Placebos – the inert substances taken by control groups in clinical trials – are often assumed to be harmless sugar pills or something along those lines. New research has found that actually it’s impossible to know what’s in placebos because there’s precious little documentation of what exactly is used in clinical trials.
Out of 176 research studies published in four of the biggest international medical journals, only one in five fully disclosed the composition of the placebo treatment. This lack of transparency suggests that all sorts of things could be being used, some of which might be having some sort of physiological effect and compromising the validity of findings on the study drug.
Placebo controlled clinical trials investigate the effects of a particular drug on a disease by comparing people who receive the treatment against patients receiving a placebo, which looks, smells, and tastes the same as the study drug but has no active ingredients. This design accounts for the placebo effect – the possibility that people in a trial may experience a health benefit because they’re taking a drug, not because the study drug they’re taking is effective.
In this study, the authors searched for randomised, placebo controlled trials published from January 2008 to December 2009 in four top medical journals – New England Journal of Medicine, JAMA, The Lancet, and Annals of Internal Medicine. A total of 176 trials were eligible for inclusion in the study – six studies of placebo pills, 65 studies of placebo injections, and 25 studies of other treatment methods (for example, nasal spray).
Only 40 (23%) of the 176 trials studied fully disclosed the composition of the placebo treatment, and 120 (68%) did not disclose any information on the placebo at all. The remainder partially disclosed what was in the placebo treatment.
Less than one in 10 (9.3%) studies that used pills disclosed the placebo, compared with 33.8% of studies that used injections and 40.0% that assessed other treatments. When papers that referred to previous publications for their primary findings or study design were excluded, these figures fell to 8.24%, 26.3%, and 27.8%, respectively.
By not paying attention to what’s in the placebo, researchers could be burying cases where the placebo has some sort of effect that’s similar to the effect of the study drug. The authors cite the example of trials of cholesterol lowering drugs that use olive oil and corn oil as the placebo. The monounsaturated and polyunsaturated fatty acids of these “placebos,” and their antioxidant and anti-inflammatory effects, could potentially reduce lipid levels and heart disease, just like the study drug, causing researchers to underestimate the effect of the cholesterol lowering drug.
“A positive or negative effect of the placebo can lead to the misleading appearance of a negative or positive effect of the drug,” author Beatrice Golomb, associate professor of medicine at the University of California, San Diego School of Medicine, told Science Daily. “And an effect in the same direction as the drug can lead a true effect of the drug to be lost. These concerns aren’t just theoretical. Where the composition has been disclosed, the ingredients of the placebo have in some instances had a likely impact on the result of the study – in either direction (obscuring a real effect, or creating a spurious one). In the cases we know about, this is not because of any willful manipulation, but because it can in fact be difficult to come up with a placebo that does not have some kind of problem.”
The authors highlight what a huge effect this lack of transparency regarding placebos could have on medical research. “Because inferences from clinical trials propagate to clinical practice, failure to report placebo composition compromises the foundation on which medical decisions are based, and on which the fate of lives may rest,” they write.
————————————————————————————————-
Golomb BA et al. (2010) What’s in placebos: who knows? Analysis of randomized, controlled trials. Annals of Internal Medicine 153 (8): 532-5 PMID: 20956710