The New England Journal of Medicine
February 3, 2000 -- Vol. 342, No. 5

Placebo Surgery in Trials of Therapy for Parkinson's Disease

To the Editor:

The Sounding Board articles by Freeman et al. (1) and Macklin (2) (Sept. 23
issue) provided an excellent discussion of sham surgery for patients with
Parkinson's disease. However, Freeman et al. seem overenthusiastic about
downplaying the risks of sham surgery. Their statement that these procedures
are easier to perform, standardize, and reproduce than traditional surgical
procedures does not do justice to the neurosurgical skills involved in
performing stereotactic brain surgery. It is probable that the wide range of
outcomes regarding the effectiveness of and morbidity associated with
pallidotomy is related to the varying skills of the operating neurosurgeon.

Freeman et al. present the cogent argument that sham surgery is the only
solution for proving the validity of the neurosurgical procedures discussed
in relation to Parkinson's disease. Not only are there important clinical
questions that need to be answered, but society also needs to know whether
any surgical therapy for Parkinson's disease provides substantial enduring
benefit.

Macklin is no doubt correct that the informed-consent process is less than
perfect. Rather than turn our backs on objectively evaluating surgical
procedures, however, we should be more diligent in obtaining consent. In the
end, the patients with Parkinson's disease who give informed and valid
consent to participate in trials in which sham surgery is used are truly
brave pioneers who are helping to advance the frontier of research. We
should not prevent such trials from taking place.

William J. Weiner, M.D.
University of Miami Medical School
Miami, FL 33155

References

1. Freeman TB, Vawter DE, Leaverton PE, et al. Use of placebo surgery in
controlled trials of a cellular-based therapy for Parkinson's disease. N
Engl J Med 1999;341:988-92.

2. Macklin R. The ethical problems with sham surgery in clinical research. N
Engl J Med 1999;341:992-6.

==================

To the Editor:

Macklin's argument against sham surgery in clinical research is
fundamentally flawed in two important respects. First, there is no ethical
requirement to minimize absolutely the risk of harm to research subjects.
The ethical requirements are that there be a reasonable balance between the
potential risks and the potential benefits to the subject or to society (1)
and that the subject make a fully informed choice about participating in the
research. (2) Within the context of a reasonable overall risk-benefit ratio,
risks should be minimized by using medically and scientifically appropriate
procedures. (1)

Second, preventing subjects from deciding what risks they are willing to
take is a direct violation of the principle of autonomy. We should not
declare research unethical because the process of informed consent can be
imperfect; instead, we should concentrate on improving that process. Medical
paternalism, in which physicians decide what risks to expose patients to, is
ethically unacceptable. A medical maternalism in which research subjects are
protected from all risks and from the autonomous exercise of their right to
choose is no better.

(The opinions expressed in this letter are those of the author and do not
necessarily represent the policy of the Department of Veterans Affairs.)

James F. Burris, M.D.
Veterans Health Administration
Washington, DC 20420

References

1. 45 CFR 46.111.

2. 45 CFR 46.116.

====================

To the Editor:

In 1993, when we proposed the idea of a double-blind design for fetal-cell
transplantation for Parkinson's disease, our plan was met with a good deal
of skepticism by members of the scientific community, including Dr.
Freeman's coauthor Dr. Olanow. (1) Because no neurosurgical trial had been
performed with a double-blind design, it was unclear whether such a protocol
could or should be carried out. Our double-blind study of 40 patients (20
patients who received implants and 20 control subjects) has been completed.
(2,3)

It is particularly challenging to do a double-blind surgical study in
patients with Parkinson's disease. Because it would be unethical to
discontinue antiparkinsonian drugs in these patients, neurotransplants have
to add substantially to the best pharmacologic treatment of each patient. It
is likely, though not sure, that the effects of transplantation would be
much easier to demonstrate in a placebo-controlled design in which the
placebo group did not receive drug therapy. Because of the great value of
levodopa and other drugs, such a placebo design could not be justified.
Double-blind studies are undertaken to demonstrate the value of a treatment
in a small group of patients with reasonable, but not zero, risk. In our
study, the patients who received implants were exposed to more risk than the
patients who had placebo surgery, because of needle penetrations of the
brain. Whether the patients who received implants benefited more than the
patients who underwent placebo surgery could be answered only with the
double-blind design.

