1526.pdf
Media
- extracted text
-
Report of tho
Public IBorrd of
on
inquiry
DEPo - PPOVEM
T f""- *
■>
GiRlFF
PAUL
D
t
Ross
ETOLLOy
0
-AV
COMMUNITY HEALTH CELL
REPORT OF THE
PUBLIC BOARD OF INQUIRY
ON DEPO-PROVERA
Judith Weisz, MB BChir.
Chairperson
Griff T. Ross, M.D., Ph.D.
Paul D. Stolley, M.D., M.P.H.
TABLE OF CONTENTS
PAGE
I.
II.
THE PUBLIC BOARD OF INQUIRY: EVENTS LEADING TO
ITS ESTABLISHMENT, ITS MANDATE, AND PROCEDURES.
1
HISTORICAL BACKGROUND ON DEPO-PROVERA
8
1.
EXPERIENCE IN THE U.S.A
8
2.
EXPERIENCE WORLDWIDE
16
III. THE SEVEN MANDATED QUESTIONS
21
THE COMMISSIONER’S QUESTION #2:
WHETHER DATA FROM BEAGLE DOG AND MONKEY STUDIES
INDICATE A POTENTIAL RISK OF BREAST OR
ENDOMETRIAL CANCER IN HUMANS FROM DMPA
2.1. EXPERIMENTAL DESIGN AND EVIDENCE:
STUDIES
ANIMAL
2.2. CRITIQUE OF THE FINDINGS IN ANIMALS
2.2.1
2.2.2
22
24
30
CAN THE MALIGNANT NEOPLASMS
IDENTIFIED IN THE MAMMARY GLANDS OF
DOGS AND IN THE UTERI OF THE TWO
MONKEYS BE ATTRIBUTED TO MPA AND IF
SO, IS THERE EVIDENCE OF A
DOSE-RESPONSE RELATIONSHIP?. . .
30
WHAT BASIC DIFFERENCES HAVE BEEN
IDENTIFIED BETWEEN THE HUMAN AND
EACH OF THE TWO ANIMAL SPECIES IN
THEIR RESPONSE TO PROGESTOGEN? .
36
2.2.2a THE RESPONSE OF DOGS TO
PROGESTOGENS - EVIDENCE FOR
SPECIES DIFFERENCES. . . .
36
SUSCEPTIBILITY OF THE
BEAGLE TO MAMMARY TUMORS.
37
DIFFERENCES BETWEEN THE
HUMAN AND THE DOG IN THE
TYPE OF MAMMARY NEOPLASIAS
DEVELOPED
38
2.2.2a.1
2.2.2a.2
P A -j E
2.2.2a.3
2.2.2a.4
2.2.2a.5
MAJOR DIFFERENCES IN THE
COMPARATIVE ENDOCRINOLOGY
OF REPRODUCTION BETWEEN
DOGS AND HUMANS
40
DIFFERENCES IN THE
RESPONSE OF DOG AND HUMAN
MAMMARY GLAND TO
PROGESTERONE IN TERMS OF
STIMULATION OF GROWTH AND
DIFFERENTIATION
41
EVIDENCE FOR DIFFERENCES
BETWEEN THE BEAGLE AND THE
HUMAN IN THE SPECIFICITY
AND REGULATION OF
PROGESTERONE RECEPTORS. .
46
2.2.2a.6
EVIDENCE FOR THE PRESENCE OF
LATENT NEOPLASTIC FOCI IN THE
CANINE MAMMARY GLAND NOT
PRESENT IN THE HUMAN MAMMARY
GLAND
48
2.2.2a. 7
COULD GROWTH HORMONE,
RELEASED BY BEAGLES IN
RESPONSE TO HIGH DOSES OF
PROGESTERONE, BE
RESPONSIBLE FOR THE
INCREASED INCIDENCE OF
MALIGNANCIES IN THE MAMMARY
GLANDS OF DMPA TREATED
DOGS?
52
IS THE FACT THAT
PROGESTOGENS CAN CAUSE
REGRESSION IN SOME CASES OF
HUMAN BREAST CANCER A
REASON TO ASSUME THAT THEY
ARE UNLIKELY TO PROMOTE THE
DISEASE?
53
2.2.2a.8
PAGE
2.2.2a.9
2.2.2b
2.2.2b.1
2.3
SUMMARY OF EVIDENCE ON
POTENTIAL RELEVANCE OF THE
FINDINGS IN THE DOG TO THE
HUMAN
53
THE RESPONSE OF THE RHESUS
MONKEY (MACACCA MULATTA) TO
PROGESTOGENS - EVIDENCE FOR
SPECIES DIFFERENCES . . . .
55
SUMMARY OF EVIDENCE ON
POTENTIAL RELEVANCE OF
THE FINDINGS IN THE
MONKEY TO THE HUMAN .
64
ANIMAL TESTING FOR CARCINOGENESIS:
APPROPRIATENESS OF GUIDELINES FOR TESTING
AND INTERPRETING RESULTS WITH STERIOD
HORMONES TO BE USED AS CONTRACEPTIVES. .
65
THE COMMISSIONER’S QUESTION #3:
CAN THE HUMAN DATA SUBMITTED BY UPJOHN
SUCCESSFULLY REFUTE THE RISK OF HUMAN CANCER
SUGGESTED BY THE ANIMAL DATA?
3. 1
3.2
EXPERIMENTAL DESIGN AND EVIDENCE:
STUDIES
80
HUMAN
85
CRITIQUE OF THE EVIDENCE ON THE HUMAN. .
85
REVIEW OF HUMAN STUDIES OF BREAST
CANCER
88
REVIEW OF HUMAN STUDIES OF
ENDOMETRIAL CANCER. . . .
95
REVIEW OF HUMAN STUDIES OF CERVICAL
CANCER
101
3.2.1
3.2.2
3.2.3
COMMISSIONER’S QUESTION #5:
WHETHER, IN THE EVENT OF CONTRACEPTIVE FAILURE,
USE OF DMPA MIGHT INCREASE THE RISK OF TERATOGENIC
EFFECTS MORE THAN OTHER CONTRACEPTIVES?
115
PAGE
5.1
5.2
3.3
5.4
EVIDENCE ON THE TERATOGENICITY OF DMPA:
SOURCES OF INFORMATION
120
INFLUENCE OF PROGESTOGENS ON SEXUAL
DIFFERENTIATION: GENITAL STRUCTURES
124
INFLUENCE OF PROGESiOGdNS ON SEauaL
DIFFERENTIATIONS:
FUNCTIONS MEDIATED BY THE
CENTRAL NERVOUS SYSTEN (CNS)
132
PROGESTOGENS AND EXTRAGENITAL ABNORMALITIES
136
THE COMMISSIONER'S QUESTION #6:
WHETHER, IN VIEW OF DEPO-PROVERA'S ADVERSE SIDE
EFFECTS OR PHARMACOLOGICAL EFFECT, ESTROGEN
THERAPY IS LIKELY TO BE PRESCRIBED, IN ADDITION TO
DEPO-PROVERA, IN A SIGNIFICANT NUMBER OF
PATIENTS?
INFLUENCE OF DMPA ON PLASMA LIPIDS AND BONE
I.
147
THE COMMISSIONER'S QUESTIONS 1, 4 AND 7 REFORMULATED:
WHETHER THE RATIO OF BENEFIT TO RISK OF
DEPO-PROVERA WHEN USED AS A CONTRACEPTIVE WARRANTS
ITS APPROVAL UNDER CONDITIONS OF GENERAL MARKETING?
WHETHER APPROVAL FOR GENERAL MARKETING AS A
CONTRACEPTIVE MIGHT INCREASE ITS USE FOR UNAPPROVED
INDICATIONS? AND, IF NOT APPROVED FOR GENERAL
MARKETING, ARE THERE CONDITIONS FOR CONTROLLING THE
LIMITED DISTRIBUTION OF THE DRUG AS A CONTRACEPTIVE
FOR CERTAIN PATIENTS WITH SPECIAL NEEDS?
163
FINDINGS OF FACT
172
DATA AVAILABLE ON THE LONG-TERM RISKS OF DMPA ARE
INSUFFICIENT AND INADEQUATE TO PROVIDE A BASIS FOR
A DECISION WHETHER THE BENEFITS OF THE DRUG AS A
CONTRACEPTIVE OUTWEIGH ITS DISADVANTAGES UNDER
CONDITIONS OF GENERAL MARKETING IN THE USA
172
PAGE
II.
III.
IV.
DATA FROM THE STUDIES OF RHESUS MONKEYS AND BEAGLE
DOGS CAN NOT BE DISMISSED AS IRRELEVANT TO THE
HUMAN WITHOUT CONCLUSIVE EVIDENCE TO THE CONTRARY.
SUCH EVIDENCE IS NOT AVAILABLE AT THIS TIME.
THEREFORE, THE FACT THAT MALIGNANT NEOPLASIAS
DEVELOPED IN TWO SPECIES IN TARGET ORGANS OF SEX
STEROIDS MUST BE CONSIDERED AS AN INDICATION OF A
POTENTIAL OF PROGESTOGENS, INCLUDING DMPA, TO
PROMOTE THE DEVELOPMENT OF MALIGNANCIES IN TARGET
ORGANS
.
........
173
THE DATA ON THE HUMAN ARE INSUFFICIENT AND
INADEQUATE TO EITHER CONFIRM OR REFUTE THE
IMPLICATION OF THE ANIMAL DATA THAT DMPA MAY
INCREASE THE RISK OF CANCER IN WOMEN USING DMPA AS
A CONTRACEPTIVE
.
175
IN CASE OF CONTRACEPTIVE FAILURE WITH DMPA, THE
RISK OF A MOTHER GIVING BIRTH TO A MALFORMED CHILD
IS UNLIKELY TO BE MEASURABLY GREATER THAN THAT
POSED BY THE ORAL CONTRACEPTIVES.
THE CHANCE IN
EACH CASE CAN BE ESTIMATED TO BE SMALL ENOUGH NOT
TO POSE AN OBSTACLE TO THE USE OF THE DRUG AS A
CONTRACEPTIVE WHEN OTHERWISE INDICATED
176
CONCLUSIONS OF LAW
1 80
1
I.
THE PUBLIC BOARD OF INQUIRY: EVENTS LEADING TO ITS
ESTABLISHMENT, ITS MANDATE, AND PROCEDURES
On July 20, 197 9 r at the request of the Upjohn Company,
the Commissioner of the Food and Drug Administration (FDA)
announced that a hearing before a Public Board of Inquiry
(PBI) would be held
to determine whether Upjohn’s supplemental new
drug application (NDA) for Depo-Provera (DMPA)
sterile aqueous suspension for intramuscular
injection as a contraceptive agent in humans
contains reports of investigation adequate to show
that the drug is safe for use under the conditions
prescribed, recommended or suggested in the
labeling as required by § 505(d)(1), (2) and (4)
of the Federal Food, Drug, and Cosmetic Act (the
Act), and whether that information combined with
other information about the drug, provides a
sufficient basis from which FDA can determine that
DMPA is safe for general marketing in the United
States [ 1 ] .
The FDA had considered a supplemental NDA for the use of DMPA
as a contraceptive on two previous occasions. in 1974 and
1978
[2, 3] .
In 1978 the request was denied.
It is to
re-evaluate this decision that the PBI was appointed at the
request of Upjohn.
pp.
[For chronology of events, see Table 1 ,
11-15.]
This was only the second time that the FDA had ordered
a public hearing before a PBI under 21 CFR Part 13 in lieu
of a formal evidentiary public hearing before an
Administrative Law Judge under 21 CFR Part 12.
The previous
PBI had been established to consider the use of Aspartame as
a sweetener and flavor enhancer in dry foods [4].
Upjohn
2
stated its reason for requesting a Public Board of Inquiry to
be that the issues under consideration are "technically
complex, require an understanding of sophisticated scientific
concepts, and necessitate the comprehension and evaluation of
a large body of scientific literature and other data"
[5] .
The PBI was established in September 1981 and mandated
to consider and render an opinion on the following issues
[6] :
1)
Whether, in comparison with other drugs
approved for contraception, the benefits of
Depo-Provera in the United States outweigh its risks
under conditions of general marketing.
2) Whether data from beagle dog and monkey
studies submitted by Upjohn indicate a potential risk
of breast or endometrial cancer in humans from
Depo-Provera.
3) Whether the human data submitted by Upjohn
can, as Upjohn claims, successfully refute the risk of
human cancer suggested by the animal data.
4) Whether an approval of Depo-Provera for
contraception under general marketing conditions is
likely to increase use of the drug as a contraceptive
under conditions not stipulated in the approved
labeling or is it likely to increase use of the" drug
for unrelated indications for which safety and
effectiveness have not been established (e.g., for
hygienic purposes in mental retardees).
Whether, in the event of contraceptive
5)
failure r use of Depo-Provera may increase the risk of
teratogenic effects to a greater extent than would
other systemic contraceptives.
6) Whether, in view of Depo-Provera’s adverse
side effects or pharmacologic effect, estrogen therapy
is likely to be prescribed in addition to Depo-Provera
in a significant number of patients.
7) Whether there are conditions of labeling and
distribution controls which would permit marketing of
3
Depo-Provera as a safe
< ~ and effective drug on a limited
basis (i.e., Whether there imay be
'
“““ patients m
certain
the U.S. for whom benefits of
r_
t
^2 Depo-Provera for
contraception outweigh potential" risks).
(Emphasis added)
All of these issues were raised by the FDA in 1978 in
course of its review of the supplemental NDA for the use of
DMPA as a contraceptive [3]. On September 10, 1981,
we were
appointed by FDA Commissioner Arthur Hull Hayes, Jr.,
serve as members of the PBI [6],
to
We were acceptable to both
the Upjohn Company and the FDA because none of us had
taken
any position on DMPA as a contraceptive in the past.
In
addition, among us, we covered the range of expertise
relevant for evaluating the issues.
The process by which we have carried out our tasks
involved :
1)
A review of the documents in the administrative
record in this matter as well as published literature on
Depo-Provera.
As new questions arose we asked one or the
other of the parties to provide additional information
for
the record including an initial statement of position
[8r 9] .
All of the literature and documents we reviewed and on which
our decision were either already in or have
1/
At the request of the Upjohn Company Dr. Stolley did not
participate either in the review of the material
concerning the potential teratogenicity of DMPA or in
making the decision on this issue [7],
4
been placed in the administrative record and are accessible
2/
to the public.—
2)
A five-day public hearing held from January 10-14,
1983, where party and nonparty participants presented
testimony and responded to our questions.
Among those
represented at the hearing. in addition to the two parties,
the Upjohn Company and the FDA’s National Center for Drugs
and Biologies (now the Center for Drugs and Biologies)
(Center), were the World Health Organization (WHO), the
International Planned Parenthood Federation, the U.S. Agency
for International Development, the Women's National Health
Mo +- wnr I; f
the Health Research Group, the Institute for the
Study of Medical Ethics, the American College of
Obstetricians and Gynecologists, and certain individuals
representing themselves [10].
In addition, on August 12, 1 983 , a second hearing was
held to receive the report of a panel of six pathologists who
had been designated by us as consultants to review the
histological slides of monkeys that developed uterine tumors
following exposure to DMPA [11].
2/
In September, 1983,
1 983, when we began, the documents in the
administrative record in this matter occupied
approximately 45 linear feet of shelf space, At the
time of the submission of our report, the documents
occupy approximately 54 linear feet.
5
3)
Periodically, meetings were held to discuss the
evidence, arrive at decisions, and outline our
report.—3/
'
Writing our report.
4
In all of these phases we were ably assisted by Ms. Rita
Schinnar, research assistant to the PBI who, like □ s, was
appointed a special Government employee.
Our report and conclusions about DMPA have evolved from
these numerous encounters and discussions.
Throughout this
process, we have considered our primary task to be to
evaluate the scientific validity of the information
available.
We attempted to determine how much of this
information qualified as facts on which definitive
conclusions could be based.
We have attempted to identify
and record the facts and to differentiate them clearly from
assumptions and hypotheses.
Assumptions are an inevitable
3/
The entire Board met in Washington, D.C. on
September 22, 1982; in Washington, D.C. on January 9,
1983; in San Francisco, CA. on March 6-7, 1983; in
Hershey, PA. on August 13-14, 1983; and in Washington,
1 984.
u.c. on August 10-11, 1984.
Dr. Weisz and Dr. Scoiley
met in Philadelphia, PA. on August 31, 1982; Dr. Weisz
and Dr. Ross in Washington, D.C., on August 10, 1983,
and in Houston, TX. on November 25-27, 1983; and
Dr. Stolley and Dr. Ross in Houston on June 13,
1 984 .
A/
Because of illness, Dr. Ross was unable to participate
in this fourth phase.
He was a full participant in the
earlier phases.
(See his letter to Dr. Weisz,
p. 181.)
6
and necessary part of science.
But, no matter how widely
accepted, they are subject to change as knowledge expands.
Hypotheses, however plausible, can not serve as evidence
unuxl tested and proven correct.
To qualify as a fact.
information and data have to pass the test of having been
generated and analyzed by adequate scientific methods.
We
recognize that the significance of facts is subject to
interpretation and that this may vary among individuals and
also change with time as knowledge expands.
In making the
recommendation on Depo-Provera, we strove to identify and
record the basis for each of the conclusions.
The report
includes, in addition to an extensive list of references, a
series of appendices (Appendix 1-5).
In these appendices are
detailed the data and the methods used to generate the data,
as distinct from their interpretation, for certain kev
publications and documents on which our conclusions are
based.
This should permit the reader to judge the factual
basis on which our evaluation of the data and our conclusions
rest.
Our findings and conclusions will have the legal sta tus
of an initial decision under 21 CFR 13.40(a)(4) and 12.120.
The initial decision becomes the final decision of the
Commissioner by operation of law unless a participant files
exceptions or the Commissioner files a notice of review.
CFR 12.120(e).
The Commissioner’s final decision is
reviewable by the courts.
21 CFR 12.140.
21
7
The structure of our report is as follows:
after a
brief historical overview of the chronology of events
relating to Depo-Provera, we address each of the questions
ented to us by the Commissioner.
We have, however,
reordered the sequence in which the questions are addressed,
so that the scientific evidence available for assessing the
risks associated with the use of the drug, required to arrive
at a regulatory decision, is presented and evaluated before
considering the regulatory decision itself.
Thus, answers
relating to the carcinogenic and teratogenic potential of the
drug (Questions #2, #3 and #5) are presented first.
These
are followed by a discussion of the issue of the possible
increased use of estrogens that introduction of DMPA as a
contraceptive may cause (Question #6) and by a separate brief
section in which we address the question of the influence of
MPA on bone and on plasma lipoproteins.
The latter two
issues, though not specifically addressed in the questions by
the Commissioner, are ones that have been raised in the
hearing process and that need to be considered in relation to
the use of DMPA as a contraceptive.
Recommendations
concerning marketing of DMPA and the consequences of making
the drug available as a contraceptive under conditions of
general marketing are presented last (combining answers to
Questions #1, #4 and #7).
After this detailed statement of
the reasoning on which our conclusions are based, we conclude
with specific findings and references supporting and
8
explaining our conclusions, as required by 21 CFR
13.?0(j).
II.
HISTORICAL BACKGROUND ON DEPO-PROVERA
1.
EXPERIENCE IN THE U.S.A.
(SEE TABLE 1r pp. 11-15)
Depo-Provera (or Depot Medroxyprogesterone Acetate or
DMPA) was developed by the Upjohn Company in 1954.
11 was
initially intended for use in the treatment of endometriosis
and for habitual or threatened abortions.
When the FDA had
approved Depo-Provera for use for these two indications in
1960 r it did so in the absence of randomized controlled
studies of efficacy.
By 1973, the FDA rescinded the earlier
approval of the drug for use in endometriosis and threatened
abortion due to lack of evidence of efficacy and concerns
about potential teratogenicity and delayed return of
fertility [12].
In 1967, the Upjohn Company submitted a supplemental new
drug application for Depo-Provera (NDA 12-541/S004) for
intramuscular injection as a contraceptive agent in humans.
In the late 1960's and early 70*s a large number of studies
were carried out to establish the efficacy of Depo-Provera
and to identify an appropriate dose level.
These early
studies were also designed to provide data on physiological
consequences and side-effects following short-term use of the
”n; e.g., weight change, prolonged amenorrhea or bleeding.
9
return of ovulation, and psychological effects [13-20].
In
these studies. observations on pregnancy outcome in women
who discontinued use of Depo-Provera, or on the pathology of
breast, cervical or endometrial biopsies in exposed women,
were made only incidentally.
Concurrent with these clinical studies, the two
long-term toxicity studies in animals (a 7-year study in
Beagle dogs and a 10-year study in Rhesus monkeys) were
initiated in 1968.
A second 7-year study in dogs was started
in 1972.
During March, April and May 19‘74, in the course of a
Congressional investigation of the use of advisory
committees by the FDA, new concerns were raised about the
drug’s safety; specifically, the possible increased risk of
cervical cancer in women using the drug for contraception
[2] .
This led to the FDA’s first announcement on
October 30, 1974, of its intention to deny approval of the
supplemental NDA for Depo-Provera [2].
Results from the two long-term studies in beagle dogs
became available in 1975 and 1979, respectively.
Both
an increased incidence of malignant mammary
tumors in dogs treated with DMPA at low and high dose levels
[22-24] .
These findings and the lack of human data to refute
them led the FDA to propose to refuse approval of the
supplemental NDA for DMPA in 1978
[3].
By 1979, the results
from the long-term study in Rhesus monkeys became available.
10
These showed endometrial carcinomas in two of the treated
monkeys [24].
Numerous reviews of the findings in the test
animals and of the published literature on Depo-Provera were
ken by various organizations in the U.S.A, and other
countries to try to assess the relevance of these findings to
humans [25-29].
As previously discussed, in March 1978 FDA
proposed to refuse to approve Upjohn's supplemental NDA for
the use of DMPA as a contraceptive, and Upjohn requested a
hearing before a Public Board of Inquiry.
It should be added that since 1972, use of Depo-Provera
has been approved by the FDA as adjunctive therapy and
palliative treatment of inoperable, recurrent and metastatic
endometrial carcinoma.
11
TABLE 1:
CHRONOLOGY
1 954 :
Depo-Provera developed by the Upjohn
Company
i 5do:
Clinical trials oegun testing its use ior
habitual or threatened abortions and
end om e t r i o s i s.
1 960 :
FDA approved the use of Depo-Provera for the
above indications.
April 1966:
FDA withdrew from the market the drug
Promone (DMPA formulated for veterinary use)
because of an increased incidence of cystic
endometrial hyperplasia in treated dogs.
Feb. 27,
1 967 :
Upjohn Co. submitted a supplemental new drug
application for Depo-Provera for
intramuscular injection as a contraceptive
agent in humans. This supplement was
submitted to allow general marketing of this
drug for contraceptive purposes.
May 1968:
Two long-term animal studies initiated: a
7-year study of Depo-Provera in beagle dogs
and a 10-year study in rhesus monkeys. At
the same time, clinical studies were
initiated in several medical settings in this
and other countries to test its efficacy and
side effects when used as a contraceptive.
1971 :
Results of the two-year toxicity studies of
Depo-Provera in mice and rats were submitted
to FDA. They were accepted by FDA although
the results on the mice were uninterpretable
because of the high mortality of animals in
both the control and experimental groups, and
inadequate pathological examination of such
animals and in spite of the small number of
rats studied [30] .
1972-1978:
Depo-Provera used at the Grady Clinic under
a formal IND status.
1972:
FDA approved the use of Depo-Provera as
adjunctive therapy in treatment of
endometrial carcinoma.
12
1 972:
A second 7-year study in beagle dogs
started because of high mortality among cogs
in first study attributed primarily to
pyometra.
Oct.
10 ,
1 973 :
FDA issued a proposal to provide patient
labeling for Depo-Provera for contraceptive
use.
Having considered the potential
complications of benign and malignant tumors
of the breast (as already demonstrated at
that time by the first studies of beagle
dogs given Depo-Provera) and of infertility,
the FDA was still of the opinion that the
benefits from use of the drug outweighed its
risks.
Oc t.
10, 1 973 :
FDA’s Obstetrics and Gynecology Committee
(meeting on Feb. 22, 1973) recommended
approval for a limited and well-defined
population; and the FDA proposed, as a
cautionary measure, to restrict the
44sibution of the drug by the manufacturer
to private practitioners, family planning
clinics, hospitals and retail pharmacies.
This was intended to maintain a registry of
physicians who utilized the drug for
contraception. The objective of this
restriction of the distribution was that " in
the event evidence appears in the future mat
the tumorigenic effect of Depo-Provera poses
an increased risk of breast tumors in the
human, direct notification of these
physicians can be made and appropriate
patient follow-up instituted [12]."
Oct.
10,
1 973 :
Concurrently with the above, the FDA
rescinded the earlier approval for use in
endometriosis and threatened abortion due to
lack of evidence of efficacy for these
indications and concerns about potential
teratogenicity, specifically, congenital
heart defects.
March, April,
May, 1974:
During Congressional Hearings on the use of
advisory committees by the FDA, the issue of
cervical cancer in users of Depo-Provera was
raised [21] .
13
Sept.
12,
1 974: FDA issued a final order providing for
patient labeling for DMPA. This order was
published in anticipation of the agency's
approval of the pending supplemental NDA for
use of Depo-Provera as a contraceptive.
Oct. 2, 1 974:
Letter sent by the Chairman of the
Subcommittee on Intergovernmental Relations,
Rep. L.H. Fountain, of the House Committee of
Government Operations to the Secretary of HEW
requesting that the FDA revoke approval of
Depo-Provera because of unresolved questions
concerning the drug’s safety and its role in
causing carcinoma of cervix [31],
Oct. 30, 1 974 :
FDA stayed the provisions of the regulation
until further notice as a question was raised
about the incidence of cervical cancer a.nong
women who par11cipated Tn studies sponsored
by Upjohn under an IND.
May 1975:
Results from the first beagle study became
available and showed development of mammary
tumors, including carcinomas, in treated
dogs [22].
March 7 , 1 978 :
FDA advised Upjohn that the supplemental NDA
was not approvable because of insufficient
information to determine whether Depo-Provera
was safe for general marketing in the United
States as a contraceptive. Specific reasons
given for this decision included the
increased incidence of mammary carcinoma in
beagle dogs; the lack of a significant
patient population in the United States in
need of this drug; and the belief that the
manufacturer would not be able to implement a
proposed post-marketing study in the United
States to assess potential serious risks of
the drug. At this time, presumably, FDA had
been informally advised of the results of the
second beagle study, but the final report was
not submitted until 1982.
14
March 28 ,
‘’
-1978: Upjohn challenged' the
FDAJs decision to
refuse approval of the supplemental NDA for
Depo-Provera and requested an opportunity for
a hearing. Upjohn waived its right to a
formal evidentiary public hearing before an
administrative law judge, and requested
instead that a Public Board of Inquiry be
established.
Oct. 24 r
1 978 :
Dec.
FDA audit of the conduct of the IND study of
Depo-Provera at the Grady Clinic [32].
This
terminated the use of the drug as a
contraceptive by the single largest known
population of subjects at a single center in
the USA.
1 978 :
May 1 r 1 979 :
July 27,
Sept.
FDA Commissioner accepted Upjohn’s request
for a hearing before a Public Board of
Inquiry.
1979:
Findings from the 10 year study of the
effects of Depo-Provera on Rhesus monkeys
became available.
It showed endometrial
in
two
of
the
monkeys in the high dose
cancer
This
was
followed
by reviews by
group.
several national and international
■organizations of the relevance to the human
of the findings in the long term studies of
monkeys and dogs [25-29].
FDA announced a hearing before a Public Board
of Inquiry to determine whether the
supplemental NDA for Depo-Provera contained
reports of investigations adequate to show
that the drug was safe for use as a
contraceptive, and whether that information
provided a sufficient basis from which FDA
could determine that Depo-Provera was safe
for general marketing in the United States
for contraception.
10, 1981: FDA formally established a Public Board of
Inquiry, consisting of a three-member panel,
assisted by a legal counselor and a research
assistant.
Sept. 23,
t
‘
*"
j was held to
1982: A prehearing
conference
establish procedures for the public hearing.
15
Jan _ 10-14,
1 983 :
Aug.
12, 1 983 :
The Public Board of Inquiry held a 5-day
public hearing on Depo-Provera, receiving
testimony and responses to questions from
Upjohn and the FDA (the parties), and from
non-party participants.
The Public Board of Inquiry held an
additional one-day public hearing on
Depo-Provera, to receive the report of an
especially-appointed panel of 6 consulting
pathologists who reviewed the histopathology
slides from the monkey study to determine the
nature and origin of the malignant tumors in
2 of 12 survivor monkeys in the high-dose
Depo-Provera group.
The appointment of the special panel of
pathologists was delayed until information
could be obtained from Upjohn about what
tissue blocks or sections from the monkey
study were available for this independent
review. In addition, slides were obtained
for review from a control monkey in the IUD
study of the Population Council that had been
diagnosed to have endometrial cancer.
Presently:
Depo-Provera is approved as adjunct therapy
for patients with metastatic endometrial
cancer.
Oct. 26, 1 984:
P.B.I. decision submitted to the
Commissioner, FDA, and placed on file in the
Dockets Management Branch. The findings and
conclusions of the Board, based on the public
hearing, will have the legal status of, and
be handled as, an initial decision.
16
2.
EXPERIENCE WORLDWIDE
Depo-Provera has been used extensively as a
rnnf-rarpptive in over seventy countries.
According to the
manufacturer’s statistics [33], an estimated 11 million
women have "ever used" DMPA; 2 million women are "current
users"; 100,000 women used it for 10 years or longer, and
1.5 million women received their first injection over 10
years ago.
We have been cognizant of the great interest by and
impact on countries overseas in a decision on this drug for
in the United States.
At the same time, we are
required by law and have the obligation to evaluate the
benefit/risk ratio that is applicable to this country only.
17
REFERENCES
1.
44 Fed. Reg. 44274 (July 27, 1979).
DMB Vol. No. 20 - Tab #1 Announcement of Hearing or.
Depo-Provera.
2.
39 Fed. Reg. 38226 (October 30, 1974)
DMB Vol. No. 188 - Tab #93, pp. 446-447
3.
43 Fed. Reg. 28555 (June 30, 1978)
DMB Vol. No. 188 - Tab #93, pp
pp.. 193-194
4.
44 Fed. Reg. 31716 (June 1, 1979)
Hearing on Aspartame.
DMB Vol. No. 20 G-1
5.
The Upjohn Company, Letters of August 25, 1978 and
October 17, 1978
DMB Vol. 1 Sup; HER # 4
6.
DMB Vol. No.
7.
Correspondence with the Upjohn Company over Dr. PaLl
Stolley's appointment. Confidential.
DMB Vol. No. 169
Tab #218
8.
PBI letter to the Upjohn Company, May 5,
DMB Vol. No. 238 - Tab #207
9.
PBI letters to the National Center for Drugs and
Biologies, September 9, 1982, and March 18, 1983.
DMB Vol. No. 245 - Tab #234
10.
Official Transcript of Proceedings, Depo-Provera
Public Board of Inquiry, January 10-14, 1983, Volumes
I, II, III, IV, V. (hereinafter, TR-I, II, III, IV, or
V) .
11.
Official Transcript of Proceedings Public Board of
1983 (hereinafter TR-2d).
J t, r-iugus t 12,
12.
38 Fed. Reg. 27940 (October 10, 1973)
DMB Vol. No. 188 - Tab #93, pp. 195-201.
13.
Coutinho EM, DeSouza JC, Csapo Al: Reversible
Sterility Induced by Medroxy-progesterone Injections.
Fert Steril 1966, 17:261-266.
DMB Vol. No. 100A G-513
Announcement of
189
1 982.
18
14 .
Azcona SC: MPA (6 Alpha-Methyl - 17
Alpha-Hydroxyprogesterone 25 mg. plus Estradiol
Cypionate 5 mg.) as Anticonceptional of Monthly
parenteral Application at Low Doses, Oct. 25, 1976.
(Unpublished)
DMB Vol. No. 68A G-512, pp. a/000056-73.
15.
Dodds GH: The Use of Sterile Medroxyprogesterone
Acetate Suspension as a Contraceptive During a
Three-year Period. Contraception, 2:15-23, 1975.
DMB Vol. No. 74A G-512, pp. 319-327.
16.
Schwallie PC, Assenzo JR: The Effect of
Depo-Medroxyprogesterone Acetate on Pituitary and
Ovarian Function, and the Return of Fertility
Following Its Discontinuation: A Review.
Contraception, 10:181-202, 1974.
DMB Vol. No."
No. 3 3 6 - Tab #461.
17.
Mishell DR, El-Habashy MA, Good RA, Moyer DL:
Contraception with an injectable progestin. Am J
Qbstet Gynecol, 101:1046, 1968
DMB Vol. No. 53A G-512
18 .
Depo-Medrox/progesterone
Seymour RJr Powell LC:
Contraceptive
. Qbstet Gynecol/
Acetate As a
1
970.
36 : 589-596,
DMB Vol. No. 176 - Tab #67/ PP- 12-19.
19.
Experience with Depo-Provera as an
Schwallie PC:
Injectable Contraceptive. J Reprod Med, 13:113-117,
1 974 .
DMB Vol. No. 1 35A
PP- 58-62.
20.
Zanartu J: Long-term contraceptive effect of
Inhibition and
injectable progestogens:
Inti J Fertil 13:415,
re-establishment of fertility.
1968.
DMB Vol. No. 53A G-512, PP- 126-151
House of Representatives, Hearings before a
Subcommittee of the Committee on Government Operations
on the Use of Advisory Committees by the FDA, March,
April, May, 1974, pp. 323-341, 353-390.
DMB Vol. No.
NO. 32 G-358
22
IRDC, Long-Term Depo-Provera Study in Dogs, Final
Report, 1976.
DMB Vol. No. 5
19
23.
Dawson Corporation, Long-Term Depo-Provera Study in
Dogs, Final Report, 1982.
DMB Vol. No. 246 - Tab #252
24.
IRDC, Long-Term Depo-Provera Study in Monkeys, Final
Report 1979
Tab #67
DMB Vol. Nos. 181-182
25.
The Upjohn Company, Report of the Workshop to
review the monkey endometrial findings and their
relevance to the contraceptive use of Depo-Provera,
Feb. 12, 1981. (See 3/8/82 PBI Update, Sec. U-13)
DMB Vol. No. 211 - Tab #192
26.
World Health Organization, Special Programme of
Research, Development and Research Training in Human
Reproduction:
Statement on Safety of the Long-Acting
Injectable Contraceptive Depo-Provera, October 1978.
DMB Vol. No. 211 - Tab #192, Sec. U-1
27.
International Planned Parenthood Federation:
Statement of the International Medical Advisory Panel
on Depo-Provera, November, 1980.
DMB Vol. No. 295 - Tab #321 Section III
28.
United States Agency for International Development:
Report of Ad-Hoc Consultative Panel on Depot
Medroxyprogesterone Acetate. July, 1980.
DMB Vol. No. 288 - Tab #321
29 .
American College of Obstetrics and Gynecologists
Subcommittee on Reproductive Endocrinology: Depot
Medroxyprogesterone Acetate Use as a Contraceptive.
Position paper submitted to the FDA for the PBI.
March 4, 1982.
DMB Vol. No. 238 G-196
30.
Depo-Provera; 18 month toxicity study on mice; 24 month
toxicity study in rats.
DMB Vol. No. 337 - Tab #461
31 .
Letter from LH Fountain, Chairman, Subcommittee on
Intergovernmental Relations of the House Committee on
Government Operations, to HEW Secretary Casper
W. Weinberger, October 2, 1 974.
DMB Vol. No. 91A - Tab #4 G-512
32.
FDA audit of Grady Clinic, Dec. 18-22, 1978.
DMB Vol. No. 305 G-625 (summary), and DMB
Vol. No. 305A (synopsis)
20
33.
The Upjohn Company, Information for PBI on
Depo-Provera: Responses to the Board’s Questions.
June 25, 1982, p. 1.
DMB Vol. No. 244 - Tab #219
21
III.
THE SEVEN MANDATED QUESTIONS
The sequence of the answers to the seven questions that
we were mandated to consider has been reordered as follows:
Questions #2, #3, #5, and #6 are addressed first. and
Questions #1 , #4 and #7 were combined and reformulated and are
addressed last.
Thus, the evidence on the potential risks on
the use of DMPA, relevant to the regulatory decision on the
approval of the drug as a contraceptive under conditions of
general marketing,
is presented before our response to the
more general questions concerning the regulatory decision
itself and its potential consequences.
A brief section on the
evidence available on the influence of MPA on bone and on
circulating lipids has also been included because of specific
questions raised during the course of the present inquiry.
The section has been inserted between the response to Question
#6 and that to Questions #1, #4 and #7.
22
THE COMMISSIONER’S QUESTION #2:
WHETHER DATA FROM BEAGLE DOG AND MONKEY
STUDIES INDICATE A POTENTIAL RISK OF
BREAST OR ENDOMETRIAL CANCER IN HUMANS
FROM DMPA.
Malignant neoplasias developed in the mammary glands and
uteri of dogs and monkeys, respectively, the two species used
for testing the consequences of long term administration of
DMPA.
To determine whether these findings are relevant to
the human,
the following three questions need to be
addressed.
D
Is there evidence to support the assumption th a t th e
malignancies that developed in secondary sex structures of
dogs and monkeys are due to factors otherthan DMPA?
In the
absence of such evidence the malignancies would have to be
considered drug-related until proven otherwise.
