Time to abandon placebo control in pivotal phase III trials?

Randomized controlled trials are the best way of testing therapeutic products to determine if they perform as expected and actually make a difference in treating a specific disease. Once preliminary evidence from phase II studies reveals that a treatment is probably effective, phase III trials are carried out to fully examine the risk/benefit profile of an experimental drug in a broader population over a longer period of time.

Hundreds of potential therapies are generated in laboratories, but only very few survive the early development stage and reach the point of human testing (1). There are many obstacles to the development of new treatments and the production of reliable evidence. These include the length of time and high financial cost involved in conducting clinical trials, the regulatory requirements for studies involving human subjects, and the difficulties in recruiting the appropriate patient population. Antidepressant trials are a good example of how difficult is to innovate in psychopharmacology, and Khan and Brown (2) discuss this issue focusing on the place of placebo-controlled studies in major depression.

A truly novel antidepressant has not been introduced for 30 years. Among multiple potential explanations, the positive impact that selective serotonin reuptake inhibitors (SSRIs) have had on clinicians and researchers may have played a key role. Notwithstanding the success of the SSRIs, however, pharmacological treatment for depression remains far from being optimal. Key challenges in depression research include the lack of objective markers for diagnosing depression (which is still largely based on subjective evaluation even in DSM-5) and the fact that trials of antidepressants are not primarily focused on answering the most important clinical issues (i.e., comparative effectiveness between interventions, long-term outcomes).

In terms of study design, the use of placebo is probably the most compelling issue (3). The majority of clinical trials in depression are placebo-controlled, because regulators require them for licensing approval (4). The rate of placebo responders in these trials, however, has added a layer of complexity and difficulty to the process of designing trials and interpreting results. Placebo responders in antidepressant trials have been increasing over time since 1981 (5). How can this be explained? Either participants are becoming more suggestible and/or placebo more effective, or the increase in placebo response must be artefactual.

In fact, the increasing placebo response rate can at least partially be explained by the phenomenon called “inflation of baseline severity” (6). Entry criteria for clinical trial participants are based not only on the categorical diagnosis, but also on the severity of the illness. Usually, the minimum criterion for enrolling an individual in an antidepressant trial is a diagnosis of major depressive disorder and a total score greater than a pre-specified threshold according to a standardized rating scale. These measures of severity of depressive symptoms, even if rated by trained assessors, are subjective and can be easily unconsciously manipulated. Researchers always struggle to find participants who meet the entry criteria, are eligible for randomization and are willing to accept randomization in a trial in which they know they may receive placebo. To recruit to time and target, investigators may tend to overemphasize some symptoms and give a higher score to some of the items to reach the minimum overall score on the rating scale and get the patient into the study. As a result, bias is introduced in the selection of participants (skewed distribution) and too mildly ill patients are enrolled who are more likely to remit “spontaneously”, which means to respond to placebo. After their high initial ratings, in fact, physicians begin to rate the condition of patients more accurately. The main consequence of inflation of baseline severity is a large drop in the severity scores between randomization and endpoint, also in the mildly ill patients who remit “spontaneously” without receiving any active treatment, thus making the placebo appear more effective.

Innovation in psychopharmacology is urgently needed not only in terms of drug discovery and development, but also in terms of the design of phase III clinical trials. The need for comparative effectiveness has been reported and highlighted many times recently (7). Therefore, the key question to ask is whether we still routinely need to have placebo-controlled phase III studies. It has been suggested that they are needed in the field of depression because findings of equivalence between a new drug and standard treatment are not evidence of efficacy unless the new drug is also significantly more effective than placebo (8). This assay sensitivity may be required in phase II studies, but in phase III one could argue that the essential question is whether a new treatment is superior to existing therapies.

We now know that some antidepressants are better than others and that individual drugs can be ranked according to their efficacy and acceptability profiles (9). Comparative effectiveness research is a key element of current efforts in health care reform in Europe and also in the U.S. (10). Prioritizing this kind of research in the field of antidepressants presents a methodological challenge. To improve treatments and patient outcomes, we need phase III trials with a superiority design against an active comparator, chosen among the most effective and better tolerated treatments already available on the market via a reliable and transparent meta-analytical process. This will set a more ambitious and clinically meaningful target, and will foster much-needed innovation in psychopharmacology to developing more effective interventions. It will also be in line with the ethical requirements of not exposing patients to placebo when an effective treatment is available.

Despite the importance of creating new therapies, investment in neuroscience is no longer a top priority for pharmaceutical industry. Drug discovery in psychiatry is just too difficult for current commercial models. It is time for academics to join forces with industry, creating new models of drug discovery (11) and driving innovation in the methodology of drug development. Abandoning the standard requirement for placebo control in pivotal phase III trials, and setting more ambitious targets for treatment advance, should be a bold first step.