Madame Chairman, Members of the Subcommittee, thank you for giving the Food
and Drug Administration (FDA or the Agency) an opportunity to testify at this
hearing. I am Mary K. Pendergast, Deputy Commissioner and Senior Advisor to the
Commissioner. With me today is Dr. D. Bruce Burlington, Director of the FDA's
Center for Devices and Radiological Health.
The focus of this hearing, technological developments in genetic testing,
and the questions raised concerning the quality, accuracy, and use of the tests,
are significant issues and of critical importance to the FDA. For purposes of
this hearing, I will focus on the principal scientific issue posed by genetic
testing, i.e., whether the genetic test is scientifically valid and the role the
FDA has in that determination.
BACKGROUND
Over the past two years, the Agency has met several times with Dr. Francis
Collins, Director of the National Center for Human Genome Research at the
National Institutes of Health, Dr. Neil A. Holtzman, Chair of the Task Force on
Genetic Testing, and others to discuss the important and challenging medical and
scientific issues surrounding genetic testing. These conversations have been
very productive and useful, as we have grappled with whether, and to what
extent, there should be oversight of genetic testing. There are no obvious, or
easy, answers to those questions.
THE QUESTIONS RAISED BY GENETIC TESTING
While there are many ethical, legal, psychological, and public policy issues
raised by genetic testing, there are three principal questions that have been
raised with respect to the performance of the tests themselves. These questions
are critical to understanding what actions the FDA has taken in the field of
genetic testing.
Does the genetic test accurately provide information about a patient's
health status? That is, is the genetic test useful to the patient and his/her
physician?
Did the clinical laboratory performing the test run the test properly? Are
we confident that the clinical laboratory got the correct answer actually
determining the presence or absence of the gene?
Is there appropriate information available so that patients can be educated
and counseled about the meaning of the genetic test?
The key question that the FDA, Dr. Collins, members of the Task Force on
Genetic Testing, and others, continue to grapple with is who should decide the
answers to these three questions. one thing is certain, there is no one group or
entity -- whether inside or outside of government -- that has the full range of
expertise and capacity to be seriously engaged in all three areas. Rather, we
must look to strategic alliances, to partnerships, and to creative approaches in
order to begin to resolve the questions posed by genetic testing.
The "who should decide" inquiry is further compounded by the fact that
genetic tests are used for a wide variety of different uses, and answers to
those three questions may differ depending on the particular use of the genetic
test. Thus, for example, genetic tests are used to collect ancillary information
in patients with an established diagnosis of cancer for tumor subtype
evaluation. In this context, the patient already has been diagnosed with a
disease and the genetic test is used to optimize treatment for the disease.
Genetic tests also are used for diagnostic testing in a symptomatic
individual. When a person goes to see his or her doctor because of symptoms,
the physician may use genetic tests to confirm or rule out certain diseases. In
this instance, the information gained from the genetic test would have an
immediate impact on both diagnosis and therapeutic decisions.
In the more complex situation, genetic tests also may be used to predict, in
healthy individuals and in fetuses through in utero testing, what their medical
problems might be in the future. The test, instead of being used to support
conclusions based on the physician's current assessment of the patient, is being
used to predict, with variable certainty, future patient outcomes. Because of
the time lag between the test and when the person gets sick, it may be decades
before the accuracy of a predictive test can be assessed. Moreover, some of the
tests are for diseases -such as Alzheimer's -- that strike people of advanced
age. If a genetic test determines that someone has a gene associated with
Alzheimer's disease and the conclusion is that person may get Alzheimer's
disease many years from now, what can we say about the likely time until disease
onset and, indeed, the chances that it will ever occur in the person tested?
In addition, in many cases genetic testing and study is initially focused on
families with a high incidence of a certain disease. It takes careful
scientific study to learn whether the results of such tests are applicable to
the general population. At present, scientists are studying whether the
correlation between the BRCA1 gene and breast cancer seen in two dozen Ashkenazi
Jewish families with strong histories of breast cancer is applicable to all
women generally.
