Authors: Christina Gonzalez-Torres (1Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, Ontario, Canada; 2Department of Psychiatry, Temerty Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada), Benoit H. Mulsant (1Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, Ontario, Canada; 2Department of Psychiatry, Temerty Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada), M. Ishrat Husain (1Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, Ontario, Canada; 2Department of Psychiatry, Temerty Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada), Martin Alda (3Department of Psychiatry, Dalhousie University, Halifax, Nova Scotia, Canada; 4National Institute of Mental Health, Klecany, Czech Republic), Robert C. Young (5Department of Psychiatry, Weil Cornell Medicine, New York, New York, US), Abigail Ortiz (1Campbell Family Mental Health Research Institute, Centre for Addiction and Mental Health, Toronto, Ontario, Canada; 2Department of Psychiatry, Temerty Faculty of Medicine, University of Toronto, Toronto, Ontario, Canada)
Categories: Article, Bipolar disorder, mania, pharmacotherapy, psychopharmacology, treatment resistance, treatment refractory, clinical trials, systematic review
Source: Bipolar disorders
Doi: 10.1111/bdi.13383
Authors: Christina Gonzalez-Torres, Benoit H. Mulsant, M. Ishrat Husain, Martin Alda, Robert C. Young, Abigail Ortiz
To review the definitions of treatment-resistant mania (TRM) in the literature and propose criteria for an operationalized definition.
Systematic search of five databases (MEDLINE, EMBASE, PsychInfo, Cochrane Central, and CINAHL) and data extraction of eligible articles.
47 articles addressing the concept of TRM were included, comprising 16 case reports, 11 case series, 3 randomized clinical trials, 8 open-label clinical trials, 1 experimental study, 7 narrative reviews, and 1 systematic review. While reviews discussed several challenges in defining TRM, definitions varied substantially based on different criteria for severity of mania, duration of mania, and use of specific therapeutic agents with minimal dosages and duration of treatment. Only a handful of the reviewed articles operationalized these criteria.
While the concept of TRM has been discussed in the literature for over three decades, we could not find an agreed-upon operationalized definition based on specific criteria. We propose and discuss a possible definition that could be used by clinicians to guide their practice and by researchers to assess the prevalence of TRM and develop and test interventions targeting TRM.
Bipolar Disorder (BD) affects at least 1% of the global population ^1^. It is characterized by relapsing and remitting episodes of mania and depression ^2^. Most patients with BD also experience inter-episodic sub-syndromal symptoms that cause substantial levels of dysfunction ^3^ and they require acute and long-term pharmacotherapy. Several guidelines have been published to guide BD treatment ^4–10^. The choice of specific psychotropic medications depends on numerous factors, including the bipolar subtype, phase of the illness, presence of psychosis, comorbidities, and history of treatment resistance. Treatment can be divided into acute and maintenance forms, with acute treatment aiming at reducing symptoms of mania or depression and maintenance treatment aiming at preventing relapses and recurrences ^9^. Acute treatment also considers the polarity of the episode. Hence, different medications alone or in combinations are typically recommended for the management of acute depressive episodes, acute manic episodes, mixed episodes, rapid-cycling, or long-term prophylaxis of these episodes.
When patients with a mental disorder fail to respond to pharmacotherapy, they are considered treatment resistant. To date, the concept of treatment–resistant depression (TRD) has been well established ^11,12^. While several publications have also explored the concept of treatment-resistant bipolar disorder (TRBD) ^10,13–17^, treatment-resistant mania (TRM), (also called treatment-refractory mania), has not received the same attention. This is in part because most patients with BD spend a greater proportion of their lives experiencing the depressive rather than manic polarity of the disorder ^18,19^. However, mania is a severe psychiatric condition that requires a swift resolution. Manic patients can remain hospitalized for long periods of time, unable to resume their normal life, and receiving high dosages of medications, often associated with adverse effects ^20^. Patients with a history of mania who do not respond to conventional pharmacotherapy are also at high risk of a substantial deterioration of their physical health ^21^. Incomplete resolution of mania puts patients at risk for relapse or recurrence, with manic recurrence estimated to occur in 40-60% of patients with BD receiving monotherapy ^22^ and 50-70% of those with more severe forms of BD receiving combination therapy ^23^. A standard definition of TRM could help improve the clinical management and outcomes of mania, and the study of its treatment.
While some definitions of TRM have been proposed in the literature, no definition has been widely adopted by the field. A well-accepted operationalized definition of TRM would allow the assessment of its true prevalence and the development and testing of interventions targeting TRM (as has been the case for TRD or treatment-resistant schizophrenia). In this context, we systematically reviewed the relevant literature, to summarize how TRM has been defined by different authors, and to inform a definition that could be adopted by clinicians and researchers.