We were not surprised that our design produced controversy seven years ago.
Even as discussion continues about the role of double-blind surgical
studies, review groups of the Food and Drug Administration and National
Institutes of Health have embraced the concept, and several such studies are
now under way.

Curt R. Freed, M.D.
Robert E. Breeze, M.D.
University of Colorado School of Medicine
Denver, CO 80262

Stanley Fahn, M.D.
Columbia University
New York, NY 10032

References

1. Cohen J. New fight over fetal tissue grafts. Science 1994;263:600-1.

2. Freed CR, Breeze RE, Greene PE, et al. Double-blind placebo-controlled
human fetal cell transplants in advanced Parkinson's disease. Soc Neurosci
1999;1:212. abstract.

3. Fahn S, Greene PE, Tsai W-Y, et al. Double-blind controlled trial of
human embryonic dopaminergic tissue transplants in advanced Parkinson's
disease: clinical outcomes. Neurology 1999;52:Suppl 2:A405.

====================

To the Editor:

Macklin states, "An alternative research design that did not involve sham
surgery... would be less rigorous from a methodologic point of view." That a
placebo procedure in a clinical trial be manifestly harmless is not only an
ethical requirement but also a fundamental requirement for clinically
relevant, generalizable inferences to be drawn from the trial. Even if the
implanted material is found to improve the patient's condition, it may not
fully counter any damage that may have occurred as a result of the surgery.
In other words, a positive effect of treatment in a trial involving a
potentially harmful control may result in the adoption into clinical
practice of a treatment that results in more harm to patients than if they
had received no intervention at all.

Some may argue that a trial that compares a treatment involving surgery and
implantation with conservative management and finds a beneficial effect also
admits the possibility that the implantation itself added nothing to, or
even detracted from, a beneficial placebo effect. If this is truly a
possibility, then perhaps the three-group trial that Macklin describes at
the end of her article is the only clinically informative option. However,
the assumption that sham surgery is likely to have a positive effect on
health requires careful examination.

Jim Slattery, M.Sc.
2 Guest Rd.
Cambridge CB1 2AL, United Kingdom

====================

To the Editor:

The advantages of a new treatment are often vastly overestimated, in spite
of the most objectively written notice of information, as a result of the
combined effects of the physician-investigator's expectations and the
patient's hopes. Therefore, many patients not only agree to participate but
also eagerly desire participation in studies of such treatments, which
represent the only hope offered to them. Under these circumstances it is
likely that the consent will not really be freely given, and therefore the
study may not be ethically acceptable.

I suggest the following solution: when participation in a randomized study
with a sham-procedure group is proposed to a patient, the patient must be
offered, in addition to the alternative of agreeing or refusing to
participate in the randomized study, the possibility of receiving the new
experimental treatment without participating in the randomized study.
Obviously, when offered this choice, fewer patients will agree to enter the
study. However, if the investigator presents the possible outcome in a
realistic way, with all the uncertainties about the outcome of the new
treatment, some patients will agree to participate for purely altruistic
reasons, and even more patients will want to participate if all the patients
are offered the possibility of receiving the new treatment later if it is
proved to be effective.

This design will force the investigator to present the possible outcomes of
the two groups of the study in a realistic and balanced way: if the
investigator exaggerates the expected benefits of the new treatment, no
patients will agree to participate in the randomized study. This approach
would make the recruitment of patients more difficult, but there are good
reasons why this type of study should not be easy to carry out.

Jean-Daniel Horisberger, M.D.
University of Lausanne Medical School
CH-1005 Lausanne, Switzerland

====================

The authors reply:
To the Editor:

It is affirming to note that all the correspondents support the use of
placebo-surgery controls under certain conditions. We agree with Burris that
it is neither required nor reasonable to expect surgical trials to involve
no risk. As with all trials, patients must be informed of the potential
risks, and risks must be reasonable in relation to the possible benefits to
the individual patient as well as to society. (1,2,3) We believe it is a
mistake to hold surgical trials to a standard more stringent than that
required for the study of nonsurgical interventions.

Horisberger recommends that patients be offered access to the experimental
treatment without participating in the study in order to minimize
exaggerated estimates of benefit and to ensure that informed consent is
freely given. Offering the patient such access to an unproved intervention
implies that the investigator condones the procedure and that it is
beneficial when this is not known to be the case. It would therefore serve
to exaggerate benefit and to induce a less informed decision about receiving
the intervention. We believe it is inappropriate for an investigator to
offer an intervention as a therapy while at the same time asserting that it
is necessary to perform a placebo-controlled trial to determine whether it
is effective.