There also is the public policy question of whether genetic tests should be
treated the same as, or differently from, other tests for the same condition or
disease. For example:
Does it matter if you learn you have a genetic disease such as sickle cell
anemia or cystic fibrosis through a test on a enzyme or a protein in your body,
or through a test on your DNA?
Is it more important for a genetic test to be better validated than a
chemical test for the same disease? s Should genetic tests be performed by a
clinical laboratory with a higher quality of testing services than those for
chemical testing?
TEST ACCURACY -- SCIENTIFIC VALIDATION
When scientists discuss the scientific validity of a genetic test, they are
talking about two different things. First, they are trying to ascertain whether
it is possible to actually test for a particular gene and its mutations and get
the correct answer. This is called the analytical sensitivity and specificity
of the test, and it addresses the question of how often there will be false
positive and false negative test results for the presence or absence of the gene
itself.
Second, and the more critical concern regarding the scientific validity of
genetic testing, is the question of whether a particular gene and its mutations
are related to a disease (i.e., whether the genetic test has clinical validity).
The mutation in a gene does not always predict disease in the tested
individual. For example, Dr. Collins, Director of the National Center for Human
Genome Research at the National Institutes of Health, led a team that discovered
the gene for cystic fibrosis. They described the most common mutation found in
cystic fibrosis patients. Further studies identified several hundred other
mutations in the cystic:: fibrosis gene. Some of these were associated with
severe cystic fibrosis, but many others were found in healthy individuals. An
accurate diagnosis of a mutation (or more accurately, a genetic polymorphism),
therefore, may not be clinically relevant and may not determine who is "sick" or
"well."
Because we do not know what a genetic test means in terms of the severity of
disease that can be expected, genetic tests have significance for what it means
to be "sick" or "well." For example, let us say that a child is tested and told
that he has a gene associated with polycystic kidney disease. But perhaps that
child has a variant of the disease that is mild and he would have lived a full
life without noticing any symptoms of the disease. If no genetic test had been
performed, he would have felt "well" because he never had any symptoms that
interfered with his life; he would not have been "sick." But, by performing the
genetic test, the person is now a person with a disease, i.e., a "sick" person.
Thus, we cannot simply intuit that a gene and its mutations are associated
with a disease; rather, the correlation between a gene and a disease requires
the collection and analysis of scientific data. Thus, there are several
critical questions, including:
- Who should gather the data and who should analyze it?
- Should the data be gathered before or after taking the test to market?
- Does the amount of data needed depend on the severity of the disease or on
whether the data is used for an immediate diagnosis or for the prediction of
future events?
- Is it possible to put a test on the market for some uses (e.g., immediate
diagnosis and treatment), but require additional data for a different use (e.g.,
widespread screening and prediction)?
If there are simple things you can do to avoid the impact of the disease,
such as diet or exercise, should more allowance be given for error and less data
than if disease avoidance would require the taking an important life step --
such as a prophylactic mastectomy or chemoprophylaxis because of a prediction of
breast cancer, foregoing children because of the risk for Huntington's disease,
or terminating a pregnancy because of a risk of congenital birth defects? What
warnings do we owe the public? Should cautionary labeling such as, "the meaning
of these test results is not yet known," be used in situations; where the data
is incomplete or insufficient?
- How can we create incentive systems to get to the correct
answers?
FDA'S ACTIVITIES REGARDING GENETIC TESTING
To date, the FDA has minimal involvement with genetic testing, although
there are a few areas where the FDA has had some activity. Although genetic
molecular diagnostic technology has been applied to a number of products,
particularly in the area of testing for infectious diseases, only two companies
have had products approved. These products are for test kits used to identify
particular genes that may provide information about the classification and
possible behavior of two cancers, lymphoma and leukemia.
A. Investigational Therapeutic Products Using Genetic Tests As Diagnostic
Criteria Currently before the FDA are a small number of requests by researchers
to treat patients with genetic diseases by novel therapies involving cells and
gene therapy products. The scientific validity of genetic tests comes into play
in two ways through these therapy protocols.