On November 25, 2022, and May 16, 2023, we searched all papers in five databases (MEDLINE, EMBASE, PsychInfo, Cochrane Central and CINAH) using a search strategy designed with the help of a professional mental health librarian (see Appendix). We complemented this search by examining the reference list of the relevant papers identified in the search. Papers were eligible for inclusion if they proposed or discussed a definition of TRM or of treatment-refractory mania (while we believe the two terms can be used interchangeably, we favor TRM because “refractory” may wrongly suggest that the disorder is irremediable). We included all types of peer-reviewed publications (i.e., case reports, case series, reports of clinical trials or experimental studies, previous narrative or systematic reviews) in English, French, or Spanish. We excluded papers solely focused bipolar depression, differential diagnosis of BD, pathophysiology or animal studies of BD; or psychotherapeutic interventions for BD. We also excluded papers that mentioned TRM but did not provide a definition, or that only cited a definition proposed in a previous publication (which was already included). Finally, we also excluded papers that discussed mania or hypomania that did not respond solely to a specific medication (e.g., “lithium-resistant mania”).
The search identified 7,370 publications reduced to 4,041 after removing 3,329 duplicates. One of the authors (CGT) screened their titles and abstracts based on the eligibility criteria listed above and retained 53 potentially eligible publications. Then, following the PRISMA guidelines ^24^, two of the authors (CGT, BHM) independently reviewed the full text of these 53 publications and excluded 6; when disagreements occurred, a discussion took place until a consensus was reached (See Figure 1). They extracted the following data from the remaining 47 publication included in the systematic authors, years of publication, number of patients with mania when relevant, criteria used (or proposed) to define TRM (i.e., severity of manic symptoms; duration of mania; pharmacotherapy to which mania had failed to respond, including agents, dosages, and duration of treatment), and, when relevant, intervention provided for TRM.
The 47 eligible publications (with the oldest eligible publication from 1993 and the most recent one from 2020) comprised 16 case reports, 11 case series, eight open-label treatment trials, three randomized controlled trials, one laboratory study, seven narrative reviews, and one systematic review that used or proposed operational or conceptual definitions of TRM. Table 1 presents the elements of these definitions organized by types and years of publication.
Our search identified 16 eligible case reports ^20,25–39^ describing patients presenting with severe mania (typically with psychosis and agitation or violence) or who had experienced multiple rapid and severe manic relapses despite pharmacotherapy for four weeks to 20 years. In seven cases ^26,29–31,33,35,39^, a duration of unremitted mania lasting longer than one year raises the issue of chronic mania. In only two reports, a scale - the Young Mania Rating Scale (YMRS) ^40^ was used to characterize the severity of mania. In all cases, one or several combinations of several psychotropic medications were used unsuccessfully, typically at least one traditional mood stabilizer and one antipsychotic (including clozapine in four cases). When levels or dosages were reported, they were typically high but the specific duration of trials of a single medication or of a single combination were mentioned in only two “at least two weeks” ^26,28,34^. In five cases ^26–28,37,38^, the patient was reported not to have responded to electroconvulsive therapy (ECT), but the treatment parameters were not described (e.g., electrode placement or number of sessions). Also, poor tolerability to psychotropic medications played a role in three cases labelled as “resistant” or “refractory” to treatment ^29,33,39^.
Our search identified eleven case series ^13,41–50^ describing two ^42,44^, three ^45,47^, ^13,41^, five ^48^, nine ^43^, 20 ^49^, 25 ^46^, and 517 patients with TRM ^50^ (see Table 1). As in case reports, patients in the smaller case series presented with severe mania that required an inpatient admission (with two exceptions) and typically did not respond to various sequences or combinations of mood stabilizers and antipsychotics, or to ECT. Two case series characterized the severity of mania with a rating scale ^17,50^. In one case series ^42^, two patients were labelled as “acute refractory mania” after either not responding to lithium and an antipsychotic for one week or not tolerating lithium in the past; otherwise, the durations of various treatments were not specified. The largest case series ^50^ reported a naturalistic follow-up of 3,373 patients with BD, of whom 517 (15.3%) were deemed to be “treatment non-responsive” based on an operationalized definition that did not include any criteria related to the type or quality of treatment (see Table 1).
Our search identified eight open label clinical trials ^51–58^ that enrolled three to 22 participants with BD, with four trials also involving participants with other diagnoses ^53,54,56,58^. Four trials required a minimum severity of mania based on a scale ^54–57^ and six trials required a minimum duration of treatment ^52–55,57^. However, the definition of TRM was blurred in five trials by the inclusion of tolerance to treatment ^52–55,57^ and in three trials by being based on pharmacotherapy during past manic episodes ^51,55,57^.
