Canadian Network for Mood and Anxiety Treatments (CANMAT) and International Society for Bipolar Disorders (ISBD) 2018 guidelines for the management of patients with bipolar disorder

2.1.1. Prevalence

Bipolar disorder is a common and disabling mental illness with significant morbidity and mortality. The estimates of prevalence of BD vary. The World Mental Health Survey Initiative reported total lifetime (and 12-month) prevalence estimates of 2.4% (1.5%) across BDI, BDII and subthreshold BD subtypes. While the prevalence rates for each subtype varied across the nine countries studied, subthreshold BD was the most common at 1.4% (0.8%), followed by BDI at 0.6% (0.4%) and BDII at 0.4% (0.3%). While Canada was not included in this study, similar results were reported from the Canadian Community Health Survey-Mental Health, which found the lifetime prevalence of BDI was 0.87% and that of BDII was 0.67%.

2.1.2. Age of onset

Bipolar disorder frequently manifests in late adolescence and young adulthood, with an overall average age of onset of 25 years. Statistical models suggest the presence of three age of onset subgroups within BDI and these can be categorized into a large early-onset group (mean ± standard deviation (SD) 17.24 ± 3.20 years), and smaller middle-onset (23.93 ± 5.12 years) and late-onset (32.20 ± 11.96 years) groups, with the proportion of individuals falling into each category being 41.7%, 24.7% and 33.6% of the total sample, respectively. However, the ages of onset tend to differ somewhat depending upon the origins of samples analysed. For instance, a recent study showed that the mean age of onset for a USA sample was 20 years, with ages of onset of 14.5 ± 4.9 years (63%), 26.5 ± 7.6 years (28.5%), and 39.5 ± 12.5 years (8.5%) for early-, middle- and late-onset groups, respectively; while a European sample showed a later mean age of onset of 29 years and a later onset in each of the three categories, with 19 ± 2.7 years (24.8%), 27.2 ± 6.3 years (50.7%), and 41.8 ± 10.7 years (24.5%) as the ages of onset for early, middle and late-onset groups, respectively. Those with an earlier age of onset tend to have a longer delay to treatment, greater depressive symptom severity, and higher levels of comorbid anxiety and substance use. While manic episodes can occur for the first time after the age of 50 years as a part of BDI, the possibility of organic mania should be considered and investigated in these cases.

2.1.3. Burden of illness

People living with BD experience substantial impairment, being symptomatic with syndromal or subsyndromal symptoms, particularly those of depression, for approximately half of their lives., Patients are unable to maintain proper work role function approximately 30% or more of the time. Quality of life is reduced in both symptomatic and non-symptomatic patients when compared to healthy controls,, , and several domains of functioning have been identified by patients as being of particular importance- including physical, sleep, mood, cognition, leisure, social, spirituality, finances, household, self-esteem, independence, identity, work, and education. For both psychosocial functioning and quality of life, impairments are more pronounced in patients with depressive symptoms,, , in those with more previous episodes/longer duration of illness,, and in those with lower cognition.

Consistent with these observations, the Global Burden of Disease Study attributed 9.9 million years lost to disability (YLD) to BD, making it the 16^th leading cause of YLD worldwide. The impact that BD has on young people is even greater, with being the sixth leading cause of disability-adjusted life years among people aged 10-24 years worldwide. The burden of this disease was further emphasized in a systematic review addressing cost of illness studies, with findings demonstrating that the worldwide annual costs per person with BD range from US $1904 to $33 090; higher per person costs associated with BDI, delayed or misdiagnosis, frequent psychiatric interventions, use of atypical antipsychotics, treatment non-adherence, poor prognosis, relapse, and comorbidity.

2.2.1. DSM-5 diagnostic criteria

Bipolar disorder encompasses a spectrum of diagnostic subgroups primarily divided according to the severity of mood elevation experienced during acute episodes On this spectrum, BDI is placed at one pole due to the presence of threshold manic episodes in which features include inflated self-esteem, decreased need for sleep, pressured speech, racing thoughts, distractibility, psychomotor agitation, and risky behaviour that leads to significant functional impairment, and may include psychotic features, and/or necessitate hospitalization. At the other end of the spectrum, cyclothymia is characterized by subthreshold presentation of hypomanic and depressive symptoms that, while chronic, do not meet diagnostic criteria for a major depressive episode or manic/hypomanic episode. BDII sits between the two conditions with hypomanic episodes qualitatively like manic periods but, although distinct and observable, are not of a sufficient duration or severity to cause significant functional impairment, hospitalization, or psychosis. Individuals with BDII also experience threshold depressive episodes.

DSM-5 has replaced the BD not otherwise specified (NOS) category in DSM-IV with two new categories; other specified bipolar and related disorder and unspecified bipolar and related disorder. Also, DSM-5 includes substance/medication-induced bipolar and related disorder and bipolar and related disorder due to another medical condition. For more detailed discussion of diagnostic categories, the reader is advised to consult DSM-5 and recent Royal Australian and New Zealand College of Psychiatrists guidelines for treatment of mood disorders.

2.2.2. DSM-5 specifiers for bipolar and related disorders

DSM-5 also includes a range of specifiers that clinicians may use to further clarify the specific course, severity, and features of BDs. While a more detailed description can be found in the DSM-5 manual, the available specifiers and their use across the spectrum are listed in Table 5 Many of these specifiers may also be used to guide treatment decisions for acute mania (Section 3) and depression (Section 4). Amongst these, the mixed features specifier, which has replaced mixed episodes, warrants consideration because of the multiple and complex presentations of mixed states it can give rise to. Furthermore, the nascency of this terminology has meant that treatment data are as yet sparse. DSM-5 has added mixed features as a specifier during an episode of major depressive disorder (MDD) as well, which will probably pose some pragmatic diagnostic challenges and management dilemmas for clinicians.

2.2.3. Staging bipolar disorder

The course of BD is heterogeneous but, on average, the risk of recurrence increases with the number of previous episodes. In addition, data examining the effect of episodes on the course of illness shows that the number of previous episodes is associated with increased duration and symptomatic severity of subsequent episodes. Moreover, the number of episodes is associated with a decreased threshold for developing further episodes and with an increased risk of dementia in the long term. The progressive course of illness in patients with multiple episodes is called clinical progression and the biological basis of clinical progression is defined as neuroprogression.,

The concepts of clinical progression and neuroprogression have provided the basis for the development of staging systems in BD. Overall, the staging models describe three broad clinical stages: (I) individuals at increased risk for developing BD due to family history as well as certain subsyndromal symptoms predictive of conversion into full-blown BD; (II) patients with fewer episodes and optimal functioning in the interepisodic periods, and (III) patients with recurrent episodes as well as decline in functioning and cognition. So far, the heterogeneity intrinsic to BD has prevented the clinical use of staging systems. In addition, the field of staging is in its infancy and the ability of staging systems to guide prognosis and treatment is still to be determined. Overall, the model of staging has helped clinicians to appreciate the importance of early identification and treatment as well as illness trajectories in BD.

