CrossFit | Sometimes a Placebo Is not a Placebo

Sometimes a Placebo Is not a Placebo

ByMaryanne Demasi, Ph.D.February 15, 2020

Placebos are used in clinical trials to demonstrate that an experimental drug is superior to the control or “inactive” pill (1). A placebo is usually defined as an “inert substance” (no effect), given to trial participants with the aim of making it impossible for them, and usually the researchers themselves, to know who is receiving an active or inactive therapy.

The exact contents of a placebo pill are often unknown; the “recipe” is not disclosed to the trial subjects, nor is it published in the peer-reviewed literature. Recently, the editor-in-chief of Clinical Therapeutics, Dr. Robert Shader, raised concerns when a 2017 study published in the New England Journal of Medicine injected one group of people with a monoclonal antibody (ocrelizumab) and the other group with a “matching placebo.”

But what was in the placebo? “Was it saline? Was it the same vehicle in which the monoclonal antibody was dissolved?” Shader rightly questioned. Researchers should provide the public with the exact ingredients contained within a placebo, but this is rarely the case.

Placebos can contain “excipients” such as chemicals, dyes, or allergens, which might unintentionally bring about symptoms in trial participants. For example, a study found 68% of people taking a cholesterol-lowering drug (cholestyramine) experienced gastrointestinal side effects (e.g., heartburn, abdominal pain, nausea, and vomiting). But then, so did 43% of people taking placebo. The study does not describe what was in the placebo pill, although, with such a high rate of side effects, it was unlikely to have been “inert.”

In trials of the human papilloma virus (HPV) vaccine, participants were told they were either receiving a “vaccine or placebo.” The vaccine manufacturer defines a placebo as an “inactive pill, liquid, or powder that has no treatment value.”

However, participants in the placebo group did not receive an inactive substance of no treatment value. “Instead,” RIAT researchers state in the BMJ, “they received an injection containing amorphous aluminium hydroxyphosphate (AAHS), a proprietary adjuvant system used in the Gardasil vaccine to boost immune response.”

Aluminium hydroxyphosphate (AAHS) structure

This raises two serious issues. First, the placebo is not inert, so the choice of adjuvant (AAHS) in the so-called “placebo control” complicates the interpretation of efficacy and safety results in trials, especially because AAHS is reported to have a harm profile.

Second, it challenges whether the researchers obtained informed consent from the participants in the clinical trial. The concept of informed consent is embedded in the principles of the Nuremberg Code, the Declaration of Helsinki, and the Belmont Report to protect rights and welfare of human research subjects.

“If trial participants were told they could receive ‘placebo’ — widely defined as an ‘inactive’ or ‘inert’ substance — without being informed of all non-inert contents of the control arm injection, this raises ethical questions about trial conduct as well,” say RIAT researchers analyzing the trial documents.

Drug regulators such as the European Medicines Agency (EMA), Health Canada, or the U.S. Food and Drug Administration (FDA) have access to this information, and for a long while, the information was protected as commercially confidential.

But changes in policy surrounding transparency have improved since the EMA, in 2010, decided it would allow independent investigators to obtain access to regulatory documents and other clinical data.

The process of obtaining regulatory documents, however, is by no means straightforward. In fact, it is often complicated and time consuming. I have made multiple appeals to a European drug regulator (Medicines Evaluation Board) to obtain information (Certificate of Analysis) regarding the ingredients of a placebo used in a controversial statin study (JUPITER trial), but so far, they have fallen on deaf ears. So, too, have my requests to the trial’s lead investigator, Dr. Paul Ridker.

Medical journals will need to take responsibility and insist that published papers report on the methodological details of “inactive” placebos. Recently, Shader of Clinical Therapeutics stated, “It will no longer be sufficient to simply indicate that a placebo was used.”

“We will require that a full description of any placebo or matched control used in a clinical trial be given in the Methods section. This means that color; type (capsule or pill or liquid); contents (e.g, lactose), including dyes; taste (if there is any); and packaging (e.g, double-dummy) must be noted,” he stated. “We are instituting this change as part of our ongoing effort to facilitate replication of findings from trials. All too often this valuable information is omitted from published trial results.”

Failing to accurately match a placebo to the experimental drug can result in underreporting of harms or misleading trial outcomes. Not only is it important to foster public confidence in therapies that rely on comprehensive and independent assessments, it is an ethical and legal requirement to provide fully informed consent for research involving human participants.

Additional Reading

Dr. Maryanne Demasi is a well-known investigative journalist whose work on scientific documentaries has been praised by the National Press Club of Australia for exhibiting “excellence in health journalism.”

Demasi earned a Ph.D. in rheumatology from the University of Adelaide in 2004. She currently works as a researcher for the Nordic Cochrane Centre.




