Debate: When Is it Ethical / Unethical to Use Placebo? – PLoS

Debate: When Is it Ethical / Unethical to Use Placebo? – PLoS

Wed, 12 Oct 2005

A debate in PLoS Medicine addresses the question: Should researchers test an experimental treatment against placebo to prove the superiority of the new treatment?

The ethical standards for research adopted by the World Medical Association as defined in the principles of the Declaration of Helsinki (DOH) state: “The benefits, risks, burdens and effectiveness of a new method should be tested against those of the best current prophylactic, diagnostic, and therapeutic methods.”

A note of clarification (in 2000) states: “in general this methodology [placebo] should only be used in the absence of existing proven therapy”

This affirms the clinical and ethical perspective of responsible practitioners and patients for whom “the crucial question in evaluating a new treatment is how it compares with the standard available treatment, and not whether it is better than placebo.”

In a debate published in PLoS, both sides acknowledge that the likelihood that–“the best current mehtods”–i.e., those approved by the FDA and recommended in physician practice guidelines–may not necessarily be safe and effective because the standard of “proof” used by the FDA is often invalid or unreliable.

Andreas Stang, Hans-Werner Hense, Karl-Heinz Jöckel argue: “the important issue is to decide when it is that we can call a therapy “standard” – that is, when can we speak of an indisputable benefit t that would make a currently available treatment¹s use in a trial control group ethically imperative?”

They acknowledge the unreliability of the evidence: “Clinical guidelines or recommendations based on high quality evidence SOMETIMES exist to support use of such a therapy.” [emphasis added].

Indeed, all too often currently marketed, widely recommended treatments, have not been proven safe and effective–certainly not in every controlled clinical trial in which they were tested. Under FDA’s lax standards drugs are approved even if they failed to demonstrate efficacy in 10, 20, 50– unlimited number of trials–as long as they can show two trials in which a new treatment shows an effect greater than placebo. And the trials are not designed to detect rare, but severe, even deadly safety hazards.

Thus, Erick H. Turner (a former FDA medical officer) and Martin R. Tramèr argue that FDA’s approval standards are not proof of efficacy–they only indicate an assumption of efficacy which all too often is proven wrong.

“If a drug with historical evidence of efficacy could be relied upon to be unfailingly effective – and placebo unfailingly ineffective – in all future clinical trials, we would readily admit that placebo is unnecessary and therefore unethical. The reality is that “proven effective therapy” – better called “assumed effective therapy” (AET) – often fails to show superiority to placebo. This is not because these drugs are in fact ineffective, but because the trials in question lack assay sensitivity . Assay sensitivity is defined as the ability of a trial to distinguish an effective from an ineffective therapy.”

Turner and Tramer make a strong case when they cite the unpublished FDA antidepression drug trial data first analyzed by Khan (2000; 2003) then by others.

“Khan et al. gained access to unpublished, as well as published, clinical trials data on antidepressants from the Food and Drug Administration (FDA) via the United States Freedom of Information Act. They obtained the FDA review documents on 51 clinical trials on nine antidepressants approved between 1985 and 2000. Of 92 active treatment arms (all involving doses that were eventually approved), 47 (51%) failed to demonstrate statistical superiority to placebo. Of these, there were seven cases (15%) in which the placebo arm was actually superior to AET. Thus, it can be seen that the phrase “proven effective therapy” should be taken with a grain of salt.”

“Now, what if the FDA had not had the benefit of looking at the placebo arms and relied on an equivalence or noninferiority design comparing study drug with AET?

Khan et al. list 12 flexile-dose studies in which (now approved) study drug outperformed AET (previously approved antidepressants). Many opponents of placebo would argue that each of these 12 trials provides ample evidence for efficacy of the study drug. However, because these trials did include placebo arms, we discover that in four of them (33%), neither AET nor study drug beat placebo. (In fact, in two of these four trials, AET was numerically inferior to placebo.) Therefore, in these four antidepressant trials, the two “active” drugs were not equally effective, but rather equally ineffective. This critical distinction would have been lost without placebo, and it would have been impossible to ascertain that these seemingly positive trials were in fact false positive trials.”

“The problem of assay sensitivity is not confined to antidepressants or even to psychotropic drugs in general. A meta-analysis by Tram,et al found that, among 52 possible comparisons between the “proven” antiemetic ondansetron and placebo, 19 (37%) failed to show a difference.”

Turner and Tramèr further argue that placebo trials are better at uncovering safety issues: “The declaration is silent on the possibility that omitting placebo can lead to problems, too, as we have now witnessed with Vioxx.”

The placebo issue clearly needs further debate focusing on the best interest of the patients who are subjects of trials. It would be wrong to defend placebo merely to overcome the confounding unreliable findings from poorly designed trials. Improving the standards and quality –and thereby the reliability of clinical trial findings–should be the first order of the day for institutional review boards (IRBs), clinical trial investigators, funding agencies, the FDA, and the journals. The research stakeholders’ financial interests have undermined public safety, resulting in preventable human casualties.

  See: the placebo debate in two issues of PLoS Medicine: March 2005 | Volume 2 | Issue 3 | and June 2005:

FDA’s drug approval process is being roundly criticized: the agency’s standards for evaluating safety and efficacy of new treatments have been shown to be demonstrably flawed; faulty judgments most often benefit commercial interests while harming consumers. It is a matter of record that the FDA has allowed blatant conflicts of interest to taint the approval process, and this tainted process led to the marketing of drugs whose safety and efficacy FDA officials knew were largely unproven–since in most of the pre-marketing trials, they failed. Such drugs were nevertheless unleashed on an unsuspecting public and allowed to be aggressively marketed for large populations. FDA’s tacit complicity in the concealment of negative data from the public, opens the FDA to charges of negligence;

Even as these major shortcomings have been uncovered, the FDA has asserted in briefs submitted to several courts – the latest in Utah–that its authority pre-empts every other authority– state and judicial. FDA’s counsel argument implies that even if the agency fails to perform its legal obligation to ensure that the public was warned about “reasonable evidence” of hazardous drug effects, manufacturers may not issue warnings on their products¹ labels without an FDA directive.

For FDA to claim that its dangerously flawed process for evaluating and monitoring the safety of drugs–leading to preventable deaths–nevertheless pre-empts all other authorities, is preposterous. The FDA is neither infallible nor given authority to exempt itself from the checks and balances of our democratic process.


Contact: Vera Hassner Sharav
veracare at ahrp org