Treatment resistant depression: Methodological overview and operational criteria

D. Souery*, J. Amsterdam, C. De Montigny, Y. Lecrubier, S. Montgomery, O. Lipp, G. Racagni, J. Zohar, J. Mendlewicz

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


A wide variety of definitions are used for Treatment Resistant Depression (TRD), considering various criteria and different concepts. Some of the key issues are: the diagnosis, the treatment adequacy in terms of dose and duration, the treatment response assessment and the number of failed therapeutic trials required. Systematic research has been characterizing the concept and criteria to define the different variables involved. Lack of consensus on these issues limits comparison across clinical trials and interpretation of treatment efficacy in the management of treatment resistant patients. Through reanalyzes of available data, we point out the limits of TRD definitions and propose conceptual and operational criteria for a collaborative research project on TRD. It appears that a number of variables commonly associated to treatment resistance are independent of patients characteristics and mainly refer to misdiagnosis and inadequate treatment. The proposed criteria are intended for therapeutic trials in TRD, combining the evaluation of treatment efficiency acid the validation of the concept of TRD itself. Major depression with poor response to two adequate trials of different classes of antidepressants is proposed for an operational definition of TRD. Rationale for this definition is discussed in contrast to alternative definitions.

Original languageEnglish
Pages (from-to)83-91
Number of pages9
JournalEuropean Neuropsychopharmacology
Issue number1-2
StatePublished - 1 Jan 1999
Externally publishedYes


  • Antidepressants
  • Definition
  • Therapeutic trials
  • Treatment resistant depression


Dive into the research topics of 'Treatment resistant depression: Methodological overview and operational criteria'. Together they form a unique fingerprint.

Cite this