Our search identified three RCTs that enrolled seven ^59^, 11 ^60^, and 39 ^61^ participants with BD. All three RCTs provided operationalized definitions of TRM (see Table 1), with two using a minimal duration of symptoms and a minimal symptom severity based on the YMRS ^60,61^. While two RCTs required a lack of improvement after being treated with several combinations of mood stabilizers or antipsychotics ^59,61^, the third RCT defined TRM based on lack of tolerance of an adequate dosage of a single mood stabilizer ^60^.
Our search identified one study that compared inflammatory biomarkers in patients with BD classified as having or not having TRM and healthy controls ^62^. The definition of TRM included a minimal symptom severity based on the YMRS, CGI-S, and improvement in YMRS score but did not specify a minimal duration of mania. The definition of TRM also required lack of response or tolerance of two mood stabilizers (one of them being lithium) and two antipsychotics but did not specify minimal duration of treatment or dosages. In addition, these treatments were provided during “previous manic episodes” rather than the current episode.
Our search identified seven narrative reviews and one systematic review proposing definitions for TRM. Most of these reviews addressed interventions for treatment-resistant BD and discussed issues related to defining treatment-resistance rather than discussing specific operational criteria (e.g., in terms of psychotropic medication, dosage, or treatment duration).
The earliest narrative review defined TRM as a “lack of return to the patient’s best previous level of functioning” without discussing required treatment(s) ^63^. A contemporaneous narrative review (focused mostly on treatment-resistant bipolar depression) proposed a more specific definition of TRM based on the absence of remission after 6 weeks of adequate treatment with at least two antimanic agents (i.e., mood stabilizers and antipsychotics) in the absence of antidepressant or other mood-elevating agents ^12^. In their review, Tohen & Gannon ^64^ identified several issues related to TRM: differentiating TRM from mania non-responsive to lithium; whether a partial response to treatment should define “partial resistance”; whether having residual subsyndromal manic symptoms should be considered part of TRM; and defining TRM based on treatment of acute mania vs. a relapse during maintenance treatment. Considering these issues, their proposed definition required a lack of response after 6 weeks of treatment with only one medication. Similarly, Gitlin ^15^ discussed issues associated with defining TRM, but did not propose operationalized criteria for TRM. The factors discussed the limitation of focusing on symptomatic rather than functional outcome; conflating treatment resistance and intolerability of treatment; whether absence of antidepressants should be required to define TRM, differentiating TRM in the context of acute and maintenance treatment; and the importance of considering non-adherence (i.e., pseudo-resistance) in patients with BD. In a subsequent review, Fountoulakis ^65^ discussed some of the same pseudo-resistance due to lack of tolerability or adherence; and whether TRM should be based on inability to achieve response or remission (while providing a definition of remission based on threshold scores on mania and depression scales). They also commented on the need to differentiate outcomes of acute or maintenance treatments and observed that a narrow definition of acute treatment-resistance focused on a specific pole (e.g., mania) would be easier to define and a necessary precursor to a broader definition of treatment-resistant BD (which in their opinion would be more “clinically oriented and meaningful”). While they did not discuss specific medications or combinations, they proposed that TRM could be defined by a lack of response after 8-10 weeks of treatment with “an effective agent”; in turn, they recommended that response should be defined according to the criteria of the International Society for Bipolar Disorders (ISBD) ^6^. Poon et al ^66^ defined TRM based on an unsatisfactory response to at least two “dissimilar” treatments with adequate dosages for at least 6 weeks, commenting on the contrast between short-term treatment efficacy vs. non-sustained long-term effectiveness when treating BD.
In contrast to these reviews, based on a review of clinical trials published between 1966 and 2001, Keck and McElroy ^67^ proposed to differentiate primary, secondary, and tertiary treatment resistance based on an “inadequate response” to one, two, or three sequential medications. Unlike most other authors, they provided specific threshold levels (e.g., lithium levels of ≥ 0.8 mmol/L) or dosages (e.g., haloperidol (≥ 0.2 mg/kg/day or risperidone ≥ 5 mg/day) to define the adequacy of treatment trials with specific medications. They also provided minimal duration of treatment (with 2 weeks being considered “probably adequate” and 3 weeks, “definitely adequate”). However, they did not describe what would constitute an “inadequate response” and did not suggest the severity or duration of mania required for TRM ^67^. Using a similar approach, Gajwani ^68^ proposed to classify TRM in four stages based on lack of response (defined as a decrease in YMRS score < 50%) to one adequate monotherapy trial (stage I), a combination of agents (Stage II), several adjunctive treatments (e.g., oxcarbazepine, or clozapine, or asenapine) (Stage III), and neurostimulation (Stage IV). While emphasizing that adequacy should be based on “adequate dose and adequate duration”, they did not provide thresholds defining adequate dosage or treatment duration for specific medications.