2.2.4. Screening and diagnosis of bipolar disorder

Due to frequent depressive onset, variable help seeking for hypomanic or manic periods, temporal instability of symptoms, and high rates of comorbidity; accurate the timely identification of BD can be difficult to achieve in many cases. Indeed, many individuals are not accurately diagnosed until up to 10 years after the onset of symptoms, with one to four alternate diagnoses typically being given prior to correct recognition and treatment, This delay has important consequences, including inadequate initial treatment and worse prognosis in terms of episode recurrence and functional outcome.,

The most frequent misdiagnosis is that of MDD, as patients are more likely to present for the treatment of depressive symptoms and may not recall periods of hypomania or mania, or may not interpret them as being pathological. Recall and insight are particularly impaired during periods of acute depression, with pronounced memory or concentration difficulties. There are several features of depression that may increase suspicion of bipolarity, and prompt more careful investigation, including earlier age of illness onset, highly recurrent depressive episodes, a family history of BD, depression with psychotic features, psychomotor agitation, atypical depressive symptoms such as hypersomnia, hyperphagia, and leaden paralysis, postpartum depression and psychosis, past suicide attempts, and antidepressant-induced manic symptoms or rapid cycling (Table 6) Given the recent change in DSM-5 to allow the possibility of depression symptoms with subthreshold simultaneous hypomanic symptoms (mixed specifier), it is also important to explore if an individual is experiencing mixed symptoms., Schizophrenia and other psychotic disorders are the second most common misdiagnosis, occurring as the initial diagnosis in as many as 30% of patients.

In addition to this under-diagnosis, there are also concerns that BD may be over-diagnosed in some circumstances. For instance, the symptoms of borderline personality disorder, substance use disorder (SUD) and attention deficit hyperactivity disorder (ADHD) overlap significantly those of hypomania/mania, and some reports suggest that patients with these conditions often get misdiagnosed with BD. These conditions also are often comorbid with BD, which makes the diagnosis of this condition often challenging.

Validated self-report instruments, such as the Mood Disorders Questionnaire (MDQ), may be used as a screening tool to flag patients for whom a more detailed assessment is needed. It is important to note, that such tools have poor sensitivity and specificity, especially in community or highly comorbid populations, and will thus have an elevated risk of also flagging those with borderline traits. As such, tools such as the MDQ should be used only as an adjunct for screening clinical populations and not for diagnostic or treatment planning purposes.

To improve the accuracy of diagnosis, it is important that clinicians strictly adhere to diagnostic criteria rather than relying on heuristics. It is important to complete a careful psychiatric history, including in first-degree relatives, with attention paid to any suspected periods of increased activity, irritability, or other change in behaviours. Collateral information from friends and family members should be included wherever possible. Ongoing monitoring of symptoms, such as mood charting, can also help to detect bipolarity that may only become apparent over time. Confirmation of the diagnosis can then be made more confidently when episodes are prospectively observed.

2.2.5. Comorbidities and mimics

As described in Section 6, patients diagnosed with BD very commonly have one or more comorbid psychiatric diagnoses, with SUDs, impulse control disorders, anxiety disorders, and personality disorders (especially cluster B disorders) particularly common. The presence of comorbidity increases the complexity of the illness and can make an accurate diagnosis even more difficult.

In addition to differentiating BD from other psychiatric diagnoses, alternative causes of mood symptoms, such as personality disorders, medical or neurological conditions, substance use, and medications must be considered in the differential diagnosis (Table 7).

2.6.1. Psychoeducation

Psychoeducation broadly includes provision of information about the nature of the illness, its treatments, and key coping strategies to the patient and family. Current psychoeducational models for BD teach skill development in detecting and managing prodromes of depression and mania, ongoing stress management, problem solving, how to diminish the effects of stigma and denial of the illness, and provide tips on enhancing medication adherence and developing healthy lifestyles (eg, minimizing the use of alcohol, tobacco, drugs, stimulants such as caffeine; getting regular exercise; and regulating sleep and wake times). A key goal is the creation of personalized coping strategies to prevent mood relapse.

Psychoeducation may be delivered individually or in group settings. Empirical models of psychoeducation involve face-to-face interaction with a therapist, but new models are being tested that involve online tools, smartphone apps, and workbooks. Consistent with broader theories of learning, it is believed that psychoeducation is enhanced when it features active learning, with attention to monitoring the development of understanding, active skill development, and homework between sessions. Peer support and group learning are also postulated to add efficacy to psychoeducation. Regardless of the type of model and content included, priority should be given to maximize the therapeutic alliance, convey empathy, and consistently monitor symptoms.

Two models of psychoeducation, both delivered in group format to individuals who are well (euthymic), have published manuals and have substantial research support. These programmes, the Barcelona BDs Program (21 sessions over 6 months) and the Life Goals Program (phase I is six weekly sessions), also have tools to aid implementation with workbooks and handout materials, and both are first-line psychoeducational interventions based on level 2 evidence for the prevention of relapse. Individual psychoeducation based on these manuals would probably be effective, and when individual trials utilizing several different approaches to psychoeducation are combined in a meta-analysis, individual psychoeducation of at least five sessions would still be a first-line intervention for relapse prevention, based on level 2 evidence., , One large study demonstrated that the six-session Life Goals Program psychoeducational intervention was equivalent in relapse prevention to 20 sessions of individual CBT, at far lower cost, with probable shared mechanisms. Furthermore, that study demonstrated that integration of best practices in medication and psychotherapy simultaneously produced striking overall improvement in course of illness. Psychoeducation does not have any significant evidence of utility in either acute depressive or manic episodes.

2.6.2. Cognitive behavioural therapy

CBT in BD is supported by several published manuals and typically is given in 20 individual sessions over 6 months, often with additional booster sessions. Despite evidence of efficacy for CBT for MDD and psychosis, the results of CBT trials for BD have been mixed. One large RCT supports it use for acute bipolar depression in a trial that compared the efficacy of up to 30 (mean 14) CBT sessions against those of FFT, IPSRT, and a three-session control intervention, but it was not possible to identify whether the benefits came from changes in the medications prescribed or the psychosocial treatments. Efficacy of CBT in relapse prevention was observed in one RCT, but not in another larger RCT, at least in patients who had multiple mood episodes. From meta-analyses, effects on either depressive symptoms or on relapse remain uncertain due to important methodological problems and study selection factors., , A promising new direction in CBT has been established by a pilot study of "recovery-focused CBT" where 33 subjects received the novel CBT intervention, with evidence of reduction of relapse in the intervention group. Group CBT in euthymic patients with BD is also a new direction and has shown to increase time in remission.