2)
If the malignancies in the secondary sex structures
are to be considered drug-related, is there evidence of a
dose-response relationship?
Dose-response relationship is an
important parameter in all toxicological tests .
It is of
special importance when testing steroid hormones since it has
been suggested that the actions of very large doses of
steroids may differ not only quantitatively but also
qualitatively from those of smaller. more physiological
amount (see testimony of Dr. C. Wayne Bardin [1]).
23
Quantitative differences refer to the usual dose-response
relationship, i.e. , a systematic change in the incidence or
In
magnitude of an effect across the whole dose range.
contrast, if differences are qualitative, certain effects
would be restricted to a certain dose range, implying a
different mechanism of action at different doses.
Within the
context of dose response relationship, it is also necessary to
consider comparative data available for the dog, the monkey
and the human pertaining to factors that influence the
effective dose, i.e., the activity and the amount of the drug
reaching target organs (e.g., metabolism of the drug,
metabolic clearance rate and steroid receptor function).
3)
Have fundamental differences been identified between
the human and the monkey, or the human and the dog in their
responses to progestogen that would warrant a conclusion that
the carcinogenic effects of DMPA seen in either of the
species of test animals are unlikely to occur in the human?
At a more general level a fourth question raised relates
to the guidelines set by the FDA for testing of contraceptive
steroidal hormonal agents in animals.
Specifically,
there is
toxicological principles that have been
to set up these guidelines may not be valid for testing
steroid hormonal agents.
The complexity of the issues
involved may be gauged from the transcripts of the
proceedings of a three day conference convened recently by
the National Institutes of Health on the subject [2].
24
2.1 . EXPERIMENTAL DESIGN AND EVIDENCE:
ANIMAL STUDIES
Two long-term toxicity studies in beagle dogs to test
the safety of DMPA were initiated in 1968 [3] and 1972 [4],
respectively.
In both studies. DMPA caused a high incidence
of mammary nodules and mammary carcinomas.
In addition,
these neoplasias tended to appear at ages younger than the
age when mammary tumors occur spontaneously in dogs (see
Table 2A and Table 2B for details).
The association between
DMPA and mammary pathology was evident in animals treated
with low as well as with high doses of the drug.
In the first study, carried out by The International
Research and Development Corporation (IRDC)
[3], a large
number of the DMPA treated animals died early in the course
of the experiment.
This was attributed primarily
to
pyometra, an expected complication of progestogen treatment
in dogs.
The major relevant finding from this study was tha t
none of the 11 controls that survived until the end of the
study developed malignant mammary tumors, whereas two of the
dogs treated with 25 times the human dose of DMPA developed
mammary adenocarcinomas, with metastases. after 40 to 42
months of treatment (Table 2A, p. 26)
[3, 5-7].
Because of
the premature death of a large number of animals, a second
study was initiated.
In the second study on dogs, carried out by the Dawson
Do c a a v
Corporation [4], the bitches used were younger and
25
were hysterectomized at the beginning of the study to avoid
the "-complication of pyometra.
The number of animals in each
group was also larger (20 per group), and the effects of DMPA
were compared not only with vehicle injected controls but also
with two groups of dogs receiving different doses of
progesterone.
As in the earlier study, high doses of
progestogens proved lethal to many dogs even when
hys terec tomi zed.
With respect to mammary tumors, in this study also, none
of the control animals had mammary cancer, whereas malignant
mammary neoplasias (some with metatases) developed in 5 dogs
in the group receiving DMPA at XI human dose.
in 8 dogs in
the group receiving DMPA of XI0 human dose. and in 8 dogs in
the group receiving DMPA at X25 human dose.
Mammary cancer
developed also in 1 dog in the group receiving low dose
progesterone and in 2 dogs receiving high dose progesterone.
Onset of the mammary tumors in the high dose DMPA group was
at 46.9 months from start of the study compared to onset at
65.6 months in the lowest dose DMPA group (Tables 2B and 2C,
pp. 27, 28)
[4-8].
finally, a ten-year study of Depo Provera in Rhesus
monkeys was initiated in 1968.
It was carried out by the
International Research Development Corporation (Table 3,
p. 29)
[5, 9].
The major finding was the occurrence of
endometrial carcinomas in two among 12 surviving monkeys in
the high-dose group.
In addition, nonmalignant mammary
TABIAJ 2A:
Suiiuiary of Animal Exper unent.s in Dogs
IRDC STJDY (1968 - 1975) [31
y Design:
Its:
-36 female beagle cogs divided
into three experinental groups:
16 controls,
4 low dose DMPA (1 X HD),
and 16 high dose I MPA (25 X IB)).
-Individual dosages were adjusted
based upon changes in body weight
prior to each injection.
-the dogs were "mature” (age unspecified),
and were not Hysterectomized at the
start of the study.
-An increased incidence of
mortality and of papable mainnary
nodules at both dose levels of
DMPA, compared to controls.
survival:
-11 of 16 controls survived the 7-year
study period.
-2 of 4 in the low-dose DMPA group
survived to the end of the study.
-All lb dogs in the high-dose DMPA group
died by 3 1/2 years from onset of tne
study.
pathology:
-5 maimiary gland adenocarcinomas,
described as being widely
metastasized, found in 2 dogs
receiving higli dose DMPA. 'lhese
dogs develo[>2d nulignant maninary
tumors idml 11 ud at 40 and 42
months from onset of the study.
-See Table 2B for details
LAWSON b'HJUY (1972 - 1979) [41
-140 female beagle dogs divided
into six experimental groups:
40 controls,
20 low dose DMPA (1 X HD),
20 mid dose DMPA (10 X HD),
20 high dose IMPA (2s X HD),
20 low dose progesterone,
and 20 in high dose progesterone.
-Individual dosages were adjusted in
the mid and high DMPA groups (from OX
to 10X-HD and from 23X to 25X-HD)
at the 5th year of the study.
-The dogs were about six nnnths
of age, and were all hysterectomized
prior to the start of the study.
•i
-High mortality in the mid and high dose
groups aixl in the high dose
progesterone group,
survival:
-30 of 40 controls survived tne 7-year study period.
-14 of 20 low-dose DMPA;
-0 of 20 mid-dose DMPA;
-1 of 20 high-dose DMPA;
-17 of 20 low-dose progesterone; and
-0 of 20 high-dose progesterone.
pathology:
-Maiiinary malignancies did not develop
in any control animal.
Mamiiary adenocarcinana and solid cell carcinoma
developed in 5 dojs in low-dose DMPA group,
8 dogs in the medium-dose DM1 ’A group,
8 dexjn in the high-dose DMPA group,
1 dog in the low-dose Prog esterone group, and
2 dojs in the niejh-dose Prog<»sterone group.
-!Dt all nulignant tumors with metastases.
•t: Tabl
<h-l
iki.
I
K)
cn
I
Table 2B:
Details from 1HPA Studies in Beagle Dojs,
(based on Testimony by Gross p. 2b) [7]
Dawson Study (1972-1979)
Controls
CMPA
Progesterone
lOx-HD 25x-HD
(Medium) (High)
Low
High
lx—I ID
(Low)
No. dogs/group
40
20
20
20
20
20
No. dead in 7 yrs.
10/40
6/20
20/20
19/20
3/20
20/20
Mean no. months to death
54.9
67.3
67.3
59.1
75.7
44.9
No. with nuimary malignancy
0
5
8
8
1
2
Mean no. months to onset
of nummary mal ignancy
0
65.6
50.1
4b.9
75.8
44.6
. i
I
NJ
Incidence of benign
inainnary tumors
10%
Mean jo. of all mainmary
and tunors per dog
0.10
95%
85%
85%
85%
40%
I
7.3
6.3
5.3
1KDC Study (1968-1975) 13]
CMPA
Controls
IxHID
(Low)
lOx-DH
(Medium)
25X-HD
(High)
No. dog s/g rotps
16
4
16
No. dead in 3.5 yrs.
5/16
2/4
16/16
No. with inaninary inal ignancy
0
0
2
F^Mn n>. months to ons<‘t
0
0
40-42
3.3
2.6
?8n-012h/SUP )32
>C
TABLE
Long-Term Depo-Proveia Study
Tab #295 Vol . 26^4
In Doga
DKG 6205
MALIGNANT MAMMARY NEOPLASMS
SUMMARY
<•
Type
of
Turaor
Doss
Group
Low Dose
Solid cell carcinoma
Depo-Provera
<
I
i
Adenoc arcinoma
Malignant mixed tumor
HediUM Dose
Depo-Provere
Solid cell carcinoma
Adenocarcinoma
Solid cell carcinoma
High Dose
Depo-Provera
Adenoca rcinoma
Low Dose
Progea t erone
High Dose
Progesterone
KEY:
A
Note:
Time
of
Onset
Anima 1
Numbe r
Spindle cell adeno
ca rc i noma
Solid cell carci noma
Spindle cell adeno
ca rc 1 noma
Xdenocat tinoma
X-41-72 9
X-43-72 $
X-44-72 9
X-47-72 9
X-59-72 9
X-47-72 9
X-47-72 9
X-61-72 9
X-62-72 9
X-6B-72 9
X-69-72 9
X-72-72 9
X-76-72 9
X-78-/2 9
X-79-72 9
X-69-72~9~~
X-79-72 9
X-R5-72 9
X-88-72 9
X-89-72 9
X-9O-72~9
5(^9 4-7 2 9
X-99-72 9
X-BI-72 9
X94-72 9
X-92-72 9
+ 2212
♦2130
♦ 2157
♦ 2403*
♦ 2 338
♦ 1940
♦ 1151
♦ 1701
♦ 1757
♦ 1276
♦ 1429
♦ 1458
♦ 1578
♦ 1821
♦ 1241
♦ 1366
♦ 1522
♦ 1394
♦ 1187
♦ 191 I
♦ 1487
♦ 1577
♦1369 I
♦ 1212
Month
of
Deal h
AS
Nalu re
of
Malignant Neoplasm
i *. n
1 i s t i in j <i< i m d I n<j
Death
♦ 1549
66
71
60
Incident a 1 Finding_
_ I ncident al F i n ding
_ I nciden t a 1 Finding
_ Mejiasl.ilic Disease
Inc>dental Finding_
_ Incidental Finding
Iucident a I Finding_
Inc 1 dent al F i nding
Incidental Finding
Met astatic D1s ea s e
Met astatic Disease
Metastatic Disease
ncident al Fin ding
Incidental F i nding
Metastatic Disease
Incidental Finding
Incidental Finding
Incidental Finding
Incidental Finding
Met as t atic Disease
Metastatic Disease
Incidental Finding_
Mel astatic Disease
Metastatic Disease
Incidental Finding_
1 nc ident a 1 Finding
X-l14-72 9
♦ 2107
"fl 2
Metastatic Disease
E (Metastatic l)i a ease)
X-122-72 9
♦ I Ifl4
60
Incident al El nd t ng
E (InIercurrent Disease)
’ x i rP72~7
♦T5 2H
TT
I n< i dent a 1
E (1 nt er cur r ent DiseaetT)
♦ nor
84
79
79
79
72
60
53
71
56
72
61
70
71
70
54
49
67
71
71
52
D ■’ Died; E " Eu I han i r. ed in Mo i i Lund Condition; S - I ii I c i i m S ai. r i t i
Seo previous listing for tluu animal.
Ibis
Type
of
i < > hy ■. t < > I «»'| I < a I
I vi»«•*
I
111.111II11.11 v
I .•
F i nd
T
T
T _________
E (Melast at ic D i a e a s e )
T
“
E*
E* _______________________
D llnlercurrent Disease)
D (I nt ercur rent Dis ease)
E (Met a a t aiic D1sease)
0 (Metastatic Disease)
E (Met aatatic Disease)
0 flntercurrenl Disease)
0 (Intercurrent Disease)
D (Met as I at ic Disease)
0*
—
—
I
N)
CO
I
—■
S ___________
~——
E_ (Intercurrent D L s e a s e)
E (Met ast at ic Di aeaseT
D (Metastatic Diseased
E (Intercurrent Disease)
E (lKetastatic Disease)
D (Metastatic Disease)
E_ ( I nt ercur rent I) i seaseV
0 ( I ntercurrent DTseaseT
J
T - Tenn mat fiai i i f ice
• •••
TABLE 3:
Experimental
Group
Summary of Animal Experiment in Rhesus Monkeys (1968-1978)
_______ Original Group
Number at
Number at
Sta rt
End
!I
[9]
+ Replacement
Monkeys
Total Numuer
Surviving at
End
With
Endomet ria1
Cancer
Controls
16
5
2
7
0
Low dose
DMPA (Ix-HD)
4
2
0
2
0
Mid-dose
DMPA (lOx-IlD)
16
7
0
7
0
High-dose
DMPA (50X-11D)
16
i
7
5
12
2
bo
I
30
nodules were observed in 3 of 7 surviving monkeys in the
mid-dose Depo- Provera group.
In this study, at autopsy,
hyperplasia of the ductal epithelium of the breast was also
noted in monkeys receiving X10 the human dose of DMPA [5,
9] -
2.2. CRITIQUE OF THE FINDINGS IN ANIMALS
2.2.1
CAN THE MALIGNANT NEOPLASMS IDENTIFIED IN
THE MAMMARY GLANDS OF DOGS AND IN THE UTERI
OF THE TWO MONKEYS BE ATTRIBUTED TO MPA AND
IF SO, .IS THERE EVIDENCE OF A DOSE-RESPONSE
RELATIONSHIP?
The data from the study by
the Dawson Research
Corporation provides evidence that the mammary carcinomas in
the dogs were drug-related.
Moreover, in this well designed
and executed study there was good indication of a
dose-response relationship:
malignancies developed both more
frequently and earlier with increasing doses of DMPA (Table
2B, p. 27)
[4] .
The study also provides evidence that the
findings in the mammary gland were due to the progestational
actions of MPA and not to some structural peculiarity of the
compound, since the effects of DMPA and progesterone were
3] .
The findings were not unexpected.
They
confirmed the association in the dog between progestational
agents and breast pathology identified in previous published
studies [3, 10-15].
However, the association between
malignant neoplasms and progestational agents is more clearly
31
evident in the Dawson study than in previous studies in which
treatment was often of shorter duration and the focus was on
mammary nodules in general with less consideration being
given to their histological characteristics.
The distinction in the progestogen treated dogs between
malignant and benign tumors, and between nodules palpated and
those characterized histologically is important.
The nodules
represent a variety of pathologies that are often reversible
and include consequences of hyperstimulation of
lobulo-alveolar growth as well as benign mixed or complex
adenomas [8 r 13].
It is, however, the histologically
characterized malignant tumors, associated with the
administration of progestogens, that are of major concern.
The relationship between these malignancies and the more
frequent benign hyperplastic lesions elicited by progestogens
in dogs has not yet been the subject of any systematic
investigation.
While the finding of neoplasias in the mammary glands of
DMPA-treated dogs could be anticipatedr the development of
uterine carcinomas in the Rhesus monkeys was unexpected.
Endometrial carcinomas are not known to occur spontaneously
in Rhesus monkeys (see testimony of Dr. S. M. Sieber [16] and
that of Dr. Norval w. King [17]).
Consequently, there is
reason to assume that the neoplasias identified in the genital
tract of the two monkeys receiving the highest doses of DMPA
were drug-related.
Whether there could be any dose-response
32
relationship, and whether neoplasias might be elicited by
lower doses of DMPA, can not be deduced from this study,
because the study was so poorly designed and executed:
the
number of controls and animals receiving lower doses of DMPA
was much too small (Table 3, P. 29), and the pathological
examinations inadequate.
Several animals died during the
course of the study and were allowed to autolyze before
undergoing pathological examination [9].
Whether uterine carcinomas occur spontaneously in Rhesus
monkeys is of obvious importance in evaluating the
significance of the findings in this small and inadequate
IRDC study in which controls were few and animals from
different sources were used.
Consequently, we requested and
were granted access to the histological slides from a control
Rhesus monkey from a study of IUD carried out by the
Population Council.
This so-called "Tatum monkey," referred
to several times at the public hearing (by Drs. Wayne Bardin
[18] and by Roy Hertz [19]), was stated to have developed a
spontaneous endometrial cancer.
It was described in the
initial pathology report as having "A superficial, 110 micron
wide, circumferential ring of endometrium" showing "atypical
hyperplasia, questionable neoplasia," interpreted as
"microfoci of non-invasive adenocarcinoma" [20].
The histological slides were examined by the consulting
pathologists to the PBI who concluded, unanimously, that the
lesion was an epithelial plaque [21].
Epithelial plaque is a
33
benign, transient structure characteristically seen in the
Rhesus in relationship to implantation and, occasionally, in
the absence of pregnancy, as a response of a progestogen-
primed uterus to trauma (see Section 2.2.2b).
The original
description of the gross appearance of the lesion (see
above), as well as the fact that it was found in a uterus,
"showing thick secretary endometrium,"
i.e., a progesterone-
stimulated endometrium, is consistent with the pathologists’
conclusion.
To date, therefore, to the best of our knowledge.
there has been no documented case in the Rhesus monkey of
spontaneous, invasive, malignant uterine carcinomas, such as
were identified in the DMPA treated monkeys.
The latter,
therefore, must be attributed to the progestogen treatment.
The conclusion that progestogens can elicit malignant
transformation in the uterus of monkeys is reinforced by a
recent report of a uterine carcinoma in a Rhesus monkey
treated for 4 3/4 years with norethisterone enanthate (200
nig/kg, corresponding to 50 times the human dose, every 8
weeks)
[22-24].
The consulting pathologists to WHO and to
the German Ministry of Health who had the opportunity to
compare histological slides from this animal with those from
the DMPA-treated monkeys, described it to be of a similar,
undifferentiated,
invasive type [22, 23].
As stated in the
report by WHO:
All three tumors arose in animals with
atrophic endometrium and mainly affected
the corpus of the uterus. All three
34
tumors were histologically similar,
poorly differentiated carcinomas .... In
all three animals the tumors invaded the
endocervix. Furthermore, this was the
first time that WHO had been able to
examine original slides from all three
tumors together, and in contrast to the
conclusions of the 1981 review, the three
tumors appeared identical in type [23].
The undifferentiated, invasive nature of the neoplasias
in the uteri of the two DMPA treated monkeys was also
confirmed by the consulting pathologists [21].
These neoplasias had no unique histological features that
would permit distinguishing them from highly undifferentiated
endometrial carcinomas in other species,
including the human.
There is only one report in the literature of a
carcinoma in situ of the endometrium in a subhuman primate
[25] .
This was in a 22-23 year old chimpanzee.
The
endometrium was described as being hyperplastic. a change
ascribed to hyperestrogenism due to ovarian pathology
resembling that in patients with Stein - Leventhal syndrome.
The authors state that the lesion might not have qualified as
a carcinoma in situ, but only as a adenomatous hyperplasia,
according to the criteria of other experienced pathologists,
e.g., Gusoerg, one of the pioneers in defining the
association between hyperestrogenism and endometrial cancer.
In any event, whether the pathology be considered neoplastic
or only potentially neoplastic, it is clearly distinct from
that found in the Rhesus monkeys treated with the
progestogens, DMPA or norethisterone enanthate.
35
It has been suggested that carcinoma of the uterus was
diagnosed also in two Rhesus monkeys receiving oral
contraceptives. According to the description of these lesions
by the pathologist only one of the two cases cited was
considered to be possibly a carcinoma in situ in a somewhat
hyperplastic endometrium, again a pathology distinct from
that found in the progestogen treated monkeys [26].
We are ,
therefore, left with the conclusion that endometrial cancer.
in particular of a highly anaplastic type, has to date been
observed in monkeys only in association with progestogen
treatment, i.e., treatment with DMPA or norethisterone
enanthate.
Consequently, the endometrial cancers found in
the IRDC monkeys treated with DMPA have to be considered drug
related.
It has been implied that the two monkeys receiving DMPA
that developed the uterine pathologies were not
representative of monkeys in general, because they were
replacement monkeys [27].
We note that no neoplasias were
identified in the uteri of two of the replacement monkeys in
the control group that were, presumably, obtained from the
Under these circumstances, we must reject
Upjohn’s conclusion that "The cause of the endometrial
carcinomas in the two monkeys is not established**
[28], and
until proven otherwise, must conclude that they were related
to DMPA.
36
2.2.2
WHAT BASIC DIFFERENCES HAVE BEEN IDENTIFIED
BETWEEN THE HUMAN AND EACH OF THE TWO
ANIMAL SPECIES IN THEIR RESPONSE TO
PROGESTOGEN?
2.2.2a THE RESPONSE OF DOGS TO
PROGESTOGENS
EVIDENCE FOR SPECIES
DIFFERENCES
It has been argued by Upjohn and others that the dog is
an inappropriate model for studies of the potential adverse
effects of the long term use of progestogens on the breast of
the human female [29-33].
A summary of this position is
presented in Upjohn’s final brief:
"Recent information about
the mechanisms underlying the mammary response of dogs to
C-21 progestogens like Depo-Provera supports the decision of
several panels of experts that the dog is an inappropriate
model for carcinogenicity testing of progestogens" [33].
On
reviewing the several arguments that have been offered to
support this proposition, we conclude that none has been
adequately substantiated.
Consequently the findings in dogs
must be considered as having potential relevance to the human.
While there are certain obvious differences at the
descriptive level between the canine and human female in the
overall regulation of reproductive function, differences in
underlying mechanisms have not been defined to any extent.
Specifically, the mechanisms by which progesterone and
progestogens cause malignant neoplasias in the canine mammary
gland have hardly been investigated.
Consequently, it is
37
premature to attribute this response of dogs to the
progestogens to some mechanism unique to the species.
The major arguments that have been presented in support
of the thesis that the dog’s response to progesterone is
uniquely different from that of the human are evaluated
below.
2.2.2a.1
SUSCEPTIBILITY OF THE BEAGLE TO
MAMMARY TUMORS
Beagles, like other dogs, have a high incidence of
mammary tumors, both benign and malignant [34-43]. 5
Of
particular importance to the question at hand are the
malignant neoplasias.
The most frequently noted malignant
neoplasias, as discussed below, appear to resemble
histologically those found in the human [34r 36-38].
A number
of similarities have been identified between the dog and the
human in the epidemiological and biological characteristics of
malignant mammary neoplasias.
These include age-specific
incidence and relationship to ovarian hormones [39-42].
Such
similarities argue for, rather than against, the use of the
5/
There are no data to support the notion that beagles are
more susceptible to neoplasias of the breast than other
strains of dogs [35]. Certain other pure bred strains
of dogs have, in fact, been reported to have a higher
incidence of such neoplasias than dogs in general,
including beagles [35, 43].
However, beagles have been
studied most extensively in an experimental setting
because of their availability and convenience of
use.
38
dog as one of the species in testing for potential promoters
i
•
6/
of mammary neoplasias.—
It is, however, obvious that
evidence available on many other biological as well as
epidemiological criteria need to be considered when trying to
evaluate the relevance of the findings in the dog to the
human.
However, the fear expressed by Harold L. Upjohn, in a
letter to the FDA Commissioner, that because "dogs develop a
variable high incidence of spontaneous mammary tumors" this
would make "proper control observations almost impossible"
[44] has not been borne out, at least in the Dawson study.
Differences between controls and progestogen treated animals
in malignant breast cancer are readily identifiable and
quantifiable in this well designed and executed study.
2.2.2a.2
DIFFERENCES BETWEEN THE HUMAN
AND THE DOG IN THE TYPE OF
MAMMARY NEOPLASIAS DEVELOPED
With respect to spontaneous neoplasias, the major
difference between the two species relates to the high
6/
We have not been able to find any comparative data on
the incidence of cancer of the breast in different
geographic locations in the world. One of the striking
characteristics of cancer of the breast in the human
female is the much higher incidence of the disease in
the developed countries in the West. These regional
differences are attributed in part to environmental
factors, including diet.
It would be of considerable
interest to know if cancer of the breast in dogs is also
subject to risk factors similar to those implicated in
the human.
39
incidence in the dog of mixed mammary tumors, a type rarely
found in the human.
In dogs, the majority of these tumors
are benign; the malignant mixed mammary tumors constitute only
a small percentage of the malignancies in the mammary gland in
this species [34].
The incidence of this type of malignancy
was not increased by prolonged treatment with DMPA.
The majority of malignant breast tumors that arise
spontaneously in the dog, as in the human, are described as
adenocarcinomas.
In the Dawson study, there were no
malignant breast tumors in the control animals.
In the
progestogen treated groups. the two most frequent types of
malignant tumors were solid cell carcinomas and
adenocarcinomas (Table 2C, p. 28)
[4] .
The question of the
cells of origin of these malignant neoplasias has not been
studied systematically, although this would be of considerable
interest and importance in attempting to understand their
histogenesis and pathophysiology.
The difficult problem of
classification of mammary neoplasias in dogs has been
approached by several groups of experienced veterinary
pathologists [34, 36, 38].
4
4 -s
They all rely on morphological
The question of cells of origin was raised by
Moulton [34] .
However, there are no data at this time to
suggest that the cells of origin of these malignant neoplasias
xn the progestogen treated beagles differ from those arising
spontaneously, either in dogs or in the human female.
In the
human, adenocarcinomas of the breast are considered to have
40
their origin in ductal epithelium, and there is no evidence to
suggest that this is different in dogs [45, 46].
Consequently, if the histological characteristics of the
malignant neoplasias are taken into account, overall species
differences in types of neoplasias can not constitute, at this
time, a reason for dismissing as irrelevant to the human the
finding of increased incidence of adeno- and solid cell
carcinomas in the progestogen treated dogs.
2.2.2a.3
MAJOR DIFFERENCES IN THE
COMPARATIVE ENDOCRINOLOGY OF
REPRODUCTION BETWEEN DOGS AND
HUMANS
The dog has a bi-annual heat period (interestrous period
approximately 200 days) followed by a prolonged luteal phase.
This luteal phase, or period of pseudopregnancy, lasts up to
60 days. and resembles, both in length and in hormonal
profile, pregnancy in the dog.
The luteal phase in the dog.
like pregnancy. is characterized by high levels of
progesterone.
However, except for a brief period before
ovulation, estradiol levels in the dog are relatively low
[14, 47] .
Thus, in the dog, during relatively long periods
of time (60 days out of 260), circulating progesterone levels
are relatively high at a time when levels of estrogen (or at
least of estradiol. since other estrogens have not been
measured) are relatively low.
Whether and how these
41
differences between the bitch and the human female might be
linked to differences in their response to progesterone and
progestogens is not known.
Certain, not very precise.
hvootheses have been advanced r e.g., "The dog has probablv
not developed a control system which counteracts harmful
effects of excessive progesterone levels"
[48] .
Studies have
not been carried out to define what such control mechanisms
might be.
of
For example, there have been no systematic studies
how progesterone receptors are regulated in the various
target organs in the dog.
"State of the art" thinking and
methods have not been applied to the question.
The limited
information available on this subject is discussed in the
next section.
2.2.2a.4
DIFFERENCES IN THE RESPONSE OF
DOG AND HUMAN MAMMARY GLAND TO
PROGESTERONE IN TERMS OF
STIMULATION OF GROWTH AND
DIFFERENTIATION
The theme of species differences in growth and
differentiation of the mammary gland in response to
progesterone has been stated in a variety ,of ways.
For
dog is the only species besides the ferret i
which progesterone alone can produce a considerable
lobulo-alveolar growth of the mammary gland"
[49], or "The
dog differs from other laboratory species in that mammary
hyperplasia can be induced with progesterone alone and does
not require estrogen priming" [50].
These statements refer
42
to the fact that, in the dog, mammary development required
for -lactation starts promptly after ovulation under the
influence of ovarian hormones and is sustained by them
throuqhout pregnancy or pseudopregnancy.
In the dog, the
predominant steroid secreted by the ovaries during pregnancy
(or pseudopregnancy) appears to be progesterone accompanied
by relatively little estrogen.
Therefore, not surprisingly,
progesterone alone can elicit in this species good, albeit
not optimal. lobulo-alveolar development even in the absence
of estrogen.
(A very similar reproductive strategy appears
to have been evolved by the ferret.
It exhibits a prolonged
luteal phase comparable in length to pregnancy, and during
such periods the breast achieves full lactational capacity
[51] ) .
In contrast, in the human female, mammary development
in preparation for lactation occurs during pregnancy under
the influence of the high levels of both progesterone and
estrogens as well as other hormones secreted by the
placenta [51].
This particular species difference, however remarkable.
can not serve as a basis for deciding on the relevance of the
in the dogs to the human without more information on
the cells of origin of the progestogen related neoplasias in
the dogs.
The known differences between the species in
hormonal response of the breast relate to lobul-alveolar
development.
In the human female. as discussed above.
adenocarcinomas of the breast are considered to have their
43
origin in cells of the ductal epithelium; there are no data to
support the notion that dogs differ in this respect.
The type
of longitudinal studies needed to establish the cells of
in of progestogen induced neoplasias in dogs have not y
been carried out.
The important considerations, then, are how ductal
epithelial cells, that presumably give rise to the
adenocarcinomas, respond to progesterone alone or, more
importantly, to progesterone or progestogens in the presence
of low levels of estrogen?
The focus in recent years has been primarily on the role
of estrogens in the development and promotion of growth of
neoplasias in secondary sex structures, including the breast.
Progesterone, in contrast, is thought of as a neutralizer of
estrogen and estrogenic action.
The notion expressed in
various ways is that estrogens act as promoters of neoplasias
because they stimulate cell division (Dr. M. Lipsett in [2]).
In so far as progesterone acts, as in the uterus,
to arrest
estrogen-induced cell division and to further cell
differentiation, it can be viewed as an antidote to the
neoplasia promoting actions of estrogens.
There is, however,
evidence that progestogens can also act on certain cell types
within target tissues, including the mammary gland, to
promote cell division and that it can do so in species other
than the dog.
First, in explanted mammary tissues from human
females, not only estradiol but also progesterone has been
44
shown to stimulate the incorporation of labelled thymidine
into- ductal epithelial cells [52, 53] .
The epithelial cells
of the interlobular ducts appeared to be particularly
onsive to progesterone.
A second perhaps more direct
line of evidence implicating progestogens with a potential to
promote neoplastic transformation of breast cells comes from
studies of mammary carcinomas induced in rats by the
carcinogen dimethyl-benzanthracene (DMBA).
Progesterone,
like estradiol, administered during the latent period
following the administration of DMBA, both reduced the time
required for the development of malignant breast tumors and
increased their size and number [54-57].
Once the tumors
were established. some of them could be shown to be hormone
dependent, requiring estrogen for their maintenance in the
castrated rats.
In contrast. progesterone was ineffective in
maintaining hormone dependent carcinomas in castrated rats
[56, 57].
Thus, while progesterone could in these
experiments promote the development of malignancy during the
latent period, it was not able to sustain malignant cells once
established.
In more general terms an absence of demonstrable
dependence of a tumor on a particular hormone does not exclude
the possibility that that hormone played a role in the
development of that tumor.
In the human female and in test animals receiving
contraceptive doses of DMPA, the progestogen, however, is not
acting alone but in the presence of at least low levels of
45
estrogen, and these two classes of steroids are known to act
synergistically on breast development in all mammalian
species studied to date.
Under these hormonal conditions
cf the monkeys treated with DMPA showed hyperplasia of
the intra-ductal epithelium [9, 58, 59].
Dr. Henry Norris, a
witness at the public hearing experienced in both human and
subhuman primate breast pathology, judged the ductal
epithelial hyperplasia to be atypical, first, because of a
resemblance to the micropapillary pattern of intraductal
carcinoma of the breast in humans and, second, because of
nuclear hyperchromatism, prominent nucleoli, and mitotic
activity [59].
Irrespective of whether these changes have the
same long term significance in monkeys as they are thought to
have in the human, the findings attest to the possibility of
ductal epithelial hyperplasia developing in a primate under
conditions of prolonged, relatively unopposed progestogen
action, the hormonal state characteristic of subjects
receiving contraceptive doses of DMPA.
The development in
DMPA treated monkeys of changes that resemble a precancerous
lesion in the human may be significant in light of the
apparent resistance of this species to agents that induce
malignancies in breast tissue of other species.
Finally, the uniqueness of the response of the
lobulo-alveolar component of the breast of dogs (and ferrets)
may be overstated.
The difference between these and other
species may be more quantitative than qualitative, as pointed
46
out by Trentin et al [60].
According to these investigators,
in castrated dogs, lobulo-alveolar growth was not elicited by
estrogen treatment alone but did occur following treatment
with estrogen + progesterone and, to a lesser degree.
following treatment with progesterone alone.
The
investigators state, "The rather good development of the
mammary glands induced by progesterone alone is not without
precedent.
Mammary growth responses to progesterone in the
absence of estrogen has been reported in the mouse, the rat
and the monkey.
The dose of progesterone required to produce
a complete alveolar type of mammary development in dogs is
relatively much less than in rats where 15 mg/day was
required"
[60] .
2.2.2a.5
EVIDENCE FOR DIFFERENCES BETWEEN
THE BEAGLE AND THE HUMAN IN THE
SPECIFICITY AND REGULATION OF
PROGESTERONE RECEPTORS
The statement repeated in various reviews of DMPA that
progesterone receptors in the dog differ from those in the
It represents the
human, is based on a single study [61].
sole attempt to examine the concentrations and regulation of
progesterone receptor in the canine.
It is a study that
would be important to follow up but which was too limited and
incomplete to permit any conclusions.
There are major
problems with the experimental design and with the
methods used in this study.
For example:
(a)
the number of
47
animals studied was too small and the number of measurements
made- too few to establish a range of values within any single
experimental group.
Consequently, significant differences
among experimental groups can not be established; (b) the
tissues sampled for comparison of concentration of receptor in
the dog with those of the human and for examining the effect
of progesterone on the receptor were limited and appear
arbitrary (e.g., myometrium, canine breast tissue before and
after administration of a single dose of progesterone, and
only at one time point, a single specimen of mixed mammary
adenoma); (c) the comparison of specificities of the receptor
in the human and the dog was made from a study of a single
sample from each.
Data on replicate determinations are not
given; (d) The sensitivity of the assay is unusually and
unaccountably low, several fold less than reported by others;
and (e) data on the characteristics of the canine progesterone
receptor under the conditions selected for the assay and the
basis for selecting the conditions used in the assay are not
reported (e.g., the reason for using a crude cytosol
preparation, results of tests to determine the stability of
the receptor under the conditions used, etc. ) .
Considerablv
more detailed studies, both of the normal and neoplastic
canine mammary tissues, are needed before it would be possible
to make any statements about species differences in the
regulation of the progesterone receptor by progesterone or in
its dhiinities for different progestogens.
Moreover,
48
information on more dynamic aspects of receptor function are
needed to provide some meaningful basis for interspecies
comparison of the role the progesterone in tissue response
<
testimony of Dr. C. Wayne Bardin [62]).
Our conclusion from the above analysis is that there is
too little comparative data on fundamental aspects of the
action of progestogens in the two species to decide which, if
e i ther, is unique in its response, much less in what way it
may be unique.
Moreover, there are clues to suggest some
common principles underlying the actions of progesterone in
different species, including the dog.
2.2.2a.6
EVIDENCE FOR THE PRESENCE OF
LATENT NEOPLASTIC FOCI IN THE
CANINE MAMMARY GLAND NOT PRESENT
IN THE HUMAN MAMMARY GLAND
Numerous references have been made various reviews on
DMPA to the presence of latent progesterone responsive
neoplastic foci in the dog [29-33, 63] .
For example, "The
evidence suggests the existence within normal beagle breast
tissues of a large pool of microscopic neoplasms, the
majority of which are benign and stationary, perhaps checked
by immunological mechanism.
As progesterone, but not
estrogen, is a potent stimulant to growth of breast tissues
in ovariectomized dogs. it is hardly surprising that
administration of large doses of progesterone-like substances
long periods induces growth of these microscopic
49
neoplasms”
[63]; "The Canine mammary gland contains latent
neoplastic foci which do not exist in the human"
[33]; or
"Careful examination...of untreated, intact, mature bitches
has revealed multiple microscopic neoplasias.
These contain
specific progesterone receptors. so that their growth is
stimulated by progestogens.
estrogens,
These effects are antagonised by
in that combinations of a progestogen with an
estrogen give less tumors than the progesterone alone"
[63] .
These statements in various reviews. like those on the
uniqueness of the canine progesterone receptor, are based on
a single publication [64].
The data presented in this one
brief publication by Cameron and Faulkin is, moreover.
too
scant and incomplete to serve as the basis for such
s ta temen ts.
Cameron and Faulkin present data on the numbers
and histological characteristics of areas of increased tissue
density they identified when trans illuminating whole-mounts
of 32 mammary glands from eight beagle dogs.
The dogs’ ages
were 7.6 - 8.5 years, i.e. an age when they become highly
prone to neoplasias [39, 40].
Any unusual structural change,
as compared to the normal lobularity of the surrounding
mammary gland, was considered by the investigators to be an
atypical nodule.
A total of 509 of these were judged to be
due to lobular hyperplasias. because they were found on
sectioning to be larger than adjacent lobules.
Proliferative
changes identified in these nodules involved either both
myoepithelial and epithelial elements or predominantly the
50
epithelial elements.
Compression of surrounding tissues was
absent and mitotic figures were rare, i.e. there was no
evidence of neoplastic changes in these nodules.
The authors
state that they also identified a total of 94 nodules that
they judged to be neoplastic, 80 of them in two of the
animals.
These, presumably, are the foci referred to in
various reports and reviews in which this study is cited.