First, because of the highly experimental nature of the protocols, you
should not treat a person for a genetic disease unless you know that the person
actually has the disease. Genetic tests are sometimes part of the entry
criteria for the gene therapy trial. Thus, the analytical sensitivity and
specificity of the test directly influences the validity and appropriateness of
the involvement of the patient in the experiment. Stated another way, the
scientific validity of a genetic test used for entry into a study is paramount
in assuring that the patient will actually have the otherwise untreatable
disease.
This issue was brought into sharp focus this past year when the FDA
authorized several protocols involving the use of cellular therapy to be
performed on fetuses with known genetic diseases. In one case, one of our
scientists had questions as to whether the genetic tests used to identify the
condition in utero were predictive. She requested a confirmatory test be
performed and learned that the fetus did not have the genetic disease and the
high risk in utero cellular therapy was avoided.
In another case, the sponsor wanted to treat fetuses with all forms of
adrenoleukodystrophy, a disease in which there can be irreversible brain damage.
The problem is that the disease has variable outcomes ranging from irreversible
brain damage to essentially normal (an asymptomatic state). An FDA scientist
worked with the sponsors to limit the studies to only those fetuses that would
develop the irreversible form of the disease.
Second, in the field of gene therapy sometimes the genetic test will play a
role in the design of the therapy for the disease. The FDA recently authorized
a protocol to use a normal BRCA1 gene to treat patients with ovarian cancer with
a diagnosed mutation in the BRCA1 gene. In other experimental protocols, a
number of cystic fibrosis patients have been treated with the normal gene CFTR
(cystic fibrosis transmembrane conductance receptor). All patients had the
delta 508 mutation that had been clearly linked with cystic fibrosis disease.
Thus, the quality of the treatment is often dependent on the quality of the
test.
Because of the importance of the analytical and clinical validity of genetic
tests used as a prelude to gene therapy, the FDA has asked sponsors of gene
therapy to provide the Agency with information that would establish the quality
of their genetic tests.
B. FDA's Proposal to Regulate Analyte Specific Reagents
Because the vast majority of genetic tests offered in the United States
today are offered by either clinical or research laboratories which have created
their own tests, the use of these in-house developed genetic tests raise many of
the questions raised by other laboratory practices These genetic tests have thus
been considered by the Agency as part of its review of laboratory tests.
Clinical and research laboratories often develop and prepare their own tests
that are intended to diagnose various medical conditions, using ingredients that
they frequently purchase from biological or chemical suppliers. The ingredients
and other materials used in developing these tests may be divided into two
groups. The first group is referred to as "general purpose reagents," which
include the laboratory apparatus, collection systems, and chemicals used broadly
in a wide variety of tests. The second group is composed of chemicals or
antibodies that may be thought of as the "active ingredients" of a test and
which are useful only in testing for one specific disease or condition.
Because, in laboratory terms, the chemical you are doing the analysis for is
called the "analyte,' these active ingredients are referred to as "analyte
specific reagents" (ASRs). These in-house developed tests (sometimes referred
to as "home brew" tests) include a wide variety of tests used in the diagnosis
of infectious diseases, cancer, genetic. and various other conditions.
The FDA currently regulates the safety and effectiveness of diagnostic tests
that are traditionally manufactured and commercially marketed as finished
products pursuant to our statutory authority to regulate medical devices. The
in-house developed tests, however, have not been actively regulated by the FDA,
and the ingredients used in them generally are not produced under FDA assured
manufacturing quality control.
Because there are no controls over the analyte specific reagents used in the
diagnostic tests, neither patients nor practitioners have assurance that all
ingredients in the laboratory-developed tests are of high quality and capable of
producing consistent results.
The FDA has been concerned that, because these ASRs are of unpredictable
quality, the present situation poses a range of risks to the public health. For
tests with impact principally on the individual, inaccurate diagnoses can result
in poor patient care and increased health care cost. In addition, as a matter
of public health, if a test for HIV, tuberculosis, or another infectious disease
is wrong, the society, not just the patient, can suffer because the patient can
unwittingly pass the disease on to others.