In MDD, CBT, interpersonal psychotherapy (IPT) and behavioural activation have been explored in multiple RCTs and in general display similar efficacies. Based on this and the findings of the study by Miklowitz and colleagues in acute bipolar depression, CBT is still recommended as an adjunctive second-line treatment for acute bipolar depression (level 2). The recommendation is also second-line for maintenance treatment (level 2) for patients with fewer episodes and less severe form of illness. No evidence exists, and hence no recommendation is made, for CBT in mania.

2.6.3. Family-focused therapy

FFT presumes that outcomes in BD may be enhanced with the support and cooperation of family or significant others, particularly in families characterized by high levels of expressed emotion. FFT focuses on communication styles between patients and their families or marital relationships, with the goal of improving relationship functioning, and is delivered to the family and patient in 21 sessions over 9 months.

For acute bipolar depression in adults, an intensive FFT (up to 30 sessions; mean 14) out-performed a three-session control condition, although this study was limited by the caveats identified for CBT and IPSRT. Given that the original creation of FFT targeted factors related to depression, it may have specific antidepressant activity, which is also suggested by reduced depression relapse in maintenance studies. For relapse prevention, four significant RCTs of varying sizes have been conducted, delivered to a mixed audience of young adults and adolescents. In these studies, FFT demonstrated efficacy in reducing recurrence of new episodes of depression, but not mania. Overall, FFT is recommended as adjunctive second-line treatment for acute depression (level 2) and for maintenance (level 2). No evidence exists, and hence no recommendation is made, for FFT for mania.

2.6.4. Interpersonal and social rhythm therapy

IPSRT expands on the IPT focus on grief, interpersonal role transition, role dispute, and interpersonal deficits by including regulation of social and sleep rhythms, specifically targeted to the bipolar population. It is typically delivered in 24 individual sessions over 9 months.,

Few controlled trials of IPSRT have been conducted, with limited evidence of acute efficacy. The first, large trial showed no effect of IPSRT compared to a control condition but did show benefit for reduction of relapse and improved occupational functioning. An acute bipolar depression study showed intensive IPSRT (up to 30 sessions; mean 14) out-performed a three-session control condition, but it is impossible to state whether the performance was related to the intensity and number of sessions, changes in medication use, or specific attributes of IPSRT. Two small studies failed to demonstrate specific benefits of IPSRT compared to control conditions., Other open studies have shown some pre-post benefits in very small samples., , Again, since many psychosocial treatments for bipolar disorder share common core elements that may be psychoeducational, it is possible that the relapse prevention aspects of psychoeducation may also result from IPSRT interventions, mediated by the same therapeutic processes.

Overall, IPSRT is recommended as an adjunctive third-line treatment for acute depression and for maintenance, based on limited (effect size and small sample size) level 2 evidence in each phase. No evidence exists, and hence no recommendation is made, for IPSRT for mania.

2.6.5. Peer interventions

Peer interventions, such as peer groups or one-on-one support, are an important strategy believed to reduce self-stigma and isolation in BD, and to help improve engagement in treatment. Some caution is needed when applying this strategy, however, as there may be risks if the peers delivering the intervention are not adequately trained or supported, and if they promote a viewpoint that does not support treatment compliance or promotes substance use.

Reviews of peer interventions for persons with serious mental illnesses, usually incorporating a small but significant number of individuals with BD, have demonstrated modest evidence from RCTs and other controlled studies suggesting that there are important improvements in self-efficacy and reduction in self-stigma., , , The largest peer intervention study involving BD allocated 153 individuals to attend 21 weekly group psychoeducation events, with another 151 assigned to attend 21 weekly group peer support events. The two programmes achieved similar outcomes in terms of time to relapse, and increased knowledge about BD, although psychoeducation was more acceptable to the subjects and worked more effectively at preventing relapse in a subset of people with fewer previous episodes.

A significant source of peer support is emerging from online resources, particularly through the websites of peer advocacy organizations such as the Depression and Bipolar Support Alliance ( http://www.dbsalliance.org/site/PageServer?pagename=peer_landing), the Mood Disorders Association of Ontario ( https://www.mooddisorders.ca/), the research and advocacy group CREST.BD ( http://www.crestbd.ca/), MoodSwings ( http://www.moodswings.net.au/), and Revivre (http:/wwww.Revivre.org). YouTube is also emerging as an important source of peer support, along with other social media.,

Overall, peer interventions receive a third-line treatment recommendation (level 2) as an adjunctive maintenance therapy.

2.6.6. Other psychosocial interventions

Various other approaches have been tried in BD, with a variety of aims, modalities, and outcome targets. None of the other interventions have been specifically targeted for bipolar depression or for mania. Some have been designed in part to reduce episode recurrence, but none have been successful in providing substantial evidence of efficacy. Because CANMAT recommendations are for the treatment of acute depression and mania, and maintenance treatment to prevent them, we do not make specific recommendations regarding these treatments. However, some of these approaches have been helpful in ameliorating some important symptoms in individuals with BD, (such as residual mood symptoms or anxiety) and so we will describe them briefly.

Although somewhat like FFT, family/caregiver interventions constitute a distinctly different psychosocial intervention in that the intervention is given to the family/caregiver, not the person with BD, and evidence exists that such interventions improve clinical outcomes in the patient., Clinical wisdom and common practice, however, support the importance of family or caregivers being included in at least some sessions with the patient (particularly for psychoeducation), both to reduce symptom burden on the individual with BD and to reduce burnout and emotional burden on the caregiver. Validated caregiver resources are available online, such as http://www.bipolarcaregivers.org.

DBT, which includes distress tolerance training, has several small studies showing its utility in the reduction of some depressive symptoms and suicidality.

One RCT of MBCT involving 95 patients did not demonstrate any difference in relapse prevention compared to a treatment-as-usual group, but did reveal fewer anxiety and depressive symptoms in the MBCT arm. Coupled with the findings of other smaller studies, this suggests that MBCT may have a role to play in anxiety reduction in BD.,

While not reviewed here, given that individuals with BD may have histories of childhood abuse and comorbid personality disorders, and experience various sequelae such as shame or conflict due to behaviours experienced during acute bipolar episodes, all of these may rightly be a target for psychosocial intervention in a very individualized manner.

2.6.7. Cognitive and functional remediation

Functional impairment as well as cognitive deficits are found in many individuals with BD, not just during an acute episode but even between episodes, prompting the evaluation of various psychosocial and biological strategies to address these problems. One intervention, functional remediation (FR), involves a 21-session group intervention over 6 months. In a large RCT, FR was shown to have a substantial impact on functioning, in comparison to treatment as usual. Coupled with the results of other small studies involving other interventions, these findings suggest that there is considerable hope in addressing cognitive and functional deficits in BD. Computer-based cognitive remediation, though, may show positive effects on cognition but not on functioning.