However, the histological characteristics of the lesions
judged to be neoplastic are not described in the original
publication , the criteria for considering them to be
neoplastic are not stated, nor do they appear to have been
reported elsewhere.
The authors’ own conclusion.
appropriately. is only that the method could be useful to
study the development of breast tumors in dogs, a suggestion
that has not been followed.
The data presented in this study can be interpreted as
evidence that, during anestrous in the dog, complete
regression of the lobulo-alveolar growth that had occurred
during the preceding pseudopregnancy does not take place.
However, whether these residual areas of persistence of
eoithelial and myoepithelial cell nests represent the
substrate for subsequent development of neoplasias, either
benign or malignant, is unknown.
Since no description
whatsoever is given of the nodules that the authors judged to
be neoplastic, it is also not possible to deduce. or even to
speculate, how these might be related to what were termed by
51
on the one hand t
the same investigators atypical hyperplasias
neoplasias found in the progestogen treated dogs on
or to the
No measurements of progesterone receptors
from
were carried out on tissue samples from this study or
Therefore z
comparable samples of canine breast tissue.
the other hand.
statements that these presumed neoplasias contain
progesterone receptors are only speculations. On the other
such atypical areas of
hand z the statements that no
hyperplasias are to be found in the human ignores the
decades concerning
voluminous literature published over many
discussions of their
comparable findings in the human and the
implications for malignant breast disease in the human
[65-73] .
Comparative studies are badly needed of early changes
in the human and canine
that may be considered precancerous
subsequent development of
breast and how these relate to the
We have only
frank malignant lesions in the two species.
been able to find one report of a study in which
investigators experienced in human and canine breast
pathology have joined forces to examine this question [38].
♦■ho histological features of noninvasive lesions in
■;
of
mastectomy specimens from 232 dogs to the clinical course
led these investigators to conclude that there
their disease
are close similarities between the two species,
This
conclusion is clearly diametrically opposite to that reached
cell
-^M-FlrstFloorTS
BAMGAlOHE ! \
52
by various readers of the report by Cameron and Faulkin
[64]\
2.2.2a.7
COULD GROWTH HORMONE, RELEASED
BY BEAGLES IN RESPONSE TO HIGH
DOSES OF PROGESTERONE, BE
RESPONSIBLE FOR THE INCREASED
INCIDENCE OF MALIGNANCIES IN THE
MAMMARY GLANDS OF DMPA TREATED
DOGS?
Beagles, unlike humans, secrete excess amounts of growth
hormone and become acromegalic when treated with high doses
of MPA [4, 8, 74] .
The hypothesis that growth hormone
mediates or potentiates the actions of progestogens on the
mammary gland of dogs is interesting and plausible.
Growth
hormone is required for mammary growth and resembles
prolactin, which in some species, notably the rat, is
implicated in the development of neoplasias of the breast.
However, the hypothesis that growth hormone, acting alone or
synergistically with progesterone, is responsible for the
increased incidence of breast cancer in progestogen treated
dogs remains to be tested.
The correlation noted between
growth hormone levels and incidence of breast nodules in no
way constitutes proof of causality.
Without direct
experimental evidence. it remains a postulate and can no t be
used to explain the findings in the mammary glands of the
MPA treated dogs.
53
2.2.2a.8
IS THE FACT THAT PROGESTOGENS
CAN CAUSE REGRESSION IN SOME
CASES OF HUMAN BREAST CANCER A
REASON TO ASSUME THAT THEY ARE
UNLIKELY TO PROMOTE THE DISEASE?
he ability of a sex steroid to
♦-o the devel
or sustain the growth of a tumor in a target organ and to
cause regression of the tumor once it is established are not
mutually exclusive.
estrogens.
This has been demonstrated in relation to
While estrogens at physiological doses may
stimulate mammary cancer growth, when given in massive,
pharmacological amounts they can cause the cancer to regress.
This apparent paradox has been known for over tnree decades
and has been applied to the treatment of mammary cancer in
women [75-79] .
By analogy, the fact that progestogens in
pharmacological doses can cause regression of some mammary
tumors does not exclude the possibility that at lower doses
they could promote the development of such tumors.
2.2.2a.9
SUMMARY OF EVIDENCE ON
POTENTIAL RELEVANCE OF THE
FINDINGS IN THE DOG TO THE
HUMAN
The canine. like the human female in the U.S.A., has a
high incidence of malignant neoplasias of the breast and. in
both , the incidence increases markedly in late adulthood.
Though dogs are subject to a type of benign tumor of the
breast rarely seen in the human, the most common types of
malignant tumors in the two species appear to be
/ \
similar.
54
Whether the histogenesis of the common malignant neoplasias
is similar in the two species is not known.
There are some
obvious and marked differences in the hormonal pattern of
women and of bitches during both the fertile and infertile
cycles as well as in the relative importance of estrogen and
progesterone in breast development.
Whether and how these
differences may be reflected in species differences in
fundamental aspects of the actions of progestogens on their
different target organs remain to be established.
The
evidence cited is inadequate to support the assertion that
there are fundamental differences between the dog and the
human in the mechanism of action of progestogens.
There is,
therefore, at this time no valid scientific basis for
dismissing as irrelevant to the human the observations made
on the carcinogenic potential of progestogen in the dog.
Whether the canine is, or is not, an appropriate species
for evaluating the potential consequences of prolonged,
relatively unopposed progestogen action has been debated
since the late 1960’s.
The same arguments. based on the same
limited, inconclusive evidence, have been repeated over and
over.
It is disappointing that there has been virtually no
systematic effort to expand the data base needed to resolve
this important issue.
55
2.2.2b
THE RESPONSE OF THE RHESUS MONKEY
(MACACCA MULATTA) TO PROGESTOGENS
-EVIDENCE FOR SPECIES
DIFFERENCES
The following are the major reasons given to support
contention that the endometrial neoplasias identified in the
Rhesus monkeys recieving DMPA are unlikely to be relevant to
the human [80]:
1 )
The neoplasias arose from a cell type present in the
uterine epithelium of the Rhesus monkeys but not in the
human, a cell type that responds to progesterone with
proli feration;
2)
The neoplasias in the Rhesus arose in an atrophic
endometerium while in women endometrial neoplasias are
associated with hyperplastic endometrium;
3)
In the human, progestogens act to counter the
development of hyperplasia and associated neoplasias of the
endometrium.
Moreover, when administered in pharmacological
doses, progestogens can cause regression of such neoplasias.
Therefore, progestogens are unlikely to promote neoplastic
transformation in the human endometrium;
4)
The neoplasias occurred only in the monkeys
receiving the highest dose of DMPA.
At these doses, actions
of the progestogen may differ not only quantitatively but
also qualitatively from those obtained with the lower doses.
The first three reasons given are closely interrelated.
They are based on current understanding and beliefs of the
56
relative roles of estrogens and progestogens in the
development of endometrial carcinomas in the human.
They
were articulated in relation to the problem of the etiology
of the endometrial cancers in the DMPA treated monkeys first
in 1979 by the expert pathologists appointed by the Upjohn
Company to review the histological slides from the monkeys in
the IRDC study [58] and subsequently restated in various
reviews and documents [28, 80, 81].
The unexpected finding of malignant neoplasias, of
presumed endometrial origin, in the uteri of two of the
DMPA treated monkeys led the Upjohn Company to request a
panel of highly experienced pathologists to review the
histological slides from the uteri and breast tissue from
these monkeys [58].
These pathologists did not question the
diagnosis of endometrial carcinoma.
However, they puzzled
over the cells of origin of the malignancy.
The pathologists
looked for, but failed to find, any evidence of hyperplastic
changes in the epithelial elements that they could interpret
as precursor lesions.
They noted decidualisation of the
stroma in the DMPA treated monkeys, but none in the glandular
"ie main expected site of decidualisation in
response to progesterone in this species (letter by Dr. Arthur
A. Hertig, Dec. 21 1979 [58]).
Neoplasias arising from the
chronically stimulated, decidualised stroma would have been
expected to be sarcomas rather than carcinomas.
But both at
the light and electron microscopic level the lesions had the
57
characteristics of carcinomas and not of sarcomas.
Absence of
epithelial hyperplasia, commonly seen in women with
endometrial carcinomas, was commented on.
To resolve the
t paradox, Dr. Valerio, one of the pathologists
consulted, offered a hypothesis linking the neoplasias to the
epithelial plaques that develop in Rhesus monkeys in relation
to implantation.
She postulated these carcinomas as "possibly
arising from the epithelial cells in the surface epithelium
and the mouths or necks of the uterine glands.
These
epithelial cells respond differently in the female Rhesus
monkey than in women to implantation or experimentally to a
lesion or trauma of the superficial endometrium in a
hormonally (progesterone) prepared uterus.
If this is the
case in these two tumors, then histogenetically they would
not be the counterpart of endometrial carcinoma in women
which is generally believed to be associated with growth
stimulation by estrogen"
April 9, 1979, p.
(letter by Dr. Marion Valerio,
101 of Reports [58]).
An alternative resolution of the apparent paradox was
proposed by Dr. Gisela Dallenberg-Hellwig, a Center witness.
and presented at the public hearing [82].
She considered the
neoplasias to have originated not from endometrial but from
endocervical cells.
Since the action of progestogens on the
latter cell type is, presumably, to promote cell division
rather than differentiation, the findings could be
accommodated into current ideas of how sex steroids may act
58
as promoters of neoplasias [2].
However, the consulting
pathologists we appointed to re-examine the histological
slides from uteri of the IRDC monkeys did not find evidence
for endocervical origin of the malignancies [21, 83].
They,
like the majority of the pathologists appointed by the German
Ministry of Health and WHO, did not find any reason to
question the endometrial origin of the tumors [22, 23].
The hypothesis that the neoplasias arising in the
progestogen treated monkeys originate in epithelial cells
that respond differently from those in the human is a
reasonable one.
It deserves and needs to be tested, as
nr, Valerio suggested when she proposed the hypothesis in
1979 [58].
Epithelial plaques were studied by pioneers in
the field of reproductive endocrinology several decades ago
[84-86], but not since.
Specifically,
there has been no
study since 1979 to test the hypothesis proposed by
Dr. Marion Valerio.
Consequently, like the hypothesis on the
role of growth hormone in the progestogen induced breast
cancer in dogs,
it is only a hypothesis.
It is worth considering what prompted the pathologists to
a hypothesis in the first place.
apparently two reasons:
There were
the absence of any obvious precursor
lesions in the uteri of the DMPA treated monkeys. and a
preoccupation with the role of estrogens in endometrial
cancer.
In the monkeys, the endometrium was obviously
atrophic, and there was no evidence of hyperplasia.
/ \
59
Consequentlyr the lesions did not resemble those most commonly
encountered in women in this country.
Endometrial cancers in
the human are most often found in association with
rplastic endometrium, and these are freque
well differentiated.
i
They are often hormone dependent and
tend to regress when estrogen is withdrawn or when patients
are given high doses of progestogens.
In many cases,
development of neoplasias with these characteristics can be
related to prolonged estrogenic stimulation inadequately
balanced by progesterone.
However, this is not the only type
of endometrial carcinoma to be found in the human, and the
view that estrogen is the only predisposing hormonal factor
for endometrial cancer may be too limited and limiting.
The consequences of hyperestrogenism in women, in
particular on the uterus, are well established.
Hyperestrogenism occurs relatively frequently in developed
countries either in association with certain diseases or as a
result of hormone treatment.
In contrast, there are no common
disease states associated with chronic unopposed, or
relatively unopposed, progestogen action.. Consequently, there
is no body of clinical observation on what the long term
consequences of such a state might be.
The use of synthetic
progestogen in treatment of certain neoplasias and, more
importantly, when used as a contraceptive, is providing the
first opportunity to observe what these consequences might be.
This opportunity has, to date, been largely neglected.
(
60
The majority of cases of endometrial carcinomas seen in
this-_ country are, indeed, associated with a hyperplastic
endometrium [87].
However, there are cases with little or no
evidence of hyperestrogenism or of endometrial hyperplasia.
There are even some instances, albeit rare. in which the
endometrium is atrophic [87-89].
These carcinomas tend to oe
less well differentiated and not hormone dependent.
In some
populations they could represent a higher proportion of the
cases.
This is suggested by a recent report from Russia in
which in 25% of the subjects the neoplasias were found in
association with an atrophic endometrium and in patients
without apparently any evidence of hyperestrogenism [90].
Therefore, neither the histological characteristics of the
malignant uterine neoplasias in the monkeys, nor of the uteri
in which the neoplasias were identified, constitute an
adequate reason for considering the findings in the monkeys
unique to that species and irrelevant to the human.
Until
adequate experimental evidence becomes available on the cells
of origin and natural history of the uterine malignancies in
progestogen treated monkeys, the proposition that these
malianancies are in a cell type unique to the Rhesus and that
it has therefore no counterpart in the human must be
considered only a hypothesis.
The etiology of this type of
neoplasia in the human is not known.
Furthermore, the
histological appearance of the uterus as well as of the
ne^yla&ias in the monkeys may have been influenced by the
61
continued administration of DMPA in pharmacological doses
To gain insight into the cells of origin of these
[89k.
neoplasias in the Rhesus, which is what the consulting
pathologists attempted to do, would require serial
observations.
In the IRDC study, regrettably, monkeys that
died in the interim and that might have provided some
information on the subject were allowed to undergo autolysis
and were unsuitable for pathological examination [9].
The proposition that progestogens are unlikely to act as
promoters of neoplasias because they can cause regression of
endometrial cancer has also been raised in relation to breast
cancer.
The therapeutic value of progestogens rests on th e i r
potential to act as anti-estrogens.
In this capacity,
progestogens can, when given in physiological amounts, reduce
the incidence of the type of endometrial neoplasias
associated with relatively unopposed action of estrogens [87,
91 , 92].
Progestogens, in pharmacological amounts, can also
cause regression in hormone-dependent neoplasias of the
endometrium [87].
Some of the biochemical phenomena
underlying these effects have been identified (e.g.. decrease
in estrogen receptor concentrations and, in the case of the
endometrium, also induction of 17B-hydroxysteroid
dehydrogenase causing increased conversion of estradiol to
the less potent metabolite estrone [87, 92] and testimony of
Dr. S.B. Gusberg [91]).
However, progestogens cannot be
exonerated from contributing to the development of endometrial
< \
62
neoplasias on this basis.
As mentioned earlier, estrogen
cannot be implicated as an etiological factor in all cases of
endometrial carcinomas.
Moreover, as discussed in relation to
breast cancer, there is no basis for excluding the possibility
of progestogens having paradoxical effects on the neoplastic
process, such as has been identified in relation to estrogens
and breast cancer (see p. 53).
Briefly, while estrogens are
implicated in promoting the development and growth of cancer
of the breast, they can, when administered in pharmacological
amounts, cause the regression of a significant proportion of
such cancers in the human.
Evidence for a similar paradoxical
effect of estrogens on the uterus is only available with
respect to the non-malignant endometrium.
It has been
reported that involution of endometrial tissue can be
accomplished in the primate by prolonged administration not
only of a progestogen or of progesterone, but also of an
estrogen [87 (p. 156)].
The relevance of the endometrial cancers identified in
the monkeys to the human has been questioned also on the
grounds that the cancers occurred only in animals receiving
the highest doses of DMPA.
The implication is that the
pathology would not occur at the lower. contraceptive doses
used in humans and that, at the high doses. new and different
mechanisms of action come into play.
This argument could have
been examined critically only if sufficient numbers of animals
had been included in the different experimental groups and
63
followed adequately to determine the nature of the
dose-response relationship; howeverf this was not the case in
the IRDC study of the monkeys (Table 3, p. 29).
The doses administered to test animals are defined
operationally and represent multiples of the human dose,
adjusted for differences in body weight.
In the absence of
adequate data on the many factors that may modify
effectiveness of the drug after its administration. this
remains a practical basis for comparison of drug effects
across species.
There is no evidence that DMPA’s potency in
the Rhesus would be greater than in the human.
On the
contrary, there are some data to suggest that progestogens
may be cleared faster in monkeys than in the human and that z
therefore, more of the drug would have to be given to achieve
the same effect [93-95].
In the frequently cited study by
Mora and Johannson monkeys were given the same dose of DMPA
as humans (150 mg total dose to each), although their body
weights were only a tenth that of the human [95]•
Nevertheless, as far as can be deduced from the minimal data
presented, after an initial, brief period during which blood
levels of immunoreactive DMPA measured in monkeys were
somewhat higher than in the human female. the levels in the
two species were stated to be comparable, and evidence of
return of ovulation occurred.
same time [95] .
in both species, at about the
The histology of the endometrium of
uinPa treated Rhesus monkeys also suggests that the potency of
64
a given amount of DMPA may be less in the Rhesus than in the
human.
Specifically, only at ten times, but not at one times
the human dose, did the changes in the endometrium of the
monkeys resemble those in the human receiving contraceptive
doses of the drug.
2.2.2b.1
SUMMARY OF EVIDENCE ON
POTENTIAL RELEVANCE OF THE
FINDINGS IN THE MONKEY TO THE
HUMAN
None of the major reasons proposed for dismissing the
relevance of the findings in the Rhesus to the human can be
substantiated by the evidence presently available.
The poor
design and execution of the IRDC study preclude the
conclusion that the endometrial cancers were a consequence of
the extreme dose of the drug used or that the two monkeys
were in some way unique and not representative of monkeys in
general.
The hypothesis that the cancers originated from an
endometrial cell type present in the monkeys but not in the
human remains to be tested.
Finally, whether the endometrial
cancers that developed in the two monkeys in the IRDC study
were the consequence only of circumstances unique to the
particular experimental conditions of the IRDC study. as
claimed by Upjohn, can not be determined without repeating the
experiment in an adequate manner.
65
2.3
ANIMAL TESTING FOR CARCINOGENESIS:
APPROPRIATENESS OF GUIDELINES FOR TESTING AND
INTERPRETING RESULTS WITH STERIOD HORMONES TO BE
USED AS CONTRACEPTIVES
The reliance on animal testing has been conventionally
applied to drugs as a prelude to their widespread
administration to humans.
Such tests are aimed at screening
for major adverse effects of the drug. its toxicity following
short term use, and its carcinogenic potential following long
term use.
The principles underlying animal tests of drugs,
in general, have been incorporated into the guidelines
established for testing of contraceptive steroidal agents
[96] .
The pitfalls and the limitations of such animal tests
have been recognized and debated since these guidelines have
been established.
The complexity of the issues involved mav
be gauged from the discussions held during a three day
workshop held by NIH in 1983 [2].
The debate among proponents and opponents of a drug that
fails the animal test screen, as in the case of DMPA, tends
to center around two theoretical problems:
the doses of the
drug used and the appropriateness of the species selected for
O U
•
The traditional approach in studies of drug safety has
been to administer a wide range of doses of the drug.
including doses much larger than those proposed for use in
the human.
This convention can be rationalized on several
grounds.
It is assumed that use of high doses will permit
66
the observations of events that otherwise may take a very
lon<£ time to develop or require a very large number of
animals to detect.
Use of a wide dose range helps delineate
a safe dose range and if a dose-response relationship exists.
it provides evidence for causal relationship between the drug
and a specific adverse effect.
In relation to hormonal steroids, the use of very high
doses has been challenged on the grounds that their effects
may differ not only quantitatively, but also qualitatively,
from those of lower, more physiological amounts.
It should
be noted that similar questions have been raised in relation
to chemical carcinogens in general [97].
One way to
compensate for this problem is to design the tests so that
the nature of dose-response relationships can be defined
rather than to abandon the use of multiples of the human
dose.
This was accomplished in the study of dogs by the
Dawson Research Corporation, but not in the IRDC study of
monkeys.
The problem of uncertainties about the suitability of
any particular species has been addressed by the requirement
rug be tested in several different animal species.
Thus, according to FDA guidelines, all contraceptive drugs
must be tested in rodents, dogs and monkeys.
If a
carcinogenic effect is demonstrated in any one species, in
particular, in a site that is expected to be affected by the
particular drug or hormonal agent (i.e., a known target
67
organ), this is considered to constitute evidence of lack of
safety.
If a drug is found to produce cancer in more than
one species, the strength of the evidence is increased.
It
is recognized that a single carcinogenic agent may cause
cancer in different organs in different species and that
cancers may be induced in unexpected sites, not previously
identified as target organs.
Only if relevant biological
responses in the species affected are known to differ in a
major way from those of the human would it seem prudent or
justified to dismiss major findings [97 (pp. 21599-21600)].
Tests on DMPA were conducted on all three species
mandated.
Neoplasms were identified in known target organs of
progestogens in two of the species, the dog and the monkey.
It would require compelling evidence that the human is
uniquely different from each of these two species to disregard
these findings.
Such compelling evidence, in our opinion, is
not now available.
Whether DMPA treatment of mice, the third
test species, was associated with any increase in neoplasias
is not known.
The results of the tests on mice presented by
Upjohn to FDA are uninterpretable, because of the very high
mortality in both the controls and the experimental groups
[98-100] .
This problem is compounded by inadequate
information on the pathology of the large number of mice that
had died and were either canibalized or autolysed
[98, 99].
Unfortunately, FDA apparently did not request that these
68
early, relatively short term experiments on the rodents be
repeated.
The World Health Organization, in its monograph on
Human Sex
"Evaluation of the Carcinogenic Risk of Chemicals:
Hormones" [101] r lists three responses in test animals that
are to be used as indicators of carcinogenicity:
D the
occurrence of types of neoplasias not observed in the
controls; 2) a significant increase in the incidence of the
same types of neoplasms as found in the controls; or 3) a
decreased latency period as compared with controls.
All three
types of responses were observed in animal studies with DMPA.
Similar and some additional criteria for evaluating the
findings from toxicity testing in animals have been proposed
by Squire:
(1) positive results obtained in two or more
animal species; (2) induction of two or more histologically
distinct types of neoplasms in one or more species; (3)
induction of tumors that are normally rare in the test
animals; (4) induction of neoplasm after administration of
low doses, or a shorter latency period required to induce a
neoplastic response; (5) induction of malignant neoplasms,
benign tumors; and (6) positive results obtained
from genotoxicity studies suggesting mutagenic potential
[102] .
All but the last of the phenomena have been observed
in the animal tests of DMPA.
According to the criteria set
by FDA, by the World Health Organization, and those proposed
by Squire, as well as by those suggested most recently in the
69
Office of Science and Technology Policy’s Report on Chemical
Carcinogens [96, 97, 101, 102], the findings must remain a
source of concern.
70
QUESTION 2 - REFERENCES
1.
TR-III, pp. 148-169 (Bardin)
2.
NIH, Workshop on Animal Testing Requirements for New
veneration Steroidal Contraceptives, April 27-29, 19b3,
Vol. I, II, III.
DMB Vols. No. 349, 350 and 351 - Tab #564
3.
IRDC, Long-Term Depo-Provera Study in Dogs, Final
Report 1976.
DMB Vol. No. 5
4.
Dawson Corporation, Long-Term Depo-Provera Study in
Dogs, Interim Findings, and also Final Report, June
1982.
273
Tab #295.
DMB Vol. Nos. 255
5.
The Upjohn Company, Information for FBI on
Responses to the Board’s Questions.
Depo-Provera:
June 25, 1982.
DMB Vol. No. 244 - Tab #219
6.
FDA Addendum to Testimony of M.A.Gross
DMB Vol. No. 311 G-650 A, Vol. 312 G-560 B, Vol. 332
G-668, Vol. 327 G-650 C
7.
TR-II, pp. 75-92
8.
Frank DW, Kirton KT, Murchison TE, Quinlan JW, Coleman
ME, Gilbertson TJ, Feenstra ES, Kimball FA: Mammary
tumors and serum hormones in the bitch treated with
medoxyprogesterone acetate and progesterone for four
years.
Fertil Steril, 31:340-346, 1979.
DMB Vol. No. 184 G-514
9.
IRDC, Long-Term Depo-Provera Study in Monkeys, Final
Report 1979.
DMB Vol. Nos. 181-182 - Tab #67
10.
Nelson LW, Carlton WW, Weikel JH: Canine neoplasms and
progestogens. JAMA, 219:1601-1606, 1972
DMB Vol. 353
Tab #570
1 1 .
Capel-Edwards K, Hall DE,.Fellows KP, Vallance
DK, Davies MJ, Lamb D, Robertson WB: Long-term
administration of progesterone to the female beagle
dog. Toxicol Appl Pharmacol 24:474-488, 1983. DRAFT
DMB Vol. No. 61A G-512, pp. 150-175
71
12.
Progesterone-related gross and
Nelson LW, Kelly WA:
microscopic changes in female beagles. Acta Pathol
13:143-156, 1976.
DMB Vol. No. 348
13.
Fowler EH, Vaughan T, Gotcsik F, Reichhart P, Reed C:
Pathologic Changes in mammary glands and uteri from
beagle bitches receiving low levels of
Medroxyprogesterone Acetate: An overview of research
in progress. In: _Pharmacology of Steroid
Contraceptive Drugs. Garattini S, and Berendes HW
(eds) (Raven Press, New York), pp. 185-210, 1977.
DMB Vol. No. 335 Tab #461
14 .
Graf, KJ, El Etreby, MF:
Endocrinology of reproduction
in the female beagle dog and its significance in
mammary gland tumorigenesis. Acta Endocrinol Suppl
(Supplement) 222:1-33, 1979.
DMB Vol. 217 - Tab #192
15.
El Etrely MF, Graf KJ, Beier S, Eiger W, Gunzel P,
Neuman F:
Suitability of the beagle dog as a test
model for the tumorigenic potential of contraceptive
steroids: a short review. Contraception, 20:237-261,
1 979 .
DMB Vol. No. 352
16.
FDA Summary of the Testimony of Witnesses for the
National Center for Drugs and Biologies, pp. 7-9
(Sieber)
DMB Vol. No. 31 1 G-646
17.
TR-2d, PP- 36-38 (King)
18.
TR-III, pp.
19.
TR-II, pp. 38-39 (Hertz)
20.
Cancer Center of the University of Rochester Medical
Center, Pathology Report of Monkey #459 submitted by
C. Wayne Bardin.
DMB Vol. No. 324 - Tab #414
21 .
Report of Group of Pathology Consultants Designated by
the Depo-Provera Public Board of Inquiry to Evaluate
and Determine the Nature of the Upjohn Company’s Rhesus
Monkey’s Uterine Tumors Observed in the Long-Term IRDC
Study, 1983.
DMB Vol. No. 346 - Tab #548 (No. G)
148-165 (Bardin)
72
22.
Report of WHO Consultation to Review Histological Data
from Monkey Studies Relating to Progestogens and
Uterine Malignancies, Geneva, February 28, 1983.
DMB Vol. No. 332 - Tab #483
23.
Report to German Ministry of Health by Consulting
Pathologists (February 11, 1983, Berlin), included in
Upjohn’s Response to Post-hearing Questions, April 14,
1983.
#506
DMB Vol. No. 343 Tab
24 .
The Upjohn Company, Response to Testimony at the Public
Board of Inquiry Hearing, Wash.,/D.C., Jan. 10-14,
1983, Vol.
I: General and Specific Rebuttals, p. 2.
DMB Vol. No. 333
25.
Hertig A, McKey JJ: Carcinoma in situ of the primate
uterus. Gynecol Oncol 1:165-183, 1973.
DMB Vol. No/ S4"6 - Tab #5 59
26.
Sobel S: FDA, Information regarding neoplasia profiles
of approved contraceptives. Memorandum to FBI, May 12,
1983.
Tab #528
DMB Vol. No. 342
27.
TR-III, P. 78 (Ory) and TR-22, p. 39 (Carlson)
28 .
The Upjohn Company, Final Brief. May 4, 1982, P- 7
DMB Vol. No. 344 - Tab #515
29.
Fraser IS, Weisberg E: A comprehensive review of
injectable contraception with special emphasis on Depot
Medroxyprogesterone Acetate, Med J Aust, special
supplement.
1:1-20, 1981, p. 10.
DMB Vol. No. 222, pp. 232-347
30.
Briggs MH: The beagle and dog contraceptive steroids.
Life Sci, 21:275,1977.
DMB Vol. No. 226, pp. 53-62.
W.Zl.G.i Ad-hoc meeting of the toxicology review panel
assessment of safety of long-acting injectable
contraceptives. Geneva, September 18-22, 1978, section
3.2.8.
DMB Vol. No. 28 G-313
32.
W.H.O.: Facts about injectable contraceptives:
memorandum from a WHO meeting. Bull Wld Hlth Org
60:199-210, 1982.
DMB Vol. No. 288
73
May 4 r 1983, p. 6.
33.
The Upjohn Company, Final Brief.
DMB Vol. No. 344 - Tab #515
34.
Moulton JE: Tumors of the mammary gland. InTumors
in Domestic Animals. Second edition, (University of
California Press. Berkley, California.) pp. 346-371,
1978.
DMB Vol. No. 348
35.
Owen LN, Briggs MH: Contraceptive steroid toxicology
in the beagle dog and its relevance to human
carcinogenicity. Current Medical Research and Opinion
4:309, 1976.
DMB Vol. No. 161A G-513, pp. 235-253
36.
Hampe JF, Misdorp W: Tumors and dysplasia of the
mammary gland. Bull Wld Hl th Org 50:113-133, 1974.
DMB Vol. No. 139A G-513 - Tab 2, pp. 33-35
37.
Misdorp W, Cotchin E, Hampe JF, Jabara AG, Sandersleben
J von: Canine malignant mammary tumors.
II.
Adenocarcinomas, solid carcinomas, and spindle cell
carcinomas. Vet Pathol 9:447-470, 1972.
DMB Vol. No. 352
38.
Gilbertson SR, Kurzman D, Zachrau RE, Hurvita Al, Black
MM: Canine mammary epithelial neoplasms: Biological
Implications of morphological characteristics assessed
in 232 dog. Vet Pathol 20:127-142, 1983.
DMB Vol. NO. 348
39.
Dorn CR, Taylor DN, Schneider R, et al.: Survey of
animal neoplasms in Alameda and Contra Costa Counties,
California. II. Cancer Morbidity in Dogs and Cats from
Alameda County. JNCI 40:307-318, 1968.
DMB Vol. No. 348
40.
Schneider R: Comparison of age, sex and incidence
rates in human and canine breast cancer. Cancer
26:419-426^, 1970
DMB Vol. No. 30 - Tab #460
41 .
Taylor GN, Shabestari L, Williams J, et al.: Mammary
neoplasia in a closed Beagle Colony. Cancer Res
36:2740-2743, 1976.
DMB Vol. No. 30 G-468
42.
Owen LN: A comparative study of canine and human
breast cancer. Invest Cell Pathol 2:257-275, 1979.
DMB Vol. No. 219 - Tab #192
74
43.
MacVean DW, Monlux AW, Anderson PS Jr, Silberg SL,
Roszel JF:
Frequency of Canine and Feline tumors on a
defined population.• Vet Pathol, 15:700-715 (1978)
DMB Vol. No. 352 - Tab #565
44.
Upjohn HL; Portion of Letter to the FDA
Commissioner, August 31, 1967, <cited
*
' \
by Upjohn in its
Responses to the PBI Questions, June 25, p. 59, 1982
DMB Vol. No. 244
45.
McDivitt RW:
Breast carcinoma.
9:3-21, 1978.
DMB Vol. No. 348
46.
Fisher ER, Gregorio RM, Fisher B, et al.: The
pathology of invasive breast cancer. Cancer 36:1-85,
1975.
-----DMB Vol. No. 348
47.
Concannon PW, Hansel W, Visek WJ: The ovarian cycle of
the bitch, plasma oestrogen, LH and progesterone.
Biol Reprod 13:112-121, 1 975.
DMB Vol. No. 348
48.
Johannson E: Testimony to the National Swedish Board
of Health and Welfare. Submitted by the Upjohn company
to the PBI, December 10, 1 982,Append ices, Vol. 11 and
Appendix XVII.
DMB Vol. No. 296
Tab #321
49.
The Upjohn Company, Information for PBI on
Depo-Provera: Responses to the Board's Questions.
June 23, 1982. p. 58.
DMB Vol. No. 244 - Tab #219
50.
TR-III, P- 7-40 (WHO submission)
51 .
Short RV, Drife J: The aetiology of mammary cancer in
man and animals. Symp Zool Soc Lond 41:211-230, 1977.
DMB Vol. No. 353 - Tab #570
52.
van Boagert LJ:
Prog Human Pathol
rne errects
The
effects or
of ovarian steroids on DNA
synthesis in explanted human resting mammary tissues
Cell Tiss Res 188:545-550, 1978.
DMB Vol. No. 352
53.
van Boagert LJ: Effect of hormone on human mammary
duct in vitro. Horm Metab Res 10:337-340, 1 978.
DMB Vol. No. 352
75
54.
Huggins C, Briziarelli G, Sutton H:
Rapid induction of
mammary carcinoma in the rat and the influence of
hormones on the tumors.
J Exp Med 109:25-51, 1958.
DMB Vol. No. 348
55.
Jabara AG:
Effects of progesterone on 9,10
dimthyl-1,2-benzanthracene-induced mammary tumors in
Gprague-Dawley rats.
Br J Cancer 21:418-429, 1967.
DMB Vol. No. 348
56.
Kelly PA, Asseline J, Labrie F, Raynaud JP: Regulation
of hormone RU16117 and other steroids in the rat.
In:
Progress in Cancer Research and Therapy
______ . McGuire WL,
Raynaud JP, Baulieu EE, (eds).
(Raven Press, New
York), 1977, pp. 85-102.
DMB Vol. No. 348
57.
Asseline J, Kelly PA, Caron MC, Labrie FE:
Control of
receptor levels and growth of
7,12-methylbenz [a]anthracene-induced mammary tumors by
estrogens, progesterone and prolactin. Endocrinology
101:666-671, 1977.
“
DMB Vol. No. 352
58.
The Upjohn Company: Roundtable discussions from
workshop of pathologists designated by Upjohn to assess
the endometrial carcinoma in monkeys receiving
Depo-Provera, February 27, 1981.
DMB Vol. No. 211 - Tab #U-14
Reports to the Upjohn Company of the Consulting
Pathologists, submitted by the Upjohn Company to the
PBI, December 10, 1982. Vol. 2, Appendix VI and
Vol. 10 of 12 submitted.
DMB Vol. No.
295
107-139 (Norris)
59.
TR-II, pp.
60.
Trentin JJ, DeVita J, Gardner WU: Effect of moderate
doses of progesterone on mammary growth and hair growth
in dogs. Anat Rec 113:163-172, 1962.
DMB Vol. No. ^46
61 .
Briggs MH:
Progestogens and mammary tumors in the
Res Vet Sci 28:199-202, 1981
beagle bitch.
DMB Vol. No. 335 - Tab #461
62.
TR-III, PP.
168-69,
(Bardin)
76
63.
Animal models and carcinogenicity in of
Briggs MH:
contraceptive steroids. In: Animal Models in Human
Reproduction.
Serio M, Martin L, (eds) (Raven Press,
New Yorkf, T980, pp. 423-431.
DMB Vol. No. 211 - Tab #192 U-15
Cameron AM, Faulkin LJ: Hyperplastic and
inflammatory nodules in the canine mammary gland.
47:1277-1283, 1971.
DMB Vol. No. 29 G-392
JNCI
65.
Foote FW, Stewart FW: Comparative studies of the
cancerous vs. non-cancerous breast. Ann Surg
121:6-222, 1 945.
DMB Vol. NO. 348
66.
Rush BE, Kramer WM: Proliferative histological
changes and occult carcinoma in the breast of the aging
Surg Gynecol Obstet 117:425-432, 1963.
female.
DMB Vol. No . 548
67.
Black MM, Barclay TH, Cutler SJ, Hankey BF, Asire AJ:
Association of atypical character sties of benign
breast lesions with subsequent risk of breast cancer.
Cancer 29:338-343, 1 972.
DMB Vol. No. 348
68.
Ashikari R, Huvos AGr Snyder RE, Lucas JC, Hutter RVP,
McDivitt RW, Schottenfeld D: A Clinicopathological
study of atypical lesions of the breast. Cancer
33:310-317, 1 974 .
DMB Vol. No. 348
69.
McCarty KS, Kesterson GHD, Wilkinson WE, Georgiade N:
Histopathological study of subcutaneous mastectomy
specimens from patients with carcinoma of the
contralateral breast. Surg Gynecol Obstet 147:682-688,
1978 .
DMB Vol. No. 348
70.
Dupont WD, Rogers LW, Vander Zwagg R, Page DL: The
epidemiological study of anatomical markers for
increased risk of mammary cancer. Path Res Pract
166:471-480, 1980.
DMB Vol. No. 348
71 .
Ashikari R, Huvos AG, Snyder RE, Sharma R, Kirsch R,
Schottenfeld D: A clinicopathologic study of atypical
lesions of the breast:
Further follow up. Path Res
Piact 166:481-490, 1980.
DMB Vol. No. 348
77
72.
Devitt JE: Clinical benign disorders of the breast and
carcinoma of the breast.
Surg Gynecol Obstet
152:347-440, 1981.
DMB Vol. NO. 348
73.
Egger H, Tulsan AH, Schneider ML: A contribution to
the natural history of breast cancer.
II Percursors
and lesions associated with small cancers of tne
breast. Arch Gynecol 231:199-208, 1982.
DMB Vol. No. 348
74.
Concannon P, Altszuler N, Hampshire J, Butler R, Hansel
W: Growth hormone, prolactin and cortisol in dogs
developing mammary nodules and an acromegaly-like
appearance during treatment with medroxyprogesterone
acetate.
Endocrinology 106:1173-1177, 1980.
