The persistence of the placebo response in antidepressant clinical trials

https://doi.org/10.1016/j.jpsychires.2007.10.004Get rights and content

Abstract

Our objective was to assess the persistence of the placebo response during at least 12 weeks of continued placebo administration in depressed patients who have responded to 6–8 weeks of acute placebo treatment. We identified 8 placebo-controlled antidepressant trials with a total of 3063 depressed patients in which, after acute phase placebo treatment, placebo was continued for more than 12 weeks. The number of patients entering the continuation phase and percentages relapsing during this phase were determined. Based on the total number of patients entering the continuation phase 79% of placebo responders remained well (did not meet relapse criteria) during this phase compared to 93% of antidepressant responders. Although significantly more patients on placebo than on antidepressants relapsed in the continuation phase, 4 out of 5 placebo responders stayed well. The widely held belief that the placebo response in depression is short-lived appears to be based largely on intuition and perhaps wishful thinking.

Introduction

The assumption that the placebo response in depression does not endure is widely held and often stated in writing. As just one example, in a recent paper reporting the results of an uncontrolled study of vagus nerve stimulation (VNS) the authors support their claim that the sustained (for one year) improvement in some of the depressed patients treated with VNS was not likely to be a placebo response on the grounds that “the placebo effect is typically transient” (George et al., 2005). Despite the absence of data in support of this concept, the belief that the placebo effect is fleeting has achieved wide acceptance. Richard Friedman (2002), writing in the New York Times, tells us that “the placebo effect is only short-lived” he goes on to say “At best, a placebo may give the patient a temporary boost if he is mildly depressed…”

Whether or not this belief is true holds more than theoretical interest. When depressed patients undergo sustained improvement with treatment, the possibility that the placebo effect may be at play is often dismissed with the caveat that since the placebo response is temporary, the observed improvement cannot be attributed to placebo. If this belief is not true, such a conclusion would not be justified.

In clinical trials research, it is often considered unethical to maintain a patient on placebo for longer than 6–12 weeks. The assumption is that a prolonged treatment with placebo leaves a patient at unnecessary risk when established treatment methods are available. As a consequence, very few trials are conducted which continue a patient on placebo for a longer period of time.

However, the idea that the placebo response is temporary is almost entirely bereft of data. No studies have explicitly looked at what happens over the long-term to patients who experience relief of a depressive episode with placebo treatment. Data that bear on this issue within the acute phase of treatment come from a study by Quitkin et al. (1984). They found that in a 6 week study, the number of depressed patients showing improvement as measured by clinical rating scales on any one occasion did not differ between those assigned to antidepressant (AD) and those assigned to placebo.

However, when they looked at persistence of improvement, only 28% of those taking placebo maintained a late onset persistent pattern of response compared to 72% of those taking the AD. Based on this analysis, they suggested that placebo response occurs early and shows a non-persistent pattern. Although these data are often pointed to as showing that the placebo response is transient, they do not address the duration of the placebo response, once it occurs.

Also touching on this matter are findings from the large scale National Institute of Mental Health (NIMH) Collaborative Research Program (Shea et al., 1992). This group of researchers found that following a 16 week treatment phase, 76% of the depressed patients relapsed over the ensuing 18 months without active treatment. More specifically, 24% of patients with follow-up data remained free of major depression relapse including 28% of patients who had received psychotherapy (Cognitive Behavior Therapy or Interpersonal Therapy), 19% of patients who had received the AD (imipramine), and 20% who had received placebo. Interestingly, the relapse rates among those depressed patients assigned to imipramine or psychotherapy were similar to that of placebo. This study suggests that after treatment ends the placebo response endures as well (or poorly) as the response to “active” treatment.

An examination of the Food and Drug Administration (FDA) database by Khan et al. (2000) revealed that magnitude of placebo response increases with the duration of the AD clinical trial. They found that in patients assigned to placebo, as duration of the AD clinical trial increased from 4 to 8 weeks, the percent reduction in Hamilton Depression Rating Scale (HAM-D) scores increased from 24.7% to 36.1% (see Table 1). Thus, these reports do not support the assumption that the placebo response is seen early during AD trials and then disappears.

The limitation of all of these reports is that they evaluate placebo response over a short period of time (i.e. 6–8 weeks). Another study by Quitkin et al. (1993) examined the placebo response beyond the acute phase trial. This study consisted of a 6 week double-blind, placebo-controlled phase followed by 6 weeks of treatment on the same regimen. In this continuation study, of the 164 patients who entered the 6 week acute phase and were assigned to placebo, 21% (n = 32) responded to placebo and went in to continuation. Of those who entered continuation, 31.3% (n = 10) relapsed. Thus, 2 out of 3 placebo responders stayed well for an additional 6 weeks. This finding raises concerns about assumptions regarding the persistence of the placebo response over time.