2.6.8. Online and digital strategies

Modern trends to rely on the internet and apps, along with access problems in mental health, have led to the study of various online tools and mobile phone apps. Such strategies also build on strong traditions of self-monitoring and self-management developed formally in traditional psychoeducational interventions. In reviews, such internet and mobile health interventions have shown good adherence to validated psychological health principles, good acceptability to patients, ease of access, and ease of use. However, research is mostly limited to pilot studies and the relatively few larger studies have not shown unequivocal benefit.,

3.3.1. Step 1: review general principles and assess medication status

Examination of a patient presenting in a manic state should include an immediate assessment for risk of aggressive behaviour, violence and safety threat to others, suicide risk especially in those with mixed features, degree of insight and the ability to adhere to treatment, comorbidity (including substance use that may be aggravating or contributing to clinical presentation), and availability of a psychosocial support network. A physical examination with laboratory investigations (described in Section 8) should be conducted, but may be deferred for patients who are uncooperative. Results of the overall assessment should be used to establish the most appropriate treatment setting (eg, ambulatory or inpatient).

Before initiating pharmacological treatment for a manic episode, it is imperative to rule out symptoms secondary to drugs of abuse, medications, other treatments, or a general medical or neurological condition (although, even in these cases, symptomatic treatment may be applied on a short-term basis). Steps should be taken to rule out any other factors that may be perpetuating symptoms such as prescribed medication, illicit drug use/abuse or an endocrine disorder. Any patients presenting with mania who have been taking antidepressants should have these medications discontinued. If there is a previous diagnosis of BD, it is appropriate to immediately commence antimanic agents. If this is the first emergence of manic symptoms, clinicians are advised to confirm the diagnosis of BD by monitoring patients for a period of time after antidepressant discontinuation and obtain collateral information to confirm whether symptoms remain and antimanic treatment is necessary. Patients should also be supported to discontinue stimulant use, including caffeine and alcohol. Current and prior therapies should be assessed, including appropriateness of medications, dosing, and trough serum levels (where indicated), as well as past response; and this should be used to direct subsequent therapeutic choices. Attention should be paid to managing withdrawal symptoms that may occur in manic patients with histories of substance abuse.

When the symptoms of mania have remitted, behavioural and educational strategies should be applied to promote ongoing medication adherence, reduce residual symptoms, help identify early signs of relapse, and support functional recovery (see Section 2).

3.3.2. Step 2: initiate or optimize therapy and check adherence

It is recommended that, for all patients (including those who are untreated as well as those receiving a non-first-line treatment), therapy be initiated with one of the available first-line monotherapy or combination treatments.

3.3.3. Step 3: add on or switch therapy (alternate first-line agents)

If therapy with one or a combination of the first-line agents (lithium, divalproex and/or an atypical antipsychotic) at optimal doses is inadequate or not tolerated, the next step is to switch to or add on an alternate first-line agent. An exception is that, despite level 1 evidence for monotherapy with paliperidone and ziprasidone, we do not recommend combination therapy with these agents due to lack of evidence for additional efficacy (see "No specific recommendation/agents that require further study" below). Because there are multiple first-line agents with substantial efficacy data and relative safety and tolerability, the use of second- and third-line agents is only recommended after unsuccessful trials of multiple first-line strategies.

3.3.4. Step 4: add on or switch therapy (second-line agents)

3.3.5. Step 5: add on or switch therapy (third-line agents)

3.3.6. Agents not recommended for the treatment of acute mania

Antimanic efficacy has not been demonstrated for allopurinol (level 1 negative), eslicarbazepine/licarbazepine (level 2 negative), gabapentin (Level 2 negative), lamotrigine (level 1 negative), omega-3 fatty acids (level 1 negative), topiramate (level 1 negative), valnoctamide (level 2 negative),, or zonisamide (level 2 negative) (Table 13).

3.3.7. No specific recommendation/agents that require further study

Trials with paliperidone (level 2 negative) and ziprasidone (level 2 negative) adjunctive therapy to lithium or divalproex showed lack of efficacy. This is surprising given that all other atypical antipsychotic agents that showed efficacy in monotherapy have also been shown to offer additional benefit when combined with lithium or divalproex. It is likely that methodological problems have contributed to failure in these studies; hence, further studies are needed before specific recommendations can be made about the use of these combinations for mania.

Studies of olanzapine (level 2 negative) or risperidone (level 3 negative) plus carbamazepine have been negative, although this is probably due to enzyme-inducing effects of carbamazepine. While this may be overcome by dosing adjustments, because such interactions are unpredictable and effective doses have not been established, we are unable to provide a specific recommendation.

Nutraceuticals such as branched chain amino acids (level 3), folic acid (level 2), and l-tryptophan (level 3), as well as other experimental agents such as medroxyprogesterone (level 3),, memantine (level 4), mexiletine (level 4), levetiracetam (level 4) and phenytoin (level 3), have all shown indications of efficacy when used adjunctively with other antimanic agents, as have glasses that block blue light (level 3). Larger controlled trials are needed, however, before a recommendation for their use in mania can be made. While an initial small RCT did not show anti-manic efficacy for verapamil, there is some evidence that it may work as an adjunctive therapy (level 4) or as monotherapy in women (level 4). Larger studies are needed before a conclusion can be made.

3.3.8. Clinical features that help direct treatment choices

Clinical features, including DSM-5 specifiers, may assist in making treatment choices between first and second-line treatment options.

In general, lithium is preferred over divalproex for individuals who display classical euphoric grandiose mania (elated mood in the absence of depressive symptoms), few prior episodes of illness, a mania-depression-euthymia course,, , and/or those with a family history of BD, especially with a family history of lithium response. Divalproex is equally effective in those with classical and dysphoric mania. Further, divalproex is recommended for those with multiple prior episodes, predominant irritable or dysphoric mood and/or comorbid substance abuse or those with a history of head trauma., , , , , Because of its teratogenic potential, however, caution should be exercised when prescribing divalproex to women of childbearing age. Patients with specific factors such as a history of head trauma, comorbid anxiety and substance abuse, schizoaffective presentations with mood-incongruent delusions, or negative history of bipolar illness in first-degree relatives may respond to carbamazepine.

Combination therapy with lithium or divalproex and an atypical antipsychotic is recommended when a response is needed faster, in patients judged at risk, who have had a previous history of partial acute or prophylactic response to monotherapy or in those with more severe manic episodes.