DMB Vol. No. 335
75.
Pearson OH, West CD, Hollander VP, Terves N:
Evaluation of endocrine therapy for advanced breast
cancer. JAMA 154:234-239, 1954.
DMB Vol. No. 348
76.
American Medical Association Council on Drugs,
Androgens and estrogens in the treatment of
disseminated mammary carcinoma.
Retrospective study of
944 patients. JAMA 172:1271-1283, 1960.
DMB Vol. No. 348
77.
Nash H: 1n: Estrogen Target Tissue and Neoplasia, Dao
TL (ed.)r (University of Chicago Press), 1972, pp. 287.
Abstract
DMB Vol. NO. 68 G-512, p. A/14
78.
Carter A, Sedransk N, Kelley RM, Ansfield FJ, Ravdin
RG, Talley RW, Potter NR: Diethylstilbestrol:
Recommended dosages for different categories of breast
cancer patients. JAMA 237:2079-2085, 1977.
DMB Vol. No. 348
2cc:: L, Pieters A, Koenders A, Benraad T, Kloppenberg:
Tamoxifen versus ethinyl estradiol in the treatment of
postmenopausal women with advanced breast cancer.
Cane Treat Reg 65:179-185, 1981.
DMB Vo l ~ "NoT
No.
48
80.
The Upjohn Company, Information for PBI on
Depo-Provera: Responses to the Board’s Questions.
June 25, 1982, pp. 41-47.
DMB Vol. No. 244 - Tab #219
78
81 .
WHO, Annual report for meeting of Advisory Group to the
WHO Special Programme of Research/ Development and
Research Training in Human Reproduction/ 1981.
DMB Vol. No. 289
82.
TR-II, pp. 142-162 (Dallenberg-Hellwig)
TR-2d (See entire transcript)
84.
Wislocki GB/ Streeter BL: On the placentation of the
macaque from the time of implantation until the
Contrib Embryol
formation of the definitive placenta.
27:1-66/ 1938.
DMB Vol. No. 212 - Tab #U-16(h)
85.
Rossman I : The deciduomal reaction in the Rhesus
Am J Anat
Epithelial proliferation.
monkey.
I.
66:277- 1 940 .
DMB Vol. No. 212 - Tab #U-16(g)
86.
Hisaw FL, Hisaw FL Jrz Dawson AB: Effects of the
relaxin on the endothelium of the endometrial blood
vessels in monkeys. Endocrinology 81:365-385/ 1967.
DMB Vol. No. 292
87.
Richardson GSr McLaughlin DH: Hormonal biology of
endometrial cancerF UICC Technical Report, 1978
DMB Vol. No. 3 52 - Tab
88.
Horwitz RI/ Feinstein ARr Vidone RAf Sommers SC, Robboy
SJ: Histopathologic distinctions in the relationship of
estrogens and endometrial cancer. JAMA 246:1425-1427/
1981.
DMB Vol. No. 348
89.
TR-2d, 39-4 2 F (Silberberg)
90.
Two pathogenic types of endometrial
Bokhman JV:
carcinoma. Gynecol Oncol 15:10-17, 1983.
DMB Vol. No. 348 - Tab #564
91 .
TR-I, PP. 73-88 (Gusberg).
92.
Gusberg SB: The individual at high risk for
endometrial carcinoma. Am J Obstet Gynecol
126:535-542/ 1976.
DMB Vol. No. 348
79
93.
metabolism and effects
Goldzieher JW, Kraemer DC: The
’-- ,-----of contraceptive steroids in primates., Acta Endocrinol
(Cited by Mora and
(Kbh) (Suppl) 1667:389, 1971-1972. ("
Johannson).
DMB Vol. No. 310
94.
Fotherby K: Metabolism of synthetic steroids by
animals and man. Acta Endocrinol (Kbh) (Suppl)
185:(75)119-147, 1974 (cited in ref. 95 by Mora and
Johannson).
DMB Vol. No. 75A G, p. 718
95.
Mora G, Johannson EDB: Plasma levels of
medroxyprogesterone□ acetate, estradiol and progesterone
in the rhesus monkey after intramuscular injection of
Depo-Provera. Contraception 14:343-349, 1976.
DMB Vol. No. 75A G, p. 718
96.
Berliner VR: U.S. Food and Drug Administration
requirements for toxicity testing of contraceptive
products. In: P h a rm a c ological Models in Contraceptive
Developments. Briggs MH, and Diczfalcusy E (eds ) / WHO
Research and Training Center, Stockholm, 1974.
DMB Vol. No. 247 G-593
/
97.
* * , Chemical
49 fed. Reg. pp. 21594-21661 (May 22, 1984).
Notice
of
Review
of
the
Science
and its
Carcinogens: 1----- -- --associated principles.
DMB Vol. No. 351
98.
The Upjohn Company, Response to Testimony at the PBI
Hearing, Vol I. General and Specific Rebuttels, March
11, 1983, p. 32.
DMB Vol. No. 333
99.
Frank DW:
1978 Review of animal safety studies with
Depo-Provera. WHO, 1978, pp. 592-593.
DMB Vol. No. 30“5
100.
“ ---- • : 18 Month Toxicity study on Mice; 24
Depo-Provera
101 .
WHO: Evaluation of the carcinogenic risk of chemicals
to human sex hormones. IARC Monpgraph, Vol. 6, Lyons,
1974.
DMB Vol. NO. 30 G-472
102.
Squire RA: Ranking animal carcinogens: a proposed
regulatory approach. Science 214:877-880, 1981.
DMB Vol. No. 348
Moduli Toxicity study in Rats.
DMB Vol. No. 337 - Tab #461
80
THE COMMISSIONER'S QUESTION #3:
CAN THE HUMAN DATA SUBMITTED BY UPJOHN SUCCESSFULLY
REFUTE THE RISK OF HUMAN CANCER SUGGESTED BY THE ANIMAL
DATA?
The data relating to the question whether in the human
the use of DMPA as a contraceptive is associated with any
change in the risk of neoplasias of
inadequate and inconclusive.
genital organs are
Most of the data are derived
from studies that were not designed as epidemiological
studies and therefore do not provide the information needed.
The few studies that specifically address the question of the
effect of DMPA on the incidence of cancer in genital organs
in the human suffer from major limitations in design and/or
execution.
Systematic investigation of the incidence of
neoplasias in long term users of DMPA as a contraceptive have
only been initiated recently.
Consequently, the data
available can not serve to resolve the questions that have
been raised concerning the carcinogenic potential of DMPA,
particular, when used over an extended period of time or
after a period of latency.
The legitimacy of these as yet
unanswered questions and the need to resolve them by
observations on the human rests not only on the findings in
the test animals.
An increased incidence of neoplasias in
in
81
genital organs, the cervix and the breast, in DMPA users was
suggested in a few studies [1-3],
Careful epidemiological observations of women using DMPA
as a contraceptive are of particular importance since they
represent the first large group of humans exposed for a
prolonged period of time to a novel hormonal environment.
This is a hormonal state characterized by relatively high
levels of progestogens in the presence of low, early
follicular phase levels or estrogens.
There is no adequate
body of knowledge that could serve as a basis for predicting
with any degree of certainty what long term consequences.
good or bad, such a hormonal state may have.
Opinion may differ on the value of animal studies ano
the interpretation that can be placed on findings derived
from them.
However, there is general agreement that with
agents such as contraceptives that are likely to be used by
large numbers of healthy women for long periods of time, it
is imperative that their long term consequences, bad or good,
be assessed by adequate human studies.
Moreover, experience
with oral contraceptives in the past has shown that there may
oe populalions of subjects at special risk for certain
adverse effects and the need to identify such subjects.
There is also a general consensus on guidelines of design and
execution of studies that need to be followed to provide
acceptable epidemiological data [4, 5].
82
There is general agreement that the data available on
the human are inadequater as discussed at the public hearing
by witnesses both for Upjohn (Drs. R. Hertz and R. Gray [6]}
(Dr. R. Hoover [7]).
Two arguments have been
presented by those who, while in agreement with this
conclusion r nevertheless feel that DMPA is unlikely to pose
any significant cause for concern as a promoter of neoplasias
in the human.
The first is that even though the human
studies individually may be inconclusive and questionable or.
scientific grounds, they are r in their totality^ reassuring
and sufficient to provide a basis for a regulatory decision.
According to this view, quantity of data can substitute for
quali ty.
This viewpoint was most graphically formulated by
the lawyer representing the Upjohn Company at the Hearing
held by the U.K., Department of Health and Social Security,
in London 1983:
If you have 30 studies...in which each has
a flaw but which flaw cannot be pointed to
as being a biased flaw, a flaw which is
likely to exclude the people who
particularly have the cancer, with any one
paper you can say that there is a flaw, and
those cancer cases might have slipped
through that hole. But if you have 30
papers, each with a theoretical flaw but no
bias, then it is inconceivable that an
increased incidence could slip through all 30.
...If you take 30 pieces of paper, each
with a random hole in it, and you put them
together, you have then got something
solid with no gap in it because it is
inconceivable that in a large number of
varied studies, presumably each with no
bias, the increased incidence could slip
unnoticed through all of them [8].
83
The second argument is that if the drug had any
significant effect on the incidence of neoplasias, this would
have made itself evident and been recognized even without any
specific studies.
The assumption underlying this argument is
that if DMPA caused an increase in the incidence of
neoplasias, this would have been noted by the health care
professionals and/or would have been brought to the attention
of the Upjohn Company through the voluntary reporting
sys tern.
Neither of these arguments is acceptable.
There are few
circumstances under which an adverse reaction to a drug will
make itself obvious in the absence of systematic study [9].
The quantitative information needed on the human to provide
some rational basis for evaluating risk versus benefit can
rarely be obtained either from casual observations or from
poor. inconclusive studies, however numerous.
In connection with DMPA, there were early indications
that there might be a special need to obtain human
epidemiological data on the carcinogenic potential of the
drug, and that concerns about its use might arise if such
data were not forth-coming.
The first suggestion that there
might be an increased incidence of cervical neoplasias was
made in 1971, by Seymour and Powell [1], and the subsequent
analysis of the data obtained from subjects receiving DMPA
under an IND between 1968-72 [2, 10].
These findings were a
major factor why questions were raised concerning the
84
approval of DMPA as a contraceptive in 1974.
The report of
the ‘first beagle study became available one year later.
It
suggested an association between DMPA and malignant
of the breast.
Even earlier, at a meeting of the
FDA’s Obstetrics and Gynecology Advisory Committee held in
1972 to assess the safety of DMPA in the human when used as a
contraceptive, concern was voiced about the risk of breast
cancer that may be associated with the use of DMPA in the
human [11].
According to the testimony given at the public hearing
by Dr. Roy Hertz, a witness for Upjohn, members of the
Obstetrics and Gynecology Advisory Committee objected when
the dog and monkey studies were first proposed as a means to
assess the potential risks of hormonal contraceptives on the
grounds that:
...the pertinence of the pre-existing data on
rats, mice, hamsters and guinea pigs could
never be resolved by additional animal
studies. On the contrary, some of us felt
that only through a carefully designed
epidemiological study could we determine the
real significance and some of the untoward
implications of earlier animal studies [12].
Large numbers of women in various countries including
the U.S.A. were receiving DMPA as a contraceptive while the
long term dog and monkey studies were being completed.
Yet
not until recently, and only after approval for general
marketing,of the drug in the United States was denied, were
1 a
c
made or acted upon to collect data in any systematic
85
manner from humans on the consequences of long term use of
DMPA on the incidence of neoplasias in the secondary sex
structures.
It will now require several years before
information will begin to become available from systematic
follow-up of users in different countries that are planned
(New Zealand and Kenya) or are ongoing [WHO]
[13] .
It is
likely to take many additional years of observation before
the full picture can emerge.
Until then, all we have are
results from studies often designed for other purposes in
which observations on neoplasias were made incidentally, and
from a few attempts at epidemiological studies based on
retrospective surveys, aptly termed as "epidemiological
archeology"
[14] .
These studies have major problems which
limit their value as a source of epidemiological data.
EXPERIMENTAL DESIGN AND EVIDENCE:
3. 1
HUMAN STUDIES
For details of the relevant studies, see Appendices
1-3.
CRITIQUE OF THE EVIDENCE ON THE HUMAN
3.2
The following are major, recurring deficiencies in the
studies reported:
1 )
Too few subjects, especially long term users of
DMPA, studied.
Yet it is prolonged use of the drug that is
thought to be required for the induction or promotion of
neoplasias;
A
86
2)
Too short a period of follow-up, thereby precluding
recognition of neoplasias in secondary sex structures that
may require a long period of latency to develop;
3)
Inadequate or inappropriate controls for comparison
of frequency of neoplasia in users versus nonusers;
4)
Failure to control or adjust for the many known
confounding risk factors for carcinomas of the cervix,
endometrium, and breast;
5)
Lack of documentation of the adequacy of data
collection on subjects.
There is. in fact, evidence both in
the United States and abroad of poor recording of information
on subjects that constitute the basis of some of the studies
[15-17] .
It is recognized that in many of the centers where
DMPA has or is being used, and that have served as a source
of published information, follow-up of patients is difficult
because of local conditions and limited resources (e.g.,
migrant population in Africa [18] or rural population in
Thailand [19]).
This conclusion is supported by the
testimony of witnesses at the public hearing and by documents
on record [20, 21]; and
oI
Lack of information on subjects in the United States
or in populations that may be expected to respond similarly
to those in the United States.
There are well known, marked
differences in the incidence of neoplasias of the breast.
endometrium, and cervix in different regions of the world
[22, 23].
They are based on, as yet, little understood
87
genetic and environmental factors.
Therefore, it is
particularly unfortunate that the opportunity was missed to
collect meaningful information at the Grady Clinic, the only
center in this country where a significant number of
subjects were receiving DMPA as a contraceptive under an IND
over a relatively long period of time (11,500 or 15,600
subjects, depending on the source)
[24-27].
This was a
setting in which. at least theoretically, resources might be
expected to have been available for adequate collection of
data relevant to the population in the United States.
That
this was not the case is evident from the findings by the FDA
audit in 1978 of the conduct of the IND at the Grady Clinic
and the information contained in various documents submitted
and published from that Clinic [15, 24-29] .
It is also
regrettable that studies were not initiated earlier in
countries such as New Zealand, where the drug is apparently
used widely as a contraceptive, where adequate data
collection should be feasible and the incidence of neoplasias
of the breast and endometrium are comparable to those in the
United States [22].
The problem of lack of information on any significant
numbers of long term users of DMPA and the lack of adequate
long term follow-up of such subjects cannot be obscured or
overcome by pooling of data from women who have used the drug
for different periods of time and presenting these data in
terms of "number of women-years or months" of DMPA use, i.e.
88
experience with the drug or surveillance following its use.
The "fallacy of this approach has been stated most aptly, by a
witness for the Women’s Health Network to the effect that
while it takes nine months to produce a baby, nine women.
each one contributing one month, can not produce a baby
[Statement not recorded in transcript].
3.2.1
REVIEW OF HUMAN STUDIES OF BREAST CANCER
There are nine publications [3, 10, 18, 27, 28, 30-33]
m which data are presented on breast pathology in subjects
using DMPA as a contraceptive (Appendix 1).
Observations made on breast pathology on 11,631 women
during studies conducted under an IND when DMPA was first
being tested as a contraceptive are summarized in two related
publications [10, 30].
Though the number of subjects is
relatively large, the data are of limited value for the
following reasons.
The information was pooled from over 80
different centers. some outside the United States.
Thus, the
subjects represent different populations which have. however,
not been identified or defined and for which there were no
t is, therefore. not possible to assess the
significance of 5 cases of breast cancer in this mixed
population of subjects.
Nor are data given on how many of
the subjects were long term users of DMPA.
Instead, the data
are pooled and presented only as number of women years of
experience.
/ >
The number given, 17,408 women years, though
89
seemingly large. represents an average of less than two years
per subject.
More importantly, data presented in this
fashion are essentially useless for trying to assess cancer
L X O A.
Xil
relation to dura Lion o£ drug use.
Neither the study reported from Chile [31] nor the one
from South Africa [18] included control groups.
There are,
moreover, no published data available for comparison on the
incidence of breast cancer in comparable populations in the
same geographic locations from which the subjects were drawn.
Both of these reports are examples of studies that were not
designed to provide epidemiological data.
They are
descriptive clinical reports on patients receiving DMPA.
Specifically,
in a prospective study in Chile, 2,418
women were followed for up to 7 years [31].
This is one
study in which a reasonable proportion of the subjects had
received DMPA for a relatively long period of time (52% of
women had taken the drug longer than three years).
Of ten
women with breast nodules, two were found to have
adenocarcinoma.
No conclusions can be drawn from this
report, not only because there were no controls, but also
because the doses of DMPA given varied, and some of the
subjects were receiving estrogens as well.
In the South African study [18], though the number of
subjects observed was much larger (19,875 women), they
included only few long term users:
seventy percent had
received DMPA for only one year or less and another 28.8% for
( \
90
between one to three years.
Moreover, since these subjects
were from a migratory population, follow-up was difficult and
incomplete, as acknowledged by the investigators.
Not a
mammary carcinoma was identified in this large group
of subjects either at initial examination or during follow-up
[18] .
The absence of breast pathology at first examination
in this number of subjects suggests that the background
incidence of breast cancer in this population is likely to be
low.
However, information on this important point is not
provided or available.
Consequently, the significance of the
data can not be evaluated.
In an update of the "Mexican Experience", Pena-Delgado
et al., describe the incidence of breast pathology in 1,025
randomly selected long-term users of DMPA, i.e.f subjects who
had used DMPA regularly for three to eight years [33].
In
this group of women. 83% of whom were between the ages of
25-39, there were no cases of mammary carcinomas.
Once
again, there was no control group for comparison. and
information is not available on the baseline incidence of
breast cancer in the type of population studied.
Therefore,
the significance of the findings can not be evaluated.
McDaniel and Pardthaisong report in a cross-sectional
study on the results of breast examination conducted at one
time point on 1,270 Thai women receiving DMPA with monthly
supplement of DES [32].
Controls comprised of 257 subjects
presenting themselves at the clinic for contraceptive
91
services for the first time.
No cases of malignant breast
lesions were identified in either group.
subjects was small.
Moreover,
The number of
it included only very few
long term users of DMPA (39.6% were using DMPA for one year
or less, 32.8% for 1-2 yrs., and, only 27.5% for longer than
2 years).
To determine what magnitude of change in the
incidence of breast cancer could have been detected in so
small a number of subjects would require information on
base-line incidence of the disease in Thailand.
This
information is not available, but it is reasonable to assume
that the incidence of breast cancer is likely to be lower in
Thailand than in the United States [22, 23].
If so, a larger
number of subjects would have had to be studied in order to
be able to detect a rare (i.e., low incidence) event.
There have been two recent reports from the Grady
Clinic, the one center in the USA where a large number of
subjects (11,400 or 15,600 depending on document) have
received DMPA as a contraceptive between the years 1968 and
1978 [27, 28]. ' Both were retrospective studies, based on
information retrievable from files and records.
Obviously,
the quality of the initial collection and recording of the
data is of crucial importance in determining the value of the
information presented.
There are reasons to question both
the accuracy and the completeness of the records kept on the
patients because of the findings by FDA in an audit of the
records, in 1978, at the Grady Clinic [15].
t
Reports,
92
published and unpublished, from the Grady Clinic also
acknowledge that in a number of cases essential information
such as parity and dates of injection were not recorded, and
that there were even uncertainties whether some of the
subjects identified through the computer as having received
DMPA had actually received the drug [24, 25, 26].
Liang et al
[28] attempted to retrieve information on
the incidence of cancer of the uterus, breast, and ovary in
5,000 black women receiving DMPA at the Grady Clinic over a
10 year period.
In this retrospective survey, computerized
information on subjects at the Grady Clinic was linked with
comoosite files of hospitalization at the Grady Hospital.
Besides the questionable accuracy and completeness of the
data base, discussed above, this study is of limited value
because of the short exposure of the majority of subjects to
DMPA (58.7% had received DMPA for less than one year, and
only 12.6% for longer than 3 years).
In these subjects, no
increased incidence of neoplasias was identified above <_hat
expected in this age-and—race-specific group.
conclusions are appropriately guarded:
The authors'
"This study indicates
c is unlikely to be a strong association between
DMPA use and cancer of the breast, uterine corpus, or ovary."
To this should be added ”at least not after a short period of
use of the drug."
Greenspan et al. reported a retrospective case-control
study of 30 subjects with breast cancer (29 black, 1 white)
/
93
and 179 matched controls [27].
Again, the duration of DMPA
use "was short (average number of injections:
2.8 for those
with cancer and 3.2 for the 179 controls).
No increased risk
of breast cancer was identified in users of DMPA compared to
nonusers in a study design that could have permitted
identifying a three-fold increase in risk.
Again r the
conclusion of the authors is appropriately modest:
"It seems
reasonable to conclude that short-term use of DMPA in black
women is not associated with any increased risk in breast
cancer" .
To this should be added "at least not a 3-fold
increased risk."
It is important to remember that with
respect to a neoplasia with an incidence as high as that of
breast cancer in the USA (1 woman in 10 is estimated to be at
risk of developing breast cancer in her lifetime), even a
small increase in risk, one that the above study could not
identify, cannot be considered insignificant.
Such risk, if
it exists. would constitute a serious individual as well as a
major public health problem.
A positive association between the use of DMPA as a
contraceptive and breast cancer was reported in one study
L
J
g a restrospective survey of nine centers for the
mentally retarded in Canada, Zarfas reported 3 deaths from
breast cancer among 533 females known to have received DMPA.
He calculated that this number of deaths due to breast cancer
in users of DMPA was much higher than the expected number of
deaths in a comparable age-cohort in the general population
/ >
94
in Canada.
But this study, too, is flawed in that these
women had received, in addition to DMPA, anticonvulsive and
psychoactive medication, and also may have had other risk
factors for breast cancer. such as nulliparity.
The Zarfas study is the only one in which DMPA use was
implicated in causing an increased incidence of breast
cancer.
Advocates of the use of DMPA tend to stress the
deficiencies of this study and to discount it on scientific
grounds.
Conversely, those opposed to the use of DMPA point
to the weaknesses of the studies in which no association
between breast cancer and the drug were identified.
But if
the same standards are applied equally to all studies, we are
left essentially without information on the effect that the
use of DMPA as a contraceptive may have on the incidence of
breast cancer.
The application of appropriate standards to
studies on which decisions are to be based is not simply an
academic exercise.
It is a fundamental requirement for
arriving at conclusions that are not subjective and
arbitrary. and conclusions that can provide an adequate basis
for a regulatory decision.
It is also required by the health
^K^fcoolouals and their clients for making an informed
decision when chosing among therapeutic options.
/
95
3.2.2
REVIEW OF HUMAN STUDIES OF ENDOMETRIAL
CANCER
Most of the evidence presented by the Upjohn Company in
relation to the endometrium and the long term use of DMPA as
a contraceptive consists of findings from endometrial
biopsies [34].
Understandably, the number of subjects that
could be studied by this invasive approach is too small to
provide any definitive data on the incidence of endometrial
cancer in subjects using DMPA as a contraceptive.
All but
the last three of the studies cited by Upjohn were carried
out before the endometrial neoplasias were identified in the
monkeys.
The primary aim of these earlier studies appears to
have been to define the morphological changes induced by the
drug as these relate to its contraceptive effects and to some
of its side effects, e.g., amenorrhoea, menorrhagia. and
return of fertility [35—38].
Therefore, not only were the
numbers of subjects studied small, but the duration of DMPA
use, as well as the follow-up after termination of use of the
drug were short.
Finding no cases of cancer of the
endometrium under these conditions can not constitute a basis
for any conclusions on the issue.
The design of these earlv
studies was appropriate for achieving the stated objectives.
but not for providing information on long term consequences
of DMPA on the incidence of endometrial cancer.
Following the discovery of endometrial cancers in the
two monkeys, five additional studies were initiated.
Three
96
of these, again, were based on endometrial biopsies.
The
last two studies represent attempts to determine.
retrospectively, whether there was any association between
the use of DMPA and cases of endometrial cancer identified in
hospital records (Thailand)
(Mexico)
[17].
[32] or death certificates
None of these latter studies contributes to
the resolution of the question, because in the biopsy studies
the number of subjects was small and the epidemiological
studies had major flaws [Appendix 2].
Pena-Delgado biopsied 92 subjects from a group of 1,025
women and found no neoplasias [34].
Cervantes and Azcona
reported on 76 biopsy specimens, 40 from long term users of
DMPA (4.5-13 years) and 36 from women who had stopped taking
DMPA for 0.6-5.4 years following long term use of the drug.
Our attempts to obtain additional information on these
subjects pointed up serious questions about the quality of
the data [16].
The deficiencies in the data base were
identified under the following circumstances:
since the
biopsy material obtained by Cervantes appeared unique but the
description of the pathological findings presented in the
report by the Upjohn Company ambiguous [39], we wanted to
have the histological slides examined by the pathologists
appointed to review the slides from the uteri of the IRDC
monkeys.
In the course of trying to obtain additional
information on the subjects. the following problems were
lueu Ln ±ed :
97
1)
According to the dates provided in the tables by the
Upjohn Company, three of the subjects receiving DMPA were
pregnant at the time when they were biopsied.
Yet the
oathologists did not report having seen any of the
characteristic changes of pregnancy in the tissue examined.
2)
The Upjohn Company confirmed that one of the
subjects was, indeed, known to be pregnant.
However, it was
unable to clarify the status of the other two subjects at the
time of the biopsy or the fate of the offspring if any were
born, because "the original clinic records apparently have
been destroyed"
3)
[16] .
Upjohn was also unable to offer any explanation for
the obvious discrepancy in at least the one case. Subject
#27, between the pregnant state of the subject and the
pathologist’s finding.
4)
Upjohn did not have and was unable to obtain
information on the dates the biopsies were taken in a number
of the subjects.
The dates provided were stated to have been
estimated from interview dates [16].
Since it was evident that accurate information could not
*_>c
obtained on these subjects, we abandoned plans to have the
histological slides examined by our panel of consulting
pathologists.
There have been only two studies designed as
epidemiological investigations of the effect of DMPA on
endometrial cancer [17, 19].
In both there is serious
98
question of ascertainment.
In the study from Thailand [19],
subjects with endometrial cancer were identified from
hospital records.
It is not known what proportion of
suhiects who become sick in rural Thailand are actually
hospitalized, but it is reasonable to assume that it is small
(see quote from Gray’s letter below).
In this study, in only
9 out of 45 probable cases of endometrial cancer did the
investigators attempt to determine whether they had been
exposed to DMPA (for details see Appendix 2).
Similarly, in
the Mexican study, it is difficult to appraise the accuracy
of the diagnoses obtained from death certificates,
on which the cases were identified
these retrospective studies,
[17].
the basis
Finally, in both of
in an undefined number of cases,
relatives rather than medical records were the source of the
information on whether the subjects had received DMPA.
Relations are unlikely to be an accurate source of
information, especially if we accept the claim that one of
the advantages of DMPA is the privacy it affords.
Some of the problems with the study from Thailand [32]
are highlighted in the letter by Dr. R. H. Gray from the
xuoudoa Scnoul of Tropical Medicine and Hygene to Dr. Colin
McCord :
The review of endometrial cancer
admissions to the Chiang Mai Hospital
between 1974-78 was not intended as a case
control study.
It is merely a first look
at available data. We are aware that case
acertainment would be incomplete and that
the data might be biased in several
99
respects, but still it seemed to be a
reasonable evaluation of the situation....
One can not interpret the findings that
none of the 10 women in the limited review
of cancer cases had used DMPA. The
relative risk would have to be increased
more than 20 fold if we were to detect any
significant effects in such a small
sample. Having said that, I think that
your calculation that only 2.2% of the
cases were reported may be a considerable
underestimate. You used U.S. data to
calculate expected deaths, but it is
known that endometrial carcinoma is far
more frequent among Americans than among
Asian women.
For example, U.S. rates are
more than 12 times greater than in Japan
and around 4.7 times greater than in the
Malay population in Singapore... We do
not know the figures for Thailand, so we
can not estimate the degree of under
reporting. Undoubtedly some cases were
missed since hospital coverage is limited
in northern Thailand. However, there is
no indication of bias since we have no
information to suggest that the use of
DMPA might be different among
non-hospitalized as compared to
hospitalized cases [21].
Finally, the unpublished study by Greenspan,
representing a further attempt to retrieve information from
the "Grady Experience"
information, this time on endometrial
cancer, serves only to demonstrate the problems of trying to
retrieve data when no appropriate record system was set up
Ldie uegxnning
[40] .
It also highlights how misleading
it may be to rely on information obtained from answers given
by patients, in particular when the questionaires are
inadequate (Appendix 2).
Thus, the issue of the risk of endometrial cancer in
long-term users of DMPA appears not to have received much
100
consideration until after such cancers were identified in the
monkeys in 1978.
In conclusion, there are no data available that could
serve as a basis for deciding whether in the human the use of
DMPA as a contraceptive has an effect on the incidence of
endometrial cancer and whether in this respect the human
female's response is similar to or differs from that of tne
Rhesus monkey.
Since 1980, the Upjohn Company has been alerted to two
ins tances in which endometrial carcinoma was diagnosed in
subjects receiving contraceptive doses of DMPA [42, 43] .
These cases are of interest primarily because they point to
problems that may arise in the diagnosis of cancer of the
endometrium in subjects receiving DMPA.
The significance of
irregular bleeding, which is the symptom of such cancers, may
not be recognized because it is an expected side effect of
DMPA use.
Thus the use of DMPA might delay the diagnosis of
endometrial carcinoma.
Both subjects in whom endometrial
cancers were diagnosed were older women.
In one of the cases
the carcinoma was undifferentiated (Class 3, Grade la) and
was diagnosed 9 months after the last injection of DMPA.
the second case the carcinoma was. apparently, well
d if ferentiated.
/ \
In
101
3.2.3
REVIEW OF HUMAN STUDIES OF CERVICAL
CANCER
Historically, the first question raised concerning
DMPA’s carcinogenic potential was in relation to carcinoma of
tne cervix.
This occurred shortly after the introduction of
DMPA for contraceptive use.
When data on the efficacy and
short-term side effects of DMPA were being collected under an
IND, a higher incidence of abnormal cytology was observed at
one of the centers in women receiving DMPA than expected in
that hospital population [1].
Specif ically. in a group of
1,123 women using DMPA, Powell and Seymour reported finding
abnormal cervical cytology in 34 women and cervical cancer in.
situ in 11 women [1] .
These findings were noted over a 66
month period; however, 60% of subjects discontinued using
DMPA over this 5 1/2 year period and only 40% remained in the
cohort at completion of the observation period.
An attempt
was made to obtain cervical smears on all continuing patients
and a "majority" of the discontinued patients; apparently
smears were obtained in 1,107 patients.
The authors
calculated a rate of about 10 cases of carcinoma in situ per
1,000 patients in the DMPA group—over a variable and unclear
time period.
They compared this to a figure of 5 per 1 ,000
"for the past 3 years" in an unexposed group of women seen in
the same hospital.
No information about age, race, etc. is
listed for the comparison group.
P J2. - C|O0
[5x6
COMMUNITY HEALTH CELL
BANGALORE - Soiree
102
Subsequently, B.D. Litt of the FDA attempted to provide
a suitable comparison group for the 10 per 1,000 figure of
Powell and Seymour, but concluded he was unable to do so
because of the incomplete description of the study population
and the inability to relate the observed rate to a defined
time period [2].
Litt used for comparison the figures
provided by the Third National Cancer Survey but pointed out
that this could be questioned since the more careful
screening of the participants in the DMPA study could inflate
the rates.
A curious feature of the observations made on these
subjects is that in many the cervical abnormalities were
diagnosed only a short time after initiation of drug use.
This raises the question whether the pathology can be
attributed to the drug, since with chemical carcinogens a
long lag time is to be expected between exposure to the
carcinogen and the development of neoplastic changes.
Under
the influence of progestogens there are hyperplastic changes
in cervical histology,
45] .
in particular in the endocervix
[44,
However, it would seem unlikely that experienced
pathologists would mistake these for cervical cancer.
In summary, the problems with the observations reported
by Seymour and Powell are:
(1 )
No verification of cytological diagnoses;
103
(2)
Rates of dysplasia and in situ cancer not related
to duration of exposure or elapsed time since
exposure;
The dysplasia and in situ cancers are not analyzed
(3)
in relation to concomitant estrogen
supplementation; and
No suitable comparison group included in the study
(4)
design.
The issue of carcinoma of the cervix in DMPA users was
brought out at a Congressional hearing [46], and was the
basis for FDA’s staying, in October 1974, the provisions of
approving the NDA for use of the drug as a contraceptive
(Table 1, p. 13) .
As a consequence of the questions raised at the
Congressional hearing, FDA obtained from Upjohn an update of
information on subjects that were diagnosed to have carcinoma
of the cervix while receiving the drug as a contraceptive
under an IND [47], and then attempted to analyse all the data
available.
The FDA noted, in addition to the problems listed
above in relation to the report by Seymour and Powell, that
the data were collected from subjects at various centers,
both in and outside the United States representing different
populations w ith unknown background incidence of cervical
cancer.
Therefore, only for the women in the United States
on whom such information was available. could some meaningful
attempt be made to analyse the data.
In both white and non
104
white subjects the age-specific and age-adjusted incidence
was higher than the expected rate based on the Third National
Cancer Survey [2].
This was evident irrespective of the
dosage of DMPA used and whether supplemental estrogen was
administered or not.
As in the study by Seymour and Powell,
the finding of an increased incidence in cervical cancer in
the DMPA treated subjects could have been due to the closer
surveillance of subjects being treated under an IND.
Appropriately designed studies needed to test this assumption
were not initiated, at least not until recently.
In their memorandum of May 17, 1974, officers of FDA who
reviewed the issue of the effect of DMPA on the incidence of
cervical cancer raised at the Congressional hearing,
concluded:
"FDA analysis of the data has been hampered by
the poor quality of the submission and the inaccessibility of
the required information" [47] .
In the same memorandum the following recommendations
were made:
1)
The FDA should undertake a more thorough evaluation
of the Depo-Provera NDA and all NDAs dealing with
contraceptives, especially with regards to adverse
reactions.
2)
An orderly procedure should be developed for
calculating event rates on person years of exposure and
characterizing the population in each clinical trial of a
contraceptive by relevant demographic characteristics.
105
3)
Standardized and age specific rates should be
computed for all important medical events observed during a
contraceptive trial.
There is no indication in the interval between 1974 and
the resubmission by Upjohn of a supplemental NDA in 1978 that
these basic considerations were incorporated into the design
of studies undertaken on the influence of DMPA on "important
medical events" or whether FDA had been successful in
implementing its own recommendations.
In the response to our questions, Upjohn cites, in
addition to Powell and Seymour [1] and Schwallie [10] , only
one other set of observations.
This report by Ide et al
[48]
is lacking in detail and the data presented cannot serve to
support the authors' conclusion that "a possible carcinogenic
effect of the hormone contraceptive on the cervix could be
related to its estrogenic component"
[48] .
There is one additional report, not cited in either
party's responses to us, that addresses the issue of DMPA and
carcinoma of the cervix [52].
Though the study was too small
to resolve the issue, it is of interest since it represents
the single attempt in the past to carry out an appropriately
designed, prospective epidemiological investigation of
carcinoma of the cervix in DMPA users.
In this study from
Chile no statistically significant differences were found in
the incidence of cervical cancer in a group of 2,239 women
receiving DMPA for up to seven years and in 2,409 women using
106
an IUD as a contraceptive .
The number of years of
observation on which the incidence rate is calculated is not
clear.
According to the tables, it was four years, while the
text implies seven years.
In either case, as the authors
acknowledge, the number of women studied and the duration of
follow up were insufficient to provide conclusive evidence
that DMPA has no influence on the incidence of cervical
cancer.
It is evident that the questions raised in 1974
concerning the incidence of the cancer of the cervix in DMPA
treated subjects could only be resolved by appropriately
designed epidemiological studies, that is, studies that
include a suitable control group for comparison.
Yet, no
such studies were initiated until very recently.
The one
started by WHO in October 1979, is not yet completed and only
preliminary data are available.
The studies of the epidemiology of carcinoma of the
cervix are known to present special difficulties.
The
problems relate not only to the numerous putative confounding
factors that need to be controlled for, but also to the lack
of agreement among pathologists on histological diagnosis
[49] .
It has even been suggested that these problems
preclude studying the effect of DMPA on cervical cancer [50].
This pessimistic view seems not to be shared by WHO.
carcinoma of the cervix is among the end points included in
its multinational case control study and is also to be
107
investigated in the study sponsored by Upjohn in New Zealand
[51] .
Thus, until the reports from the above studies are
published, the early suggestions that the drug may increase
the incidence of cervical neoplasia can not be dismissed.
We
have examined the preliminary, confidential report by WHO of
its initial findings, and found no basis for changing this
conclusion .
An examination of the documents and events relating to
the question of cervical cancer and DMPA provide an insight
into the basis of the problems with the evidence submitted on
the safety of DMPA.
The initial collection of data under an
IND was not adequate to provide a basis for evaluating
whether the drug had any influence on "important medical
events" whether carcinoma of the cervix, breast or
endome tr i urn.
This is perhaps not surprising, since at the
early phases of research into the use of the drug as a
contraceptive, the major objective was, understandably, to
establish efficacy, decide on dosage and identify immediate,
short term side effects.
