Placebo effect size could be used as a benchmark in TRD trials

Treatment-resistant depression (TRD) was associated with high placebo response rates and research used variable definitions of TRD, a meta-analysis of randomized controlled trials (RCTs) found.

The pooled placebo effect size for all treatment modalities was large (g=1.05, 95% CI, 0.91-1.1) and placebo effect size in RCTs of specific modalities did not significantly differ.

This finding “may help to interpret the results of past and future RCTs,” wrote Zafiris Daskalakis, MD, PhD, of the University of California San Diego, and co-authors in JAMA Network Open. “Researchers who conduct clinical trials may now compare their results with a benchmark for expected placebo effect in TRD.”

“For instance, a placebo-controlled RCT that reports negative findings but had a placebo effect size greater than g=1.05 could be interpreted as a false-negative; conversely, an RCT that reports positive findings with a placebo effect size less than g=1.05 could be a false-positive,” Daskalakis and colleagues pointed out.

“Our results also provide a context for noninferiority trials that do not use a placebo or sham control intervention,” they continued. “Treatments that differentiate from this benchmark (g=1.05) by a predetermined margin would be expected to be superior to placebo.”

The group used literature published through June 2021, screening trials with 10 or more participants that had recruited patients with TRD according to a common, but not universal, definition as “no response to two separate antidepressant trials of adequate dose and duration.”

Included trial participants were randomized to a placebo/sham arm versus a pharmacotherapy, brain stimulation, or psychotherapy arm, though no studies of psychotherapy met inclusion criteria for the meta-analysis. Placebo arm was defined as an inert pill, liquid, parenteral placebo, or sham device or therapy (such as repetitive transcranial magnetic stimulation, transcranial direct current stimulation, or invasive brain stimulation) that did not include any theoretical active property to induce the proposed therapeutic effect.

The researchers noted that prior trials of repetitive transcranial magnetic stimulation and escitalopram (Lexapro) have suggested that TRD is associated with smaller placebo effect than non-TRD. “To our knowledge, no analysis has assessed the placebo effect in patients with TRD receiving other treatment modalities,” they wrote.

Trials with wait list and treatment-as-usual placebos were not included, but those with an open-label phase before randomization were allowed. Fifty RCTs were ultimately included in the meta-analysis (n=3,228; mean age 45.8; 55% women).

Effect sizes for each treatment modality were:

  • Pill placebo g=1.14, 95% CI 0.99-1.30.
  • Parenteral placebo g=1.33, 95% CI 0.63-2.04.
  • Liquid placebo g=0.45, 95% CI −0.26 to 1.15.
  • Repetitive transcranial magnetic stimulation g=0.89, 95% CI 0.63-1.15.
  • Transcranial direct current stimulation g=1.32, 95% CI 0.53-2.11.
  • Invasive brain stimulation g=0.86, 95% CI 0.58-1.14.

Response and remission rates associated with placebo were comparable across modalities. Variables associated with a larger placebo effect size were open-label prospective treatment before double-blind placebo randomization (β=0.35, 95% CI 0.11-0.59, P=0.004), later year of publication (β=0.03, 95% CI 0.003-0.05, P=0.03), and industry-sponsored trials (β=0.34, 95% CI 0.09-0.58, P=0.007).

In addition, the number of unsuccessful treatment interventions was associated with a smaller placebo effect size (β= −0.12, 95% CI −0.23 to −0.01, P=0.03).

In an accompanying editorial, Cristina Cusin, MD, of Massachusetts General Hospital in Boston and Luana Colloca, MD, PhD, of the University of Maryland in Baltimore, agreed that the present meta-analysis “sets a benchmark for future clinical trials in TRD regarding the expected placebo response rate,” and noted that the overall rate of positive response of TRD to placebo was 23.5%, and the remission rate was 15.5%.

These are “substantially lower than the average response rate to placebo in major depressive disorder trials (35-40%),” Cusin and Collaca added. “Still, the reported overall placebo or sham response rate in TRD may be greater than what many clinicians would expect based on their own practices, which may reflect differences between clinical trials and real-world outcomes.”

A 2016 review of pharmacologic antidepressants in non-TRD depression found that average placebo response rates were relatively constant at 35-40% from 1991-2015, and that length of the study and multicenter design influenced placebo response rates. A 2009 meta-analysis considered data from 2002 to 2008 for repetitive transcranial magnetic stimulation and escitalopram RCTs and concluded that placebo response was large for both interventions.

There are many reasons clinical trials may fail to demonstrate benefit even for potentially efficacious interventions, Cusin and Colloca observed. “Among the greatest challenges is a high placebo response rate,” they wrote. “The placebo phenomenon per se encompasses expectations (possibly higher in a novel treatment or with a more invasive procedure), relationship between clinician and patient, and other contextual or environmental factors.”

Limitations of the meta-analysis include the lack of psychotherapy trials meeting inclusion criteria and lack of a consistently used TRD definition.

“There is no standard TRD definition,” Daskalakis and co-authors wrote. “A large proportion of studies that were excluded reported including patients with TRD, even though they defined TRD based on only one failed antidepressant trial. For this reason, we were unable to include two large industry-sponsored trials for repetitive transcranial magnetic stimulation.”

“Future research would benefit from an improved and consistent definition of TRD,” they advised.

  1. Treatment-resistant depression (TRD) is associated with high placebo response rates and research uses inconsistent definitions of TRD, a meta-analysis found.

  2. Response and remission rates associated with placebo were comparable across treatment modalities.

Paul Smyth, MD, Contributing Writer, BreakingMED™

Daskalakis reported grants from Brainsway Inc, nonfinancial support from Magventure Inc, and nonfinancial support from Brainsway during the conduct of the study.

Cusin reported personal fees from Janssen, Perception, and Takeda; and grants from Clexio, Livanova, AFSP, and the National Institute of Mental Health outside the submitted work. Colloca had no disclosures to report.

Cat ID: 55

Topic ID: 87,55,730,192,55,925