4.2.1. Misdiagnosis and delayed diagnosis

Patients with depression occurring in the context of BD are frequently misdiagnosed as having MDD, since the presence of mania or hypomania (particularly mild or moderate episodes which do not require hospitalization) may be challenging to establish retrospectively. This is especially true in the absence of a comprehensive diagnostic interview or collateral information, as patients may often lack basic knowledge of what hypomania/mania is, and/or have limited insight into these symptoms; and thus may not disclose this information unless specifically asked. Alternatively, patients who will ultimately present with hypomanic or manic episodes may only have experienced episodes of depression. Thus, clinicians must be vigilant for a diagnosis of BD, and routinely ask for symptoms of a previous manic/hypomanic episode in every patient presenting with a major depressive episode. A diagnosis of MDD should be made only after excluding the possibility of BD.

In addition to overt manic/hypomanic symptoms, there are numerous features that increase the likelihood of a diagnosis of BD in depressed individuals. These include earlier age of illness onset (before 25 years), brief, highly recurrent depressive episodes, a family history of BD, depression with psychotic features, atypical features such as reverse vegetative symptoms of hypersomnia and hyperphagia, leaden paralysis, psychomotor agitation, postpartum depression or psychosis, and antidepressant-induced irritability, manic symptoms or rapid cycling, (Table 6).

Individuals with depression who are at high risk for BD, particularly those with a strong family history of BD, should be closely monitored for emergence of manic or mixed symptoms. Consideration should also be given to applying the BD depression treatment recommendations amongst those at very high risk, rather than risk potential iatrogenic effects of antidepressant monotherapy, although this recommendation is based on clinical experience as there is a lack of sufficient research addressing this issue. As discussed in Section 2, there are also several useful psychosocial interventions, such as individual and family psychoeducation and FFT, that have been shown to have some benefit in this population.

4.2.2. Suicide risk

Principles related to management of suicidal ideation and risk (see Section 2 and) are of utmost importance during depressive episodes, as >70% of suicide deaths and suicide attempts in patients with BD occur during this phase., Depressive episodes with mixed features are a particularly dangerous period associated with even higher short-term risks of suicide attempts or death. Overall, it is imperative for clinicians to review risk factors (Table 9) and determine an appropriate treatment setting to address any safety issues. All patients at risk should be encouraged to develop and share a written safety plan listing coping strategies and sources of support which may be applied during times of crisis. As described in Section 2, the most common method of suicide in this population is self-poisoning, and so potential benefits of various treatments should be considered against their risk of toxicity and lethality. One study found that there were fewer deaths due to lethal lithium levels compared to carbamazepine, and that opioids and benzodiazepines were the most common medication classes ingested at lethal levels--noteworthy given the lack of efficacy of these agents in the disorder.

4.2.3. Cognitive and functional impairment

Part of the impact of acute and subsyndromal depressive symptoms on functional impairment is thought to be mediated through cognitive performance, which is both subjectively and objectively impaired in bipolar depression and linked to poor psychosocial function., , ,

Because of the important link between cognition and functioning, attention should be paid to avoiding treatments that may further exacerbate cognitive difficulties (see Section 8). Although evidence for their efficacy is limited, cognitive enhancement therapies can be considered experimental in this population., ,

4.4.1. Step 1: review general principles and assess medication status

Examination of a patient presenting in a depressed state should include an assessment of the nature and severity of depression and associated symptoms, risk of suicide/self-harm behaviour, ability to adhere to a treatment plan, availability of a psychosocial support network, and functional impairment. Laboratory investigations (described in Section 8) should also be completed. Results of the overall assessment should be used to establish the most appropriate treatment setting (eg, ambulatory or inpatient), with consideration given to management of safety risks. Before initiating pharmacological treatment for a depressive episode, it is imperative to rule out symptoms secondary to alcohol/drug use, medications, other treatments, or a general medical condition. Patients should be supported to discontinue stimulant use and limit nicotine, caffeine, drug, and alcohol use. Course of illness and treatments used in current and prior episodes should be assessed, including past response to and tolerability of specific medications and doses, and used to direct subsequent therapeutic choices. Consideration should be given to restarting medications if their recent discontinuation appeared to coincide with a depressive relapse.

Psychoeducation and other psychosocial strategies should also be offered alongside pharmacological treatment to promote ongoing medication adherence, reduce residual symptoms and suicidal behaviour, help identify early signs of relapse, and support functional recovery (see Section 2).

4.4.2. Step 2: initiate or optimize therapy and check adherence

It is recommended for all patients that pharmacotherapy be initiated with one or more of the available first-line agents. The choice of agent or agents to manage an acute bipolar depressive episode should be discussed with the patient and their supporters (as appropriate) and take into account current and prior medication use and response, personal preference, and the safety and tolerability of each agent, as well as clinical features that may influence prognosis (see "Clinical features that help direct treatment choices").

4.4.3. Step 3: add on or switch therapy (alternate first-line agents)

Across several different medications for bipolar depression, early improvement (after 2 weeks) has been found to be a reasonable predictor of overall response, whereas lack of early improvement is a more robust predictor of non-response. Lamotrigine is the exception to this rule, given a necessary slow titration initiating the medication. In the case of non-response, dosing should be optimized and issues of non-adherence identified and addressed (see Section 2) prior to adjusting treatment strategies.

When determining whether an agent should be switched or another first-line agent be added on to any current treatment, the effectiveness of each of the medications needs to be understood in the context of all the goals of managing BD. It is often the case that a medication may be selected to address several goals; for instance, lithium could be added for acute depression with intent to also bolster anti-manic prophylaxis. In this scenario, if lithium is ineffective in the individual patient for an acute bipolar depression but is also being used over the long term for anti-manic prophylaxis, then an "add-on" intervention should be the next treatment for the acute bipolar depression. If, for instance, the anti-manic prophylaxis is already being fully provided by an atypical antipsychotic, then the new medication could replace lithium via a switch strategy. Decision-making must also address efficacy for comorbid conditions, as well as tolerability concerns. In principle, all things being equal, a switch is preferred over add-on to limit the degree of polypharmacy, but the clinical reality is that medications may be helpful for some but not all components of the illness, and using rational polypharmacy via add-on treatments is often required. For situations in which patients experience a depressive episode while already receiving an adequately dosed antidepressant, strong consideration should be given to discontinuing or switching the class of antidepressant, unless clear benefits are apparent in reducing the severity or frequency of depressive episodes. Switch of medications should be done in an overlap and taper manner unless there is medical necessity for abrupt discontinuation.

All first-line options should be tried in adequate doses for an adequate duration of time before considering second-line options either as an add-on or switch strategy.

