In recent decades, the risk seems to have increased for new antidepressive compounds to fail to obtain approval from the Food and Drug Administration (FDA) or European Medicines Agency (EMA). One might wonder why this is the case in times of more sophisticated trial methodology, which could theoretically be expected to increase the likelihood of obtaining valid results. However, improving the validity of results does not necessarily increase the likelihood of obtaining positive results and may actually have the opposite effect, that is, more valid results could also mean more negative results.
If one wants to better understand the complex background of why antidepressants fail to obtain approval, one has to consider several hypothetical scenarios:
Some or all new antidepressive compounds might simply not be as effective as approved antidepressants (ADs) and consequently may not be able to demonstrate superiority over placebo as ‘easily’ as approved ADs did.
The efficacy of ADs in general is so low that it might be difficult in principle to show better efficacy than placebo in randomized controlled trials (RCTs).
The efficacy of new antidepressive compounds may be difficult to prove in RCTs because the number of participants showing a placebo response may have increased in recent decades or other design issues may present obstacles.
Studies may prove efficacy, but the effect size may be so small that there is doubt as to its clinical relevance.
In addition to these potential efficacy-related explanations, tolerability- or safety-related issues may be the reason for failure to gain approval; also, the risk-benefit ratio may be deemed to be unacceptable. These aspects seem to be relatively rare among ADs, but they can sometimes be the reason why antipsychotics are not approved (mostly as add-on treatment) for the indication major depressive disorder (MDD). Finally, ADs may fail to be approved for reasons related to the pharmacological characteristics of the compound.
Among the above aspects, the increased difficulty of proving efficacy and the associated risk of negative and failed studies appear to be the key points to be discussed here as reasons why new ADs are failing to gain approval. As elucidated below, additional study design issues and in particular the problems of the placebo effect and heterogeneity of AD studies are relevant in this context.
Regulatory authorities demand as the most pragmatic and adequate proof of efficacy that an experimental AD demonstrates superiority over placebo in double-blind placebo-controlled studies. The EMA also asks for a third arm with a standard AD as a validator or comparator (i.e. the experimental AD versus a standard AD versus placebo) [Citation1]. However, such a recommendation might result in higher responses in the placebo group, perhaps caused by participants having higher expectations that they will be in an active drug group [Citation2,Citation3].
Each of the components in the design of a double-blind placebo-controlled study of an AD is well known to be a potential source of problems and errors, which in turn can affect the validity and robustness of the study and its ability to differentiate between the active compound and placebo. However, even well-designed studies are quite likely to be negative studies or, in the case of 3-arm studies, failed studies, in which neither the experimental nor the standard AD is statistically significantly superior to placebo.
In general, placebo response rates are high in drug studies, including studies of ADs [Citation4,Citation5]. Overall, in recent decades, about one-third of placebo-controlled double-blind RCTs of ADs were positive, about one-third negative and about one-third failed studies [Citation6,Citation7], meaning that the proportion of negative/failed studies is high. In addition, the difference between ADs and placebo is small in terms of differences in both pre-post scores and remitter rates; while in earlier years the difference in remitter rates was about 20%, in the recent past it was only about 15% at the most [Citation7]. The discussion about the meta-analysis by Kirsch [Citation8,Citation9], who was overly critical in his claim that ADs generally act on the placebo level, underlines the idea that the window of opportunity to demonstrate superiority over placebo is indeed very small, given a placebo-verum difference of about 2–3 points in the change on the Hamilton Depression Rating Scale (HAMD).
Several methodological problems are hidden behind these unsatisfactory results, including the following: insufficient reliability in making diagnoses and assessing symptom severity, especially in multicenter studies; increasing inclusion of ‘symptom carriers’ found through newspaper advertisements (especially in studies conducted in the USA); inclusion in earlier studies of too high a proportion of patients with mild depression; the level of depression being rated as too high before inclusion into the study (overrating); multi-arm studies; high number of patients receiving additional psychopharmaceuticals (e.g. hypnotics, anxiolytics); and inclusion of study centers offering too much supportive psychotherapeutic treatment. It has been discussed whether expert ratings are a problem because they tend to overestimate the effect of interventions in both the drug and placebo arms, which could lower the trial sensitivity; however, one must question whether self-ratings, with their associated problems, represent a viable alternative [Citation10].
Some authors suggest that the poor outcome of current AD studies can partially be explained by an increase in the placebo response over recent years [Citation2,Citation11–Citation13]. However, other authors have observed an increase of the response in both the experimental and placebo groups and doubt that the placebo response itself has increased [Citation8,Citation11,Citation14]. Beyond this controversy, the problem of placebo response is clearly of great relevance in current AD studies.
A randomized 5-arm study in patients with MDD provides an interesting basis for a discussion of the problem of placebo response [Citation15]. This study included two parallel parts in which (1) sertraline was compared with a pill placebo under double-blind conditions and (2) manual-guided cognitive behavioral therapy (CBT, 10 sessions) was compared with a pseudo-placebo (guided self-help group) under conditions that were not double blind but in which the outcome assessment was performed by blind raters; patients randomized to the 5th arm could choose to receive CBT or sertraline [Citation15]. The interesting, key finding was that although sertraline showed statistically significant superiority over placebo, the difference of 2.3 points in the HAMD (17 items) was small. This small difference was because of an impressive placebo response. CBT, on the other hand, demonstrated superiority over the pseudo-placebo much more clearly because there was almost no response to the pseudo-placebo. The study found no relevant difference between the mean pre-post HAMD-17 score of sertraline and that of psychotherapy. The results of this study show impressively how important the placebo response is for a positive or negative outcome of a study. All possible measures have to be taken to avoid inadvertently increasing the placebo response, for example through study design aspects.
