FR 51132, 51135 (November 2,
1978)
attempt to answer the first, and simpler, predictably will-seek to maximize their
(eridrin).) The proponents of DES have
of these questions is unconvincing. They profits by turning to alternatives.
provided very little information to this
The most likely alternative to the use
rely soley on the statement of Dr. Jukes
record about the availability of
of DES would be the use, in its stead, of
(M-99 at 15-16) that a decrease in fat in
alternatives to DES.
alternative growth promotants. The
the diet reduces humeri exposure to
Information about alternative growth
government's environmental Impact
diseases such as cancer, heart disease
promotants is not readily available from
analysis (G-116) bases its conclusions
and diabetes (Manufacturing Parties'
sources of which I could appropriately
on the assumption that producers now
Exceptions at 177). Dr. Jukes referred to
take official notice. While NADA's
an article (M-107) that reviews a
using DES would switch to other
approved after 1969 are required to be
number of epidemiological reports
available growth promotants. (Cf. Gmade the subject of a'published
115, discussed below.) The
dealing with various cancers and-their
regulation, see 21 U.S.C. 360b(i), not all
-possible causes. The thesis of this
environmental impact statement (issued
previously approved drugs are the
in 1976) assumes the use of the two
review is that "over nutrition" is a.
subject of such regulations. Some animal prominent cause of cancer. The author,
Synovex products (under their chemical
drugs may,.in addition, be exempt from
Ernest L. Wynder, suggests that the
names-estradiol benzoate plus
the definition of "new animal drugs" or American public should consume a diet
testosterone'proprionate and estradiol
subject to its "grandfather" clauses, see. lower in calories, total fats, saturated
benzoate plus progesterone), Ralgro by
21 U.S.C. 321(w); Pub. L No. 90-399,
fats, and cholesterol than-its present
its chemical name (Zeranol),
Section 108(3)(1969). Such drugs need
diet. The basis for this recommendation
melengestrol acetate (MGA,, and
is apparently the differing incidence of
not be covered by approved NADA's
monensin.

Federal Register / Vol. 44, No. 185
A large number of alternative growth
promotants are mentioned in the record.
These include: Synovex-S implant (200
mg progesterone and 20 mg estradiol
benzoate) (PS-15, PS-20, PS-25);
Synovex-H implant (200 mg testosterone
propionate and 20 mg estradiol
benzoate) (PS-16, PS-44); Ralgro implant
(resorcylic acid lactone), 36"mg (PS-20,
PA-25. M-125 at 1419); monensinsodium (PA-31 at 6); Rumensin
(monensin) (an antibiotic) (PA-23 at
- 453); a feed additive consisting of
microencapsulated animal fats (not
approved by the FDA as of February
1976) (id.); an intravaginal device to
stimulate the expression of estrus in
heifers [id.; cf. M-51 at 30); estradiol 17b (PS-12); melengestrol acetate (MGA)
(PS-16, PS-44); dienestrol diacetate (PS19); hexestrol (dihydrodiethylstilbestrol) (id.); coumestrol (an
"isoflavonic estrogen" found in alfalfa)
(PS-25); zeranol in lambs (metabolic
effects) (PS-30); testosterone propionate
in lambs (PS-34]; chlortetracycline in
lambs {id.); reserpine in lambs (id.);
Smilagenin (a nonestrogenic substance)
(M-125 at 1419). The record does not
show that any of the above (other than
those products referred to in the
ehvironmental impactstatement) is or is
not now available or likely to be
available in the future as an alternative
-to DES. As discussed above, a notice of
opportunity for hearing has issued for
withdrawal of approval of both Synovex
products (i.e., Synovex-S and SynovexH).
DES is generally used in the raising of
steers (castrated male cattle), which are
easier to deal with than bulls and have,
in the past, been thought to provide
better tasting beef. One alternative to
the use of DES is a change in cattleraising practices. In the European
countries in which DES has been
banned, meat producers apparently do
not castrate bull calves; thus they raise
bulls rather than steers (M-64 at 24).
The bulls have available, as growth
promotants, natural hormones provided
by their testes that are comparable to
the amount of growth promotant added
to steers by the administration of DES
(id.). An expert witness for the
intervening parties, Dr. Donald R. Gill,
stated that his university had produced
publications favorable to the raising of
bulls (as opposed to steers), but that he
personally had had bad experiences
with large numbers of bulls fed in
commercial feed lots (Tr. at 2006-7).
Nevertheless, the raising of bulls is yet
another alternative that might be
utilized by cattle producers wishing to
maximize the growth of their cattle if
,DES were banned.