4.4.4. Step 4: add on or switch therapy (second-line agents)

4.4.5. Step 5: add on or switch therapy (third-line agents)

4.4.6. Agents not recommended for the treatment of acute bipolar depression

Antidepressants should not be used as monotherapy in patients with BDI depression, as available trials do not support their efficacy and there are concerns about their safety in terms of mood switching (level 2 negative)., , ,

Aripiprazole monotherapy failed to separate from placebo in two bipolar depression trials. Although the pooled analysis reported separation, the mean difference in Montgomery-Åsberg Depression Rating Scale (MADRS) change score was only 1.12 points, which is not clinically meaningful and it is thus not recommended (level 1 negative). Ziprasidone monotherapy or adjunctive therapy (level 1 negative),, lamotrigine in combination with folic acid (level 2 negative), and mifepristone (adjunctive) (Leve 2 negative) are also not recommended due to evidence for lack of antidepressant efficacy (Table 15).

4.4.7. No specific recommendation/agents that require further study

There are insufficient data to make a recommendation regarding the use of aspirin (adjunctive) (level 3 negative), celecoxib (adjunctive) (level 3 negative), gabapentin (monotherapy) (level 3 negative), leviteracitam (adjunctive) (level 3 negative), lisdexamfetamine (adjunctive) (level 3 negative), memantine (adjunctive) (level 3), pioglitazone (adjunctive) (level 3),, riluzole (level 4 negative), and risperidone (adjunctive) (level 3). Although adjunctive therapy with pregnenolone separated from placebo at week 6, the change in depressive symptoms was not significantly different from week 8 to week 12 between the two groups (level 2).

Need for rapid response

Amongst the first-line options recommended, quetiapine and lurasidone have separated from placebo in clinical trials at as early as week 1., , Thus, these medications may be preferable when a rapid response is required, for example in patients who are at increased risk of suicide or who have medical complications, including dehydration. While ECT is recommended as a second-line option, this may also be used earlier when a rapid response is imperative. Second-line options such as cariprazine and olanzapine-fluoxetine have also separated from placebo at as early as week 1 and may also be considered when a rapid response is desirable, but this needs to be balanced against the potential side effects. Lamotrigine administration requires slower titration due to the risk of skin rashes, Stevens-Johnson syndrome, and toxic epidermal necrolysis, and is thus not ideal for patients requiring a rapid response. Lamotrigine, however, is well tolerated, and there is some evidence that its effectiveness may be more pronounced in patients experiencing depressive cognitions and psychomotor slowing.

Previous treatment response

Adjunctive antidepressant use may be appropriate in those with prior antidepressant response if there was no history of treatment-emergent switch.

Anxious distress

Symptoms of anxiety are often experienced during a depressive episode, and are predictive of more persistent depressive symptoms and increased suicidal ideation. A pooled analysis of two double-blind RCTs demonstrated that quetiapine is more effective than placebo in relieving symptoms of anxiety co-occurring alongside bipolar depression, and olanzapine-fluoxetine combination has also been shown to be effective. In a post hoc analysis, lurasidone was effective in improving depressive as well as anxiety symptoms in patients with MDD who had mixed features and anxiety. The anxiolytic effects of divalproex, risperidone, and lamotrigine appear to be limited.,

Mixed features

Many patients with bipolar depression will also experience at least subsyndromal hypomanic or manic features, and this presentation is associated with more severe depressive symptoms, as well as a higher rate of substance use and cardiovascular disease. For many of these patients, combination therapy will be necessary to adequately address symptoms. Pooled analysis indicates that atypical antipsychotics show a class effect in alleviating mixed features in bipolar depression, with olanzapine-fluoxetine combination, asenapine, and lurasidone all demonstrating efficacy. Lurasidone has further been shown to have efficacy in treating both depressive and hypomanic symptoms in MDD with mixed features. The ISBD Task Force recommends avoiding antidepressants in patients with mixed features and the CANMAT/ISBD group concurs with this recommendation.

Melancholic features

No specific studies assessed the predictive ability of melancholic features; however, clinical experience suggests that ECT is very effective in this population.

Atypical features

There is some evidence for efficacy of tranylcypromine in patients with anergic bipolar depression. However, given the risks of potential manic switch, this agent should only be used in conjunction with lithium or divalproex or an atypical antipsychotic. Clinicians also must consider adverse events of this agent related to its interactions with food and other medications.

Psychotic features (mood congruent or incongruent)

Up to 20% of inpatients experience psychosis in the context of an acute bipolar depressive episode. The relative efficacy of various medications to treat these features in this phase of illness has not been examined, although clinical experience suggests that ECT and antipsychotics are highly effective for this population.

Rapid cycling

As described in Section 3, hypothyroidism, antidepressants and substance abuse may be associated with rapid cycling, thus making assessment of thyroid function and discontinuation of antidepressants, drugs of abuse, stimulants, and other psychotropic agents imperative. As there is no evidence to support any specific agent to treat acute depression during a rapid cycling phase, appropriate pharmacotherapy should be selected based on effectiveness in the acute and maintenance phases. Lithium, divalproex, olanzapine, and quetiapine all appear to have comparable maintenance efficacies in these patients. In contrast, lamotrigine did not separate from placebo in maintenance treatment in patients with rapid cycling BDI. Antidepressants are not recommended, as they have been shown to destabilize patients, even with concurrent mood stabilizer use.

Seasonal pattern

While some individual patients may show a seasonal pattern, Canadian data are mixed as to whether episodes of mania or depression in BD follow a consistent seasonal variation. There is no evidence for the superiority of any agent in patients with an observed seasonal pattern of manic episodes.

5.5.1. Step 1: review general principles and assess medication status

Many agents recommended for management of acute manic or depressive episodes have prophylactic efficacy. Generally, medications that have been found to be effective in the acute phase should be continued during the maintenance phase. However, there are exceptions to this: the efficacy of adjunctive antidepressant therapy has not been examined systematically in large double-blind placebo-controlled trials; hence, long-term antidepressant use is not recommended, especially in light of the concerns about potential risk of manic/hypomanic switch and mood instability. However, in the subgroups of patients who have responded to combination treatment and are stable, preliminary evidence suggests that withdrawal of antidepressants may contribute to destabilization.

Clinical trials have shown that many atypical antipsychotics are effective in preventing relapse of mood episodes; with many agents, this efficacy is related to prevention of manic episodes but not depressive episodes. However, many of these trials have been conducted in patients with an index manic episode and, given that the polarity of an index episode predicts the polarity of relapse, depressive relapse rates in placebo groups in such studies have been low- compromising the statistical power to test the efficacy of these agents in prevention of depressive relapses. Thus, the efficacy of many of these agents in preventing depressive relapses remains unknown.

For patients who are currently not receiving or responding to pharmacological treatment, a careful history including details of clinical course, response (or lack thereof) to previously used medications, and family history should be collected. Other variables to be considered include psychiatric comorbidity (including substance use), the predominant illness polarity, and the polarity of the most recent episode.

