Perhaps I was naive, but even though I have written about placebos numerous times here and elsewhere I have always assumed that placebo pills were literally sugar pills. It had not occurred to me that placebo content needs to be specifically disclosed, and that their composition might not be as standard as I had assumed.

A new study in the Annals of Internal Medicine reviews clinical trials over the last two years. They found that only 8.2% of clinical trials with a placebo pill as a control specifically disclosed the placebo content. Meanwhile 26.7% of trials involving injections and procedures disclosed the precise nature of the treatment. This difference makes sense, in that there is much more interest in the nature of a placebo treatment when it is more complex than just taking a pill.

The concept of a placebo is that it is completely physiologically inert, therefore any response to the placebo is due to other factors (other than a physiological response to an active intervention). Typically, in pharmaceutical trials, the company that makes the drug being studied will also manufacture look-alike placebos. These can contain the sugar base and any other fillers, the same coating, etc. – but not the active drug. But according to the study authors perhaps we cannot assume that these placebo pills are completely inert.

It is possible that some of the other ingredients (fillers) in the pill may have biological activity. They may produce effects or side effects that can obscure the effect of the drug they are being compared to. Or, negative side effects in the placebo may produce the illusion of a beneficial effect in the drug being studied, even if it does not work.

All of this seems like common sense to me, and therefore nothing that could have a significant physiological effect should be an undisclosed component of a placebo. The authors do not imply that there is deliberate manipulation of the placebo, but if the contents are not disclosed this cannot be ruled out. More likely there is unintentional use of a placebo that is not entirely inert. Either way, requiring full disclosure of placebo content will go a long way to addressing this issue.

I should mention that there is also the concept of an active placebo – a placebo treatment that is deliberately not inert but meant to mimic the side effects of the active treatment. The purpose of this is to prevent subjects in a clinical trial from being unblinded. Getting side effects can clue someone in to the fact that they are on the active treatment, while the absence of side effects may be a hint that someone is getting the placebo. So a placebo that does nothing but produce a mild side effect  can trick subjects in the placebo arm into thinking they are getting the active treatment. But of course, this introduces the problem of obscuring the results if the active placebo may accidentally provide a beneficial effect.

This problem is more of a factor with non-drug trials. In acupuncture trials, for example, sham acupuncture (at non-acupuncture points) or placebo acupuncture (without skin penetration) is often given as an active placebo. Of course, when these trials are negative proponents can claim that the placebo form of acupuncture was not inert and accidentally provided a clinical effect.

Conclusion

The take home message from this new study is that designing the placebo arm of a clinical trial is not always as straightforward as might be assumed. In some types of trials great attention is paid to designing the placebo treatment, but in drug trials it is often assumed that the sugar pills provided by the company performing the study are simply “placebos” without further thought being given.

I agree with the authors that it should become standard for clinical trials to disclose the exact content of the placebo treatment. This should be a trivial matter for those making the placebos. When it comes to clinical trials (as with science in general) transparency is vital.