We agree with Weiner that differences in surgical methods and skills can
affect the outcome of surgical procedures. However, it is easier to
standardize surgical methods used with nonablative and minimally invasive
procedures, as in the case of cellular transplantation. It is because of the
variability that can occur even with a single surgeon that randomized
clinical trials are necessary. (4)

Freed et al. assert that we were skeptical about the use of a placebo
control in their trial design. Rather, we argued for multiple trials in
which different variables with regard to transplantation were used, so that
negative or inconclusive results with a single protocol would not adversely
affect the entire field. (5) We were not opposed to the use of placebo
controls. In our article, we cited numerous examples of such studies
(including neurosurgical trials) that predated both of our trials.

Our concern is that patients may undergo operations performed by surgeons
who stand to benefit financially, academically, and by reputation but who
fail to disclose that the operations have unproved benefits and undefined
risks. If the safety and efficacy of a new surgical procedure are not
established in controlled trials involving patients who give informed
consent, thousands of future patients may be subjected to the risks and
costs of an unproved operation in uncontrolled studies, without their full
knowledge and consent.

Thomas B. Freeman, M.D.
University of South Florida
Tampa, FL 33606

Dorothy E. Vawter, Ph.D.
Minnesota Center for Health Care Ethics
Minneapolis, MN 55454

C. Warren Olanow, M.D.
Mount Sinai School of Medicine
New York, NY 10029

References

1. Department of Health and Human Services, National Institutes of Health,
Office for Protection from Research Risks. Protection of human subjects. 45
CFR 46 (1991).

2. National Commission for the Protection of Human Subjects of Biomedical
and Behavioral Research. The Belmont report: ethical principles and
guidelines for the protection of human subjects of research. Washington,
D.C.: Government Printing Office, 1978. (DHEW publication no. (OS) 78-0012.)

3. Levine RJ. The use of placebos in randomized clinical trials. IRB
1985;7(2):1-4.

4. Baum M. Reflections on randomised controlled trials in surgery. Lancet
1999;353:Suppl 1:SI6-SI8.

5. Cohen J. New fight over fetal tissue grafts. Science 1994;263:600-1.

====================

To the Editor:

The claim that my argument against sham surgery in clinical research is
"fundamentally flawed in two important respects" is mistaken on both counts.
Burris contends that "there is no ethical requirement to minimize absolutely
the risk of harm to research subjects." I do not know what he intends by
"absolutely"; it is a term without meaning in this context. However, a look
at the U.S. Code of Federal Regulations reveals the following as the first
of the requirements that institutional review boards must determine are
satisfied in order to approve research: "Risks to subjects are minimized."
(1)

The second alleged flaw is that "preventing subjects from deciding what
risks they are willing to take is a direct violation of the principle of
autonomy." But the role of the institutional review board in our system for
the protection of human subjects is precisely to determine the acceptability
of risks. Once the institutional review board has determined that risks to
subjects are minimized and that "risks to subjects are reasonable in
relation to anticipated benefits," (1) then potential subjects are asked to
provide their voluntary, informed consent. If subjects' informed consent
were the only ethical requirement for research, there would be no need for
institutional review boards to make risk-benefit determinations before
approving research protocols.

Weiner endorses the view that "sham surgery is the only solution for proving
the validity of the neurosurgical procedures" and concludes that we should
not prevent such trials from taking place. Although I acknowledge that using
sham surgery as a control may be the best way to seek proof, it is not the
only "objective" way to demonstrate efficacy. There are other, albeit less
rigorous, ways. Slattery questions whether it is even the best way, since
harm produced in a control group of a trial prevents investigators from
determining whether there is a benefit from the experimental treatment.

Freed et al. note that review groups of the Food and Drug Administration and
National Institutes of Health have embraced the concept of trials that
involve sham surgery and that several such studies are under way. This means
that the trials were approved by institutional review boards, an important
procedural requirement in our system for the protection of human subjects.
However, approval by an institutional review board does not make a study
ethically acceptable. It is evidence only that the institutional review
board found the research to be ethically acceptable, a judgment with which
another institutional review board might reasonably disagree.

Ruth Macklin, Ph.D.
Albert Einstein College of Medicine
Bronx, NY 10461

References

1. 45 CFR 46.111.