Finally, a recent systematic review conducted by a working group commissioned by the International College of Neuropsychopharmacology (CINP) ^10^ proposed a definition of TRM that echoes the earlier definition by Fountoulakis ^65^ discussed above. TRM would be defined by a lack of response to 8-10 weeks of treatment. In this updated definition, treatment has to be evidence-based according to the CINP or Canadian Network of Mood and Anxiety Treatments (CANMAT) most recent guideline, while response remains based on the criteria of the ISBD ^6^, which requires both a reduction ≥ 50% in mania symptom severity using the YMRS or MRS scale and “a lack of exacerbation of depressive symptoms” (operationalized in the CINP definition of TRM as an increase in MADRS or HDRS scores, or these scores exceeding 6).
We identified 47 publications over the past three decades that have proposed or discussed definitions of TRM. Case reports and case series described a wide range of severity or duration of manic symptoms and of medications used to attempt to control manic symptoms. Similarly, open or controlled trials for TRM used eligibility criteria based on various levels of severity, duration of symptoms, and prior ineffective treatments. Also, some of the trials enrolled patients based on intolerance to treatment or a history of non-response to treatment during previous episodes. Several narrative reviews addressed these (and other) conceptual issues related to TRM, with a consensus that lack of response to treatment due to non-adherence or intolerance should be considered “pseudo-resistance” rather than TRM.
One early review offered a definition of TRM requiring an “inadequate response” to adequate medications trials ^67^. While the “inadequate response” was not further defined, the authors provided a list of medications with specific blood level or dosage thresholds corresponding to adequate trials. They also proposed that a minimal duration of 2 or 3 weeks of treatment would correspond to probable or definite adequacy, respectively, and that failure to respond to one, two, or three adequate treatments would define primary, secondary, or tertiary TRM. Building on this and other work, the most recent review ^10^ advised that lack of response be defined based on the published ISBD criteria for acute response in the treatment of mania ^6^. This recent review also addressed the challenge of the evolving definition of adequate pharmacotherapy for acute mania by recommending that it be based on the most recent available guidelines.
The result of our review points out that the definition of TRM has evolved. As a group, contemporary authors suggest that TRM should be distinguished from secondary mania, chronic mania (which is very rare), and pseudo-resistant mania (due to treatment intolerance or non-adherence). They also suggest that TRM in the context of a lack of response to acute treatment of mania should be differentiated from a broader concept of treatment-resistant bipolar disorder encompassing not just TRM but also other poor treatment outcomes, such as relapse or recurrence despite continuation or maintenance long-term treatment, switch to a depressive or mixed episode, or rapid cycling. Despite this conceptual progress, TRM remains less clearly defined in comparison with treatment-resistant unipolar depression ^11,69^, schizophrenia ^69,70^, obsessive-compulsive disorder ^71^, or even bipolar depression ^10^. Unlike with these other disorders, we could not identify one published operational definition simple enough to be used by a trained clinician having access to the patient’s chart documenting past treatment and their outcomes. We propose the elements of such a definition in Table 2 and discuss our rationale below.
As shown in Table 1, a growing number of definitions of TRM required a minimal symptomatic severity of mania based on a scale. We favor this approach (i.e., an absolute score) over a “lack of response” defined by a relative reduction in symptoms (e.g., 30% or 50%) because an absolute score can be determined cross-sectionally (i.e., after completion of treatment) while a relative change in scores require a baseline measure before treatment is initiated. Thus, we propose the following minimal severity criterion, which can be assessed cross-sectionally: meeting the DSM-5 criteria for a manic episode with a YMRS score ≥18 or a MRS score ≥13. Congruent with other authors (e.g., ^66^), this criterion excludes subsyndromal residual symptoms (or a switch to a depressive episode) from being part of TRM. As discussed above, these important clinical challenges can be considered part of a broader concept of treatment-resistant bipolar disorder.
According to some definitions, a manic episode as short as two weeks could be qualified as TRM; other definitions require a duration of 12 weeks or longer. We propose a minimal duration of 6 weeks because it is the shortest duration that would allow enough time for an adequate pharmacotherapy trial (see below).