Therefore, on March 14, 1996 the Agency proposed a regulation under its
medical device statutory authority that would require, among other things, that:
- ASRs be made under good manufacturing practices;
- ASRs be used to create in-house tests only by laboratories certified as
"high complexity laboratories" under the Clinical Laboratories Improvement
Amendments of 1988 (CLIA);
- Notice be provided to practitioners that the scientific validity of the
tests, using ASRs, are not reviewed by the FDA.
(61 Fed. Reg. 10484). Under this proposed rule, for most tests, including
genetic tests, the FDA would not do any assessment as to whether any particular
test offered by any laboratory had any analytical or clinical validity. Rather,
the proposed rule is intended to limit the use of ASRs to laboratories that are
required under CLIA to do their own analytical validation, and to keep the
quality of the ASRs consistent. Thus, if a laboratory has developed a test that
has analytical utility, then the test will not be rendered meaningless by poor
quality reagents.
The Agency's proposed rule does not address whether any particular genetic
test might be clinically valid, i.e., that it might accurately forecast a
person's disease state. The proposed rule reserved more rigorous effectiveness
requirements only for ASRs to be used in tests that are intended to diagnose
contagious conditions likely to result in fatal outcomes (e.g., HIV and
tuberculosis), and in tests used to determine the safe use of blood and blood
products, safeguarding the blood supply. The proposed rule also would not
regulate the claims made by the testing companies, clinical laboratories, or
physicians for the use of the tests, although it does require that there be a
disclaimer that the test has not been reviewed by the FDA.
As part of that ASR rulemaking, the Agency asked the public to comment on
whether there should be any further regulatory requirements imposed on the ASRs
used for human genetic testing, 1 and whether distinctions could be made
depending on the type of genetic test that was offered. Specifically, the
Agency asked whether more stringent controls, if thought necessary, could be
limited to "only those ASR's used in tests intended for use in overtly healthy
people to identify a genetic predisposition to a dementing disease, or to fatal
or potentially fatal medical disorders (e.g., cancers or Alzheimer's disease),
in situations where penetrance is poorly defined or variable and latency is long
(5 years or longer)." 1 The analytes for most of the new genetic tests are
segments of DNA that contain the genetic mutations that are either known or
thought to cause the disease.
The comments to the proposed rule were mixed. Some commentators supported
the ASR proposed rule, with its limited oversight of the ASRs used in genetic
testing. Other commentators -including a majority of the members on the Task
Force on Genetic Testing -- favored a more extensive regulatory approach for
genetic tests. The Agency is reviewing the comments received.
CLINICAL LABORATORY COMPETENCE
Although I have focused my remarks on the analytical and clinical validity
of genetic testing, I would Like to briefly address the issue of laboratory
competence. The Uncertainties surrounding the validity of genetic tests are
magnified because genetic testing is very complex and requires extreme care, and
there are few laboratories which have extensive experience in performing DNA
analyses. If history is any guide, without some type of quality control system
in place, many, if not most, of the laboratories that offer these tests will
perform them competently, but some will not. We, collectively, need to learn
from all of the various ways laboratories and other diagnostic providers have
had oversight in the past -- under such models as CLIA and the Mammography
Quality Standards Act. We need to look at the problems and successes under
these models to identify a model of oversight that might be the most appropriate
for this new technology.
POSSIBLE IMPACT OF ADDITIONAL REGULATORY OVERSIGHT
At present, we estimate that there are, or soon will be, dozens of companies
or laboratories offering hundreds of different
CONCLUSION
The future of genetic testing is filled with both great hope and some
trepidation. The FDA is carefully considering what, if any, additional role
it should play in the oversight of genetic testing. To that end, the Agency
looks forward to the guidance on these issues which will be offered by the Task
Force on Genetic Testing which will be presented to the Secretary of the
Department of Health and Human Services.
Thank you again for the opportunity to testify, I will be glad to answer any
questions.
