This a column from Dana's print publication Brain in the News
is increasingly becoming a treatment field. The past 10 years have brought
about significant therapeutic advances for diseases such as multiple sclerosis,
stroke, Parkinson’s disease, migraine, epilepsy, schizophrenia, and depression.
To be sure, there are others such as Alzheimer’s disease, brain tumors, and
spinal cord injury that have held out, and we are no further ahead than we were
10 years ago. But even in these diseases, there are significant leads that are
being, or on the verge of being, tested.
evaluation involves clinical trials in which a new approach—usually a
medication, but sometimes a procedure—is compared with existing
approaches. But what if there are no
existing therapies? Or the existing therapies are only partially successful? In
these cases the new therapy is compared to a “placebo.” A placebo is an inert
substance or sham procedure that should have no therapeutic effects. What makes things complicated is that in some
cases there is a positive, the so-called “placebo response,” or less commonly a
negative “nocebo” response. In the New England Journal of Medicine there is a lengthy discussion of the placebo response by doctors Ted Kaptchuk and Franklin Miller
like to confine my remarks to a common situation: where there is no effective
therapy or only partially effective therapies. This situation requires a different
approach: a new drug. In the simplest clinical trial, researchers divide the
group of subjects (patients) randomly into two groups. In this way factors such
a gender, age, ethnicity, and education should be cancelled out between the
groups. One group gets the “real” drug and another gets a placebo. The
researchers select some outcome measures—indications that there is some
positive effect on the disease, such as rate of progression, evidence of
recovery, survival, or quality of life. In addition, they select the duration
for the trial (six months, one year, or beyond). The trial is “double-blinded,”
meaning that the subjects don’t know which group they are in, and the investigators,
examining the subjects at intervals, don’t know either. At the end of the trial
they break the code and compare the two groups.
results should be straightforward: If the people getting the drug do better
than the placebo group, researchers move to the next step to get the therapy
approved. It is not always that simple. Sometimes some of the participants
receiving the placebo show a positive response on at least a few of the chosen
measures. Now what do you do?
solution is to kick the placebo responders out of the study. However, the
research consultant probably won’t allow it. These studies are usually set up
on an “intent to treat” basis. Once a patient is enrolled, he or she stays in.
Researchers can’t change the criteria for being included after the study has
Could the placebo
response have been avoided in the first place? Perhaps—not everyone is a
placebo responder. In most studies it is 20 to 30 percent of the participants.
It is particularly an issue in studies involving people with problems involving
pain, as in back pain or migraine. A colleague of mine, Tony Ho, used to work
at Johns Hopkins and now is in the pharmaceutical industry. Tony was leading a
study of a new agent for migraine and in preliminary studies was having a
serious problem with placebo responders—at least 30 percent in the placebo
group were showing a positive response. He came up with a unique solution. He
started his trial as described above, but put both groups on the placebo to
start. After a month, he then restarted his trial, excluding the placebo
responders, and performed a new randomization. The result was a much lower
placebo response in the final trial.
the mechanism of this placebo response? It is a complex subject outlined by doctors
Kaptchuk and Miller in their article, and varies between studies. I will
discuss it in more detail in an upcoming column.