Ongoing clinical monitoring, including medication blood levels as appropriate, is also a crucial part of maintenance treatment that should be used to support enhanced medication adherence, detection of early symptoms of recurrence, and monitoring of side effects (see Section 8).

5.5.2. Step 2: initiate or optimize therapy and check adherence

The choice of agent or agents used in maintenance treatment should be discussed with the patient and their caregivers (as appropriate) and, based on knowledge of current and prior medication use and response, safety and tolerability of each agent, predominant episode polarity, and clinical features that may influence prognosis (see "Clinical features that help direct treatment choices"). As with treatment for mania and acute depression, we recommend that treatment choices for maintenance treatment of BD should follow the hierarchy listed in Table 17 unless patient preference or other considerations such as previous response/non-response, tolerability or predominant polarity justify other choices. Similarly, as a general rule, if a patient has been treated for an acute mood episode and responded to a first-line maintenance treatment, we recommend continuing that treatment for maintenance even if lower down in the hierarchy. As an example, if a patient responded to asenapine in an acute manic episode, asenapine should be continued, even if it is lower down in the hierarchy for maintenance treatment. It may be necessary to lower the dose to some degree once in maintenance treatment as patients often experience greater side effects once out of the acute episode.

There is evidence that the risk of recurrence is reduced when an antipsychotic is combined with lithium/divalproex. When a combination therapy of an atypical antipsychotic with lithium/divalproex was used to treat acute mania, continuing the atypical antipsychotic for the first 6 months following response offered clear benefit in reducing risk of mood episode recurrence (level 2), but the benefits beyond 6 months remain uncertain. Therefore, clinicians are advised to re-evaluate risks and benefits after 6 months of sustained response to determine whether maintenance combination therapy with an atypical antipsychotic is justified.

5.5.3. Step 3: add on or switch therapy (alternate first-line agents)

If therapy with one or a combination of the first-line agents at optimal doses is inadequate or not tolerated, the next step is to switch to or add on an alternate first-line agent. Because there are multiple first-line agents with substantial efficacy data and relative safety and tolerability, the use of second-line agents is only recommended after unsuccessful trials of multiple first-line strategies.

5.5.4. Step 4: add on or switch therapy (second-line agents)

5.5.5. Step 5: add on or switch therapy (third-line agents)

5.5.6. No specific recommendation/agents that require further study

We do not provide specific recommendations for the use of cariprazine, as there is currently only evidence for efficacy in acute manic and depressive episodes, and not yet for maintenance treatment (level 4). While a small RCT suggests a lack of efficacy for flupenthixol as maintenance treatment, larger studies are needed before definite conclusions can be drawn (level 3 negative). Likewise, oxcarbazepine adjunctive therapy requires further evaluation (level 4)., , No recommendation is made for topiramate as there is an absence of controlled data supporting its efficacy in maintenance (level 4 negative), and a lack of efficacy in acute mania; however, its use may be indicated as it has efficacy for many syndromes that are often comorbid with BD (Section 7).

5.5.7. Agents not recommended for maintenance treatment

Perphenazine is not recommended for maintenance based on evidence that patients treated with perphenazine and a mood stabilizer following an episode are more likely to have emergent depressive symptoms or intolerable side effects, compared to those maintained on the mood stabilizer alone (level 2 negative). Tricyclic antidepressant mono or adjunctive maintenance therapy is not recommended due to an increased risk of manic switch (level 2 negative), , (Table 18).

5.5.8. Clinical features that direct treatment choices

Clinical trials tell us how efficacious one drug is in comparison with another (or placebo) in groups of patients. To determine the degree of long-term response in an individual patient requires a different evaluation and may take a considerable amount of time. Few patients manage a lifetime of BD with monotherapy--most will require short- or long-term combination therapies to address acute or subsyndromal symptoms as well as to reduce rates of recurrence. Some,, , but not all,, reports suggest that long-term treatment becomes less effective with longer duration of untreated illness, an argument for finding an effective treatment as early as possible.

In most instances, it is difficult to differentiate nonspecific correlates of good prognosis of the illness from factors specific to the response to a particular mood stabilizer. Available data come mostly from naturalistic/cohort studies and few randomized trials. Nevertheless, several tentative predictors are emerging from the available data.

Factors associated with overall good prognosis of BD include good treatment adherence, lack of early adversity, intermediate age at onset, good social support, and the absence of spontaneous rapid cycling, or features of a personality disorder.

In general, lithium is the gold standard for maintenance treatment, as it is effective in preventing both manic and depressive episodes (magnitude of prophylactic efficacy greater against mania vs depression) and appears to have a degree of anti-suicidal effects., , , , , , Patients who respond well to lithium treatment usually have an episodic remitting pre-treatment clinical course, a family history of BD (especially BD responsive to lithium), low rates of comorbidity (especially anxiety and substance abuse disorders), and a pattern of mania-depression-euthymia in biphasic episodes, as well as a typical clinical presentation., , Responsiveness may also be a familial trait, with a study showing that patients who have a lithium-responsive relative have a 67% likelihood of also being lithium responsive, versus 35% of those without a responsive relative. Among biological measures, lack of electroencephalogram (EEG) abnormalities, higher brain lithium concentration, increased N-acetyl aspartate and lower myo-inositol peaks on magnetic resonance spectroscopy, as well as several variants in candidate gene studies, may predict response, but these studies require confirmation. Response to lithium in particular seems to be quite specific, as shown in a study of neurons derived from induced pluripotent stem cells. The neurons from people with BD were hyperexcitable and their activity was selectively modified by in vitro lithium in accordance with clinical response.

Responders to lamotrigine have a predominantly depressive polarity as well as comorbid anxiety., Lamotrigine monotherapy is not appropriate for patients with frequent manic episodes, as it has limited efficacy in preventing mania.

Quetiapine has been shown to be effective in preventing manic, depressive and mixed episodes in patients with index manic, depressive and mixed episodes, and thus may be particularly valuable in those with mixed features. Asenapine appears to be effective in preventing both mania and depression, although the magnitude of prophylactic efficacy is greater for mania relative to depression. In a randomized open trial of carbamazepine versus lithium, responders to carbamazepine were more likely to have an atypical illness, BDII or schizoaffective disorder.

Data to differentiate anti-psychotic medication responders from non-responders are lacking.

Overall, some of these possible predictors can have clinical utility, but not all are practical. For instance, it is difficult to evaluate a pre-treatment course in patients who started their treatment after one or two episodes (practice recommended by most treatment guidelines), and biomarkers are intriguing but lack sufficient replication and are not readily available.

In patients with a history of a rapid cycling course, as indicated in previous sections, factors associated with rapid cycling must be addressed. These include discontinuation of stimulants and antidepressants and treating hypothyroidism if present. With regard to treatment options, the evidence suggests that monotherapy with a single mood stabilizer is often ineffective and patients may require a combination of mood stabilizers to achieve mood stability.

