Consensus Definition Of ‘Treatment-Resistant Depression’ For Regulatory Trials
By Deborah Borfitz
March 23, 2022 | Hundreds of clinical trials are underway testing therapies for treatment-resistant depression (TRD), yet until very recently there was not a clear definition of what is even being treated. But with newly proposed criteria to describe the condition, researchers and regulatory agencies have a basis for agreement that could help inform study design and potentially improve the lives of millions of people worldwide, according to Luca Sforzini, M.D., a psychiatrist and researcher with the Institute of Psychiatry, Psychology & Neuroscience (IoPPN) at King’s College London.
The consensus definition is contained in a report published recently in Molecular Psychiatry (DOI: 10.1038/s41380-021-01381-x) and reflects areas of agreement between over 60 mental health experts from research, industry, and regulatory bodies as well as a patient representative. The input from a person with lived experience was unique and invaluable to the exercise that used a Delphi-method-based consensus approach to define both TRD and the less-commonly studied partially responsive depression (PRD), Sforzini says.
Researchers from IoPPN and the National Institute for Health Research Maudsley Biomedical Research Centre (South London) co-led the report.
The definitions are intended to serve as operational criteria for future clinical studies, especially if conducted for regulatory purposes. They also are the first step toward one of the main objectives of the EU Patient-Centric Clinical Trial Platform (EU-PEARL), a project of the Innovative Medicines Initiative to design a protocol for platform trials of new medications for TRD/PRD, he says.
To date, the definition of TRD in published studies has been highly variable, Sforzini says. For example, clinical trials often have samples with a single antidepressant failure—a “less restrictive” approach than the consensus opinion in scientific papers that there be two antidepressant failures—simply to qualify more individuals for participation.
“If we are talking about treatment-resistant depression, it’s important that everyone means the same thing,” he continues. Study results will be more reliable, and of greater use in real-world treatment scenarios, if consensus criteria are followed. Some antidepressants being approved by regulatory agencies are specifically indicated for TRD, he notes.
The approval process itself can have a certain variability without a clear-cut definition of the condition, continues Sforzini. The definitions for TRD and PRD used by the U.S. Food and Drug Administration (FDA), for example, are “not exactly the same” as the ones adopted by the European Medicines Agency.
TRD affects more than 75 million (up to 30%) of the estimated 230 million people worldwide with major depressive disorder and, according to the World Health Organization, is one of the leading causes of disability. The most recent FDA-approved drug for its treatment is the oral antidepressant Spravato (esketamine).
As with many clinical studies for TRD, the trials for Spravato included people with major depressive disorder who had not responded to at least two antidepressant medications administered at an adequate dose and duration—which definitionally aligns with a key recommendation in the newly published consensus document, Sforzini says. But in most studies on TRD, these “classic criteria” have not been used.
In intervention trials, TRD has been most often defined as a minimum of only one previous failed treatment. Additionally, only one in five studies enroll participants who meet the additional criteria of adequate dose and duration of medications, and definitions of response, non-response and partial response are not univocal, Sforzini says.
The new guideline offers multiple specific recommendations for TRD/PRD regulatory clinical trials, grouped by the level of consensus (strong, moderate, or weak) reached. Among the recommendations with a strong level of consensus (95% to 100% agreement) are that a definition of TRD—a minimum of two failed treatments with 25% or less reduction in symptoms with adequate dosing and duration—is necessary, discontinuation of treatment before the completion of the fourth week without clear evidence of lack of response should not be considered treatment failure, previous psychotherapy failing to improve major depressive disorder symptoms should not be an exclusion criterion, and all specifiers of depression (excepting bipolar depression) should be considered within the TRD/PRD definition.
Two-thirds of recommendations had a moderate level of consensus, meaning they were supported by a substantial majority (61% to 94%) of the expert panel. In this group was the recommendation that PRD, as distinct from TRD, can be defined by a 25% to 50% reduction in symptoms to at least one antidepressant. The differentiation between these levels of treatment resistance will be useful to identify individuals appropriate for certain trials and treatments, Sforzini says.
Definitional issues can arise with medical conditions if the definitions are not “unequivocal,” says Sforzini. “In psychiatry, we use the clinician’s assessment and rating scales or structured interviews that have good validity and reliability, but we need to use this information in a way that generates the same conclusions across different cultural and geographic settings.”
Multiple assessment questionnaires might be used to diagnose TRD and PRD, although the experts achieved moderate consensus around the idea of using “staging models” (notably, the Maudsley Staging Model) that also assess the level of treatment resistance as the preferred instrument. Lacking approved biomarkers discoverable with a blood or genetic test, clear definitions are critical to minimizing subjectivity bias and “may help to advance psychiatry to be precision medicine,” Sforzini says.
Preliminary data suggest it may be possible to genetically discern responders and nonresponders to treatment, he adds. Deep learning approaches using both clinical and genetic biomarkers have also been brought into the mix (Frontiers in Psychiatry, DOI: 10.3389/fpsyt.2018.00290).
In their consensus recommendations, the researchers strongly agree that the collection of biological data, such as blood samples and brain scans, should be done consistently with the aim of identifying possible markers or measures that could identify people with different forms of depression that may respond to different types of treatment.
Areas Of Uncertainty
All the consensus recommendations on TRD/PRD represent the majority opinion of the multi-stakeholder group, says Sforzini. Opinions were only slightly favoring the idea of having the definitions of TRD (and PRD) include the current episode—as well as only considering treatment failures within the last two years for long current episodes.
As the sole recommendation with a weak level of consensus, this represents an area of uncertainty requiring further research and discussion. This position sits in the middle of a continuum of different opinions ranging from who wanted to evaluate the response in the same episode and within a period shorter than two years to those who wanted to consider the lifetime clinical history. “However, we reached a consensus on all the major areas of debate, and these are widely discussed throughout the document, so that even minority views are adequately represented,” Sforzini says.
The consensus recommendations are now ready for global use with no expected obstacles to their adoption, he says, although time will tell their uptake by regulatory agencies. The guideline document will need periodic updating to stay current with emerging evidence around the diagnostic characteristics of TRD and PRD, he additionally notes.