Hence the different sources of
subjects, the range of doses given with or without estrogen
supplementation, and the relatively short and variable period
of observation.
The attempts to derive epidemiological data
from such observations, as in the case of carcinoma of the
cervix, only demonstrate the difficulties inherent in trying
to derive such data from studies not designed to provide
108
them.
Even when these early studies did point to the
possibility that DMPA might increase the incidence of a major
medical event, such as carcinoma of the cervix in the human
(not in a test animal), no attempt was made to initiate
appropriately designed studies to resolve the issue.
Consequently, the NDA submitted in 1978 did not contain any
new, substantial evidence on this or any other of the major
outstanding issues.
109
QUESTION 3 - REFERENCES—^
*1.
Powell LC, Seymour RJ: Effects of
depo-medroxyprogesterone acetate as a contraceptive
agent. Am J Obstet Gynecol 110:36, 1971.
DMB Vol. No. ‘116A G-513
*2.
Litt BD: Statistical review of carcinoma in situ
reported among contraceptive users of Depo-Provers,
1 974 .
DMB Vol. No. 25 G-161 , G-161A through 161N G-148 ,
G-148AZ G-148B, G-170B, G-180, G-180A, G-151, G-17 1 .
*3.
’
~:
The
Zarfas DE, Fyfe I, Gorodzinsky
F:
Utilization of Depo-Provera in
i.. the Ontario Government
Facilities for the Mentally Retarded: A pilot project.
October 1981 (Unpublished).
DMB Vol. No. 203 G-550.
4.
Ory
Ory H:
H: Minutes of a meeting of the Obstetrics and
Gynecology Advisory Committee, January 6, 1977, pp. 9
(H. Ory).
Ory).
(C. Anello) and 41 (H.
DMB Vol. No. 252
5.
Committee of the Epidemiology Work Group of the
Interagency Regulatory Liason Group of the Society for
Epidemiologic Research: Guidelines for Documentation
1 14:609,
of Epidemiologic Studies. Am J Epidemiol
1981 .
DMB Vol. No. 348
6.
TR-I, pp. 192-195 (Gray); pp. 62, 65 (Hertz) and
The Upjohn Company, Text of Witnesses Oral
Presentation (Gray RH, Hertz R)
Tab
DMB Vol. No. 289 - Tab #321, DMB Vol. No. 291
#321
7.
Tr-IIr pp. 184-197 (Hoover)
8.
"
’‘:
Official
U.K., Dept, of Health and Social’ Security:
public
hearing
under
the
provisions
of
transcript of a i
the Medicines Act 1968, Section 21(5), held in London,
April 25-29, 1983. Day 1, p. 9.
(Bathhurst)
DMB Vol. No. 347 - Tab #540
*/
Abstracts of references marked with an asterisk are
found in the appendices.
110
N
9.
Jick H: The discovery of drug-induced illness.
Engl J Med 296:481, 1977.
DMB Vol. No. 352
10.
Experience with Depo-Provera as> an
Schwallie PC:
injectable contraceptive. J Reprod Med t 13:113,
1 974 .
pp. 58-62
DMB Vol. No. 1 35A
11 .
FDA, Obstetrics and Gynecology Advisory Committee,
Transcript of Proceedings, June 2, 1972.
DMB Vol. No. 209 G-564
12.
Tr-I, pp. 38-39 (Hertz).
13.
Tne Upjohn Company, Information for PBI on
Depo-Provera: Responses to the Board's Questions June 25, 1982. pp. 86-98.
Tab #219
DMB Vol. No. 244
14.
TR-III, pp. 95-112 (Potts).
15.
FDA, Data audit of IND #9693 (Depo-Provera as an
injectable contraceptive) at Grady Memorial Hospital,
Dec. 18-22, 1978.
DMB Vol. No. 305 G-652 (Summary)
DMB Vol. No. 305A G-652 (Confidential)
16.
-- j to
Questions on
The Upjohn Company, Response
the
Effect
of
Depo-Provera on
Cervantes Study on 1-1 983 , pp. 2-4.
Endometrium, April 14, 1983,
DMB Vol. No. 343 - Tab #506
*17.
al . : Effect
Cervantes A, Azcona SC, Bribiesca LB, et al.:
of medroxyprogesterone ace ta te on human endometrium
after five or more years of use as a contraceptive.
(Unpublished)
DMB Vol. No. 213 - Tab #U 20(a)
*18.
Rail HJS, et al.: Comparative contraceptive
experience with three—month and six—month
medroxyprogesterone acetate regimens. J Reprod Med
18:55, 1977.
pp. 268-273
DMB Vol. No. 161 A
*19.
McDaniel EB, Potts M: International Forum Update:
Depot medroxyprogesterone acetate and endometrial
carcinoma.
Int J Gynecol Obstet 17:297, 1979.
DMB Vol. No. 222 pp. 43-45
20.
TR-III, pp. 87-88, 89, 90.
111
21 .
Letter from RH Gray, London School of Hygiene and
Tropical Medicine to Dr. Colin McCord, Chief Technical
Advisor at MCH/PP/UNFPA in Dacca, Bangladesh,
April 30, 1980.
DMB Vol. No. 302 G-325
22 .
Waterhouse J, Muir J, Correa P, et al (eds) : Cancer
Incidence in Five Continents, Vol. Ill/ Lyon: I ARC
Publication No. 15, 1976.
DMB Vol. No.
335
23.
Mahboub E, Eyler N, Wynder EL:: Epidemiology of Cancer
Clin
Gynecol, 25:5-17,
of the Endometrium. Cl
--- Obstet
2'-- 1 982 .
DMB Vol. No . 348
24 .
Hatcher RA, Greenspan A, Long WN, et al.:
Experience at the
Depo-medroxyprogesterone Acetate:
Grady Memorial Hospital Family Planning Program in
1967-1979 (Unpublished)
Atlanta, Georgia:
Tab #295
DMB Vol. No. 279
25.
Greenspan AR, Hatcher RA: The prevalence of estrogen
use and pregnancy in women using
Depo-medroxyprogesterone acetate for contraception at
Grady Memorial Hospital Family Planning Clinic.
(Unpublished) Sept. 1979.
DMB Vol. No. 285 - Tab #320
26.
A:: Depo-Provera and Exogenous
Mohberg NR, Greenspan
the
Grady
Memorial Family Planning Clinic:
Estrogen at
Technical Report, September 26, 1 980.
DMB Vol. No. 214 - Tab #U-29
★ 27 .
Greenspan AR, Hatcher RA, Moore M, et al.: The
Association of Depo-medroxyprogesterone acetate and
breast cancer. Contraception 21:563, 1980.
DMB Vol. No. 222 pp. 84-90
*28.
Risk of
Liang AP, Greenspan A, Layde PM, et al.:
Breast, Uterine Corpus, and Ovarian Cancer in Women
Receiving Medroxyprogesterone Injections. JAMA
249 : 2909 , 1983. (unpublished version)
DMB Vol. No. 283 - Tab G 628
*29.
al. : Mortality among
Ory HW, Rubin G, Jones U, et al.:
young black contraceptive users
users..
(Unpublished) 1982
DMB Vol. No. 283 - Tab #190C, G-629
112
*30.
Schwallie PC, Mohberg NR: Medroxyproges terone
Acetate: An Injectable Contraceptive. Advances in
Planned Parenthood, 12:36-43, 1977.
DMB Vol. No. 161A G-513
*31 .
Mammary
Zanartu J, Onetto E, Medina E, Dabances A:
to
under
continuous
exposure
i
women
gland nodules in
1
973
.
Contraception
7
:
2
03
,
progestagens.
DMB Vol. No. 93A, pp. 635-644
*32.
McDaniel EB, Pardthaisong T: Incidence of breast
nodules in women receiving multiple doses of
medroxyprogesterone acetate. J Biosoc Sei 5:83,
1 97 3 .
Tab #66, pp. 31-33
DMB Vol. No. 170
*33.
Pena-Delgado J, Aleman-Herrera M, Baez-Reyes A:
Long-term use of medroxyprogesterone acetate in
contraception.
Sem Med Mex 98:331, 1981.
DMB Vol. No. 221, p. 95
34 .
Tne Upjohn Company, Information for PBI on
Responses to the Board’s Questions Depo-Provera:
pp. 70-81.
1
982.
June 25,
- Tab #219
No
.
244
DMB Vol.
35.
Lee RA: Contraceptive and Endometrial Effects of
Medroxyprogesterone Acetate. Am J Obstet Gynecol,
104:130-133, 1 969 .
DMB Vol. No. 336 - Tab #461
36.
Maqueo MT, Gorodovsky J, Rice-Wrey E, Goldzieber JW:
Endometrial Changes in Women Using Hormonal
Contraceptives for Periods up to Ten Years.
Contraception, 1:115-120, 1970.
DMB Vol. No. 336 - Tab #461
37 .
Roland M, Clyman MJ, Decker A, Ober WB:
Classification of Endometrial Response to Synthetic
Progestogen-Estrogen Compounds. Pertil Steril,
15:143-163, 1964.
DMB Vol. No. 336
38.
Rice-Wray E, Aranda-Rosel1 A,, Maqueo M, Goldzieher DW:
effects of
Comparison of the long-term endometrial
c
synthetic progestins used in fertility control. Am J
Obstet Gynecol, 87:429-433, 1963.
DMB Vol. No. 336
113
39.
jy Information for PBI on
The Upjohn Company,
Responses
to the Board’s Questions Depo-Provera: 1
June 25, 1982, p.
P- 73.
DMB Vol. No. 244 - Tab #219
*40.
Greenspan
AR: a follow-up study of D&C or endometrial
biopsy in Depo-Provera users, December 1978.
(Unpublished, Chap. 8 in Ref. #24)
DMB Vol. No. 221, pp. 388-390
41 .
The Upjohn Company, Information for PBI on
Responses to the Board’s Questions Depo-Provera:
June 25, 1982. p. 76.
Tab #219
DMB Vol. No. 244
42.
The Upjohn Company, Information for PBI on
Responses to the Board’s Questions Depo-Provera:
June 25, 1 982, P- 74.
244 - Tab #219
DMB Vol. No.
*43.
Stamm H, DeGrandi P: Adenocarcinome de 1’endometre
sous traitement contraceptif par medroxyprogesterone.
Rev Med Suisse Romande 101 : 9 1 3, 1981.
DMB Vol. No. 234 - Tab #1
44 .
Maqueo M, Azuela JC, Calderon JJ, Goldzieher JW:
Morphology of the Cervix in Women treated with
synthetic progestins. M J Qbstet Gynecol,
Gynecol, 96
f Vol,
pp. 994-998. 1966
- Tab #461
DMB Vol. No. 339
45.
Candy I, Abell MR: Progestogen-induced Adenomatous
Hyperplasia of the Uterine Cervic. JAMA, 203
p. 323-326
Tab #461
DMB Vol. No. 338
46.
U.S. House of Representatives, Hearings before
oetore a
Subcommittee of the Committee on Government Operations
on the use of advisory committees by the FDA, March,
April, May, 1974.
G-358
DMB Vol. No. 32
47.
Letter from Dr. Dubey Satya for Dr. Charles Annello,
Division of Statistics, Bureau of Drugs, to Dr. Marion
J. Finkel, Acting Director, Office of Scientific
Evaluation, Bureau of Drugs, May 17, 1974.
DMB Vol. No. 25 G-153
114
48.
observations of
Ide Pr Wijants P, Bonte J: Cytological observati
cervico-vaginal smears on hormonal contraception. Rev
Cytol Clin 5:105, 1972. (Cited in Upjohn's response to
the FBI, p. 82)
Tab #219
DMB Vol. No. 244
49.
TR-I pp. 81-82 (Weid, GL)
50.
United States Agency for International
Development: Report on the Ad-Hoc Consultative Panel on
Depo Medrot Medroxyprogesterone Acetate, July 1980.
DMB Vol. No. 2 09 - Tab # 190(e)
51 .
The Upjohn Company, Information for the FBI on
Depo-Provera: Response to the Board’s Questions.
25 1982, pp. 88-94
Tab #219
DMB Vol. No. 244
52.
Dabancens A, Prado R, Larraguibel R, Zanartu J: Intra
Epithelial Cervical Neoplasia in Women Using
Intra-Uterine Devices and Long-Acting Injectable
Progestogens as Contraceptives, Am J Qbstet Gynecol
119:1052-1056 (1974)
DMB Vol. No. 353 - Tab #570
June
115
THE COMMISSIONER’S QUESTION #5:
WHETHER, IN THE EVENT OF CONTRACEPTIVE FAILURE, USE OF
DMPA MIGHT INCREASE THE RISK OF TERATOGENIC EFFECTS
MORE THAN OTHER CONTRACEPTIVES?
There is no evidence to suggest that MPA, intrinsically,
would be more teratogenic for humans than other progestogens
incorporated into oral contraceptives.
Like other
progestogens, MPA may influence sexual differentiation. most
obviously that of the external genitalia and, possibly also
affect the incidence of some rare but serious malformations
of other organ systems.
However, there are two reasons why,
in practice, MPA used as a depot preparation might increase
the risk of any adverse effect the compound may have on the
developing organism.
First, the action of DMPA, once
injected, can not be terminated even if pregnancy is
d iagnosed .
Second, since the use of DMPA is associated with
amenorrhea or irregular bleeding, one of the best and
earliest indicators of pregnancy is lost.
Consequently,
pregnancy may go undiagnosed and the mother may receive
additional injections of the drug.
(This potentially
important problem is not mentioned in the literature and it
is not clear how and when such patients get alerted to the
fact that they are pregnant.)
These two features of DMPA
will tend to increase the duration of exposure of the fetus
116
to the drug and, thereby.
the number and range of critical
events that might be affected by it.
On the other hand, provided that the injections are
administered correctly and the schedule is adhered to,
In
contraceptive failure with DMPA should occur only rarely.
so far as use failure can be reduced or eliminated, overall
failure rate should be lower than that achieved in the
general population with oral contraceptives.
Similarly, if
-guidelines for avoiding injecting patients who may be
pregnant are strictly followed. and, in case of doubt,
pregnancy tests applied, the incidence of inadvertent
exposure of the fetus to DMPA should be reducible still
further.
Finally, judging from the experience gained when
progestogens were used during pregnancy for therapeutic
purposes, the frequency with which the teratogenicity of the
drug is likely to express itself can be estimated to be
small.
Therefore, the chance of an individual using DMPA as
a contraceptive giving birth to an abnormal offspring is very
small and, within reasonable limits, quantifiable.
Consequently, the teratogenic potential of DMPA should,
in
itself, not constitute a reason against using DMPA as a
contraceptive, when otherwise indicated.
The above conclusion should. however, not be equated
with the view that progestogens such as DMPA, whether
administered alone or in combination with estrogens, have
little or no potential to harm the developing fetus [1,2].
117
There is. in our opinion, inadequate basis for rejecting the
possibility that progestogens may be teratogenic, although
this potential action of this class of steroids may be
relevant to only a few, susceptible individuals.
The
d i f ference in viewpoint and interpretation of the data are
not simply of academic interest.
It is likely to influence
the degree of care taken to avoid injecting women with DMPA
who may be pregnant and, therefore, the number of fetuses
exposed to the drug unnecessarily.
The fact that the increased risk indicated in some
studies is not confirmed by others can not be explained
simply on the basis of differences in quality of positive
7/ Rather, it could reflect the
versus negative studies.—
fact that when an agent poses only a small added risk, as is
the case in the positive studies implicating progestogens as
teratogens, and the base line incidence of the abnormality is
low, one is operating close to the limit of sensitivity of
the epidemiological methods used.
It would take very large
studies of many subjects and accurate collection of data to
detect consistently any changes in incidence [6].
Small
risks identified in epidemiological studies could be
7/
Details of the design and the findings in the various
studies have been tabulated in the testimony submitted
by Dr. Grey, witness for Upjohn [3] and will not be
restated. A similar analysis of the substantive
portions of the studies is found in the submission of
Dr. Done, witness for The Women's National Health
Network [4].
118
artifacts and there may, indeed, be no causal relationship.
However, the small size of the risk attributable to
proges togens in the positive studies could be due to the
presence of a small number of susceptible individuals in the
large, heterogeneous population of subjects studied in the
epidemiological surveys.
Factors, including some that are
clearly genetically determined, are being identified that can
account for marked differences in individual susceptibility
-to specific xenobiotics, environmental agents as well as
drugs [7].
Under these circumstances, while the risk may
appear negligible, at least for the population at large. for
the few that are susceptible, it is not small.
Therefore, we
think it is unwarranted to deny the possibility that
exogenous progestogens can cause significant harm to any
fetus on the grounds that epidemiological findings are
contradictory.
We also reject the argument that the
progestogens are unlikely to cause extragenital abnormalities
because no mechanisms have been identified by which they
might do so.
This argument assumes a completeness of
knowledge about the mechanisms of action of steroid hormones
in differentiation and on their target cells that is
unwarranted.
Accepting either of these premises is likely to
encourage a decrease in precautions taken to avoid injecting
DMPA to pregnant women.
That this may, in fact, have
happened is suggested by Upjohn ’s estimate that approximately
one woman in 200-250 receives her first injection of DMPA
119
when already pregnant [1].
It is reinforced by such finding.
as those in the report by Cervantes, in which certainly one.
and possibly three out of 40 long term users of DMPA were
apparently pregnant, presumably unbeknown to themselves or
their physicians, when biopsied for endometrial specimens
(see Question 3).
Proponents of the use of DMPA as a contraceptive have
stressed the very low incidence of any malformations that
might be attributed to DMPA have tended to focus on the
genital abnormalities, specifically in females, and to
minimize their severity.
Opponents of the use of DMPA, on
the other hand, have concentrated on the range and severity
of the congenital malformations that DMPA might cause, with
little regard for the frequency with which such effects may
be expected to occur.
However, both of these factors, the
nature of the abnormalities and the frequency with which they
may be expected to occur following use of DMPA as a
contraceptive need to be evaluated to arrive at an estimate
of risk.
In summary, the two primary concerns in this evaluation
have been to try to determine whether sufficient data are
available to arrive at a conclusion on the nature of the
abnormalities that may be caused by MPA, and to make a
reasonable estimate of the magnitude of risk to the
individual, that is, the chance of her bearing a child with a
congenital abnormality when using DMPA as a contraceptive.
A
120
detailed comparison with the risk posed by other
contraceptive agents was not attempted.
5.1
EVIDENCE ON THE TERATOGENICITY OF DMPA:
OF INFORMATION
SOURCES
The rate of inadvertent exposure of fetuses to
contraceptive doses of DMPA can be deduced from data
available on contraceptive failure.
However z there have been
no systematic studies to define the circumstances under which
inadvertent exposure tends to occur and, most importantly.
the fate of such exposed infants.
That there must be several
thousand such offspring can be deduced from the estimates of
the number of women years of experience with DMPA as a
contraceptive (stated by Upjohn to be to be over 11 million)
and the rates of contraceptive failure reported in the
various studies (generally below 1 but up to 1.5/100
women years of use in some series)
ID .
Yet specific
reference to the fate of fewer than 50 such infants can be
found in the literature.
(See Appendix for summary of some
representive publications in which these cases are
mentioned.)
For the most part. even the sex of the baby is
not stated.
Yet altered sexual differentiation of the
external genitalia is the best documented and understood
congential malformation associated with the use of
progestational agents during early pregnancy.
There is also
no evidence that the infants exposed inadvertently to DMPA in
( '•
121
utero were reexamined at any time after the immediate post
partum period, though it is recognized that many birth
defects, including severe abnormalities. go undetected if the
babies are not followed after birth.
For example, in the
prospective study of 20,000 deliveries by women who were
members of the Kaiser Foundation Health Plan, the incidence
of serious congenital malformates diagnosed at birth was
1.1%, rose to 2.5% at year 1 and to 3.7% at five years [8].
The corresponding figures for non-severe abnormalities in the
same study were 1.8%, 6.2% and 11.3%.
Serious anomalies were
considered ones that. if not corrected, would interfere with
the child’s development and/or well being. e.g., congenital
heart defect, limb reduction, cleft lip [8] .
Consequently,
neither the scant information contained in the few reports in
which mention is made of a few infants born to mothers who
had received DMPA inadvertantly while pregnant nor the fact
that only a single case of mild virilization was reported to
the Upjohn Company under a voluntary reporting system [1],
can provide the data base needed to try to quantify the risk
of teratogenicity of DMPA when used as a contraceptive or to
identify the nature of the abnormalities that might be
attributed to the drug under these conditions.
In responding to our questions Upjohn stated that "After
15 years of postmarketing experience with Depo-Provera,
sample sizes are just becoming adequate to allow systematic
epidemiological studies of children exposed in utero to
/
122
contraceptive doses, particularly those that have reached at
least five years of age"
[1] .
This statement implies that
information on such children might be retrievable.
Documentation has, however, not been provided that adequate
records have been kept on a sufficient number of cases to
make such retrospective survey meaningful.
In addition.
Upjohn refers to plans for one retrospective and one
prospective study in Thailand and New Zealand, respectively.
Until these studies are completed and the data published,
they clearly can not serve as a source of information on the
effects of exposure of fetuses in utero to contraceptive
doses of DMPA.
Since contraceptive failure rate with DMPA is low, a
substantial number of subjects would have had to be followed
in order to accumulate information on an adequate number of
infants born to mothers receiving the drug as a
contraceptive.
Theoretically, such information would have
been simpler to accumulate than that on the incidence of
neoplasias in individuals using DMPA as a contraceptive.
The
risk of exposing an infant to DMPA inadvertently and the
effect the drug might have on the fetus may be assumed not to
be effected by the duration of use of the drug.
Consequently, information can be pooled from subjects
irrespective of duration of treatment and the terms "women
years of experience" can be used legitimately and
meaningfully.
/
This is in contrast to the use of the term in
123
relation to the development of neoplasias, in which case the
length of exposure of each individual has to be taken into
consideration if the data are to be meaningful (see
p. 88) .
In the absence of adequate information on the incidence
of birth defects in infants born to DMPA users as a result of
contraceptive failure, one has to fall back on data available
on the teratogenic effects of MPA and other progestogens when
used in the human for purposes other than contraception.
There is an extensive literature on the effect on fetuses of
progestogens used therapeutically, or to diagnose pregnancy
and of progestogens combined with an estrogen used as a
contraceptive or to diagnose pregnancy [2, 5).
The problems
associated with these studies and with extrapolating from
these data are well recognized.
They include the following:
(1) The estimates of risk are based on pooled data on
the outcome of pregnancy in women who have received different
progestogens, frequently in combination with estrogens and
other drugs.
Though synthetic progestogens share many
actions in common, there are important differences among
them.
These are reflected in differences in relative potency
with respect to various biological end points that
progestogens may effect and. hence, in the spectrum of
salient actions characteristic for each.
Progestogens
are
known to differ markedly in their androgenic potency, the
aspect of their action through which they cause
124
masculinization of female fetuses.
Their antiandrogenic
potency, the basis for their ability to interfere with
masculinization of male fetuses, is also likely to vary.
But
this aspect of their action has been less well characterized.
Thus there is inadequate basis for comparing the relative
antiandrogenic potency of the different progestogens that
have been administered over the years to pregnant women.
(2) There has been a wide variation in the nature and
doses of progestogens given and the time in pregnancy when
they were administered.
(3) The progestogens were given therapeutically to women
suspected of being at risk for spontaneous abortion and
therefore r possibly, already at increased risk for giving
birth to offspring with deformities.
(4) Lack of adequate controls constitutes a major
problem in a number of studies.
(5) A key problem referred to previously is the small
size of the increase in risk attributable to progestogen even
in the positive studies.
Consequently, any effect is likely
to be close to the limit of sensitivity of the epidemiologic
methodology used.
5.2
INFLUENCE OF PROGESTOGENS ON SEXUAL
DIFFERENTIAffON; GENITAL STRUCTURES
In experimental animals progestogens can masculinize the
external genitalia of females by acting as an androgen.
125
However, because of their antiadrogenic potential.
progestogens can also interfere with the normal
masculinization of these structures in males [6,9].
A number
of characteristics of progestogens* actions on its target
organs have been identified that help explain this apparent
paradox (see below).
There is general agreement that in the humanr as in
experimental animals, progestogens may masculinize the
external genitalia of females.
There is, however,
disagreement whether in the human the converse. defective
masculinization of the external genitalia, manifested as
hypospadias, can be attributed to progestogens.
A causal relationship between exogenous sex steroid
administration and masculinization of external genitalia of
the human female could be established relatively easily.
This abnormality occurs only rarely and, when it does, it can
be related in most instances to congenital adrenal
hyperplasis or adrenogenital syndrome, an enzyme defect
causing abnormal androgen secretion by the adrenal cortex.
When masculinization of the female fetus occurs in the
absence of any such abnormal endogenous sburce of androgen
secretion, there is little difficulty in assigning the cause
to the exposure of the fetus during the critical early stages
of its development to an exogenous agent of known androgenic
potential, such as a progestogen [10],
126
The ability of progestogens to cause permanent
masculinization of the external genitalia of the human female
was recognized when these agents were widely used
therapeutically for recurrent or threatened abortions [5,
10] .
However, as was also noted during this period,
masculinization occured only in a small proportion of the
fetuses of mothers given the progestogen.
Moreover, only in
a few of these was the masculinization severe enough to
result in labioscrotal fusion.
In mild cases, clitoral
enlargement, the most frequent abnormality, was noted to
be come less obvious with time, leading to the impression that
the defect is reversible. However, true reversibility of the
abnormality, as suggested by Upjohn [1], would be most
surprising if masculinization occurred during the period of
organogenesis, before the 12th week of life.
Only if the
androgen or androgenic progestogen acted during the later
stages of pregnancy could the clitoral enlargement be
expected to regress.
Fusion of the labioscrotal fold requires corrective
surgery.
Whether the abnormality is considered slight and
the requirement for surgery a relatively trivial matter, as
claimed by Upjohn [1], would depend on the context in which
it develops.
Such abnormalities were initially considered
trivial, because the investigators believed in the efficacy
of the progestogen treatment and that its benefits far
outweighed this risk or side effect.
As stated by Wilkins et
127
al in their first publication on the subject, "It is most
probable that most of the patients we have studied would
never have been born without their use" i.e., without the
progestogen [10].
Under those circumstances, not
surprisingly, they considered even surgical correction to be
a simple and relatively trivial matter.
However,
underplaying this side effect. in relation to the use of
progestogens as a contraceptive, may mislead physicians and
result in less care being taken to avoid injecting women who
may be pregnant.
We have only been able to identify four studies that
provide information on the actual incidence of
masculinization of female offspring of mothers receiving
progestogens therapeutically during pregnancy.
According to
a retrospective survey by Bongiovanni and McPadden [11], from
a total of 650 female offspring of mothers receiving 17
ethinyltestosterone or its nor derivative starting before the
7th week of pregnancy, only 2 females were noted by their
attending physician to exhibit signs of virilization of the
urogenital sinus at birth and on follow up.
In a prospective
study, Burstein and Wasserman also reported a low incidence
of masculinization, 2 cases in 172 mothers receiving Provera,
an incidence of 0.6% [12].
In contrast, Jacobson identified
signs of virilization in 18% of female offspring of mothers
receiving noreth indrone [13].
One reason for the much higher
incidence in this study could be a more focused examination
128
of the offspring for signs of masculinization by a single
physician.
Finally, in a Japanese study an incidence of
2.25% virilization was recorded in 888 female offspring of
mothers treated with progestogens [14].
It should be noted
that in none of these publications is there a definition
provided of the criteria used to consider the clitoris to be
pathological in size.
Since DMPA is relatively less androgenic than many of
the progestogens used therapeutically, the frequency and the
magnitude of the clitoral enlargement is likely to be less in
susceptible fetuses exposed to contraceptive doses of DMPA
than recorded in these surveys.
The reason(s) why only a few
fetuses are susceptible to maculinization, while others are
not, is not known.
tested .
Hypotheses have been offered but none
Th us there is no way of predicting who the few
susceptible individuals might be.
Similar surveys on the incidence of hypospadias in male
offspring of mothers treated with progestogens for pregnancy
salvage are not available in the published literature.
However, there must have been some cases of hypospadias among
such offspring, since the background incidence of this
abnormality in the United States is relatively high (15).
Currently, approximately one in 200-250 males are noted to
have hypospadias at birth.
Some cases of hypospadias in
offspring of mothers given DMPA were, in fact, noted in a
small survey conducted by the Upjohn Company and submitted as
< \
129
part of the application for an NDA for the use of
Depo-Provera as a contraceptive in 1967 [16].
Six out of 204
male offspring of mothers who received DMPA for pregnancy
salvage exhibited some abnormality of the genitalia.
Four of
these were in the 182 males that had been exposed to the DMPA
during the first trimester, the period critical for
masculinization of the external genitalia.
Of these, two had
isolated hypospadias, one bilateral undescended testes and
one bilateral hydrocele.
The two remaining cases of genital
abnormalities, bilateral hydrocoles, occurred in the group of
55 males whose mothers had been treated with DMPA during the
second trimester. that is, after the period of organogenesis.
The incidence of 2 cases of hypospadias in the DMPA treated
group would appear to be high.
However, whether this
represents a statistically significant increase cannot be
deduced from this report because there were no controls and
the sample size was small.
By the same token these data
cannot serve to support the reason why they had been
submitted in the first place, that is. to provide evidence
that DMPA used as a contraceptive is unlikely to increase the
frequency of congenital malformations in general.
Whether progestogens do or do not increase the incidence
of hypospadias is still a subject of controversy, 14 years
after such an association was first suggested by Aarskog
[17] .
Progestogens have been identified as a significant
risk factor in some studies but not in others [17-23].
Tnus,
130
and failure of
a causal association between progestogens
masculinization is less clear cut and, in fact, is likely to
virilization in the
prove more difficult to establish than
female.
Unlike the latter abnormality, minor degrees of
hypospadias, without indentifiable cause, occur relatively
frequently in the U.S.A,
[15].
The number of etiological
of masculine
factors that could interfere with the process
differentiation are also much greater than those that can be
implicated in abnormal masculinization of females.
Masculinization requires adequate production of testosterone
functioning of the mediators
by the fetal testes and normal
its target
of the differentiating action of testosterone on
includes the androgen receptor as well as
organs. The latter
Several mechanisms have
steroid metabolizing enzymes [9].
been identified through which progestogens may act as
antiandrogens and, thereby, intervene in the normal process
of masculinization [24 29].
These include, inhibition of
[24] or in
enzymes involved in the synthesis of testosterone
metabolites in target cells
i ts conversion to active
including 5 alpha reductase, the enzyme required for the
formation of dihydrotestosterone, the mediator of the
differentiating action of testosterone on the external
genitalia [9, 26].
Progestogens can also interfere with and
influence testosterone * s interaction with its receptor
[27-29].
131
Interestingly several congenital defects have also been
ident if ied that could be responsible for defective
masculinization [9] .
There are some clues on factors that
may make some fetuses susceptible to the antiandrogenic
actions of progestogens.
For example, several genetic
defects have been identified that. like progestogens, effect
specific enzymes in the pathway of biosynthesis of
testosterone or enzymes responsible for its conversion to
active metabolites at its target organs [9].
These
abnormalities may be responsible for some cases of congenital
failure of masculinization. such as is manifested by
hypospad ias.
Since these genetic defects appear to be
autosomal recessive in nature, they would find overt
expression only in individuals that inherit a defective gene
from both parents.
The superimposition of a progestogen on
such an abnormality could make overt an abnormality in a
heterozygous individual in whom it would otherwise remain
latent and could exaggerate the abnormality in a homozygous
ind iv idual.
Similarly synergism between the progestogen and
other xenobiotics effecting one of the many steps in the
process of masculinization could be the basis of
susceptibility in others.
These are clearly only hypotheses.
They are advanced only to provide an example of the types of
underlying factors that may make some male fetuses
susceptible to exogenous progestogens.
Like the hypotheses
advanced by others to explain the susceptibility of a small
(>
132
number of female fetuses to the masculinizing actions of
progestogens, they need to be tested.
However, since there
is no way to identify individuals who may be susceptible to
progestogens for whatever reason, it is essential that all
precautions be taken to avoid administering DMPA to women who
may be pregnant and not to underplay the teratogenic
potential of the progestogens.
5.3
INFLUENCE OF PROGESTOGENS_ON SEXUAL
[HF F E R E NT I AT IONS: FUNCTIONS ^MEDIATED BY THE
CENTRAL NERVOUS SYSTEN (CNS)
A legitimate concern with respect to the exposure of
developing humans to drugs and other xenobiotics has to be
whether they effect the future functional capacity of the
individualr in particular the expression of CNS mediated
functions.
Traditionally, teratology has been concerned with
structural abnormalities affecting major organ systems.
Such’
abnormalities are readily identifiable within the first few
years of birth.
Information available on human embryology
also makes it possible to pinpoint times in development when
specific events in the formation of major organ systems occur
and could be disrupted by teratological agents,
Structural
malformations of major organ systems can only be induced by
teratological agents during the period of organogenesis,
within the first trimester.
However, there is increasing
recognition of important functional changes and defects that
I >
133
may be induced by agents acting on the developing organism
after completion of organogenesis when more subtle anatomical
and biochemical developmental events take place.
Interference with these so called organizational events may
lead to permanent functional changes, deficits and
abnormalities in the offspring.
This phenomenon has been
identified and defined in some detail relation to the CNS and
has given rise to the concept of behavioral or
neurob io1og ical teratology.
Sex steroids have been among the first agents recognized
as having important organizational effects on the developing
mammalian CNS.
They are known to act as determinants of the
functional capacity of the individual with respect to a wide
range of sexually dimorphic behavioral functions [30].
Specif ically, in males masculinization of sexually dimorphic
functions mediated by the CNS require that testosterone acts
on the developing brain during critical periods of its
development.
Progestogens, by acting as androgens, could
cause inappropriate masculinization of these functions in
females.
Conversely, since progestogens are also
antiandrogenic, they could interfere with the normal
masculinization of brain by endogenous testosterone in males.
The extensive literature on findings on experimental animals,
including subhuman primates, provides the basis for the
questions that need to be raised concerning the long term
effects that DMPA may have on the developing CNS in the human
I >
134
during later stages of fetal development.
Specifically, can
they cause any masculinization of CNS mediated functions in
females or interfere with normal masculinization of such
functions in males?
As in case of the genital structures, it
will be necessary to keep in mind that there may be certain
individuals who, either because of genetic or environmental
factors, may be particularly susceptible to the androgenic
and antiandrogenic actions of progestogens.
There are
considerable problems involved in obtaining meaningful data
to answer these questions [30].
They include selection of
appropriate behavioral end points for study, their the
quantification, and the need for detailed, long term follow-
up of affected individuals and matched controls by scientists
experienced in the field.
There is one ongoing study of the behavioral and
intellectual development of a small group of male and female
offspring of mothers who have received DMPA therapeutically
during pregnancy [30].
The females are ones born with
virilized external genitalia as a result of this treatment.
The tests being administered aim to identify consequences of
both the androgenic and antiandrogenic effects of the
progestogen.
This study, though small. is of considerable
value because of the care taken in its design and execution.
The findings to date have been reassuring in that no major
d i f ferences from controls have been detected in the indices
of behavioral, psychosexual or cognitive functions selected
135
for study.
Considering the difficulties attending such
stud ies t it is unlikely that very much additional information
will become available in the near future. in particular,
since no adequate provisions appear to have been made to
obtain detailed information and to follow prospectively
offspring exposed in utero to contraceptive doses of DMPA.
It is only through adequately designed prospective studies
that the more subtle effects that drugs may have on CNS
-mediated functions in the human may be identified.
That
these may be unexpected and likely to remain unrecognized
without adequately designed studies is illustrated by recent
findings on the influence of exposure to DES in utero on the
incidence of psychiatric illness in later life [31].
The
uncertainties that are likely to remain with respect to
whether DMPA can cause any permanent functional changes in
the CNS of fetuses and that these may be variations in
individual susceptibility to its actions, underscore once
more the importance of taking all precautions to avoid
injecting women who may be pregnant with contraceptive doses
of the drug.
The concern over potential functional changes that might
be caused by DMPA in the CNS of the developing human has to
be extended to the postnatal period.
This issue and the need
for additional studies to identify any such effects has been
addressed in a report by WHO [32].
One specific area of
concern is a continued role that androgens may have in the
/ \
136
postnatal period in shaping sexually dimorphic functions of
males.
This suggested by the marked elevation in plasma
testosterone in boys during the first three months of life
[33, 34].
Medroxyprogesterone acetate is known to be
excreted into milk were its concentration is comparable to
that in maternal serum [32, 35].
It has been estimated that
the amount absorbed by "a theoretically average infant” and
the levels of the drug in its circulation are likely to be
small [ 35].
However r direct measurements still have not been
carried out to test the appropriateness of the assumptions on
which these calculations are based.
Nor have testosterone
levels been measured in blood of males ingesting MPA and its
metabolites via the maternal milk to determine if the amount
of the drug or its active metabolites transferred to the
neona te is sufficient to have a biological effect.
Such
information is needed to provide a more rational basis for
choices among contraceptives during the period of
lacta t ion.
5.4
PROGESTOGENS AND EXTRAGENITAL ABNORMALITIES
The epidemiological data relating to the question
whether progestogens can be held responsible for causing
congenital abnormalities in extragenital structures is far
from conclusive.
The problem is not so much in the quality
of the date but in the nature of the problems associated with
investigating this question.
