Defining Treatment Resistance

Although clinicians and researchers have not reached consensus on diagnosing treatment-resistant schizophrenia (TRS), most of us use the same working definition. For instance, the Canadian Psychiatric Association treatment guidelines suggest that patients who fail to respond to two antipsychotic trials are treatment resistant. Persistent positive symptoms and functional impairment are the essential factors in TRS, and the significance of the diagnosis lies in determining when a clozapine trial is appropriate, a high-stakes decision for an individual patient.

In a paper published in Psychiatry Research (ref), a group of Canadian and Japanese psychiatrists have proposed a definition that takes into account medication unresponsiveness and functional impairment. Based on a literature review, they endorse the widespread criterion for TRS of failure of two different antipsychotics, each attaining a chlorpromazine-equivalent daily dose of at least 600 mg for a duration at least six consecutive weeks. The authors do not specify any particular classes of antipsychotics given the limited evidence that non-clozapine antipsychotics differ in efficacy.

An important factor is how well previous treatment response is documented. The authors indicate that the failure to respond to past trials should be “clearly documented and unequivocal.” The use of quantitative measures such as the PANSS and BPRS is ideal but is rare in routine practice, so instead the authors suggest treatment failure as end-point of CGI-Severity scale of 4 or greater and Global Assessment of Function (GAF) of 50 or less. If the past trials are not adequately documented, failure of one prospective trial would be required to satisfy this criterion.

Once the presence of TRS is confirmed, the authors propose that response to treatment would require a score on CGI-Improvement scale of 1 or 2 or a decrease of 20% on the PANSS or BPRS, along with and an 20-point or greater increase in GAF. They define a partial response as a CGI-Improvement score of 3 or a 10-19% improvement in PANSS or BPRS, along with GAF improvement of 10-19 points.

At the BC Psychosis Program, we are fortunate to have the resources that allow us to employ standardized instruments such as the PANSS, CGI scale and GAF at baseline and at discharge to document patients’ response to treatment. One point of this article is that simply using the CGI and GAF, which take little time, would be very helpful for subsequent care givers in understanding a patient’s response to prior medication trials.


Suzuki T, Remington G, Mulsant BH et al. Defining treatment-resistant schizophrenia and response to antipsychotics: A review and recommendation. Psychiatry Res. 2012;197(1-2):1-6. Abstract