6.2.1. General considerations for interpreting recommendations

The treatment of BDII has been understudied relative to BDI. This is probably due to the long-standing but discredited impression of BDII as a less severe form of BD. The number of RCTs in BDII is substantially smaller than that in BDI, and those studies that do exist are frequently under-powered. It also remains common for trials to enrol patients with both BDII and BDI without reporting results separately, making it difficult to determine if there are clinically meaningful differences in treatment response between the two illnesses. This is important because, while clinical experience and the results of many studies suggest that response to mood stabilizers and antipsychotics is similar in BDII and BDI, there are enough exceptions to suggest this should not be taken for granted., , This is also the case for antidepressants, which may have a more favourable risk-benefit ratio in BDII (reviewed below). Therefore, in formulating our recommendations, studies that enrolled patients with both BDII and BDI but did not report results for BDII separately are assigned level 4 status (expert opinion) if the proportion of patients with BDII was less than 50%.

The relative paucity of large, methodologically sound clinical trials in BDII creates challenges in formulating evidence-based guidelines. As will be seen, there are fewer treatments with high-quality evidence in BDII compared to BDI and fewer first-line treatment recommendations. The limitations of the evidence base necessitate an awareness of the nuances of the available studies, and a greater reliance on clinical experience. We thus have endeavoured to be clear in outlining the rationale for selecting first, second, and third-line treatments for hypomania, depression, and maintenance treatment of this important illness. There is clearly a pressing need for adequately powered trials in BDII across all illness phases.

6.2.2. Acute management of hypomania

The general principles for assessing mania apply to hypomania. For some patients, hypomania causes no to minimal functional impairment and may even be associated with brief periods of above-normal functioning. However, prolonged, relatively severe, or mixed or irritable hypomania may be impairing. Treatment should include discontinuing agents that can worsen or prolong symptoms, including antidepressants and stimulants, and initiating appropriate pharmacotherapy.

Unfortunately, many standard medications for mania, including lithium and most atypical antipsychotics, have not been studied in hypomania. There are four placebo-controlled trials that investigated divalproex (level 4), N-acetylcysteine (level 4), and quetiapine (level 4), ; and one open-label study of risperidone (level 4) in acute hypomania. The studies generally suggested efficacy but all had significant weaknesses, including one or more of: (i) small sample sizes, (ii) mixed samples with BDI, BDII, and BD NOS; (iii) mixed samples with hypomania and mania, and (iv) positive findings on some but not all outcomes. The small numbers of patients and mixed samples mean that even the placebo-controlled trials only met criteria for level 4 evidence (Table 1).

These methodological limitations, coupled with the lack of clinical trial evidence for many medications, make it difficult to make specific suggestions for the treatment of hypomania. Clinical experience suggests that all anti-manic medications are also efficacious in hypomania. Thus, when hypomania is frequent, severe, or impairing enough to require treatment, clinicians should consider mood stabilizers such as lithium or divalproex and/or atypical antipsychotics. N-acetylcysteine may also be of benefit, but further studies are needed.

6.2.3. Acute management of bipolar II depression

The general principles for assessing depression in patients with BDI apply to those with BDII. First, second, and third-line treatment options are listed below and shown in Table 19. Specific considerations regarding each treatment are highlighted in the relevant sections.

6.2.4. Maintenance treatment

Maintenance treatment is important to prevent relapse, reduce subsyndromal symptoms, and improve quality of life. As with BDI, selection of an agent should be informed by acute phase treatment. Recommended agents and their evidence ratings are listed in Table 20.

7.1.1. Pre-conception, psychoeducation and contraceptive counselling

The importance of pre-conception counselling should be raised with all women of child bearing age. It should be provided for all patients at least 3 months prior to considering pregnancy or immediately for those already pregnant. The issues most frequently raised are fear of adverse effects of medications on the fetus, fear of illness recurrence, and genetic transmission to offspring. Other important topics to review include the effect of BD on risk for gestational hypertension, antepartum haemorrhage, induction of labour, caesarean section, instrumental delivery, preterm delivery, and neonatal size., Discussion of modifiable risk factors is critical in pre-conception management of BD. For instance, pregnant women with BD are more often overweight, more often smoke tobacco during pregnancy, have poorer diet quality, and present more often with a history of drug and alcohol misuse in pregnancy. Modification of these risk factors may have a significant positive impact on outcomes for both the mother and child.

Decisions should be made collaboratively on whether medications should be continued, discontinued, or switched; and whether any dosage changes are needed. Conventional antipsychotics and risperidone may need to be discontinued to increase the likelihood of conception, as these medications often increase serum prolactin levels and thus interfere with ovulation and decrease fertility. For women who wish to have a medication-free pregnancy, it might be appropriate to have one or more psychotropic medications gradually tapered off prior to conception provided they have been clinically stable for a minimum of 4-6 months and are considered at low risk of relapse. Information regarding potential teratogenic effects of different psychotropic medications, as well as limitations of the scientific evidence, should be discussed and carefully considered. The decision to stop medications pre-conceptually should ideally occur only after careful individualized risk-benefit analysis for a given patient., , If pharmacotherapy is required, monotherapy at minimum effective dose is recommended whenever possible.,

Contraceptive counselling, including emphasis on its effectiveness in reducing the likelihood of unintended pregnancies, should be included as part of a comprehensive treatment plan for women with BD.

Several anticonvulsants, including carbamazepine, topiramate, and lamotrigine, can affect the pharmacokinetics of oral contraceptives and some might significantly reduce the effectiveness of oral contraceptives, and this should be considered when making treatment decisions., Oral contraceptives might also have effects on the efficacy of lamotrigine via reduction in lamotrigine levels.

While folic acid supplementation is protective against spontaneous spina bifida, there is not enough evidence to indicate that folic acid, even in high doses, protects against spina bifida following the use of anticonvulsants. In addition, Health Canada recommends that "Valproate products (valproic acid, divalproex sodium) should not be used in female children, in female adolescents, in women of childbearing potential and pregnant women unless alternative treatments are ineffective or not tolerated because of its high teratogenic potential and risk of developmental disorders in infants exposed in utero to valproate." Health Canada also recommends that women of childbearing potential must use effective contraception during treatment and be informed of the risks associated with the use of valproate products during pregnancy, and in women planning to become pregnant every effort should be made to switch to appropriate alternative treatment prior to conception.

Regardless of treatment decisions made prior to or following conception, it is also important to work with the patient to develop and agree upon a monitoring schedule and treatment plan to be implemented should clinically significant symptoms emerge. Education about the risks of psychotropic medications is critical, and careful discussion is needed regarding the magnitude of risks and benefits and limitations with the data. It is important to acknowledge the patient's desire to do what is right for the child and support the decision made. