The use of placebo in clinical trials on bipolar disorder: a new approach for an old debate

Abstract

Background: Placebo response is high in bipolar disorder trials. This is a challenge for researchers and an enigma for clinicians. This article will attempt to aid in the comprehension of the ethical questions and methodologies related to the use of placebo in bipolar disorder.

Methods: A systematic review of the main computerized data bases (Medline, Embase, Psychlit, Current Contents) was carried out, using the following key words: placebo, bipolar disorder, mania, depression, clinical trial.

Results: Placebo has a very important role in the study of bipolar disorders. From a methodological point of view, and given the limitations of non-inferiority designs, the use of placebo is necessary to evaluate the sensitivity of the study and ensure that the drug under investigation is really effective. For this reason, several procedures may be used to satisfactorily solve the ethical objections. Placebo response in bipolar disorder trials is more likely to occur in patients who are mildly ill, bipolar II, mixed-episode, first-episode, rapid cycling, atypical, non-psychotic, substance abusers and medically ill. The use of concomitant medication such as benzodiazepines, a high frequency of visits, a high number of treatment groups and sites, fixed-dose designs, and the concomitant use of psychotherapy are likely to increase placebo response.

Conclusion: The subject of the use of placebo in psychiatric clinical trials leads to very relevant ethical, methodological and clinical dilemmas. Many reasons support the use of placebo in acute bipolar studies, whereas in maintenance the length of the treatment with placebo makes the decision more difficult. Ethical concerns can be addressed through operational definitions of outcome measures, add-on and discontinuation designs, early escape strategies, and processes designed to make sure that fully informed consent was fairly obtained. There are a number of clinical and methodological variables that are associated with placebo response.