1 Friday,

September 21. 1979 / Notices

SM87

The next subquestion is what will be
MGA-DES-treated cattle are
the extent of the difference in fat
reported as having had significantly
consumed by the public If DES is
lower marbling scores than MGAreplaced by any of the alternative
treated groups (PS-16). (The decrease in
growth promotants. The record has little fat in the edible tissues of DES-treated
information on this question. Data on
animals apparently decreases what is
referred to as the "marbling score.' The
the following alternatives do appear in
the record:
decrease in the marbling score, in turn.
No growth promofantat all-Dr.
decreases the Department of Agriculture
Rodney L. Preston, a manufacturing
grade assigned to the meat products
parties' witness, testified that among the (PS-20 and 1211; see, generally, for
positive effects of the use of DES is the
present USDA grading regulations, 9
production of meat with more protein
CFR Part 53). Studies relevant to the fat
and less fat, a result that he
question thus sometime speak of
characterizes as "in harmony with
lowered marbling scores or lolvered
proper human nutrition" (M-124 at 3).
carcass grades.
Dr. Preston made no attempt to quantify
Dienestroldiacetate-A1955 report
the increase in protein or reduction in
states that DES-fed steers produced
fat to be expected in either cattle or
carcasses that were rated under federal
sheep.
carcass grades as slightly inferior to the
A review article by Dr. Preston states
carcasses from dienestrol-fed steers
that the effect of DES on carcass
(and particularly inferior to control
composition is related to the ratio
animals) (PS-19 at 332-33).
between dietary proteinoand dietary
Hexestrol-The same 1955 report
energy (apparently, calories). At a
found that DES-fed steers produced
certain ratio. DES can be expected, he
carcasses slightly inferior in federal
stated, to decrease the deposit of fat in
carcass grade to the carcasses of
the carcasses of lambs (M-125 at 1416hexestrol-fed steers (id.].
17). Again, no amount of decrease is
Ralgro-Onestudy showed that
given.
carcass grades with Ralgro treatment
The Administrative Law Judge cited
were similar to those resulting from DES
M-109 at 700 for the proposition that the treatment (PS-20).
reduction in fat content in treated steers
Testosteronepropionate--One
study
is less than I percent (I.D. at 19). He
showed that DES treatment of lambs
caused significantly lower carcass
apparently relied upon the estimated fat
in.total carcass composition reflected on grades than treatment with testosterone
Table 2 of that reporL The
propionate (PS-34).
Chlortetracyclineplus reserpinemanufacturing parties take the position.
which seems to be reasonable, that tle
These drugs, when administered
amount of fat in the muscle, as opposed
together, produced significantly higher
to the total amount of fat in the animal,
grades of carcasses of lambs than did
is important (Manufacturing Parties'
DES treatment PS-34).
Bulls as alternativesto steers-Bulls
Exceptions at 176). They go on'to argue
are reported as having less marbling in
that this report, because it shows
the lean meat than DES-treated steers in
increased body fat thickness (citing Mone study (PS-4). In another study, bulls
109 at 700. 701) and no increase in
overall body fat, demonstrated that DES were compared with steers in a test in
which half of the bulls and half of the
use resulted in decreased intramuscular
steers were treated with DES (24 ing in
fat (Manufacturing Parties' Exceptions
pellets for the steers and 60 mg in pellets
at 176). *
A large number of articles detailing
for the bulls]. The report states that the
tests with various levels of DES were
carcasses of both the treated and the
untreated steers were significantly
submitted to the record by the
higher in fat content than the carcasses
intervening parties (see, e.g., PS-16; PSof the untreated bulls (PS-35). A table in
17). Review of those articles shows that
the study shows that the carcass grades
DES does appear to decrease the fat
of the treated steers were higher than
content of the edible tissues of treated
animals, though the amount of decrease
the carcass grades of the untreated bulls
varies with the amount of DES used, the
and that th- percentage of carcass fat in
form in which it is used, the amount and the treated steers was greater than the
kind of feed provided to the animals and percentage of fat in the treated bulls (id.
the age at which they are slaughtered:
at Table 3). A subsequent evaluation of
Because the studies reported involved
animals from this study also found that
use of DES under conditions of use
steers generally had more abundant
different from the approved conditions.
marbling than did bulls (PS-36).
it is not possible to determine from these
None of the cited information gives a
articles Ifow much of a saving of fat in
real basis for a calculation of how much.
edible tissues occurs when DES is used
if any, saving in the tat content of meat
in accordance with its approved usies.
would result from the continued use of