The Large Placebo Effect in Cough Presents Major Challenges for Antitussive Trials

There are a number of clinical trials going on to test treatments for chronic cough, but they have a big challenge: cough clinical trials have a very big placebo effect.

During a session on the placebo effect in antitussive clinical trials, Ron Eccles, BSc, PhD, DSc, emeritus professor at Cardiff University, explained the issue, why it is such a problem, and some strategies to get around the placebo effect.

“Cough has a very big placebo effect, because it's under voluntary control,” he explained. “And coughing is driven by a sensation and the urge to cough.”

The placebo effect occurs because of the anticipation or expectancy of a treatment having a big effect, but it can also be influenced by the taste, smell, or color of a treatment.

A blinded study should be balanced, meaning that the placebo effects on each side are balanced. Patients on the active treatment are trying to guess what they’re on, which influences the placebo effect, but their guesses should be no better than chance: 50/50. The patients who guess they are on active treatment will anticipate the benefit of the treatment, which gives an enhanced placebo effect, Eccles explained. Patients in the placebo group also have a 50/50 chance of guessing correctly what they’re on.

When the placebo effect is subtracted from the active medicine, the pharmacological effect is left. This principle of additivity is important for clinical trials.

“Because if additivity is not working, if it is invalid, then it is impossible, I say impossible, to properly determine the efficacy of a medicine,” Eccles said.

Side effects can unblind, unbalance, and confound a clinical trial, which is an issue in all clinical trials, not just antitussive trials. Without effective blinding, it is unclear if the difference between the active and placebo is real or a manifestation of patients knowing what treatment they’re receiving.

In addition, side effects can indicate to a patient that they are on an active medicine, which in turn can amplify the placebo effect. The anticipation of a benefit may increase the effect of the medication if patients are experiencing side effects and correctly guess they are on the active treatment.

This can be a problem if a medicine has little or no pharmacological effect, but the side effects make patients anticipate a benefit and it seems there is efficacy when there is none, Eccles explained.

There are some ways to improve trials to reduce or remove the placebo effect.

Change the shared participant information. If the information at the start of the trial outlines the side effects, patients not experiencing them will know they are on the placebo. Eccles suggests informing the participants that drug side effects are also commonly associated with placebo treatments and the presence of side effects does not necessarily mean the participant is receiving the active treatment. “This will help in some ways to blind the study,” he said.

Another change could be to only inform participants about serious side effects that may cause permanent harm and not inform them about side effects that may terminate once the trial is finished. However, there are ethical and legal issues to such changes in participant information.

Active placebos. This has been tried with psychoactive medicines. If the active medicine causes dry mouth, the researchers might use a placebo such as atropine, which causes dry mouth, to blind the study.

Crossover design. For medicines that are effective, but maybe not very effective, the crossover design may be useful. This allows patients to compare the placebo and the active medicine, but patients are unblinded to some effect. This is not a useful design for medicines that have side effects because the side effects will only amplify the placebo effect during the crossover.

Patient selection. The researchers may choose to include only patients who have a high cough severity and are less likely to be able to control their cough. There’s a proposal to eliminate placebo responders in order to favor the active treatment, although this is not easy to do.

Finally, the researchers could avoid patients who have participated in similar clinical trials, because these may be unblinded. It’s not uncommon for researchers running a trial to try and recruit patients who have been used in the past, because they are easy to get a hold of, Eccles explained.

However, “they may have a lot of history and memories about previous clinical trials. And this often confounds studies,” he said. But he admitted it’s very difficult to recruit totally naïve patients to clinical trials.

All this means that trials for new antitussive therapies are likely to be unblinded due to side effects, and most of the investigators have said the new trials may be unblinded, Eccles said.

“This is fundamental issue in these clinical trials,” he said. “If there is unblinding, then these trials are invalid, as far as regarding measurement of efficacy. Lack of information about blinding in these studies makes it impossible—and I really reiterate, impossible—to determine the efficacy of new antitussives that have significant side effects.”