The case of the bad placebo

While studies sometimes reach erroneous conclusions, researchers can help set the record straight.

When it comes to clinical research, the most powerful type of study is a randomized, double-blind, placebo-controlled trial.

But even a well-designed study can produce questionable results. The recent follow-up of a 2019 cardiovascular trial called REDUCE-IT is an example that offers a great lesson. While innovative treatments are the focus of many clinical trials like this, the choice of placebo is also crucial.

What made this a meaningful study?

In this type of study, subjects are randomly assigned to two groups: one group receives the treatment being evaluated (e.g. a new drug), while the other group receives a dummy treatment called a placebo.

Neither study participants nor researchers know who is receiving active treatment and who is receiving a placebo. This means they are both blind to the group assignment - that's why it's called double blind. Treatment allocation will be encrypted and kept secret until the end of the study or decrypted at earlier, scheduled intervals to monitor effectiveness or safety.

This reduces the likelihood that researchers' or participants' expectations will influence study results. This means that any differences in health or side effects can reasonably be attributed to the treatment – ​​or lack of treatment.

What you should know about placebo treatment

Ideally, study participants and researchers cannot tell who is receiving active treatment and who is receiving a placebo. But sometimes participants can say what they received. For example, active treatment may have a bitter taste or a noticeable side effect such as diarrhea.

In this case, the study is no longer double-blind. This means that expectations can influence results. Studies can assess this by asking participants during or after the study whether they thought they were taking active treatment or a placebo. If the answers appear random or the subjects answer “I don’t know,” blinding was successful.

While placebo treatment should have no effect, that's not always true:

• Der bekannte Placebo-Effekt ist ein positiver Effekt, der mit einer Nutzenerwartung verbunden ist: Wenn Sie jemandem sagen, dass eine Pille Schmerzen lindern kann, werden manche Menschen eine Schmerzlinderung erfahren, selbst wenn diese Pille ein Placebo war.
• Eine negative Nebenwirkung aufgrund eines Placebos wird als Nocebo-Effekt bezeichnet: Wenn Sie jemandem sagen, dass er durch die Placebo-Pille, die er einnimmt, Durchfall bekommen könnte, kann die Erwartung bei einigen Menschen dazu führen, dass dies der Fall ist. (Dasselbe Placebo, das in einer anderen Studie verwendet wird, kann Kopfschmerzen auslösen, wenn dies die Nebenwirkung ist, vor der die Studienteilnehmer gewarnt werden.)

Finally, a placebo should have no direct biological effects on the person taking it. And this is where REDUCE-IT seems to have gone wrong.

REDUCE-IT shows how important it is to carefully select a placebo

The full name of REDUCE-IT is Reduction of Cardiovascular Events With Icosapent Ethyl–Intervention Trial). It was designed to determine whether the drug icosapent ethyl could lower triglyceride levels to reduce cardiovascular diseases such as heart attack or stroke.

Triglycerides are a type of fat in the blood. High levels may increase cardiovascular risk, but experts aren't sure whether treatments to lower triglyceride levels lead to fewer heart attacks or strokes.

Triglyceride levels decreased in participants who received the active drug. Rates of cardiovascular problems, including heart attack or stroke, were a whopping 25% lower than those who received a placebo. There was even a 20% reduction in cardiovascular deaths in the treatment group.

Based on these results, the FDA approved a drug label claiming that icosapent ethyl benefits certain people at high risk of cardiovascular disease.

But questions arose shortly after the study was published in 2019. The treatment group performed better than the placebo group. However, a careful reading of the results suggested that this may have been because people in the placebo group had more heart attacks and strokes over time, not because the treatment group had fewer.

A follow-up study shows a different result

In response to these questions, the study authors conducted additional analyses. This time they examined substances in the blood, called biomarkers, that are associated with cardiovascular risk. They found little change in biomarker results in participants who received the active drug. But the biomarkers worsened in the placebo group, suggesting that the apparent benefit conferred by the drug could be due to thatNegativeEffect of the placebo!

How can a placebo worsen cardiovascular risk? One possibility is that the mineral oil placebo used in this study reduced the absorption of statin drugs that participants were taking to lower their cholesterol levels, which also affects heart and blood vessel health. Regardless, this new analysis suggests that skepticism about the original study's dramatic results was appropriate and additional studies are warranted.

The end result

For me, this story has three take-home points:

• Es gibt viele Möglichkeiten für die Forschung, zu falschen Schlussfolgerungen zu kommen; Eine unglückliche Placebo-Wahl ist ungewöhnlich, scheint aber hier wahr zu sein.
• Damit medizinische Forschung vertrauenswürdig ist, müssen Forscher bereit sein, Kritik anzunehmen, Ergebnisse neu zu bewerten und gegebenenfalls zusätzliche Analysen durchzuführen.
• Im Fall von REDUCE-IT scheint dieser Selbstkorrekturprozess funktioniert zu haben.

After the first study in 2019, there was great enthusiasm for the drug icosapent ethyl. However, after this latest analysis, that excitement may be fading. But one thing should be clear: this is not scientific indecision, as is sometimes claimed. Reassessment and correction when warranted is the way science should work.