Of great relevance in this context are the problems of multicenter studies and the involvement of study sites in different countries. The problem of heterogeneity between centers and countries is of great importance and is receiving an increasing amount of attention; it becomes apparent when comparing the results of AD studies from different regions of the world [Citation16]. Recently, the vortioxetine database provided an interesting example of regional heterogeneity [Citation17]: the data clearly indicated that the studies performed in the USA had a tendency to achieve much smaller pre-post differences in the Montgomery-Asberg Depression Rating Scale (MADRS) score between vortioxetine and placebo than the non-US studies. Reasons for this regional heterogeneity include the group of methodological problems mentioned above, that is, the inclusion of participants found through advertisements, who tend to include higher proportions of ‘symptom carriers’, patients with milder degrees of depression and patients with more chronic depression, etc. Insufficient data are available to draw conclusions about differences in the reliability of ratings and the general quality of study implementation. Of interest is whether other studies performed in regions of the world outside the USA and western Europe can differentiate more clearly between an experimental compound and placebo and, if so, whether the findings of these studies are valid.
Expert opinion
There appears to be only a narrow window of opportunity to demonstrate the efficacy of ADs in placebo-controlled studies, given a verum-placebo difference of about 15% in the remitter rate or 2–3 points in the pre-post HAMD change score differences.
Although the methodology of randomized placebo-controlled AD studies is sophisticated, the risk of negative or failed studies is quite high.
Several problems and methodological issues in the implementation of AD studies can lead to poor results of individual studies and inconsistency of a whole data set on an individual AD.
A high placebo response and heterogeneity of studies are among the most important reasons for the failure to prove efficacy of an AD to a degree that is sufficient to gain approval from regulatory authorities.
Declaration of interest
In the last three years HJ Möller has received honoraria for lectures or expert meetings from the following pharmaceutical companies: Lilly, Lundbeck, Schwabe, and Servier. The author has no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
Acknowledgment
The author thanks Jacquie Klesing, Board-certified Editor in the Life Sciences (ELS), for editing assistance with the manuscript. Ms. Klesing has no conflicts of interest to declare.
Additional information
Funding
References
- Moller HJ, Broich K. Principle standards and problems regarding proof of efficacy in clinical psychopharmacology. Eur Arch Psychiatry Clin Neurosci. 2010 Feb;260(1):3–16.
- Papakostas GI, Fava M. Does the probability of receiving placebo influence clinical trial outcome? A meta-regression of double-blind, randomized clinical trials in MDD. Eur Neuropsychopharmacol. 2009 Jan;19(1):34–40.
- Rutherford BR, Rose SA, Sneed JR, et al. Study design affects participant expectations: a survey. J Clin Psychopharmacol. 2009 Apr;29(2):179–181.
- Schedlowski M, Enck P, Rief W, et al. Neuro-bio-behavioral mechanisms of placebo and nocebo responses: implications for clinical trials and clinical practice. Pharmacol Rev. 2015 Jul;67(3):697–730.
- Leuchter AF, Hunter AM, Tartter M, et al. Role of pill-taking, expectation and therapeutic alliance in the placebo response in clinical trials for major depression. Br J Psychiatry. 2014 Dec;205(6):443–449.
- Baghai TC, Blier P, Baldwin DS, et al. General and comparative efficacy and effectiveness of antidepressants in the acute treatment of depressive disorders: a report by the WPA section of pharmacopsychiatry. Eur Arch Psychiatry Clin Neurosci. 2011 Nov;261(Suppl 3):207–245.
- Moller HJ. Antidepressants: controversies about their efficacy in depression, their effect on suicidality and their place in a complex psychiatric treatment approach. World J Biol Psychiatry. 2009;10(3):180–195.
- Fountoulakis KN, Moller HJ. Efficacy of antidepressants: a re-analysis and re-interpretation of the Kirsch data. Int J Neuropsychopharmacol. 2011 Apr;14(3):405–412.
- Kirsch I, Deacon BJ, Huedo-Medina TB, et al. Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration. Plos Med. 2008 Feb;5(2):e45.
- Moller HJ. Standardised rating scales in psychiatry: methodological basis, their possibilities and limitations and descriptions of important rating scales. World J Biol Psychiatry. 2009;10(1):6–26.
- Walsh BT, Seidman SN, Sysko R, et al. Placebo response in studies of major depression: variable, substantial, and growing. JAMA. 2002 Apr 10;287(14):1840–1847.
- Fava M. The role of regulators, investigators, and patient participants in the rise of the placebo response in major depressive disorder. World Psychiatry. 2015 Oct;14(3):307–308.
- Kasper S, Dold M. Factors contributing to the increasing placebo response in antidepressant trials. World Psychiatry. 2015 Oct;14(3):304–306.
- Furukawa TA, Cipriani A, Atkinson LZ, et al. Placebo response rates in antidepressant trials: a systematic review of published and unpublished double-blind randomised controlled studies. Lancet Psychiatry. 2016 Nov;3(11):1059–1066.
- Hegerl U, Hautzinger M, Mergl R, et al. Effects of pharmacotherapy and psychotherapy in depressed primary-care patients: a randomized, controlled trial including a patients’ choice arm. Int J Neuropsychopharmacol. 2010 Feb;13(1):31–44.
- Niklson IA, Reimitz PE. Baseline characteristics of major depressive disorder patients in clinical trials in Europe and United States: is there a transatlantic difference? J Psychiatr Res. 2001 Mar–Apr;35(2):71–81.
- Pae CU, Wang SM, Han C, et al. Vortioxetine: a meta-analysis of 12 short-term, randomized, placebo-controlled clinical trials for the treatment of major depressive disorder. J Psychiatry Neurosci. 2015 May;40(3):174–186.