It is likely to prove difficult
137
to establish by means of standard epidemiological approaches
does not exist between
whether a causal relationship does or
The
progestogens and the various abnormalities implicated.
number of organ systems implicated in some reports is large,
and without an understanding of underlying mechanisms'there
is no rational basis for grouping the various abnormalities.
The background incidence of the different abnormalities
implicated is low, considerably lower than that of
hypospadias [15] and the increase in risk that could be
attributed to the steriods, even in the most positive
stud ies, is small [ 2-5] .
A wide v ariety of progestogens have
and at different times
been administered at different doses
during pregnancy.
In re trospective studies there have been
serious problems of ascertainment of drug exposure.
Under
the circumstances, it is doubtful if further analyses of the
existing literature and data base can shed any new light on
the subject.
Experimental data on animals are also scant, as
ex tragenitai
is our knowledge of how progestogens might cause
the situation with
defects. This is in clear contrast to
genital structures
respect to the effect of progestogens on
and on sexual differentiation.
However, even if one accepts
the possibility that progestogens are teratogenic for
extragenital structures, it is evident that the risk of DMPA
causing the birth of a child with extragenital abnormalities
as a
is small, provided the drug is used appropriately
contraceptive. The risk is not a sufficient reason to reject
138
using the drug as a contraceptive when indicated or for
aborting a fetus that has been exposed to it inadvertently.
However r the seriousness of the abnormalities implicated
reinforces the need to ensure that all precautions are taken
to avoid administering DMPA to women who may be pregnant.
The absence of known mechanisms by which progestogens
might cause abnormalities in extragenital structures has led
some reviewers to consider such risks to be negligible or
nonex is tent [2].
Specifically, it has been stated that only
genital structures should be considered targets for the
teratogenic effects of progestogens, since only in these
structures have receptors, through which progestogens might
exert an effect, been identified during the period of
organogenesis.
several grounds.
This proposition needs to be questioned on
First, the mechanisms through which
progestogens may act are no t restricted to those mediated by
receptors.
For example, as discussed in relation to the
affects of progestogens on sexual differentiation, these
steroids can also act via non receptor mediated mechanism
by a direct interaction with the enzymes [24-26].
It is also
difficult to explain on the basis of receptor mediated
mechanisms alone the remarkably high concentrations of
labeled steroid found in the adrenal gland of human fetuses
following the administration of radioactive MPA to the mother
[36] .
The significance of this observation, made close to
two decades ago, is still obscure.
Second, progestogen can
139
interact not only with the progesterone receptor but also
In
with the receptors for glucocorticoids and androgens.
fact, the interaction of MPA with the glucocorticoid receptor
has been held responsible for the ability of progestogens to
cause cleft palate in rabbits [37] and its interaction with
the androgen receptor for the ability of progestogens to
masculinize female fetuses, as discussed previously.
It
would require detailed knowledge of the ontogenic pattern of
all three classes of receptors in the various extragenital
organs to exclude the possibility of progestogens having any
receptor mediated teratogenic effects on organogenesis.
The
application of sensitive and discriminating techniques, both
biochemical and autoradiographic, have revealed the presence
of receptors for sex steroids in a number of unexpected sites
in adults (e.g., androgen receptors in the heart muscle,
kidney and fibroblasts [27-29, 41]).
Similar, sensitive
methods have not yet been applied in any systematic manner to
study steroid receptors in various organs of the mammalian
fetus throughout the period of organogenesis.
Thus there is
no theoretical basis for discounting the possibility of
progestogens influencing organogenesis.
! \
140
*/
QUESTION 5 - REFERENCES—
1 .
The Upjohn Company, Information for PBI on
Board
Depo-Provera : Response to the I
-- ”' ’s 2Questions.
June 23, 1982.
DMB Vol. No.
NO. 244 - Tab #219
2.
sex ---------------------------hormones
Brant RL: Are female --------WllbUll UO,
Wilson
JG OLCUIU
teratogenic? Am J Obstet Gynecol 14:567-580 (1981)
DMB Vol. No. 283 G-633
3.
Gray RH: Progestins in therapy: Teratogenes is.
Written testimony submitted by R.H. Grey, witness for
Upjohn at the public hearing
DMB Vol. No. 283 - Tab #321
4.
Written testimony submitted by A. K. Done .r witnesses
for the Women’s Health Network
DMB Vol. No. 302 - Tab #326(c)
5.
Schardein JL: Congenital abnormalities and hormones
during pregnancy. Teratology 22:251-270 (1980)
DMB Vol. No. 246 G-610
6.
Aarskog D: Maternal progestins as a possible cause
of hypospadias. N Eng l J Med 300 :75-78 ( 1 979)
DMB Vol. No. 311 G-654
7.
Omenn GS:
Environmental risk assessment: Relation to
mutagenesis, teratogenesis and reproductive effects.
J Am Coll Toxicol 2:113-123 (1983)
DMB Vol. No. 352 - Tab #565
8.
Torfs CP, Milkovich L, Van Den Berg BJ: The
Relationship between hormonal pregnancy tests and
congenital anomalies: A prospective study. Am J
Epidomiol 113:563-574 (1981)
DMB Vol. No. 289 - Tab #321
9.
Imperato-McGinley J, Peterson RE: Male
pseudohermaphroditism: complexities of male
phenotypic development. Am J Med 61:251-172 (1976)
DMB Vol. No. 352 - Tab #565 pp. 181-202
Abstracts of references marked with an asterisk are
found in the appendices.
141
10.
Wilkins L, Jones HW, Holman GH, Stempfel RS:
Masculinization of female fetus associated with
administration of oral and intramuscular progestins
during gestation: Non-adrenal pseuudohermaphroditism.
J ClinEndocrinol Metab 18:559 (1958)
DMB Vol. No320 - Tab #368
11.
Bongiovanni AM, McFadden AJ: Steroids during
pregnancy and possible fetal consequences. Fertil
Steril 11:181-186, (1960).
DMB Vol. No. 287 - Tab #321
12.
Burstein R, Wasserman HC: The effect of Provera on
the fetus. Obstet Gynecol 23 : 931-934 (1964).
DMB Vol. No. 309
13.
Jacobson BD: The hazards of noreth indrone therapy
during pregnancy. Am J Obstet Gynecol 84:962-968
(1962).
Tab #570
DMB Vol. No. 353
14.
Ishizuka N, Kawashima Y, Nakanishi T, Sugawa T.:
Statistical observation on genital abnormalities of
newborns following the administration of synthetic
progestins to their mothers. J Jap Obstet Gynecol Soc
9:271, 1962.
Tab #570
DMB Vol. NO. 353
15.
Centers for Disease Control: Congenital Malformations
Surveillance Report, January-December 1980, Issued,
February 1982.
Tab #565
DMB Vol. No. 352
16.
Submission by Upjohn 12/26/67
DMB Vol. No. 42 G-512 (confidential)
DMB Vol. No. 42A G-512
*17.
Aarskog D: Clincal and cytogenetic studies in
203:1,
hypospad i as. Acta Paediat Scand (suppl.)
1970 .
DMB Vol. No. 239A A-1
*18.
Czeizel A, Toth J, Erodi E: Aetiological studies of
hypospadias in Hungary. Human Hered 29:166-171,
1979 .
DMB Vol. No. 283 G-624
142
19.
Sweet RA, Schrott HG, Kurland R: Study of the
incidence of hypospadias in Rochester, 1940-1970 and a
case control comparison of possible etiological
factors. Mayo Clin Proc. 49:52, 1974.
DMB Vol. No. 240 A-78
*20.
Mau G: iProgestins during pregnancy and hypospadias,
ogy 24:285, 1981.
Teratoldgy
DMB Vol. No. 248 G-620
21 .
Lorber CA, Cassidy SD, Engel“ E: Is there an
embryofetal exogenous sex i_steroid exposure syndrome?
Fertil Steril, 31:21, 1979.
DMB Vol. No.
No. 239B A-50
22.
Heinonen OP, Slone J, Shapiro: Birth defects and
drugs in pregnancy. Littleton. Mass Pub Science
pp.. 392, 1 977 .
Group. pp
No. 234A - Tab #193(16)
DMB Vol. No.
*23.
Monteleone-Neto R, Castilla EE, Paz JE: Hypospadias:
An epidemiological study in Latin America. Am Med
Genet 10:5-19 1981.
DMB Vol. No.
No . 353 - Tab #570
24.
Goldman AS, Bongiovanni AM: Induced genital
abnormali ties. Ann NY Acad Sci 142:755 (1967).
DMB Vol. No. 239A A-24
25.
Barbieri RL, Ryan KJ:
KJ : Direct effects of
medroxyprogesterone acetate and megasterol acetate on
rat testicular steroidogenesis. Acta Endocrinol
94:419-425, 1980.
DMB Vol. No. 248 G-615
26.
Wright Fr Giacomini* M, Riahi M, Mowszowicz I:
Antihormone activity of progesterone and progestins.
Progesterone and progestins, Bardin C. , Milgrome
I n :__________________
E., and Mauvais-Jarvis, P. (eds), pp. 135-147, 1983
DMB Vol. No. 352 - Tab #565
27 .
Gupta C, Bullock LP, Bardin CW: Further studies on
the antiandrogenic and synandrogenic actions of
progestagens. Endrocrinology 102:736 (1977).
DMB Vol. No. 352 - Tab #565
28.
Brown TR, Bullock LP, Bardin WC: The biological
actions and metabolism of methylprogesterone: a
progestin that mimics and modifies the effects of
Endocr inology 114-1820 (1981)
testosterone.
352
- Tab #565
DMB Vol. No.
143
29.
Bardin WC: The androgenic, an tiandrogenic and
In:
synandrogenic actions of progestins.
Progesterone and progestins, Bardin C, Milgrom E,
Mauvais-Jarvis P. (eds), pp. 135-157 1983 Wright F,
Giacomini M, Riahi M, Mowszowicz I: Antihormone
activity of progesterone and progestins. In:
Progesterone and progestins, Bardin C., Milgrom E.,
and Mauvais-Jarvis, P. (eds), pp. 121-134 1983
DMB Vol. No. 352
30.
Heino HL,
Meyers-Bahlburg H.R.L., Ehrhardt A.A.:
HL
Neurobehavioral effects of prenatal origin.
In:
Drugs and Chemicals; Risks to the fetus and newborn,
pp.' ’gJ^I 07. Allan R. Lisk Inc., N.Y. 1980
DMB Vol. No. 352 - Tab #565
31 .
Vessey MP, Fairweather DVI, Norman-Smith B, Buckley
JA: A randomized double-blind controlled trial of the
long term
value of stilbestrol therapy in pregnancy:
Br
J Obstet
follow up of mothers and their offspring.
Gvnecol 90:1007-1017, 1983
DMB "Vol. No. 352 - Tab #56 5
32.
WHO Technical Report series, The effect of female sex
hormones on fetal development and infant health.
Geneva, Dec. 10th-14th 1979
DMB Vol. No. 248 G-621
33.
Forest MG, Cathiard AM: Pattern of plasma
4-androstenedione in normal
testosterone and
Evidence
for testicular activity at birth.
newborns:
Clin
Endocrino
l
Metab
41:977-987 1975
J_____
Tab
#565
DMB Vol. No. 352 -
34 .
Forest MG, Peretti E Bertrand J: Testicular and
adrenal androgens and
t... their binding to plasma proteins
‘
: Developmental pattern of
in the developmental period:
’
4-androstenedione
,
plasma testosterone,
< dehydroepianrostenedione and its sulphate in premature
and small for date infants as compared with that of
full term infants. J Steroid Biochem 12:25-36 1980
DMB Vol. No. 352 - Tab #565
35.
Schwallie PC: The Effect of Depo-Medroxyprogesterone
acetate on the fetus and nursing infant: a review.
Con tracept ion 23:375-386, 1981
DMB Vol. No. 302 - Tab #326(D4)
144
36.
Besch PK,
In
N Ullery JC, Barry RD, Couri D:
PK Vorys N,
vivo metabolism of 3H-medroxyprogesterone acetate in
pregnant and nonpregnant women and in the fetus. Am J
Obstet Gynecol 95:228-237 1966
DMB Vol. NoT T84 G-515
*37.
Kimmel GL, Hartwell BS, Andrews FD: A potential
mechanism in medroxyprogesterone acetate
teratogenesis. Teratology 19-171-176 (1979)
DMB Vol. No. 184 G-515
38.
Bartsch W, Krieg M, Voight KD: Quantification of
endogenous testosterone, 5 -dihydrotestosterone and 5
-andostane-3 ,17 -diol in subcellular fractions of
the prostate, bulbocavernosus/levator ani muscles,
skeletal muscle and heart muscle of the rat. J
Steroid Biochem 13:259-264 (1980)
DMB Vol. No. 352 - Tab #565
145
THE COMMISSIONER’S QUESTION #6:
WHETHER, IN VIEW OF DEPO-PROVERA’S ADVERSE SIDE EFFECTS
OR PHARMACOLOGICAL EFFECT, ESTROGEN THERAPY IS LIKELY
TO BE PRESCRIBED, IN ADDITION TO DEPO-PROVERA, IN A
SIGNIFICANT NUMBER OF PATIENTS?
Estrogens, including DES, have been used in an attempt
to counter irregular bleeding. the most frequent short-term
side effect of DMPA.
In add it ion , they have been used to
arrest heavy bleeding, a rare but serious complication of the
drug.
There is no adequate documentation of how effective or
ineffective estrogen has proved to be in controlling these
undesirable side effects.
However, there appears to be a
concensus among those with extensive experience in the use of
DMPA as a contraceptive that estrogen is ineffective in
either regulating or arresting uterine bleeding caused by
DMPA [1-4].
Consequently, it is unlikely that estrogens
would be used in conjunction with DMPA in any significant
number of subjects.
Moreover, estrogen is not likely to be
prescribed for the particular groups of subjects for whom the
use of DMPA is indicated, that is, for those with
contra-indication to the use of estrogens, and in those who
do not wish or cannot be relied on to take a medication on a
daily basis.
( \
146
QUESTION 6 ■ REFERENCES
1 .
Long acting progestins - promise and prospects.
Population Reports 9:18-55 1983. Population information
program, The Johns Hopkins University.
DMB Vol. No. 353 - Tab #570
2.
World Health Organization, Injectable hormonal
contraceptives: technical and safety aspects, WHO
offset publication No. 65 1982
DMB Vol. No. 210 - Tab #191
3.
In:
Vecchio TJ: Long acting injectable contraceptives.
Adv Steroid Biochem Pharmacol, Briggs MH, Christie GA
(eds.) Academic 'Press, London, 1 976, pp. 1-64
DMB Vol. No. 299 - Tab #322
4.
Fraser IS, Weisberg, E, A comprehensive review of
injectible contraception with special emphasis on depot
medroxyprogesterone acetate Med J Austral 191, Special
supp:
1-20 1981
Tab #159
DMB Vol. No. 199A
147
INFLUENCE OF DMPA ON PLASMA LIPIDS AND BONE
The focus in the controversy surrounding the general
marketing of DMPA as a contraceptive has been on the
carcinogenic potential of the drug.
In this section we
address briefly two additional issues relevant to an
assessment of the long term safety of the drug, both of which
The
have been raised during the course of this inquiry.
first of these, the effect of MPA on plasma lipoproteins. was
the subject of an extensive discussion at the public hearing
[b 2] .
Its importance relates to findings that some
synthetic progestogens can cause the types of changes in the
pattern of plasma lipoproteins that have been found in
epidemiological studies to be associated with an increased
risk of atherosclerotic cardiovascular disease [3, 5].
The
second, the influence of prolonged use of DMPA on bone
integrity, relates to one of the questions we addressed to
the parties before the public hearing [6].
Specifically, we
requested information on the status of estrogen dependent
structures in long term users of DMPA and whether there is
any evidence of hypoestrogenism in such subjects.
The
consequence of hypoestrogenism on bone is of particular
concern because it could predispose the individual to
osteoporosis.
Both atherosclerosis and osteoporosis are
serious, essentially irreversible conditions.
Therefore, for
the sake of completeness we are including a brief section on
/ \
148
these two issues, although they were not specifically
addressed by the questions of the Commissioner, and our
findings on these issues have not been considered by us in
our decision.
As indicated above, some synthetic progestogens have
been found to alter the profile of plasma lipoproteins [3-
5] .
Specifically, they may reduce the concentrations in
blood of high density lipoproteins (HDL) and elevate the
.concentration low density lipoproteins.
In epidemiological
studies, decreased levels of high density lipoproteins and
increased levels of low density lipoproteins have been found
to be associated with an increased risk of atherosclerotic
heart disease.
These observations provide the basis for the
concern that synthetic progestogens may be a predisposing
factor to atherosclerotic cardiovascular disease andr
therefore, the need to evaluate the effects of MPA on plasma
1ipoprote ins.
The direction of the changes caused by the progestogens
is similar to that caused by androgens [3, 5], specifically
by testosterone, the hormone considered to be one of the
factors responsible for the higher incidence of
atherosclerotic cardiovascular disease in males.
Estrogens,
in contrast, change plasma lipoprotein profiles in the
opposite direction.
Consequently, they are considered to be
one of the factors protecting females from such diseases.
Based on these considerations it is reasonable to propose
149
that synthetic progestogens effect plasma lipoproteins by
virtue of their potential to act as androgens and/or as
antiestrogens, and in case of DMPA, also by causing a
decrease in the levels of estrogens in blood.
However,
progestogens could influence plasma lipoproteins not only by
acting as androgen agonists but also by acting as progestins
or progesterone agonists.
This has been proposed by some
investigators on the basis of a comparison of the influence
of various progestogens w ith different relative androgenic,
antiandrogenic and progestational potencies (see discussion
in [7] ) .
That progesterone may act directly on fat cells and
influence the disposition of lipids is suggested by findings
in experimental animals [8].
In the rat, fat cells have been
found to have both estrogen and progesterone receptors, and
both these steroids appear to play a role in the regulation
and distribution of the bodies’ metabolic resources in part
by acting on peripheral metabolic tissues including the lipid
compartment [8].
It is also of interest to note that the
weight gain experienced by many women receiving DMPA appears
to be attributable to changes in the lipid compartment [9].
The statement found in reviews that progesterone itself
appears to lack effects on lipid metabolism [4] are
referenced by publications that pre-date some of the more
sophisticated methods that have since become available for
establishing plasma lipoprotein profiles.
150
As discussed in relation to the teratogenic effects of
progestogens, the various synthetic progestogens differ
greatly in their relative potencies with respect to their
actions as androgens, antiandrogens, antiestrogens and
progestins.
There have been insufficient systematic studies
of the influence of different progestogens on plasma
lipoprotein patterns to evaluate the relative importance of
these different potential mechanisms of action of these
drugs.
Consequently, there is no adequate basis for
predicting what the consequences of the long term use of DMPA
on plasma lipoprotein profiles might be.
The question needs
to be resolved by direct experimentation, specifically, by
systematic, longitudional collection of data from women
receiving DMPA as a contraceptive.
There have been two reports on plasma lipoproteins in
women receiving DMPA as a contraceptive
[10, 11].
In a cross
sectional study from Holland, 23 women receiving three
monthly injections of 150 mg DMPA for one year or longer had
significantly lower concentrations of plasma high-density
cholesterol levels than a matched group of 23 women using IUD
[7] .
The time elapsed since the last injection did not
appear to effect the findings.
Plasma triglyceride levels
were not significantly different in the two groups.
In the
second study. in 12 Thai women receiving three monthly
injections of 150 mg DMPA, there were no changes in the
concentrations of serum cholesterol, phospholipids or
151
triglycerides either three weeks after initiation of
treatment or after three and 12 months of treatment [10].
The number of subjects in both studies was small and in
neither were all the relevant lipoproteins measured.
Consequently, they do not provide an adequate basis for
evaluating whether the long term use of DMPA affects the
profile of plasma lipoproteins.
A much more detailed analysis of the influence of MPA on
different plasma lipoprotein fractions is to be found in a
report from Finland [12] .
Oral administration of MPA to 8
healthy women for two weeks during the luteal phase of the
cycle was found to have no effect on plasma high density
1ipoprotein fractions.
In contrast, the more androgenic
progestin, levonorgestrel, a nortestosterone derived steroid,
caused under the same conditions a significant reduction in
HDL.
However, the value of this more detailed study for
resolving the issue at hand is limited for two reasons.
First, the subjects were exposed to the progestins for only
two weeks and second, the drug was administered orally rather
than parenterally.
Duration of treatment is an important
consideration since certain effects of progestins on
antiandrogens lipoprotein profiles may take time to become
man i fest [3].
The route of administration of the progestogen
is important since it can affect the potency and actions of
the drug.
When the drug is administered orally, it is acted
on by the liver before reaching other target organs, and it
WITY
C^LL
152
also acts in a high concentration on the liver, one of its
major target organs with respect to metabolic effects (first
pass effect).
The same limitations apply to some of the
other studies in which MPA was reported to have statistically
significant effect on plasma lipoproteins when given orally
either alone or in combination with estrogens [12, 13] .
Clearly, additional studies are needed to determine what
changes may be associated specifically with the use of MPA in
these putative indicators of atherosclerotic cardiovascular
disease, when the drug is administered parenterally and over
at least one year.
This should be possible to accomplish
relatively easily.
More long term epidemiological studies
will be needed to find out how any changes that may be
identified relate to the long term risk of atherosclerotic
carciovascular disease.
Clarification of these issues is
especially important, in particular, since one of the
proposed indications for the use of DMPA is for subjects with
other risk factors for cardiovascular disease, such as
hypertension, which might be exacerbated by estrogens.
The need to evaluate the consequences of the long term
use of DMPA on bone is based on the following considerations.
In women receiving 150 mg DMPA at three monthly intervals the
concentration of estrogens in blood is maintained at the
level found during the early follicular phase in women with
normal menstrual cycles [14].
These levels, though higher
than those found after menopause, are lower than the
153
concentration of estrogens at any other stage of the
menstrual cycle.
The net effect is that estrogen dependent
tissues in women using DMPA as a contraceptive are exposed to
a lower integrated dose of estrogens than normal.
There is
insufficient information to predict whether these levels are
sufficient to maintain the integrity of bone.
If not, this
could result in diminished bone mass and an increased
predisposition to osteoporosis in susceptible individuals.
There is some evidence to suggest that this effect of DMPA
might be counteracted by the drug having, like other
synthetic progestins, a direct anabolic action on bone [15,
18] .
The data available is, however, insufficient to provide
the definitive evidence needed to resolve this issue.
There
have been only few attempts to evaluate the influence of
progestogens in general on calcium balance and on bone, and
we have only been able to identify one very limited study in
which women actually using parenteral DMPA as a contraceptive
have been the subject of such an investigation [19].
Altogether, in only two studies was an attempt made to
evaluate directly the consequences of the administration of
progestogens on bone, in addition to indirect indices of
calcium balance, that is, blood levels or urinary excretion
of calcium, urinary excretion of proline.
To evaluate the influence of MPA on bone, Simpson and
Dale compared serum levels of phosphorus, calcium and
magnesium in women using DMPA with those of subjects using
154
Ovulen, or mechanical devices for contraception and with
pregnant women in 35th week of gestation [19].
In the women
receiving DMPA, calcium levels were reported to be elevated
in the first three months of treatment but then returned to
levels comparable to those not receiving any steroids.
The
DMPA treated women also had higher than normal levels of
serum phosphorus and magnesium levels that did not vary with
duration of use of the drug.
Clearly this study does not
provide much insight into what may be happening to bone
integrity in these subjects.
In another study, the influence
on bone of the prolonged use (3-13 years) as a contraceptive
of another progestogen, lynestrenol, was evaluated in 104
women by applying radiogrammetric methods to x-rays of the
hand [17].
Tne authors report finding an increase in bone
mass in those using lynestrenol and conclude that this
progestin, like estrogens, has a protective effect on bone
mass.
It should be noted that norethisterone is one of the
major metabolites of lynestrenol.
This raises the question
whether the anabolic effect observed in this study might not
be attributable to the androgenic aspect of noreth indrone ’ s
action.
A related steroid, norethisterone. was found to
lower plasma calcium concentrations as well as urinary
calcium/creatinine and urinary hydroxyproline/creatinine
ratios in hypoparathyroid subjects, a finding which led the
investigator to propose a protective role of progestogens on
bone [15].
155
Finally, there have been two studies in which the
influence of medroxyprogesterone acetate on bone integrity
was examined [16, 18].
In both only indirect indices were
used and only the acute effects of the drug evaluated.
Both
studies provide an indication that DMPA may have an anabolic
effect on bone.
This conclusion is based, in the earlier
study by Dompani et al [16], on a decrease in calcium
excretion following acute administration of MPA to patients
with senile osteoporosis and in the recent one by Mandel et
al
[18] on the basis of a decrease in ratios of calcium and
hydroxyproline to creatinine ratios in postmenopausal women
following four weeks treatment with MPA administered orally.
As Mandel et al state in their concluding remark to the
abstract of their publication. "Long term studies of bone
density will be required to confirm these apparent.
beneficial effects of DMPA on bone metabolism in
postmenopausal women."
Such long term studies are equally
necessary to establish definitively that bone integrity is
maintained in subjects using DMPA as a contraceptive for an
extended period of time.
In conclusion, the questions concerning MPA’s influence
on plasma lipoproteins and bone have a legitimate basis.
Women using DMPA as a contraceptive have not been the subject
of any systematic investigation with respect to these issues.
The available data on the effect of various progestogens on
plasma lipoproteins is too incomplete and contradictory to
156
provide an adequate basis for concluding whether DMPA is or
is not likely to effect the incidence of atherosclerotic
cardiovascular disease.
The data on the influence of
synthetic progestogens on bone may be considered encouraging
in that, at least according to certain indirect indices of
bone metabolisms, they appear to have an anabolic effect on
bone.
To establish this definitively will however, require
additional studies.
As stated previously, our analysis of the data of the
influence of DMPA on plasma lipoprotein profiles and on bone
integrity has not been considered in making our final
decision.
There are two reasons for this.
First, these were
not issues addressed directly by the Commissioner’s
questions.
Second, their importance in contrast to that of
DMPA’s carcinogenic potential, has only come to be recognized
relatively recently.
The data from large scale
epidemiological studies, establishing an association between
high and low density plasma lipoprotein concentration in
blood and atherosclerotic cardio-vascular disease, only
became available in the late 1970’s.
The prevelance of post
menopausal osteoporosis in the U.S.A. has also become
recognized only relatively recently.
We have addressed these
issues since they were raised during the inquiry.
Moreover,
they represent valid questions and concerns that, now that
they have become clearly identified, will need to be resolved
promptly by appropriately designed studies.
157
REFERENCES
1.
Written testimony of Dr. JC LaRosa, witness for FDA.
DMB Vol. No. 311 G-643
2.
Tr-II, pp. 42-58 (JC LaRosa)
3.
Lithell H, Ahren T, Odlind V, Weiner E, Vessby B,
Victor A, Johannson EB: Effects of progestins on
lipoprotein patterns. In: Progesterone and
progestins, C.W. Bardin C.W. Milgrom, E. and
Mauvais-Jarvis P. (eds.). Raven press, New York, N.Y.
1 983.
DMB Vol. No. 352
Tab #565
4.
Miller, G.J.: Clinical implications of hormonal
changes in lipoprotein metabolism, with special
reference to the progestins.
In: Progesterone and
progestins, C.W. Bardin C.W. Milgrom, E. and
Mauvais-Jarvis P. (eds.). Raven press. New York, N.Y.
1983, pp. 433-446.
DMB Vol. No. 352
Tab #565
5.
Krauss RM: Effects of progestational agents on serum
lipids and lipoproteins. J Reprod Med 27:510 1982
DMB Vol. No. 352 - Tab #565
6.
The Upjohn Company, Information for PBI on Depo-Provera
Response to the Board's Questions, June 25, 1982
DMB Vol. No. 224 - Tab #219
7.
Kremer J, Bruijn HWA, Hindriks FR: Serum high density
lipoprotein levels in women using contraceptive
injection of depo-medroxyprogesterone acetate
Contraception, 22:359-367 1980.
DMB Vol. No. 224 - Tab #219
8.
Gray JM, Wade GN: Food intake, body weight, and
adiposity in female rats: actions and interactions of
progestins and antiandrogens. Am J Physiol
2 40: E474-E48 1 , 1 98 1
’
DMB Vol. No. 352 - Tab #565
9.
Amatayakul K, Sivascomboon B, Thanangkul O: A study of
the mechanism of weight gain in medroxyprogesterone
acetate users. Con tracept ion 22:605-622 1980
DMB Vol. No. 353“-“Tab T570
158
10.
Amatayakul K, Silvassomboon B, Sinkamani R: Effects of
‘
* j on serum lipid, protein,
medroxyprogesterone acetate
glucose"tolerance and liver function in Thai women,
Con traception 21 :83 1 980
DMB Vol. No. 353 - Tab #570
11.
Tikkanen MJ, Nikkala EA, Kuusi T, Sipinen S: Different
effects of two progestins on plasma high density
lipoprotein (HDL 2) and postheparin plasma hepatic
lipase activity, Ath e rosclerosis 40 : 365-369 , 1 981
DMB Vol. No. 273 G-589
12.
Bradley DD, Wingard J, Pettiti DB, Krauss RM,
Ramacharan S: Serum-high-density-1ipoprotein
cholesterol in women using oral contraceptives,
estrogen and progestins, N Engl J Med, 299:17-20, 1 978
DMB Vol. No. 206 G-552
13.
Silverstolpe G, Gustafson A, Samsioe G, Svanborg A:
Lipid metabolism studies in oophorectomized women.
Effects of three different progestogens, Acta Obstet
Gynecol Scand (suppl), 88:89-95, 1979
DMB Vol. No. 235A - Tab #193
14.
Jeppson S, Johansson EDB: Medroxyprogesterone acetate,
estradiol, FSH and LH in peripheral blood after
intramuscular administration of Depo-Provera to women,
Contraception, 14:461-479, 1976.
DMB Vol. No. 161A G-513
15.
Gallagher JC: Biochemical Effects of Estrogens and
Progesterone on Calcium Metabolism, Osteoporosis:
Recent Advances in Pathogenesis and Treatment, Univ.
Press, Baltimore, 1981
DMB Vol. No. 352 - Tab #565
16.
Molinis G, Bompani R, Sacli G, Gavazzoli L:
Modificazioni indotte dal trattamento con
medrossiprogesterone acetato sull'escrezione urinaria
del calcio dopo carico per os e per vena in soggetti
anziani. G Gerentol 18:361-372 1970 English
translation on record.
DMB Vol. No. 205A - Tab #183
17.
Dequeker J, de Muylder E, Fer in J: Effect of
Lynestrenol treatment on bone mass in cycling women.
Con t r a c e p t i o n, 15:717-722 1977
Tab #322
DMB Vol. No. 298
159
18.
Mandell FP, Davidson BJ, Erlik Y, Judd HL, Meldrum DR:
Effects of Progestins on Bone Metabolism in
Postmenopausal Women. J Reprod Med, 27:511-514, 1982
DMB Vol. No. 352 - Tab #565
160
THE COMMISSIONER'S QUESTIONS 1, 4 AND 7 REFORMULATED:
WHETHER THE RATIO OF BENEFIT TO RISK OF DEPO-PROVERA
WHEN USED AS A CONTRACEPTIVE WARRANTS ITS APPROVAL UNDER
CONDITIONS OF GENERAL MARKETING? WHETHER APPROVAL FOR
GENERAL MARKETING AS A CONTRACEPTIVE MIGHT INCREASE ITS
USE FOR UNAPPROVED INDICATIONS? AND, IF NOT APPROVED
FOR GENERAL MARKETING, ARE THERE CONDITIONS FOR
CONTROLLING THE LIMITED DISTRIBUTION OF THE DRUG AS A
CONTRACEPTIVE FOR CERTAIN PATIENTS WITH SPECIAL NEEDS?
The available evidence presented fails to provide an
adequate, scientifically justifiable basis for concluding
whether the use of DMPA as a contraceptive does or does not
pose any long term risks.
The collection of data from the
women who have been using DMPA as a contraceptive world wide
over the last 15 years has been too haphazard and
uncoordinated to provide evidence of the nature and quality
required to resolve major outstanding questions concerning
the drug's long term safety.
The theoretical arguments and
hypotheses offered why MPA is unlikely to be carcinogenic in
the human or why the findings of neoplasias in the test
animals are not to be considered relevant to the human do not
substitute for the lack of data.
In the absence of adequate
data there is no basis for concluding that the benefits of
the drug as a contraceptive outweigh its risks, a requirement
which is inherent for approval of an NDA in the U.S.A.
[1] .
Neither the obvious benefits offered by the drug nor the fact
that it appears to be devoid of short term serious and
161
irreversible adverse effects can compensate for the lack of
data on the long term consequences of its use as a
contraceptive.
Depo-Provera represents the first successful large scale
application to contraception of the concept of using a pure
progestational agent, and of a formulation of a chemical
contraceptive that provides convenient long-term protection
from pregnancy.
DMPA is a highly effective contraceptive.
Method failure appears to be as low as that with oral
contraceptives, but with DMPA, failures due to lack of
patient compliance can be reduced or eliminated.
Consequently, it is possible to achieve with DMPA lower
overall lower failure rates than with any other currently
available reversible contraceptive.
The convenience and the
privacy afforded to the individual user, as well as the
relief from the burden of daily pill taking are some
advantages of DMPA that have been stressed.
The absence of
any estrogen in the formulation and the lack of any obvious
estrogenic action of MPA can also be of importance to certain
subj ects.
The short term, immed iate t and reversible side effects
of DMPA have been well studied and documented.
While some of
these side effects, such as excessive bleeding or spotting,
depression, headaches, weight gain, and loss of libido, are
not trivial, the drug has proved acceptable to a large number
of women.
/ A
Some of the side effects may prove intolerable to
162
certain subjects leading to an overall discontinuation rate
for DMPA at least as great as that for oral contraceptives.
One obvious, regrettable omission in the literature on the
immediate side effects of DMPA is information on how long it
takes for potentially serious complications, such as
However,
depression, to disappear after the last injection.
there is no evidence that following short term use DMPA
causes serious, irreversible side effects, such as the
thromboembolic complications associated with oral
contraceptives.
Serious bleeding requiring surgical
intervention is reported to be rare.
Fetal abnormality is
the one potentially serious and irreversible consequence of
DMPA that may occur at any time during DMPA use.
But, as
discussed in relation to question 5, this can be considered a
rare and largely avoidable risk.
Thus, neither the
short term side effects of the drug nor its teratogenic
potential should constitute a reason for not proposing to use
DMPA as a contraceptive, provided that these risks are
communicated clearly to potential users.
They do not
constitute significantly greater problems than are posed by
oral contraceptives.
In contrast, the possibility that DMPA may have serious
irreversible effects that become evident only after its
long-term use or after a period of latency must continue to
pose a problem to the approval of the drug for general
marketing as a contraceptive.
Concerns over DMPA’s
163
carcinogenic potential was why in 1974 questions were raised
about the advisability of approving the supplemental NDA for
use of the drug as a contraceptive.
The questions arose
because an analysis of the data then available from the
initial IND studies indicated increased incidence of
carcinoma of the cervix in women receiving DMPA as a
contraceptive [see pp.
101-108].
While the findings can not
be considered conclusive, since data on appropriate controls
we re not available, they did provide a clear indication for a
need to carry out controlled studies.
Yet no such studies
were undertaken, at least not until recently.
Therefore,
there are no additional data at this time to either confirm
or dispell the suspicions raised by these early observations
that DMPA may be carcinogenic in the human.
In the interim,
the suspicion that DMPA may be carcinogenic has been
reinforced by the findings of neoplasias in DMPA treated dogs
and monkeys, the two species selected to study the long term
effects of the drug.
During the same period of time, data
that could have served to evaluate the significance to the
human of the findings in test animals were not being
collected in any purposeful and systematic manner.
Only
recently have appropriately designed epidemiological studies
been initiated to follow the very large number of women
receiving contraceptive doses of DMPA.
Nor has any
significant new information been added to help resolve the
long-standing controversy whether the dog is, or is not, an
164
appropriate species for studying the influence of
progestogens on the human breast (see pp. 36-54 ) .
This
failure to accumulate relevant, scientifically valid, new
information is unfortunate and frustrating.
If the fear of
carcinogenesis is unfounded, this needs to be established
lest an important addition to the options available for
fertility control be condemned unnecessarily.
If; on the
other hand, such a risk does exist, its magnitude must be
def ined, and the use of the drug limited accordingly.
Data
are also inadequate on the effect of DMPA on bone and on
plasma lipoprotein patterns to provide a basis for deciding
whether the long term use of the drug does or does not
increase the subsequent risk of osteoporosis or
atherosclerosis.
Thus , while the benefits of the drug are
clear, there is insufficient basis for evaluating the risks
associated with its use.
The questions raised concerning the long term risks of
DMPA relate, for the most part, to diseases to which women in
the United States are already at increased risk.
Tn is
applies to carcinoma of the breast and uterus, postmenopausal
osteoporosis and atherosclerosis.
Should DMPA prove to cause
even a relatively small increase in the risk of any one of
these diseases, this would carry a different significance in
the United States than in countries where the background
incidence of these diseases is markedly lower.
Consideration
of the specific circumstances of each country and population
165
of subjects can, we think, effect not only the benefit but
also the risk side of the equation on which the decision
whether or not to use the drug must depend.
Approval of a drug for general marketing implies that
questions raised concerning major side effects and risks
associated with its use have been adequately investigated and
resolved.
In this context, resolution, like the term safety,
does not necessarily mean that the drug has been shown to be
entirely without major undesirable effects, but that these
.side effects have been defined and the frequency of th e i r
occurrence established within reasonable limits by
scientifically acceptable methods.