  1. Sometimes researchers deliberately use an “active” placebo to mimic the side effects of a drug in cases where the side effects are well known (e.g., dry mouth with antidepressants).

Comments on Sometimes a Placebo Is not a Placebo


Comment thread URL copied!
Back to 200216
Philip Tempest
March 9th, 2020 at 4:45 pm
Commented on: Sometimes a Placebo Is not a Placebo

Regarding your interest in the placebo use in JUPITER trial, have you made a request to AstraZeneca?  This is more likely to be successful than regulators or investigators since these will be under some sort of confidentiality agreement.  In my experience pharma companies will respond positively to requests from Cochrane.   No doubt you are aware that the rosuvastatin and placebo formulation numbers are respectively F12673 and F12832.

Comment URL copied!
Keith Crossfit
February 27th, 2020 at 1:37 pm
Commented on: Sometimes a Placebo Is not a Placebo

Currently, most vaccine trials use an adjuvant as the placebo, but not all, since some vaccines don’t contain an adjuvant. So they use another vaccine as the comparator.

The FDA issued guidelines years ago about using non-inert or “active” placebos in clinical trials. The guidelines were initially developed for “regular” drug trials and were done at the request of the medical community. Physicians were unable to determine whether a new drug was better than an old drug when the new drug had only been compared to an inert placebo during the clinical trials. What they wanted was for drug makers to do head-to-head comparative efficacy and safety trials. Thus, the FDA changed the requirements for clinical trials and this ushered in the era of using “active” placebos.

Under these guidelines, vaccine manufacturers were allowed to use vaccine antigen+adjuvant versus adjuvant trials. This permits them to assess antibody production but, of course, it obscures any safety issues solely due to the adjuvant which are, by design, bio-reactive. This applies to active placebos in regular drug trials, as well.

Comment URL copied!
Emily Kaplan
February 26th, 2020 at 2:36 am
Commented on: Sometimes a Placebo Is not a Placebo

This is fascinating and deeply upsetting. You have to wonder how any scientists designing a study could claim ignorance on this massive effect. Thank you for sharing!

Comment URL copied!
Malcolm Kendrick
February 17th, 2020 at 8:12 am
Commented on: Sometimes a Placebo Is not a Placebo

Nice succinct article MaryAnne. My main issue with the 'non-inert' placebo is that it helps to drive the Nocebo arguement. The statin trials have found virtually identical adverse effects between the statin and the placebo, and stated that adverse effects found in observational studies are due to the nocebo effect (you think you are going to get an adverse effect, so you get one). However, adverse effects in the statin trials have ranged from 6%, up to 80%. 6% in WOSCOPS, 80% in METEOR. It is hard to believe that a placebo can cause 6% adverse effect in one trial and 80% in another?

Comment URL copied!
Maryanne Demasi
February 18th, 2020 at 6:46 am

Yep, I agree!

Comment URL copied!
David Cumming
February 17th, 2020 at 1:37 am
Commented on: Sometimes a Placebo Is not a Placebo

Hi, Dr. Demasi. I liked your article, and I have a few questions. It seems clear that knowing what's in the Placebo is essential, and the ingredients should be published. As I clicked through to some of the links in the article, the discussion of the topic deepened and I got confused. Particularly the link referenced in the Notes regarding how "active" placebos are selected to mimic the side effects of the drug under investigation. The article says that this is to prevent "unblinding" the study. I can't tell from your essay what you are objecting to specifically. Is it the fact that the Placebo is a secret, the fact that the Placebo deliberately mimics the side effects of the drug? In the abstract, the author argues that "active" placebos may be advantageous and should be used more often in certain situations. Both? Do you believe this is incorrect and that "active" placebos should not be used?

Please forgive the remedial nature of my questions; I don't have any medical training, just an interest in understanding. Thank you for contributing to CrossFit Health. I learn tons from it!

Thank you!

Comment URL copied!
Maryanne Demasi
February 17th, 2020 at 2:34 am

Thanks David. Correct, sometimes the ingredients of a placebo are not disclosed and that concerns researchers, journal editors etc. If its not disclosed to trial subjects, then you have an issue of whether they gave "informed consent" to the experiment. On the other hand, sometimes you need an 'active placebo' to mimic the known side effects of a drug so that people in the control group can't work out whether they are on the drug or placebo (see the note). This article is more about an issue of disclosure rather than whether a placebo is active or not.

Comment URL copied!
Zdb Crossfit
February 16th, 2020 at 4:49 am
Commented on: Sometimes a Placebo Is not a Placebo

Thank you. Please keep doing this work and publishing these types of articles. Thank you much.

Comment URL copied!