Congruent with most other definitions, we propose that TRM requires treatment with at least two medications, with one being a traditional mood stabilizer (i.e., lithium, divalproex, or carbamazepine) and the other one being an antipsychotic. If these two medications are used sequentially, therapeutic dosages of each should be taken for at least three weeks. The therapeutic dosage for a mood stabilizer would be a dosage associated with a level within an established therapeutic range; for an antipsychotic, it would be at or above the lowest dosage used in registration trials e.g., the lowest dosage defining the therapeutic dosage range approved by the US Food and Drug Administration. While some antipsychotics may not be approved for the treatment of mania, dosage equivalency with antipsychotics that are approved can be used. If two medications are used concurrently, then the duration of the trial should be at least six weeks and a therapeutic level or dosage are required for both medications. Finally, while a relapse or recurrence of acute mania while taking a therapeutic dosage of one medication would not meet our definition of TRM, this medication would “count” as one failed trial (i.e., if a medication did not prevent a relapse or recurrence, it would not be expected to be efficacious to treat an acute episode). Following this logic, a relapse or recurrence of acute mania while taking therapeutic dosages of both a mood stabilizer and an antipsychotic would meet our definition of TRM.
Our definition does not distinguish between different antipsychotics (e.g., first- or second-generation antipsychotics) because we are not aware of any convincing evidence associating specific classes or mechanisms of action (MOA) of antipsychotics with specific outcomes in the treatment of mania ^72–74^. Thus, as most (but not all) reviewed authors, our proposed definition requires persistence of mania despite treatment with (at least) one antipsychotic (in addition to one traditional mood stabilizer) and we do not differentiate between antipsychotics of different classes or with putative different MOAs. Similarly, we do not differentiate antipsychotics that have been approved or not approved by the US Food and Drug Agency or the European Regulatory Agency for the treatment of mania because these approvals have been related to historical or commercial reasons rather than scientific or clinical ones (e.g., no first-generation antipsychotics -including clozapine- is approved, while almost all second-generation antipsychotics are approved).
Also, according to our definition of TRM, a mood stabilizer and an antipsychotic can, but do not have to, have been prescribed concurrently. This is a compromise based on the literature we reviewed (see Table I): focusing on the 18 papers published since 2000 that are not case reports, seven define TRM based on one failed medication trial (i.e., Askoy Poyraz et al, 2015, Evins et al, 2006; Gajwani 2009; Fountoulakis 2012; Fountoulakis et al, 2020; Keck & McElroy, 2001; Vik et al, 2013); five define TRM based on two failed medication trials (i.e., Gitlin 2006; Green et al, 2000; Kumar et al, 2015; Poon et al, 2012; Vieta et al, 2002); four define TRM based on three or more failed medication trials (i.e., Benzoni et al, 2015; Britto de Macedo-Soares et al, 2005; Bulut et al, 2019; Chen et al, 2011); and, while some mention that medications can be given sequentially or in combination (like we do), only two require a failed combination (e.g., Ifteni et al, 2014; Manhas et al, 2016). In this context, we considered whether we should require that a mood stabilizer and an antipsychotic had been used in combination to define TRM. We did not do so because only two authors have this requirement (see above) and this requirement would define a higher degree of resistance as done by one author (Gajwani, 2009) who distinguished two stages of TRM based on whether or not a combination of a mood stabilizer and an atypical antipsychotic had been used. Similarly, we only require that the failed medications have been prescribed within the recommended therapeutic range (for level or dosage) as established in registration trials. Thus, some patients who do not respond to the lowest recommended levels or dosages may be classified as TRM even though they may have responded to higher dosages. Implicit in this approach, patients who require high therapeutic (or supra-therapeutic) dosages would be classified as TRM (i.e., TRM does not equate total treatment refractoriness and the simplest strategy when treating a patient with TRM could be to increase to the maximum level or dosage tolerated).
In conclusion, our review emphasizes the possibility of operationalizing a definition of TRM based on specific criteria including the severity and duration of mania, the type of medications that have been used, and their duration and dosages. This basic definition of TRM can lay the foundations for a broader concept of treatment-resistant BD, taking into account longitudinal symptomatic and functional outcomes. However, TRM should be considered a modifier (i.e., a dynamic state) rather than a diagnostic subcategory (i.e., a fixed trait). Also, we are proposing a categorical definition of TRM (e.g., criteria that could be used as eligibility criteria in a clinical trial) rather than a more complex dimensional description of non-responsiveness (e.g., stages of treatment-resistance that could be used to define and contrast subgroups of patients). We view this “simple” definition as a starting point and we expect it to be refined and changed based on new empirical data (e.g., results of treatment trials).