This information is
needed to enable health care professionals and the subjects
they treat to arrive at an informed decision concerning the
risk vs benefits associated with the use of a drug and to
make an informed choice among drugs.
This is a reasonable
expectation for a drug that has been approved for general
marketing and one that also conforms to the legal
requirements .
Such information is not available for DMPA and
in its absence there is no valid basis for comparing the
risks of DMPA with those of other contraceptives.
The immediate benefits of DMPA as a contraceptive.
combined with the fact that no immediate life-threatening
complications appear to be associated with its use, may make
questions of possible delayed, serious adverse effects of the
drug appear, to some, a theoretical exercise.
We reject this
166
view.
In certain situations the immediate benefits may.
indeed, outweigh any unknown long term risks.
However, the
long term risks need to be defined before a decision can be
made on the level of safety of a contraceptive intended for
use by essentially healthy women, possibly for long periods
of time.
The immediate risks posed by pregnancy, the problems of
unwanted pregnancies and the risks of abortion, are well
documented (testimony of E. Connell
[1]).
The importance for
women of achieving control over their fertility is also
recogn i zed .
These considerations have been rightly stressed
by those involved in the important social task of family
planning .
Additional options for contraception are
unquestionably needed.
But the risks of unwanted pregnancy
and the benefits of DMPA as a contraceptive can not obscure
the fact that the factual information available on DMPA fails
to provide adequate, scientifically valid information on the
major outstanding question of the drug s long term side
effects.
To imply by approval of the drug for general
marketing that such data exist would be misleading.
The fact that the available scientific evidence on DMPA
is, in our opinion, insufficient to warrant its approval as a
contraceptive under conditions of general marketing in the
United States makes the question whether such an approval
would lead to increased use of the drug for unapproved
indications largely irrelevant.
In any case, it would seem
167
that the decision on approval of DMPA as a contraceptive
should r in any case, not be influenced by what consequence a
decision would have on the use of the drug for otherr
unapproved indications.
Physicians in the United States may
and are using DMPA for a variety of unapproved indications
either as part of an approved new drug investigation or as
part of their own practice of medicine.
It is, we
understand , FDA policy not to regulate the physician's
practice of medicine in prescribing approved drugs for
unapproved indications.
Since our analysis leads us to conclude that the data
available are insufficient to warrant approval of DMPA for
general marketing in the United States, the next
consideration is whether the drug should or could be made
available as an approved drug for certain patients with
special needs.
There are, unquestionably, individuals with
special needs for whom using DMPA as a contraceptive may
indeed provide a reasonable and desirable option.
They
include those in need of the degree of protection from
pregnancy offered by chemical contraception but for whom the
use of oral contraceptives is contraindicated for medical
reasons.
Similarly, there may be compelling personal reasons
why DMPA may need to be considered as a contraceptive for
some individuals.
These include subjects who can not rely on
themselves or can not be relied on to take a pill regularly
or need the privacy offered by the drug.
f \
168
Upjohn provides in its final brief a more extensive list
Of 10 categories of patients for whom it considers
the
risk/benefit ratio to be particularly favorable [2].—/
Seven of these 10 categories of indications (#’s 1,
2, 3, 5,
7 and 8) would essentially cover the subjects
we have
identified, and listed above, as having special needs
that
would warrant considering using DMPA as
a contraceptive. Two
of the categories refer to subjects for whom
the use of DMPA
may be indicated for very special reasons and which,
in our
opinion, are not relevant to the
more general issue of the
use of the drug for contraception,
specifically, subjects
with sickle cell anemia (# 4) and those
for whom amenorrhea
would be advantageous (# 10).
In persons with sickle cell anemia,
the indication for
□sing DMPA rests on a study in which DMPA
was found to
improve the hematological status of patients with
th is
disease [3]. Confirmation of these
interesting observations
8/
D
2)
3)
4)
5)
6)
7)
8)
9)
10)
f A
Patients who have medical contraindications to other
highly effective methods and who are unwilling or
unable to terminate pregnancy due to method
failure;
who have a high
1
risk of infection from IUDs;
who have a risk of cardiovascular
----- : problems,
regardless of age and
.2 smoking habits;
who have sickle cell anemia;
who can not tolerate tthe
’
side effects of estrogens;
who have had repeated failures
------------’ with other methods;
who have had repeated abortions;
who need a rcontraceptive method that
provides a high
degree of privacy;
for whom anovulation would provide relief
dysemenorrhoea and premenstrual syndrome; from
for whom amenorrhea would be advantageous.
169
is needed on a larger number of subjects.
Subjects for whom
amenorrhea would be advantageous presumably refers to those
mentally retarded for whom the drug would be used for hygenic
reasons.
Finally, it is not clear why Upjohn considers that
DMPA would be "uniquely well suited n for women needing relief
from dysmenorrhoea or premenstrual tension, since oral
contraceptives may be equally effective in such cases.
Whether it is desirable and feasible for the FDA to
approve the use of DMPA as a contraceptive for a lim i ted ,
defined population of subjects under controlled conditions,
is a subject on which there was some divergence of opinion
among the three members of the PBI.
Two of us (Drs. Stolley
and Weisz) do not think it desirable that the FDA set
up broad categories of indications for use of DMPA as a
contraceptive since the use of the drug is likely to be
appropriate only for selected patients within each category.
It is, in our opinion, preferable if the indications and the
risk/benefit evaluation are arrived at on an individual
basis, and with informed choice and consent being arrived af
by the physician and the patient or those responsible for
her.
We do not think that indications should be considered
routine.
In those instances in which the physician and his
or her patient should come to the conclusion that DMPA is the
drug of choice for that patient, this option is available
under the practice of medicine without any additional action
on the part of FDA since DMPA is currently approved in the
170
United States for use for other indications.
A minor
inconvenience is that the dose of the injectable formulation
on the market for the approved indication, i.e., for
endometrial cancer. is larger than that needed for
contraception and needs to be adjusted accordingly.
Two of the Board members (Drs. Stolley and Weisz) also
question whether there is, at this time. any effective
mechanism by which FDA can limit the distribution of the drug
to a specific population of subjects or by which information
on subjects receiving the drug as a contraceptive can be
collected in a systematic fashion in the U.S.A.
Dr. Ross, on the other hand, suggests that, if feasible,
approval should be given for use of the drug for two broad
categories of patients, the mentally retarded and drug
addicts (see Dr. Ross' letter, P.
181).
A summary of our Findings of Fact and our Conclusions of
Law follow.
171
REFERENCES
( A
1.
Tr-I (Dr. E. Connell)
2.
The Upjohn Company, Final Brief.
DMB Vol. No. 344 - Tab #515
3.
Ceulaer K, Hayes R, Gruber C, Serjeant GR:
Medroxyprogesterone Acetate and Homozygous sickle-cell
disease. Lancet 2:229-231, July 31, 1982
DMB Vol. No. 282 - Tab #317
May 2 r 1983, P- 3
-4
, Ke^cvH
"TlMt
r
e■
'A
172
(A;-
/7 Oc'^
FINDINGS OF FACT
DATA AVAILABLE ON THE LONG-TERM RISKS OF DMPA ARE
INSUFFICIENT AND INADEQUATE TO PROVIDE A BASIS FOR A
DECISION WHETHER THE BENEFITS OF THE DRUG AS A
CONTRACEPTIVE OUTWEIGH ITS DISADVANTAGES UNDER
CONDITIONS OF GENERAL MARKETING IN THE USA.
I.
There are adequate data to assess the efficacy and
benefits of DMPA as a contraceptive.
There is also
sufficient information on its short term side effects and
risks.
The drug is clearly a highly effective contraceptive
with certain specific advantages, and it does not appear to
pose any immediate irreversible serious side effects.
However , the facts relating to the long term consequences of
the use of the drug are inadequate and insufficient to
provide a basis for risk assessment.
This is a serious
deficiency in light of the specific questions that have been
raised that the drug may have major adverse effects following
its long term use or that may become evident only after a
Most important among these has been the
. . < 9/
concern over the drug’s carcinogenic potential.—'
latent period.
The long term consequences of the use of DMPA on
neoplasias, in particular of the breast.and uterus, as well
2/
Data are also inadequate to establish effect of MPA on
bone and on the profile of plasma lipoproteins,
information needed to evaluate whether the long term use
of the drug will or will not predispose the individual
to osteoporosis or to atherosclerosis. Our Conclusions
of Law do not rely on this finding.
173
as osteoporosis and atherosclerosis are of particular
relevance for any risk/benefit assessment of the drug’s use
in the United States because of the susceptibility of the
population in this country to these diseases.
In the absence of adequate data on the long term
consequences of the drug it is not possible to arrive at any
scientifically defensible conclusion whether or not the
benefits of the drug, when used as a contraceptive, outweigh
its risks for the average healthy individual in the United
States.
It also makes it impossible to compare the
r isk/benefit ratio of DMPA with that of other drugs availablefor contraception.
II.
DATA FROM THE STUDIES OF RHESUS MONKEYS AND BEAGLE
DOGS CAN NOT BE DISMISSED AS IRRELEVANT TO THE HUMAN
WITHOUT CONCLUSIVE EVIDENCE TO THE CONTRARY. SUCH
EVIDENCE IS NOT AVAILABLE AT THIS TIME. THEREFORE,
THE FACT THAT MALIGNANT NEOPLASIAS DEVELOPED IN TWO
SPECIES IN TARGET ORGANS OF SEX STEROIDS MUST BE
CONSIDERED AS AN INDICATION OF A POTENTIAL OF
PROGESTOGENS, INCLUDING DMPA, TO PROMOTE THE
DEVELOPMENT OF MALIGNANCIES IN TARGET ORGANS.
The findings from animal tests implicate DMPA as a
potential promoter of neoplasias because:
D
Chronic administration of DMPA was associated with
the development of malignant neoplasias in two mammalian
species.
2)
steroids.
The neoplasias developed in target organs of sex
174
3)
There is good evidence to support the conclusion
that in both species the malignancies were drug related.
4)
There is no evidence to support the conclusion that
the effect of the drug is to be attributed only to the
administration of excessively high doses and that the effect
of lower doses would differ qualitatively from those of
higher doses.
Therefore t DMPA in these experiments exhibited the
characteristics of a potential carcinogen according to
generally accepted criteria.
Under the circumstances, to
dismiss the findings as irrelevant to the human would require
conclusive experimental evidence of fundamental differences
among the species in the basic mechanisms of action of the
hormone or in the responses of target cells.
no such evidence at hand.
There is as yet
Specifically, there are no data on
the histogenesis of the neoplasias nor on the mechanism of
action of progestogens on the presumed cells of origin of the
neoplasias in the test animals.
Therefore, there is no
evidence to support the claim that the malignancies developed
either in cell types unique to the species or as a result of
a species specific response of target cells to progestogens.
Conversely, data on women who have been exposed for prolonged
periods to the relatively unopposed action of progestogens
are inadequate to warrant the conclusion that their response
to this hormonal state in terms of neoplasias would differ in
175
some fundamental way from the two species of test
animals.
THE DATA ON THE HUMAN ARE INSUFFICIENT AND
INADEQUATE TO EITHER CONFIRM OR REFUTE THE
IMPLICATION OF THE ANIMAL DATA THAT DMPA MAY
INCREASE THE RISK OF CANCER IN WOMEN USING DMPA
AS A CONTRACEPTIVE.
Ill.
The available data on the human can not provide a basis
for concluding whether DMPA, used as a contraceptive, does or
does not influence the incidence of carcinomas in general or
Of the accessory organs of reproduction in particular,
because:
1 )
They fail to provide information on an adequate
number of long term users of DMPA, or on ex-users who have
been followed for a long enough period of time.
There are
only minimal data on subjects that have used DMPA for 5 years
or longer with most of the data reported having been obtained
from women who have used the drug for 2 years or less.
2)
In the majority of the studies there were no
controls followed in parallel with those using DMPA.
In many
studies from developing countries there is not even
information on the background incidence of the diseases being
studied in DMPA users that could serve as a basis for
comparison.
3)
In a number of the retrospective studies there is
reason to question the adequacy of the record keeping on
176
subjects receiving DMPA and. therefore, of the possibility of
retrieving the data subsequently for any valid analysis.
To obtain the direct evidence needed to resolve the
issue would have required purposeful, systematic collection
and recording of data on users of DMPA and appropriate
controls with consideration of the natural history of the
diseases being monitored.
Not until recently have such
studies been initiated.
Until they are completed and full
reports of them available their value as evidence is
-1 imi ted.
IN CASE OF CONTRACEPTIVE FAILURE WITH DMPA, THE
RISK OF A MOTHER GIVING BIRTH TO A MALFORMED
CHILD IS UNLIKELY TO BE MEASURABLY GREATER THAN
THAT POSED BY THE ORAL CONTRACEPTIVES. THE
CHANCE IN EACH CASE CAN BE ESTIMATED TO BE SMALL
ENOUGH NOT TO POSE AN OBSTACLE TO THE USE OF THE
DRUG AS A CONTRACEPTIVE WHEN OTHERWISE INDICATED.
IV.
Data have not been systematically collected on offspring
that have been inadvertently exposed to DMPA in utero.
Conclusions, therefore, can only be based on the body of
epidemiological data obtained on the effects of a variety of
sex steroids, including progestogens , on the developing human
fetus.
In these cases, the drugs had been administered for a
variety of indications and at various times during pregnancy.
This is clearly a less than ideal data base.
Nonetheless it
can provide some general estimate of the magnitude of the
risk.
! \
177
According to these data the risk of various
malformations attributable to protestogens for the various
malformations implicated is low.
The rate of contraceptive
failure with DMPA when used appropriately is also low.
Consequently, the chance of a mother bearing a malformed
child following contraceptive failure can be estimated to be
small.
However, because DMPA is a long acting depot
preparation, the exposure of any susceptible fetus to the
drug is likely to be more prolonged than with oral
contraceptives.
Consequently, tne range of critical and
vulnerable events that may come under the drug’s influence
may also be expected to be greater than with oral
contraceptives.
It should be possible to counter balance
this disadvantage of DMPA oy ensuring that contraceptive
failure is kept at a minimum and taking the necessary steps
to avoid injecting women already pregnant.
As with oral
contraceptives this risk should not, in itself, constitute a
reason for not using the drug if otherwise indicated.
There have been no direct determinations of the
concentrations of MPA in the blood of breast fed infants of
mothers receiving DMPA as a contraceptive nor if the amount
of the drug transferred passed onto the infant is sufficient
to have a biological effect.
This information is needed
before advocating the use of DMPA as a contraceptive to
lactating mothers in the postnatal period and before it is
178
possible to conclude that the drug does not pose any risk of
functional teratogenicity.
179
CONCLUSIONS OF LAW
Accordingly, Upjohn's supplemental new drug application
for Depo-Provera (DMPA) sterile aqueous suspension for
intramuscular injection as a contraceptive agent in humans
does not contain reports of investigation adequate to show
that the drug is safe for use under the conditions
prescribed, recommended or suggested in the labeling as
required by § 505(d)(1), (2) and (4) of the Federal Food,
.Drug , and Cosmetic Act, and the information contained in the
supplemental new drug application, combined with other
information about the drug, does^
not provide sufficient basis
from which FDA can determine that DMPA is safe for general
marketing in the United States.
y/
.1
V
6
I
< 5
180
Dated this
/z^
day of
, 1984.
P —
Judith WeisZf MB BUhir.
Chairperson
2>.
/nj>
Paul D. Stolley? M.D^/^M.P.H.
- 181 The University of Texas
Health Science Center at Houston
MEDICAL SCHOOL
Department of
Obstetrics, Gynecology and
Reproductive Sciences
6431 Fannin, Suite 3.270
Houston, Texas 77030
(713) 792-5360
July 23, 1984
Dr. Judith Weisz
The Pennsylvania State University
College of Medicine
500 University Drive
Hershey, Pennsylvania 17033
Dear Judith:
I am disappointed to inform you that deterioration in my health prevents
my participation in drafting and parsing the final report of the Public Board
of Inquiry. - Since my appointment to the Board in September, 1982, I have
participated in both of the public hearings and in the various meetings we have
held to review the evidence presented at the hearings, As a result, I have
formed an opinion about the quality of the information in the record, which I
wish to summarize briefly, and suggest a it"Bottom Line" for the decision.
The quality of the data submitted in writing and orally is inadequate to
provide a scientifically valid basis to either affirm or deny the long term
safety of the drug for use in human subjects, As I understand the law, approval
of a drug for marketing requires test results scientifically adequate to show
that the drug is safe for use.
While I concur with you and Paul that the data submitted are inadequate to
show that Depo Provera is safe, these also fail to show that the drug is unsafe
for human use. Accordingly, as an alternative to outright disapproval, I would
recommend that the Commissioner consider approval, if practicable, for limited
use in two populations of patients for whom satisfactory alternatives do not
exist. These populations are:
1.
2.
Oliogophrenic (mentally retarded) women and
Women using drugs abusively.
Moreover, such a course of action would make it possible for physicians to
participate with these persons or their agents in reaching decisions about the
use of the drug for contraception.
This course of action would be consistent with the fact that nothing we have
seen or heard established the concept of intolerable risks following use on human
subjects, and allow time for further tests to be performed and results analyzed.
Sincerely,
Griff T. Ross, M.D., Ph.D.
GTR: bmc
•nu.BFjnch • Dimwi o( Continue Eflucitxxi • GrMJuita Sc hod of Bomajoi Satncw • Sena* of »umc hmm • Moocai Schoa • SpMch »no HMnng imtmm • ScW of Muwng . Seftoa of A»*o Hmm Scwncos
(\
f.
M. *'
•- ’ 'r'j
ri X
APPENDIX-!
DMPA USE AND RISK OF BREAST CANCER_:
Reference
Greenspan, A.R.,
Hatcher,R.A., Moore, M.,
et al.
The Association of
Depo-Medroxyprogesterone
Acetate and Breast
Cancer. Contraception
21:563, 1980
DMB Vol. 285 Tab #320
HUMANOATA
Number of Women Studied,
Dose and Duration of DMPA Use
Case-control study of breast cancer in the Grady
Memorial Hospital Family Planning Clinic, Atlanta,
Georgia.
Thirty women with breast cancer, age 55 or younger,
who were enrolled in the Clinic between 1967 and
1979 and 6 controls per case, not known to have
breast cancer, and matched to the case for age
(+2 years), race, and date of visit to the clinic
(±6 mos), had their contraceptive history ascertain
ed via the medical records in the Clinic.
Findings
1)
Used
DMPA
The mean number of DMPA Injections (150 mg) was
2.8 In cases and 3.3 in controls (i.e., short term
users).
Cases with
Controls
breast
wl thout
Cancer
| breast cancer
(16.7%) |
(17.9%) |
I
I
5
Did not |
use DMPA |
25
L
1
Total
Average parity was 3.3 in cases and 4.6 In controls.
70% were older than 30 years of age
Comments
2) Short latency period.
I
32
|
J_________ I 3) The study did have the
power to detect a RR of
I
I
147
I
|
3-fold or 4-fold (with 80%
J________ I
30
1) Short average exposure
to DMPA.
and 95% chance,
respectively).
179
Matched relative risk estimate = 0.92.
(95% CI:0.33-2.58). Controlling for
parity did not alter this result.
2) For all 5 exposed cases, DMPA
exposure preceded diagnosis of
breast cancer by an average of
4 years (range 2-5 yrs) (i.e.,
short latency).
I
00
NJ
I
Liang, A.P., Greenspan,
A.R., Layde, P.M., et al.
"The Risk of Breast,
Uterine Corpus, and
Ovarian Cancer In
Women using DMPA"
JAMA, 249:3909, 1983
0MB Vol. 283 Tab G-628
5000 Black women attending the Grady Clinic during
1967-1976 were followed up through 1980 for occur
rence of cancer.
The followup period ranged from 4 years (for women
whose initial Injection was in 1976) to 13 years
(for women whose initial injection was in 1967).
The median length of followup was 9 years. In the
1981 stratified random sample of GMH Clinic attendees,
72% of DMPA users could be contacted. (28% not
contacted).
Duration of Use
< 13 mos.
13- 36 mos.
37- 60 mos.
61-114 mos.
Number_of_Women
2,934
(58.7%)
1.438
(28.7%)
425
( 8.5%)
_206
( 4 1%)
N-5.0U3
(100%)
1) Based on the National Cancer
Institute SEER age-specific
incidence data for Black women
1973-1980, the expected number
of breast cancer cases was
10.07. This compares to 7
cases of breast cancer actually
observed at Grady.
The RR = 0.69 (95% CI: 0.3-1.4)
1) Short average duration of
DMPA use.
2) Incomplete ascertainment
of occurrence of cancer.
Deceased persons obviously
not contacted.
3) Authors state that
■although these data are
not conclusive* they are
reassuring. This study
Indicates that there Is
unlikely to be a strong
association between DMPA
use and breast cancer."
2) 6 out of 9 cases with cancer (7
breast cancer, 1 uterine
corpus, 1 ovary) In this
series of women had only 1
or 2 Injections of DMPA.
Authors suggest it is unlikely that
4) Evidence exists from other
the cancers of these women were
sources of Incomplete
causally linked to DMPA.
Initial recording of data
(See FDA Audit).
J. A
Reference
McDaniel, E.B. and
Pardthalsong, T.
Incidence of Breast
Nodules In Women
Receiving Multiple Doses
of Medroxyprogesterone
Acetate J. Biosoc, Sci.
5:83, 1973
DMB Vol. 170
Tab #66
A
r
Number of Women Studied,
Dose and Duration of DMPA Use
i
-A 1
1
d
Findings
Cross-sectional study of 1527 Thai women:
1) the Incidence of mammary nodules
257 were new patients presenting for contraceptive
was 1.2% In the controls (3 out
of 257 IUD patients or firstservices (l.e. controls).
1270 already received 1 to 25 consecutive 3-month
timers OCS or DMPA) vs. 0.8% in
DMPA (150 mg) together with monthly oral DES supple
DMPA users (8 out of 1270).
ment .
2) on re-examination of 6 of the 8
women, 3 had definite breast
Number of Homen
masses (1 galactocele; 1 scar
Duration of DMPA Use:
503 (39.6%)
tissue from old breast abscess;
1- 4 consecutive Injections (1 yr.)
417 (32.8%)
1 large mammary dysplasia with
5- 9 consecutive injections (2 yrs.)
315 (24.8%)
chronic inflammation).
17-20 consecut ive 1njec t i ons (5 yrs.)
35 ( 2.7%)
21-25 consecut i ve inject ions (6 yrs.)
N=127O
100%
3) Authors noted that in over 6
years of experience with DMPA
and estrogen supplement In
13,418 patients (with 18,890
women yrs. of use and 74,086
injections), the formation of
breast nodules has never been
a patient complaint.
Comments
1) No Information on how
controls on IUD
compare with the DMPA
users on risk factors such
as, age, parity, previous
oral contraceptive use.
2) A latency period of a
maximum of 6 years may not
be sufficiently long for a
breast tumor to develope.
3) The use of estrogen
supplement (DES) Is a
confounder.
4) No nodules In the breast were
observed In a series of
13,418 women.
5) If the Incidence of breast
cancer is low in Thai women,
these data may not be
relevant to U.S. women in
whom the Incidence of breast
cancer is much higher. There
are marked international
variations in the rates of
cancer,
e.g. the age-standardized
Incidence rate for breast
cancer is:
19.4/100,000 Chinese women
in Singapore.
17.6/100,000 Malayan women
in Singapore.
25.5/100,000 Indian women
in Singapore,
n-
76.1/100,000 white women
in U.S.
56.5/100,000 black women
in U.S.
44.0/100,000 Chinese women
in U.S.
(based on Waterhouse J,
Muir C, Correa P. et al.
Eds. Cancer Incidence in 5
Continents, Leon, IARC, 1976)
.1
6) Sampling bias because
patients too 111 to travel
to the city or who had died,
did not have an equal chance
of being sampled.
I
oo
uj
I
APPENDIX 1
DMPA USE AND RISK OF BREAST CANCER:
Referenc e
Ory, H.W., Rubin, G.,
Jones, V., et al.
"Mortality among young
black contraceptive
users" 1982.
(unpubli shed)
DM8 Vol. 283
TAB #190C G-629
Number of Women Studied,
Dose and Duration of DMPA Use
Comparison of mortality rates in black women using
one of four contraceptive methods (OCS, DMPA, IUD,
Barrier).
Death certificates for 218 women whose names,
addresses and birthdates matched the Grady family
planning records were analyzed for cause of death,
(ages 10-44 yrs. old).
Person-months of observation were computed from
the date of first visit to the family planning
clinic to the date of death or to Dec. 1977 (which
ever came f irst).
HUMANL_DAlTA
Findings
Comments
1) The rates of death per 10,000
years of follow up from all
causes among the 4 groups were:
12, 30, 30, 0 for users of IUDs,
OCs, DMPA and barrier methods,
respect ively.
1) Misclassification bias
because women were
considered users on only
1 method of contraception,
the one chosen at the first
clinic visit. If switched
to other contraceptive,
women were still counted in
original grouping.
2) The cause specific death-rates
show a risk of death from cancer
for 0C users 3 times higher than
the risk for ever use of IUD or
DMPA.
2) No information on former
contraceptive use, before
attending the clinic.
On average, women were followed up for 8.2 years.
3) No information on duration
of contraceptive use.
Person-years of contraceptive use were computed from
the date of first clinic visit to the date of last
visit for that method + 6 months (ignoring switching
between methods).
4) Type of cancer not
spec i fled.
All rates were age-standardized because of different
age distributions in the 4 groups.
5) The computerized name
matching system may not have
Identified the total number
of deaths.
I
Total Number
Cause of Death
1 1
Cancer
Circulatory
51
134
Total Medicine
84
Accident & Violence
All Deaths
218
.1
Death rate/10,000 women per yr.
Use Ever of___________
DMPA
Barrier
IUD Ever
PC
0.0
0.3
0.3
1 .1
1 .4
2.0
2.5
2.1
6.7
4.5
5.3
5.4
2.9
3.1
2.9
3.5
7.6
9.6
8.1
8.9
co
4^
I
r
APPENDIX 1
DMPA USE AND RISK OF BREAST CANCER:
Reference
Pena-Delgado, J.,
Aleman-Herrera, M., and
Baez-Reyes, A. Long term
Use of Medroxyprogesterone
Acetate in Contraception
Sem, Med. Mex.
98:331, 1981
DMB Vol. 221 :95)
Number of Women Studied,
Dose and Duration of DMPA Use
Study of IX random sample (N-1025 cases) out of
71.188 new clients who chose DMPA (150 mg) and who
used the method continuously for 3 to 8 1/2 years.
Duration of DMPA Use
< 12 doses
13-24 doses
25-36 doses
37-38 doses
Number of Women
(8.3X)
85
(72X)
738
199
(19.4X)
LO.JX}
3
(100%)
N=1025
HUMAN DATA
F Hid i ngs
1) "No sign of any greater
tendency to occurrence of
benign breast nodules or
mammary cancer under DMPA"
(compared to what??)
2) 6 out of 1025 had mammary
nodules with 5 breast
biopsies negative. No
information on the sixth.
3) "In all 6 cases, it was not
possible to verify that the
presence of nodules was due
to use of the drug."
Comments
1) Basically, a follow-up study
of long-term users of DMPA
without controls.
2) No information on previous
contraceptive use, parity,
etc .
3) No Information on background
incidence of breast cancer
in this population of
Mexican women.
4) In some Instances, manrnary
tension was decreased by
the medication.
I
Rail, H.J.S.,
Van Niererk,W.A., et al.,
Comparative Contraceptive
Experience with ThreeMonth and Six Month
Medroxyprogesterone
Acetate Regimens
J. Reprod. Med.
18:55, 1977
Cross-sectional study of 19,875 women (migratory,
1ow-to-middle socioeconomic population) in South
Africa, treated in 2 hospitals between 1970-1975
with 150 mg/3 mos or 450 mg/6 mos DMPA.
Duration of DMPA Use
< 13 mos.
13-36 mos.
> 37.mos.
DMB Vol. 161A
Number of Users
13,922
(/0.0X)
(28.8%)
5.715
238
L_L2%1
(100X)
N-19,8/5
1) No mammary tumors were diag
nosed in the 3-month or
6-month DMPA groups.
(i.e., 0 out of 19875 women
seen over 5 years )
1) Complete ascertainment of
mammary tumors is
questionable in this
migratory population.
2) No information available
on the incidence rate of
mammary tumors in a
comparison group in South
Africa.
3) Short-term use and short
follow-up.
.1
co
CH
I
APPENDIX 1
DMPA USE AND RISK OF BREAST CANCER:
Reference
Schwallie, P.C.
Experience with DepoProvera as a Injectable
Contraceptive
J Reprod Med.
13:113, 1974
Number of Women Studied,
Dose and Duration of DMPA Use
Pooled data from 11,500 women worldwide (all ages,
all dose-regimen, U.S. and non-U.S., black and
white) who were treated with OMPA during 1963-1973.
H^MAN_DAT_A
1) 14 women among the 11,500
studied either developed
mammary nodules or had an
enlargement of previously
existing nodules.
8 of the 14 received
2
received
2
received
1
rece ived
1
rece1ved
DMB Vol. 135A pp. 58-62
Comments
Findings
150 mg DMPA
250 mg
300 mg
400 mg
500 mg
Their age range was 20-45,
para 1-7, and number of
Injections 1-24.
and See Also
1) No control data available
and therefore cannot estimate degree of excess
risk posed by DMPA. Note:
heterogeneity of subjects
from different countries
with different breast cancer
incidence.
2) No adjustment for
confounding risk factors
(e.g., parity, obesity,
etc.)
2) 3 of the 14 had also received
estrogen supplements.
Schwal11e P.C., and
Mohberg N.R.
Medroxyprogesterone
Acetate: an injectable
contracept1ve
Review of same data on women as in the above.
1) 24 women either developed
mammary nodules or had an
enlargement of preexistent
nodules.
2) 5 of the 24 had care inoma,
but 3 of these received also
concomitant estrogen therapy.
Adv. Planned Parenthood
12:35, 1977
DMB Vol. 161A G-513
.1
I
co
cn
I
f
APPENDIX 1
DMPAJJSL-AND RI SK _OF_ BRF ASJ_CANC£R:
Reference
Zanartu, J., Onetto E.,
Medina, E., Dabancens A.
Mammary Glands Nodules in
Women Under Continuous
Exposure to Progestagens
Contraception
7:203, 1973
DMB 93A pp. 635-644
Number of Women Studied,
Dose and Duration of DMPA_Use
Prospective study of the mammary gland in 2418 women
using DMPA (250, 300, 500 mg/3 mos. or 1000 mg/6 mos.)
and 932 women using chlormadinonp acetate (CA) in
Chile. Women were given estrogen supplement in cases
of amenorrhea
Duratlon of DMPA Use
< 13 mos.
13 -36 mos.
37 60 mos.
61-84 mos. (5-7 yrs.)
OMPA_Usej^s
491 (20.3X)
6/1 (2/.8%)
923 (38.2%)
_333 (13,8%)
N-2418
100%
HUMAN DATA
F i nd Ings
Comments
1) the incidence of breast nodules
was 10 out of 3350 women treated
with DMPA or CA between 1965 and
1971. (i.e. 3 per 1000)
1) An adequate control group of
similar age and parity,
never exposed to
progestagens is not
available for comparison.
2) Of the 10 cases with mammary
nodules,
2 received 150 mg. DMPA,
6 received 250 mg OMPA,
1 received 1000 mg. DMPA,
1 got CA.
2) Use of estrogen is a
confounder as is the high
number of pregnancies in
these women.
3) 7 of the 10 received also
cyrlic estrogen during 1
month to 3 years.
4) 2 cases of these 10, (the
250 and 1000 mg. OMPA)
were mammary care inomas.
2 was chronic mastitis
1 fibroadenoma and
5 adenosis.
5) All 10 cases with nodules
were long-term users of DMPA
or CA (46 to 12 months); the 2
carcinomas occurred in users
for 60 and 66 months.
6) These 10 cases ranged in age
from 23 to 37 years old, and
had 1 to 15 pregnancies.
,1
3) The follow up study
comprised breast examination
twice a year. It is not
clear If, and how, women who
stopped using the drug were
examined if they didn't
return to the clinic.x i.e.,
question of complete
followup and ascertainment.
I
co
I
1
APPENDIX 1
DMPA_USE AND RISK OF 8REAST_CANCER:
Reference
Zarfas, D.E., Fyfe, I.,
Gorodzinky, I., The
Utilization of Depo
Provera in the Ontario
GovernmentFac11ities for
the Mentally Retarded:
A Pilot Project,
1981. (unpublished)
DMB Vol. No. 203 G-550
Number of Women Studied,
Dose and Dura t i on_ of DMPA Use
Questionnaires were completed by the staff of 9
facilities for the mentally retarded on all 533
woman for whom DMPA had been prescribed.
HUMAN DATA
Lj Dd in-gs
Consents
(1) Of 533 females who received
DMPA. 21 had died.
1) The information based
entirely on reports from the
staff of the facilities for
the mentally retarded; no
independent verification of
the data attempted by the
authors.
(2) Of the 21 deaths, 3 died of
breast cancer.
Duration of DMPA Use
No. Women
< 3 yrs.
3-5 yrs.
> 5 yrs.
Unknown Duration
Tota 1
218 (41%)
67 (13%)
208 (39%)
40 ( 8%)
N-533
(3) Information on all deaths of
female residents in those
facilities since 1965 for
ages 15 and over produced an
additional cohort of 342, 3
of which were due to breast
L§.nc e£.
Breast Cancer Mortality
in DMPA
in DMPA
Nonusers
Users
3/342
3/21
(ages 32,33,40)
(ages 53,66,77)
2) Duplication of cases may
have occurred since most
Centers use case numbers
rather than names to
identify individual
questionnaires, and record
linkage between Centers was
impossible.
3) No adjustment made for^
confounders such as
nulliparity, use of other
drugs, etc., which may
increase risk of breast
cancer in this group.
4) Complete ascertainment of
breast cancer questionable
because post-mortem
examination of breasts is
not required unless there is
apriori evidence of breast
cancer. Therefore, some
cases may have been missed.
5) NOTE the younger ages at
which breast cancer occurred
among the DMPA users!
.1
I
co
co
I
APPENDIX 1
DMPA USE AND RISK OF BREAST CANCER■
Reference
Spellacy, W.N., Buhl,
e W.C.,
Birk, S.A., Stimulated
Plasma Prolactin Levels
in Women Using
Medroxyprogesterone
Acetate or an
Intrauterine Device
for Contraception
Fertil. Steril.
26:970, 1975
Number of Women Studied,
Dose and Duration of DMPA Use
Study of plasma prolactin levels in women using DMPA.
12 women who had been receiving MPA (150 mg.) for
4-5 years were compared to 9 women who had been
using an IUD for
1 year) and none of these women
was taking other drugs.
MPA women were older than the IUD women
(mean age: 33.3 yrs. vs. 26.4 yrs.)
MPA women has a higher parity (mean:
4.5 vs. 1.5).
IWMAN-DATA
Findings
Comments
1) Although less prolactin was
relpased after the
chlorpromazine stimulation
in the MPA group, the
differences were not signifi
cant .
1) Authors* conclusion was that
since breast neoplasia may
be related to chronically
elevated prolactin levels,
these data are reassuring.
But, relationship of
prolactin to breast cancer
in humans, unlike in the
rodent, is not known.
2) There was no significant
difference also in basal blood
prolactin concentrations be
tween the two groups.
Blood samples were drawn before and after
chlorpromazine stimulation.
9
Zanartu, J., Aguilera,
E., et al.. Effect of LongActing Contraceptive
Progestogen on Lactation
Obstet. Gynecol.
47:174,1976
0MB Vol. »
Prospective study on the duration of lactation in 406
mothers injected postpartum for contraception with
DMPA (150 mg./3 mos. or P5O-3OO mg./6 mos.) compared
with 173 nontreated controls.
The follow up period was 18 months.
Average age and parity reported to be comparable in
both groups.
.1
1) no pathologic breast conditions
were observed in mothers during
the study.
1) Only 1 or 2 injections,
1.e. short-term use.
2) The objective of the study
was to examine effect on
lactation. The observation
on breast pathology was
Incidental.
I
o
vx
I
e
I
appendix.2
REFERENCE
STUDY DESIGN
HNOINGS
COMMENTS
Cervantes, A., Azcona,
Bribiesca. Aguirre, Velasco.
"Effect of
Medroxyprogesterone
acetate on human
endometrium after
five or more years
of use as a contra
ceptive," 1982
(unpubli shed)
DMB Vol. 213 Tab #U-20(a)
Phase I of the study
involved histophatological
examination of biopsies
taken from 76 women who
used DMPA (150 mg) for >5 years.
(40 active pat ients
and 36 who discontinued use).
The histophatology findings:
1) Question concerning
confirmation of
histopathology (for detailed
discussion see text).
see also Upjohn’s response
to questions on this study
April 15, 1983. pp. 2-4
DMB Vol. 343 Tab #506
1)
No anaplasia was found
in any of 40 active patients:
g^simple hyperplasia"
21 "slight to severe
atrophy of endometrium"
4 "normal secretory or
proliferative endometrium"
Phase II of the study
involved the examination of
all death certificates of
women above 20 years of age
between 1967-1978 in Coahuila,
Mexico, to identify the
cause of death, and then
inquire from relatives or
last attending physician
about prior contraceptive use.