To our knowledge, our review is unique in that it looks specifically at relapse in patients responding to placebo during the acute phase of treatment, who continue to receive placebo in the continuation phase of the study. In order to assess the persistence of the placebo response over an extended period of time, we searched the literature base of existing AD clinical trials. We sought randomized double-blind, placebo-controlled trials with an acute phase of 6–8 weeks followed by a double-blind continuation phase of at least 12 weeks in which the acute phase placebo control group continued to receive placebo. We hypothesized that the placebo response, once it occurs, is likely to persist beyond 12 weeks of continued treatment for the majority of depressed patients assigned to placebo.

Section snippets

Methods

In order to assess the persistence of the placebo response, we evaluated randomized, double-blind placebo-controlled trials following an extensive Medline search. The following inclusion/exclusion criteria were used: (1) randomized double-blind placebo-controlled study for an established AD, (2) acute double-blind, placebo-controlled phase lasted a minimum of 6 weeks, (3) responders to placebo, as well as ADs were allowed to continue into the continuation phase on the same treatment

Results

The total number of patients initially randomized to acute treatment across seven of the studies was 3063 with 1015 (33%) randomized to the placebo group and 1755 (66%) randomized to the AD group (Table 3). For one study (Montgomery et al., 1993), the number of patients randomized to AD or placebo in the acute phase was not reported, and therefore could not be included when calculating the overall percent of patients entering the continuation phase from the acute phase of treatment.

Table 4

Discussion

Our aim was to assess the persistence of placebo response in randomized double-blind, placebo-controlled antidepressant clinical trials in which placebo responders in the acute phase of the trial went into the continuation phase of the study. Among the depressed outpatients assigned to antidepressants, 51% responded and went on to the continuation phase. Ninety three percent of these patients did not relapse during this phase (see Table 3, Table 5).

On the other hand, of the depressed patients

References (19)

There are more references available in the full text version of this article.

Cited by (63)

  • Prospective study of nocebo effects related to symptoms of idiopathic environmental intolerance attributed to electromagnetic fields (IEI-EMF)

    2020, Environmental Research
    Citation Excerpt :

    We are not aware of any other study investigating the longitudinal development of induced nocebo effects in the context of IEI-EMF in healthy participants. While few clinical studies (coronary heart disease, Rana et al., 2005; depression, Khan et al., 2008) and one experimental study (pain perception, Colloca and Benedetti, 2006) indicate that a placebo effect can be stable across several days or even longer time periods (Kaptchuk et al., 2008, for review), we are only aware of one clinical study investigating the longevity of the nocebo effect. That study demonstrated that patients taking finasteride due to benign prostatic hyperplasia reported significantly more sexual side effects up to twelve months when they previously received information that sexual side effects could occur compared to not receiving this information (Mondaini et al., 2007).

  • Differing antidepressant maintenance methodologies

    2017, Contemporary Clinical Trials
    Citation Excerpt :

    Most of the findings from these responder extension trials were reported in review articles by Zimmerman et al. [32] and Khan et al. [33]. The rate of relapse by trial responders receiving extended placebo treatment for 4–12 months was 21%–25%, which was distinctly lower than the 46% placebo rate in the placebo substitution (6 to 12 month) trials reported in Table 1, and was evidence that a placebo response has a sizable degree of persistence [33]. The relapse rate for ADM responders in these extended (4 to 12 month) responder trials averaged 7%–8% (Table 2) compared to the 23% rate for continuing ADM patients in the placebo substitution (6–12 month) trials listed on Table 1.

  • Psychiatric Disorders

    2017, Clinical and Translational Science: Principles of Human Research: Second Edition
  • Predicting placebo response in adolescents with major depressive disorder: The Adolescent Placebo Impact Composite Score (APICS)

    2015, Journal of Psychiatric Research
    Citation Excerpt :

    In clinical practice, depressed youth who are highly likely to respond to placebo could be offered a psychotherapeutic intervention such as cognitive behavioral therapy as a first line treatment strategy (Birmaher et al., 2007; Cheung et al., 2007; Goodyer et al., 2007; Renaud et al., 1998; Rutherford and Roose, 2013). Kahn et al. (2008), in examining the persistence of placebo response in antidepressant clinical trials, found that 75% of placebo responders stay well during the continuation phase of treatment. The perspective here is that being assigned to placebo in an antidepressant clinical trial represents an intensive form of clinical management, since it entails greater contact with health care staff, which has many supportive and therapeutic components.

View all citing articles on Scopus
View full text