More Placebo Clinical Trials Predicted for Cancer Patients. After the controversy created by the use of a placebo in an early clinical trial for GIST patients, we hoped this issue would have gone away. Last year at the American Society of Clinical Oncology (ASCO) conference, it was raised again by an official of the Food and Drug Administration (FDA) who chose to advocate the use of placebos in clinical trials for cancer patients. This year at ASCO we were greeted by a paper called: “Research, Ethical and Regulatory Perspectives Regarding the Use of Placebos for Terminally Ill Patients with Cancer”.
After noting that clinical trials for cancer have not historically involved placebo controls, the authors address whether that precedent should change with the relatively recent development of novel targeted cancer drugs such as Gleevec. They observe that these new drugs are different from more traditional cancer therapies in that they may produce disease stabilization rather than tumor reduction and that they may do so with comparatively reduced toxicities. They reason that the use of placebos may help tease out the efficacy of these drugs when stability-rather than tumor shrinkage- is the end point and, further, that the lower toxicity of the drug “may allow patients, investigators and involved clinicians to presumably remain blinded to the use of a placebo.” They predict that placebo-controlled trials involving patients with terminal disease are expected to become increasingly common.
Amongst the criteria they cite for a placebo-based clinical trial to be considered ethical is that the trial must have a methodological justification and must fulfill the following ethical consideration regarding risk: “a patient randomly assigned to placebo should not be substantially more likely than those in active treatment group(s) to: die; suffer irreversible morbidity, disability, or other substantial harms; suffer reversible but serious harms; or suffer severe discomfort.” They further note that for a trial to be ethical, the placebo arm must also include the best supportive care.
The authors consider that the use of historical data as a substitute for randomized placebo-controlled trials is problematic due to differences in patient populations, unrecognized prognostic factors, changes in supportive care and subsequent therapies over time.
They further discuss the differences in various placebo trial designs, the major one being whether the trial protocol permits a cross-over to the treatment arm should the patient on a placebo demonstrate disease progression.
They recommend (“encourage”) that the choice of placebo trial design should be based upon a “dialogue between investigators, sponsors, and the FDA,” although they also stress “that the FDA does not mandate the use of placebocontrolled trials.”
Comment on Placebo Clinical Trials:
Although the authors seem well intentioned, it is striking that all four are physicians and never include within their perspective the participation of patients. Can one imagine in comparison, an article, entitled Research, Ethical and Regulatory Perspectives Regarding the Reimbursement of Physicians for the Care of Terminally Ill Patients with Cancer, discussing profound changes in such reimbursement without inviting the point of view of a single physician? It seems to me that whomever decides whether to permit a placebo-based clinical trial for terminally ill cancer patients is as important as the criteria for such a decision. Under current circumstances a person in a criminal trial would receive much more due process, certainly including the right to speak and be represented, than he would in a clinical trial.
Had the authors asked us for our opinion we would have told them the following:
1. You may not make decisions about us without us, stealing a mantra from our colleagues at ECPC, the European Cancer Patient Coalition. Current FDA procedures not only do not mandate the participation of patients in the development of clinical trial protocols, they will not permit patients to see the proposed protocol unless the pharmaceutical company involved consents, something the company will generally not do unless the patient agrees in advance not to disclose what he/she sees. In other words, the patient must agree not to publicly discuss with other patients what he has seen in order to get permission to see it.
2. The burden of proof for dismissing the use of historical data must be upon those proposing to use a placebo and that burden must pass the objective and transparent review of a panel that does not include those proposing the trial protocol. That has not happened in the past.
3. There must be an objective and transparent assessment of the short and long-term consequences of any placebobased trial to those that were placed on such a placebo, including a follow-up and careful documentation regarding their survival and well-being. That has not happened in the past. Particular attention should be paid to the following question: if a drug being tested is likely to produce stability rather than tumor shrinkage, then how could any progression which occurs due to a placebo be reversible?
4. Patients are not idiots and can figure out whether they are on a placebo. In fact this has happened in the past when patients on a placebo-based clinical trial discovered that not only did they have different side effects than those on the actual drug (a decrease in toxicity does not mean the absence of recognizable side effects) but that when they opened the pill, the placebo was a white powder and the drug was orange. Some of the patients involved began a discussion over the internet regarding how to decipher whether one had been given a placebo, thus affecting the trial blind. When we presented this situation to the drug company we were dismissed with the observation: “so what, the patient has to stay on the placebo anyway in order to be able to cross-over to the actual drug.” Aside from the fact that this is simply not true (what prevents the patients from just not taking what they know is a placebo, or opting for surgery, or just leaving the trial?), it certainly raises a serious issue regarding being able to maintain a blinded placebo in a clinical trial for savvy and internet connected patients. Finally, I would add the provocative note that a group of patients may be capable of going one step further by creating a mechanism to test whether a pill is real or a placebo. While one might legitimately criticize such a step as a threat to the integrity of the research process, others might understand that terminally ill patients fighting for survival may have competing interests.
5. Finally, we submit that the search for bias free clinical trial protocols for new targeted drugs must also consider that should patients remain on these drugs for some time that changes in actual drug dosage must be considered. Current clinical trials make no attempt to factor this in and instead consider that the dosage initially prescribed should be exclusively used for analysis regardless of how many patients are no longer on that regimen when they demonstrate disease progression.
Let’s end with a story:
A major patient advocacy organization has gathered a clinical trial group consisting of cancer researchers, pharmaceutical company executives and FDA officials on the roof of a 50 story building. At issue is whether a new and novel fast opening parachute would permit the survival of someone falling from such a height. Historical data is deemed by the patient advocacy organization to insufficient to answer this question.
Half the clinical trial group is fitted with the new parachute on their back. The other half is fitted with a placebo parachute on their back. The patient advocacy organization concludes that neither group is smart enough to figure out whether they have the real parachute by temporarily removing what is on their back and carefully examining it.
They then set a fire on the first floor of the building and, after the flames have reached the 49th floor, they ask the clinical trial group to sign an informed consent form agreeing to participate in the trial. One hundred percent of the group signs, closely matching the consent rates of terminally ill cancer patients being asked to agree to a placebo-based clinical trial. Patients then assemble on the street a safe distance from the building and observe the clinical trial participants leaping off the roof.
Of the 50 participants half were randomly assigned to the actual parachute group and half to the placebo group. Thirty-three percent of the actual parachute group survived the fall as compared to zero percent of the placebo group! Publication of the data will be presented at a meeting of PASCO (Patient Advocacy Society Cancer Organization).
From the Section of Hematology/ Oncology and the MacLean Center for Clinical Medical Ethics, University of Chicago; Section of Hematology/ Oncology and the Cancer Research Center, University of Chicago, Chicago, Il; Division of Bioethics, National Institutes of Health, Bethesda, MD; U.S. FDA, Washington D.C.