2)
Among the 36 inactive patients
(on average discontinued use
for more than 1 1/2 years):
17 "normal active endometrium"
11 "simple hyperplasia"
8-"slight to severe atrophy
of endometrium"
The epidemiological findings:
1) 12,356 total deaths in the
12-year period 1967-1978.
Of these. 172 women died of
genital cancer during this
period:
146-cervical cancer
12-ovarian cancer
6-endometrla 1 cancer
8-und iagnosed
2) The history of contraceptive
use in these women was:
.1
6-cases of endometrial cancer
who never used any kind of
steroid contraceptives.
12-cases of ovarian cancer
who never used any kind of
steroid contraceptives.
6-of 146 cases of cervical
cancer used combine_d oral
contracept1ves at one time
during 11fetime.
2) Possibility of difference in
coding cause of death.
3) Ascertainment of drug use
questionable. Interview
with next-of-kin about
contraceptive use may cause
mis-classification.
4) No information on background
incidence of either breast or
endometrial cancer in
comparable Mexican population.
Note high degree of
international variations in
Incidence rates of cancer.
e.g., the age-standarized
incidence rate of cancer of
the corpus uteri for Spanish
women 1s:
9.5/100,000 in El Paso
8.4/100,000 in New Mexico
6.1/100,000 in Puerto Rico
The incidence rate for white
women is:
33.3/100,000 in Alameda, CA
20.1/100,000 in Detroit
For black women, the rate is
13.6/100.000 in Alameda. CA
10.5/100,000 in Detroit
(Based on Waterhouse and Muir,
1976).
|
o
I
• •••
. *
1
APPENDIX 2
THE ASSOCIATIOfLBETWEEN_P”!A_ySL_LMQ
0EVEL0PMENT OE ENDOMETRIAL CANCER
STUDY DESIGN
FINDINGS
COMMENTS
REFERENCE
Greenspan, A.R.
Chap. 8: "The endo
metrium and DMPA,"
in The Grady
Experience, 1978
(unpublished)
832 current users of DMPA at
the Grady Clinic were asked
to respond to the question:
"If you have used the shot
before, have you had a D&C
since you started using it?"
34 replied "yes."
1) No patients were found to have
had cancer based on the
pathology reports.
1)
Serious misclassification
problems (e.g. only 19%
agreement between medical
record and patient report).
2)
Serious problems of
ascertainment, since only
the charts of those who
responded "yes" were
reviewed while among those
responding "no" there may
have actually been some
DiCs.
0MB Vol. 221 and
DMB Vol. 279 Tab 295
Out of 34 responders, the medical
charts of 26 were reviewed for a
history of D8tC or endometrial
biopsy done anytime after
having at least 1 DMPA
injection.
21 of the 26 patients
answering "yes" didn't have
history of D&C or endometrial
biopsy. Patients misin
terpreted question or did not
know what D&C meant.
Pitts, H.,
Endometrial biopsy series, 1979
(cited in above)
(1) 6 had inadequate
endometrial tissue for
biopsy;
2 had evidence
of atrophic endometrium
Examination of endometrial
biopsies from 8 patients
who had used DMPA for > 3 years.
(2) These results suggest
hypo rather than hyper
plastic changes in the
endometrium following
long-term administration
of DMPA
.1
1) Study of consequences of
DMPA on endometrial
histology, not of incidence
of endometrial cancer.
A,
I
APPENDIX.?
THE ASSOCIATION BETWEEN DMPA USE_AND
DEVELOPMENT OF ENDOMETRIAL CANCER
REFERENCE
STUDY DESIGN
hndings
COMMENTS
Williams, 0.
The endometrium after prolonged
use of DMPA, 1975.
(cited in above)
The names of all women
having hysterectomies at
Grady Memorial Hospital
between 1967 and 1974 were
obtained and their drug
histories were reviewed.
1) 7 women (out of how many?)
who had hysterectomy
were found to have a
history of DMPA use
of 1 year or longer.
1)
Study uninterpretable.
2)
No information provided on
the size of the denominator
in the study.
2) duration of DMPA use
by these patients was
19 39 months; mean
duration of use was
25 months.
3)
Cervical dyplasia is not a
usual indication for DMPA
use.
4)
Uncertainty about reason
for administration of DMPA
in these women suggests
serious problems in record
keeping.
3) 6 of the 7 patients had
received DMPA at 60 or
fewer days prior to
surgery; 1 patient had
received DMPA 120 days
prior to surgery.
©
4) No cases of endometrial
cancer were found in
this group. Author states
that these patients had
cervical dysplasia and DMPA
may have been used as
palliative treatment.
.1
I
I
APPENDIX 2
THE ASSOCIATION BETWEEN DMPA_USE AND
DEVELOPMENT OF ENDOMETRIAL CANCER
REFERENCE
STUDY DESIGN
FINDINGS
COMMENTS
Stamm, H., deGrandi, P.
Adenocarcinoma of the
Endometrium Following
Treatment with
Medroxyprogesterone
For Contraception
Rev. Med. Suisse Romande
101:913, 1981
Case report of endometrial
cancer In a 44-year-old
woman after use of DMPA
(150 mg) for 2-1/2 years.
1) Stage I endometrial carcinoma
diagnosed In 1980 and a
hysterectomy performed.
1) Endometrial cancer diagnosed
while on DMPA.
DMB Vol. 234 Tab #1
Behary, C.M., Dollberg, L.,
Czernobllsky, B.
Endolymphatic Stromal Myosts
In Two Patients on Progestagen
Therapy Contraception
13:1-6, 1976
DMB Vol. #
Patient history:
- father who died from
cancer of the esophagus
- patient smoked 1 pack of
cigarettes dally
- had menarche at age 14,
2 pregnancies
- at age 38 (1975) started
oral contraceptive use
for 2 months
- then IUD use for 2 years
- In 1978 started DMPA use
- severe bleeding appeared
In 1980.
I
(1) The authors speculate that
growth of these lesions was
stimulated by the progestagen
therapy.
Case report of 2 women, in
Israel, who were found to
Shave massive endolymphatic
stromal myosis after
receiving DMPA (50 mg and
ethinyl estradiol) for > 1
year and for > 8 years,
respectively.
Case 1:
20 year-old, unmarried,
given progestagen for
severe menorrhagia and
dysmenorrhagla.
Case 2:
44 year-old, 2 pregnancies,
suffering from menometrorrhagla for 8 years
while receiving Metrulen
(ethynodlol diacetate and
mesteranol).
.1
l
e
rtHrt
. 2
I
THE ASSOCIATION BETWEEN_DMPA.USE AND
DEVELOPMENT OF ENDOMETR_IAL CANC£R
REFERENCE
STUDY DESIGN
FINDINGS
COMMENTS
McDaniel,E.B., Potts, M.
International forum update:
depot medroxyprogesterone
acetate and endometrial
cancer.
Inti. J. Gynecol. Obstet.
17:297, 1979
Retrospective survey of all
hospital admissions for
proven endometrial cancer
in two Thai provinces where
DMPA (150 mg or 450 mg/4 mo.)
has been widely used since
1965.
1) Total number with discharge
diagnosis of endometrial
cancer, 1974-1978, in seven
hospitals in the two
provinces with over four
million population....^ 49
1) Numbers dwindled fast from
49 to 9.
2) Total number with confirmed
pathology of endometrial
cancer..... ..... = 39
27 proven diagnosis
12 presumptive diagnosis
3) Ascertainment of drug use
questionable.
DMB Vol. 222
2) Assumes that all cases of
endometrial cancer reach
the hospital.
4) Incidence of endometrial
cancer in this population
is not known.
3) Of 27 with proven diagnosis,
only 16 came from the two
provinces where DMPA was
widely in use
(27-16=9 were not followed
up because assumed to be
nonusers of DMPA.)
4) Of the 16 assumed to have
had an opportunity to be
exposed to DMPA, another 7
were not followed up
because:
- 4 were assumed too old
(>50 years) to have
used contraceptives
- 1 had never married or
borne children, so
assumed not to have used
DMPA.
- 1 gave false address and
could not be located
- 1 had changed address and
moved too far.
ie
.1
5) Of the 9 successful follow
ups :
- 4 were alive
- 5 were deceased
and the Information on drug
use was obtained either by
mail, or by personal inter
views, either with the patient
or with family members.
(Why could not the clinic
records be checked for
verification??)
6) None of the 9 contacted cases
reoorted use of DMPA ever.
i
J:
i
(
• ••••
•••••
APPENDIX 3
The Association between DMPA Use and Risk of Cervical Carcinoma 1n-S1tu
*
Reference
Oabances, A.,Prado, R.,
Larragulbel, R., Zanartu, J.
Intraepithelial cervical
neoplasia in women using
intrauterine devices and
long-acting Injectable
progestogens as contraceptives.
£
Number Women Studied
and Follow-up Period
Findings
1) The group of women receiving
the injectable contraception
had a higher prevalence rate
of preclinlcal cervical
neoplasia at baseline as
compared with women using IUDs
The follow-up period was (1965-1971),
(13.1/1000 vs. 7.3/1000 in the
with women regularly evaluated
two groups, respectively).
for cervical neoplastic changes.
A controlled prospective study of
2684 women receiving either DMPA
or chlormadinone acetate, and of
2409 women given IUDs, in Chile.
Am. J. Obstet. Gynecol.
119:1052, 1974
2) The Incidence rate (i.e. new
cases) of preclinlcal cervical
neoplasia were not different In
the two groups after 4 years of
followup. (1.26/1000 women-years
with 95% CI: 0.4-2.06 in the
OMPA/CA group vs. 1.82/1000
women-years with 95% CI: 04-3.33
in the IUD group).
0MB Vol. #
3
1) The only study with relatively long
follow-up of women receiving DMPA
(4 years of prospective observations
on these women).
2) The study Includes a relatively larg<
number of long-term users of DMPA.
(1428 women were taking DMPA or CA f<
3 years or longer. Of these, 443
women were taking Injections for 5
years or longer).
3) Authors concluded that no significant
difference seen In the risk of
developing cervical carcinoma betweer
the two cohorts of women taking DMPA
or IUD.
i
Ide P, Wljnants P, Bonte J:
Cytological observations of
cervlco-vaglnal smears on
hormone contraception.
Revue Cytologic Clinique
5:105, 1972
DMB Vol. 244 Tab #219
Microscopic and cytologic examina
tion of the cervix In Belgian
women participating in a program
for early cancer detection.
482 women were using continuous
progesterone contraception
(duration unspecified), 2,349
progesterone contraception, and
2,134 randomly chosen controls,
matched on age and parity.
©
■©
Comments
.1
(1) 34% of the controls had cervical
abnormalities (Pap I and II
smears), compared to 47% in the
continuous progesterone group,
and 60% In the combined
estrogen progesterone group,
based on clinical pathology.
1)
Lacking details on duration of use <
contraception In the respective
groups.
2)
Demographic Information on the
controls not provided.
3)
Authors attribute pathology only to
effect of estrogens.
4)
It is difficult to deduce either
the study design or findings from
this report.
• •••
1
APPENIDX 3
The Association between DMPA Use and Risk of Cervical Carcinoma In-Situ
Reference
e
Litt 80., FOA Statistical Review
of CareInoma-in-Situ Among
Contraceptive Users of OepoProvera.
June 17. 1974
DMB Vol. G-161
Number Women Studied
and Fol low-up Period
A re-analysis of the Upjohn data
for 127 U.S. women with a Grade HI
Pap smear.
FDA Vol. 116A
1,123 women In a hospital In
Galveston, Texas given DMPA
(150 mg) for contraception,
together with estrogen supplement.
Cervical cytology was done on
1,107 patients.
Duration of use
No. of Women
< 12 months
13-36 months
37-54 months
> 54 months
756 (67.3%)
300 (26.7%)
66 ( 5.9%)
1
tota I
N-l,123
100.0%
Hospital population Is approximately
80% Black; older age-group; higher
parity; lower socio-economic status
group with a higher than usual
incidence of cervical malignancy.
.1
©
Comments
1) 11 (of 127 women with Grade III
Pap smear) had carelnoma-1n~S1tu.
Specifically, the 11 cases of
carcinoma in situ reported by
Powell and Seymour.
1) Source of denominator is ambiguous.
2) Comparison of rates for cervical
cancer In Depo Provers users with
those of the third National Cancer
Survey showed for DMPA users a
rate 4.9 times higher than the
general population In white women,
and 3.1 times higher In non-wh1te
women.
e
Powell, L.C. and Seymour, R.J.
Effects of DepoMedroxyprogesterone Acetate
as a Contraceptive Agent.
Am. J. Obstet. Gynecol.
110:36, 1971
Findings
1) 23 (out of 1107) had suspicious
and positive smears.
(Abnormal cytology rate of
21/1,000 patients In DMPA group).
This compared to an abnormal
cytology rate of 12/1,000 patients
with repeat smears over 3 years
in general population In this
hosplta1.
2) In 82 patients who had the
cervix evaluated histologically:
Histopathology
No. women
58
no malignancy
11 (13.4%)
dysplasia
11 (13.4%)
carcinoma In-sltu
0
Invasive carcinoma
other pelvic malignancy 0
_2
Unknown
Total
82
3) Authors observed that the rate
of carcinoma In-sltu In the DMPA
group (9.8/1,000) was twice as
high as the' f^ate occurring In
previously screened patients In
past 3 years In their hospital
(5.02/1.000).
2) The effect of detection in a1 screened
population may account for the
apparent tripling of rate in the
treatment group compared to that
reported by the 3NCS which draws from
the general population.
3) Illustrates the difficulties of
attempting to derive epidemiological
Information from a study that was not
designed as an epidemiological study
In the first place.
i)
Observations made In course of a
study carried out under an IND.
primarily to establish efficacy and
side effects. No controls Included
In the protocol.
2)
Confounders for risk of cervical
cancer not controlled for (e.g. age,
parity, sexual activity).
3)
The detection rate of cervical
cancer In a screened population may
be expected to be higher than that
In the general population, l.e.
close surveillance under an IND.
v
c
L0
, 3
i
APPENDIX 3
The Association between DMPA Use and Risk of Cervical Carcinoma In-Sltu
Reference
Schwallle PC. Experience with DepoProvera as an Injectable
Contracept1ve
J. Reprod. Med.
13:113, 1974
DMB Vol. 135A pp. 58-62
Number Women Studied
and Follow-up Period
Pooled data from U.S. and non-U.S.
studies on DMPA.
11,500 women treated with various
dosage regimens during 1963-1973.
and Upjohn Responses to the PBI
Questions 1982; p. 81
DMB Vol. 244 Tab #219
.1
Findings
Comments
1) 35 patients (out of 11,500)
diagnosed cervical carcinoma
in-situ.
(Papanicolaou smears were done
at 6 month to yearly Intervals
on all pat ients).
1) Information on dose and duration not
avaIlable.
2) Upjohn Report to the PBI,
referring to Schwallle's data,
adds that on further analysis
it was shown that 33 cases were
diagnosed among 6,378 U.S.
women studied for a total of
8504 woman-years.
(does it mean that only 2 cases
were observed In 5122 subjects
from other countries?)
3) Race - specific rates not provided.
?) “
No3 controls and no Information on
background incidence for the different
populations.
4) International variation In the rates
of cancer not controlled for.
(e.g., age-standardized Incidence
rates for cervix uteri are:
27.5/100,000 In Brazil,
62.8/100,000 In Colombia, /
25.5/100,000 in Puerto Rico,
16.2/100,000 In Osaka, Japan
compared to:
14.0/100,000 for White females
32.1/100,000 for Black females
in Detroit.
(based on Waterhouse and Muir, 1976).
I
I
I
APPENDIX 4
REVIEW OF TERATOGENICITY ASSOCIATED WITH DMPA EXPOSURE.:
Reference
3
Andrew, F.D. and Staples, R.E.
Prenatal toxicity of
medroxyprogesterone acetate
in rabbits, rats, and mice.
Teratology
15:25, 1976.
ANIMAL STUDIES
Study Design
Post-implantation study:
MPA given daily for 3, 6 or 9 consecutive days during gestation days
7-15 to mice and rabbits and to rats on days 8-16.
(at least 6 animals were included in each group).
Glucocorticoids cause cleft palate In rabbits.
FDA Vol. 302
Tab - 01
Cbang, M.C.
Effect of medroxyprogesterone
acetate and estrogen on the
development of the early
rabbit embryos.
Findings
1) Malformations attributable to MPA did not
occur in fetuses of mice or rats exposed to
the largest dosage tested.
2) But 1. 3 or 10 mg/kg on days 13-15 to rabbits
resulted in 6, 28, and 42% cleft palate,
respectively in each dose. Comparable cleft
palate frequencies were not seen in untreated
control fetuses.
3) A significant dose-related Increase In fetal
mortality occurred In rabbits given MPA at
all periods tested, but not in mice or rats.
Injection of high dose OMPA given to artificially inseminated
mature female rabbits on Day 1. Al I 6 rabbits (with MPA treatment
before or after insemination) were examined on Day 11.
1) The Injection of MPA on Day 1, 4, or 5 had no
effect on the development of early rabbit
embryos.
Contraception.
10:405, 1974.
c
Vol. 239(A)
Tab - Al 5
e
e
Eibs, H.G., Splelmann, H.
Hagele, M.
Teratogenic effects of
cyproterone acetate and
medroxyprogesterone treatment
during the pre - and
postImplantation period of
mouse embryos. I.
Pregnant mice treated with a single Injection of MPA on Day 2 of
pregnancy, or days 1-12 of gestation.
1) 1st experiment: MPA treatment on day 2 was
followed by sporadic Increases In dead and
resorbed fetuses, a decrease in fetal weight,
an increase in the rates of cleft palate and
malformed or abnormally developed fetuses.
2) None of these effects, however, were dose
dependent.
3) 22nd■ experiment: MPA treatment on one day
between 1-12 days revealed a high rate of
respiratory and urinary tract abnormalities
only on day 9.
Teratology,
25:27, 1982.
FDA Vol. 247
Tab - G598
4) Cleft palate was significantly more frequent
In all treated groups, though days of peak
sensitivity were not detected.
e
.i
5) All these effects were observed at massive
doses, i.e. the lowest dose in this study is
2 times the human dose.
6) It differs from the Andrew Study above In
strain of mice studied, time of Injections In
relation to implantation and dose level.
3
APPENDIXJ.
REVIEW OF TERATOGENICITY ASSOCIATED WITH DMPA__W0^^-:
Reference
Kimmel, G.L., Hartwell, B.S.,
Andrew, F.D.
A potential mechanism in
medroxyprogesterone acetate
teratogenes is.
ANIMAL STUDIES
Findings
Study Design
Female rats and rabbits studied to show a distinct species difference
in the binding of MPA to gluco-corticoid receptor.
1) Conclude that MPA may be teratogenic in
rabbits by binding with specific
glucocorticoid receptors.
Teratology,
19:171, 1979.
FDA Vol. 302
Tab - D25
Revesz, C., Chappel, L. I.,
Gaudry, R.
Masculinization of female
fetuses in the rat by
progestational compounds
(abstract)
Rats treated with MPA during 15th - 20th days of pregnancy.
4 received daily dose of 0.25 mg.
4 - 1 mg daily.
5 - 2 mg daily.
4 - 2.5 mg daily.
3 - 5 mg daily.
(small numbers of animals per treatment group).
Endocrinology
66:140, 1960.
FDA Vol. 239(B)
TAB - A67
Satayasthit, N., et al.
The effect of
medroxyprogesterone acetate,
administered to the lactating
ration the subsequent growth,
maturation, and reproductive
function of the litter.
e
28 control and 32 experimental female rats plus 19 control and 19
experimental male ra_ts. The experimental animals were produced
by rats given 5 mg. MPA/g body weight On Day 3 after parturition
(1 e study the effect on lactating young rats.)
2) MasculinizatIon of female fetuses In the rat
was observed.
There were 151 males, 18 females, and 2^5
intersexuals born to the 17 MPA treated
rats.
At the lowest dose of MPA, masculinization
was seen In 2 out of 17 females at 20 days of
age.
At 2.0 and 2.5 mg, all females were
pseudohermaphrodites; externally all appeared
1 Ike ma les.
The effects are related to the androgenec1ty
of the compounds.
1) No difference In growth rates between experl
mental and control animals observed until Day
60.
2) All the experimental rats became pregnant,
l.e. MPA
and no abnormalities seen at birth. 1.-.
Is not obviously androgenic.
3) Treatment of lactating females with MPA did
J. Reprod. Fert.
46:411, 1976.
ignlficantly
sL
. delay the onset of vaglnaj.
•
.J of •/
opening
and
the? first oestrus cycle In the
jOung
delay In sexual maturity).
young..(i.e.
(
This effect was observed at 2 1/2 the human
dose.
FDA Vol. 30
G-458
Abstract Vol. 6BA
.i
o
1) Delivery was delayed after large doses
of these compounds and the pups were removed
by Cessarian Section on Day 22.
4) What is missing In this study is a
sufficiently long follow up to see if
polycystic ovary develops, a sign of
androgen 1zatIon.
There is no page 200
I
5
I
APPENDIX 5A
EFF EXT_0F_MPA_0N_0FFSPRING £
EXPOSURE OF FETUSES TO CONIRACEPT IVE _OOSES_0L MPA
Reference
Dodds, G.H.
The use of sterile
medroxyprogesterone
acetate suspension as
a contraceptive during
a three-year period.
Number of Women and Condition Studied,
Dose and Duration of MPAJKe-
3 year follow-up of 1883 women
treated with DMPA (250 mg/3 mos)
in Hong-Kong.
Pregnancy .History
1) Out of 736 women who discontinued
DMPA. 279 (38X) were lost to
follow up or used other
contraception during the 10
month fo1 low-up.
2) One contraceptive failure:
injection given between 6-10
weeks of conception (i.e. high
dose early in gestation period).
Contraception
2: 15, 1975
FDA Vol. 74A
p. 319-323
3) 40 out of 60 women who discontin
ued DMPA to get pregnant, con
ceived within 18 months after
last injection.
Pregnancy Outcone and Effects on Fetus
1) Out of 184 pregnancies (at tail-end of
injection):
137 were live births
3 were stillbirths
26 abortions
3 ectopic pregnancies
9 currently pregnant
6 lost-to-follow-up, unk. outcome
2) In the one case of drug failure, a living
infant with no abnormalities was born,
sex not stated.
4) 144 planned i unplanned preg
nancies occurred within 1-28
months after last injection in
women receiving 1-9 injections.
Mohberg, N.R. and
Greenspan. A.
Depo Provera and
exogenous estrogen at
the Grady Memorial
Family Planning Clinic.
Technical Report,
Sept., 1980
Family planning clinic chart review
of random sample of 708 DMPA contra
ceptive users from the Grady Hospital
in Atlanta, GA. [Mean months of DMPA
use was 24 ].
1) 167 DMPA users (24%) of the
sample became pregnant sub
sequent to DMPA use.
2) Mean time elapsed from last DMPA
injection to conception was 18.2
months median time was 16.0 months.
3) 2 possible contraceptive
failures, i.e. injection
given after conception.
FOA Vol. 214 Tab U-29
.1
I
c
1) No Information provided on pregnancy outcome
because "the records of the status of the
babies born to former DMPA users were sparse
and cryptic. We would need a more thorough
follow-up system to allow claims to be made
concerning these infants" (p. 9).
I
1
APPENDIX 5A
EFFECL0f MPA ON OFFSPRINGi
EXPOSURE OFTEWSES^O CONTRACEPTIVE DOSES OF MPA
Reference
Parveen, L. et al.
Injectable contraception
In rural Bangladesh
Lancet,
1:946, 1977
Number of Women and Condition Studied.
Dose and Duration of MPA Use
1601 women using DMPA for up to
3 yrs. in Bangladesh. 1020 of
these were breast-feeding at time
of first Injection.
(150 mg. with estrogen supplement).
FDA Vol. 82A
pp. 129-130
Pregnancy History
1) only one pregnancy due to method
fal lure.
2) 3 women received first injection
while already pregnant.
*
3) 5 women received Injection while
already pregnant expecting It
would cause an abortion.
Pregnancy Outcome and Effects on Fetus
1) Of the total 30 pregnancies. 10 had normal
deliveries; 2 had spontaneous abortions;
18 currently pregnant.
2) No specific Information on the fate of
offspring exposed early in fetal life to
DMPA.
3) No information on effect on the breast-fed
children.
4) Therefore, a total of 9 cases
with exposure to maximal dose of
DMPA early in gestation.
5) 21 women became pregnant after
dropping out, i.e. tail-end of
the injection.
I
Powell and Seymour
Effect on
depo-medroxyprogesterone
acetate as a
contraceptive agent.
Am J Ostet Gynec
110: 36. 1971
FDA Vol. 126A
pp. 150-155
and see also
Seymour, R.J., Powell, L.C.
Depot-medroxyprogesterone
acetate: A
contraceptive.
752 patients followed over 44 months
In Galveston, Texas.
1) 4 patients had already
conceived when first Injection
was given.
2) 2 additional patients missed an
Injection and were exposed to DMPA
when already pregnant.
3) 1 case of method failure.
4) Therefore, a total of 7 patients
where maximal dose (150 mg/3 mos.)
occurred during early pregnancy.
5) 47 additional pregnancies occurred
after DMPA had been discontinued,
I.e. tail-end of the Injection.
Obst Gynecol.
36:589. 1970.
FDA Vol. 126A
pp. 202-209
.1
1) Among the 7 patients who Inadvertantly
received DMPA while pregnant: 4 had full
term, normal deliveries: 1 normal premature;
2 abortions. Sex of Infants not stated.
2) The outcome of the 47 pregnancies after DMPA
discontinuance (I.e. at recovery of
fertility) was:
29 full-term
5 premature
1 immature
1 ectopic pregnancy
8 abortions
3 lost to follow-up
N>
o
to
I
I
APPENDIX 58
TRANSFER OF MPA TO FETUS__OB_NEONATE
Reference
Besch, P.K., Vorys, N.,
Ullery, J.C.
In vivo metabolism of
H’-medroxyprogesterone
acetate In pregnant and
nonpregnant women and
In the fetus.
Number of Women and Condition Studied,
Dose and Duration of MPA Use
4 pregnant patients were administered
40 mg. MPA 8 hours before therapeutic
abortion to ascertain transplancental
mechanism and to study fetal
metabolIsm of MPA.
Pregnancy Outcome and Effects on Fetus
Pregnancy History
1) 4 male fetuses delivered
at 11, 12, 16, and 22 weeks
were removed by standard
procedures.
1) The fetal adrenal contained the bulk of the
radioactivity recovered. Larger amounts of
radioactivity were in the adrenals compared
to other tissues.
2) There appears to be a doubling in concent
ration between the 16th and 22nd wks. in the
amniotic fluid to blood ratio and In kidney
to blood ratio for MPA.
Am. J. Obstet. Gynec.
95:220, 1966.
OM0 Vol. 302
Tab 04
Crist, R.D., Krantz, K.E.,
Warren, J.C.
Placental transfer of
synthetic progestins
1) Significant quantities of MPA were found to
cross the in vitro perfused term placenta.
Term placentas obtained at normal
deliveries were exposed to samples of
progestins including MPA (110 mc/mg)
to study placental transfer.
2) Since the quantity of unchanged MPA
recovered was never as much as originally
added, this suggests alteration of the MPA
to unrecognized metabolites or binding
within the placenta.
e.g. the mataboltes of MPA contained less
than 5% of the radioactivity on the initial
chromatograms. Five metabolites of MPA were
present.
Obstet. Gynecol.
25:89, 1965.
DMB Vol.
Saxena, B.N. ,
Shrlmanker, K.,
Grudzinkas, J.G.
Blood and milk samples were collected
at regular intervals from 7 lactating
women given DMPA (150mg) one week
after delivery for contraception.
Levels of contraceptive
steroids in breast milk
and plasma of lactating
women.
Contraception
16:605, 1977
DMB Vol. 82A
pp. 167-175
.1
1)
In all 7 women the MPA levels in breast milk
were similar to plasma concentrations. The
ratio of MPA Inplasma to milk was 1:1
throughout the study period up to 87 days.
2)
No Information given on effects on the
suckling Infants.
I
K)
o
UJ
I
t
APPENDIX 58
TRANSFER OF HPA EQ PETO DR-NEONATE
Reference
Turner, S.J., Mizock, G.B.,
Feldman, G.L.:
Prolonged gynecologic
and endocrine
manifestations
subsequent to
administration of
medroxyprogesterone
acetate during pregnancy.
Number of Women and Condition Studied,
Dose and Duration of HPA Use
Pregnancy History
Pregnancy Outcome, and Effects on Fetus
1) 32 of the 77 patients conceived again
subsequently: Of these, 28 had normal
deliveries and 4 had abortions in 1st
trimester. 4 (out of the 77) were lost to
follow-up. No information on offspring
exposed to DMPA in utero.
20 ward service patients (ages 15-25)
10 being controls and 10 given DMPA
(100 mg. weekly) from 30th to 36th
week of gestation (cum. total of 500
mg DMPA). Additionally, 77 patients
treated with DMPA during pregnancy.
2) The objective was to study MPA effects on
fertility, not effects on the fetus.
Am J Obst Gynecol
95: 222, 1966.
I
FDA Vol. 16
pp. 1888-1893
Zanartu, J., Aguilera, E.,
Munoz, H., et al
Effect on a long-acting
contraceptive progestogen
on lactation.
Obstet Gynecol
47: 174, 1976
406 nursing mother receiving DMPA
every 3 or 6 months (150 mg or
250-300 mg, respectively) for
contraception.
1) Two mothers became pregnant
at the 12th and 21st months
postpartum while under the
regimen of 250 mg OMPA/6
months.
339 mothers were given DMPA within
3 months postpartum; 133 within the
first 30 days postpartum; 67 between
3 and 6 months postpartum.
FDA Vol. 16
pp. 2049-2051
a
2) Authors caution that a more careful
evaluation is required on its effects on
weight, growth, health, since MPA or its
metabolites are expected to be excreted in
the milk.
3) No information is provided on the two
pregnancies where fetuses were exposed to
high dose MPA.
.1
B
1) No adverse iatrogenic effects were observed
in the breast-fed children from mothers
on MPA. No information on how newborn were
tested or duration of follow-up.
I
to
o
I
f
APPENDIX 5C
RELATIONSHIP BETWEEN PROGESTINS AND HYPOSPADIAS
Reference
e
Aarskog, 0.
Maternal Progestins
as a Possible Cause
of Hypospadias.
New Eng. J. Med.
300:75-78. 1979.
See Also
Hndinis
Study Design
Case-control study. 130 hypospadic
patients and 111 patients with oral
clefts born during the same period
were studied retrospectively with
reference to pathogenetic mechanisms
that might have interferred with
testicular differentiation of function
during fetal life.
Aarskog, 0.
Clinical and cytogenlc
studies In hypospadias.
Acta Paedlatr. Scand.
(Suppl) 203:1-62. 1970.
1) In 11 out of 130 cases of hypospadias
(8.5%) there was a history of maternal
progestin Intake in early
pregnancy:
6 - for threatened abortion.
5
for pregnancy test (with estrogen).
2) In contrast, 2 mothers of 111
infants with oral clefts (1.8%)
had a history of progestin intake
in early pregnancy.
Conclusions and Comments
1) There appeared to be a relationship
between time of exposure to progestin
treatment (t.e. week of gestation) and
the degree of hypospadias.
I
■o
11
I
________ L
I
- I
119
I
I
1
I
130
(1.0%) I
2
I
______ L
I
109
I
1
in
RR = 5.04
(95% CI: 1.09-23.24)
Case-control study: 294 cases with
simple isolated hypospadias in
Budapest. Hungary born during
1970-1972 and a comparison group
of healthy children matched on
place and time of birth.
1) Sex hormone therapy for threatened
abortion was significantly more
frequent among mothers of cases
of hypospadias than among controls
(x3 = 9.00, p < 0.01).
28 of 294 cases with hypospadias
(9.5%) vs. 12 of 294 controls
(4.0%) were exposed In utero to
progestogens (RR - 2.4)
Hum. Hered.
29:166-171, 1979
0MB Vol.
2) More mothers of cases of
hypospadias took also sedatives.
.1
0
controls
+ I
progestins 1
0MB Vol .
Czeizel, A.
Toth, J.,
Erodl, E.
Aetlologlcal studies
of hypospadias In
Hungary
hypospadic
cases__
(8.5%) |
1) Significant seasonality was documented,
with hypospadias more frequently
occurring in children born between
August and December. [Roberts and
Lloyd. 1973, and Neto et al. reported
similar f indings. ]
sex
hormones
I
* I
1
I
- I
1
cases_____
(9.5%) |
28
I
266
controls
(4.0%) |
12
|
1 ______ L
I
I
282
I
I
1
I
294
RR = 2.47
(95% CI: 1.23-4.96)
294
I
1
e
APPENDIX 5C
RELATIONSHIP BETWEEN PROGESTINS AND HYPOSPADIAS
Reference
Mau, G.
Progestins during
pregnancy and hypospadias
Teratology
24:285-287. 1981
Cohort study: A sample of 3,602 male
newborns who were examined five times
(at birth, at 6 weeks, and at 9, 18
and 36 months). Detailed information
was available about the Intake of
drugs throughout pregnancy by the
mothers of these newborn.
1)
33 of the total cohort of
3602 male newborn had
hypospadias (1%).
2)
Mothers of this cohort of
newborns were exposed
to progestins either
for hormonal pregnancy
test (4.2% or 151/3602)
or for treatment for
threatened abortion
(11.3% or 408/3602).
3)
Of the 33 cases with
hypospadias, 13 mothers
had used in addition to
progestins other drugs
during the first trimester.
0MB Vol.
e
Conclusions and Comments
Findings
Study Design
A)
The frequency of exposure
to high doses of progestins
during early pregnancy was
15% in the group with
hypospadias (5 of 33)
compared to 11.3% in the
group without hypospadias
(403 of 3569). This high
dose intake was due to
treatment for threatened
abortion.
5)
The frequency of exposure
to progestins used as a
pregnancy test was 9% in
the group with hypospadias
(3 of 33) compared to 4.1%
in the group without
hypospadias (148 of 3,569).
6)
The combined exposure to
progestins was 24% in the
group with hypospadias (8/33)
compared to 15% in the group
without hypospadias (551/3569).
The relative risk = 1.75
(95% CI: 0.5-4.4).
.1
1)
Author concludes that if a risk exists, it
is a low risk, but nonetheless a distinct
poss ibi1i ty.
I
o
i
*
*
••••
•
I
I
•
APPENDIX 5C
RELATIONSHIP BETWEEN PROGESTINS AND HYPOSPADIAS
Reference
Study Design
Neto, R.M., Monteleon, R.R.
Castilla, E.E.,
Paz, J.E.
Case-control study of liveborn
in six Latin American countries
during 1967-1975.
Hypospadias:
An Epidemiological
Study In Latin America
324 male fetuses with hypospadias
were compared to a control group
consisting of the first nonmalformed
child born after each "case",
matched for sex. place, and time of
birth.
Am. J. Med. Genet.
10:1-10, 1981
DM8 Vol.
Drug intake during the first
trimester of pregnancy was
higher among mothers with
hypospadias than among control
mothers (34.4% vs. 20.4%).
1)
2)
Among the ingested drugs were
sex hormones. Specifically,
24 of 314 cases (7.6%) vs. 12
of 319 controls (3.8%) had been
exposed to sex hormones in utero.
3)
The overall frequency of
hypospadias was 7.6 per 10,000
live births. In 13.6% (44/324)
of cases, hypospadias were
associated with another
malformation, most often in
the genital area.
The frequency of hypospadias
by country was:
17.7/10,000
BraziI
11.5/10,000
Venezuela
6.6/10,000
Chi le
6.2/10,000
Argentina
5.0/10,000
Ecuador
Sweet, R.A.,
Schrott, H.G., et al.
Study of the Incidence
of hypospadias in
Rochester, Minnesota,
1940-1970, and a case
control comparison of
possible etiologic
factors.
1)
Case-control study: 113 cases of
hypospadias were identified in a
review of records of 13,776 live
male infants (8.2 per 100 live male
births) and for each case two controls
were matched on hospital of delivery,
last menstrual period of mother
(±1 month), and maternal age
(±1 year).
2)
Mayo Clin. Proc.
49:52-58, 1974.
0MB Vol.
Conclusions and Cownents
Findings
.1
Hydroxyprogesterone caproate,
intramuscularly, was given to
4 mothers of cases of hypospadias
but to only 1 control mother.
3.7% (4/107) VS. 0.4 (1/226).
9 mothers of cases of hypospadias
vs. 5 mothers of controls had
taken combination oral
contraceptives within 9 months
of their last menstrual period.
1)
No information Is provided on dose and
timing of exposure, nor Is any progestin
uniquely Identified.
2)
Authors state that the difficulty in
detecting an association between exposure
to sex hormones and hypospadias may be due
to the fact that the relative risk Is low
(but statistically significant) whereas the
Is
spontaneous incidence of hypospadias is
high.
Cases with
hypospadias
controls
|
(7.6%) |
(3.8%) I
Yes I
24
|
12
I
Sex
hormones
J_________1________ L
I
I
’
307
No |
290
I
I
290
I______ L_____ L
314
319
RR - 2.1
(95% CI: 1.1-4.3)
Cases with
hypospadias
|
(3.5%) I
+ I
<
!
progestins J
_____ L
'! _ i
112
113
controls
(0.4%) |
1
I
______ L
I
225
I
I
226
RR «= 8.04
(95% CI: 0.89 - 72.74)
Position: 769 (6 views)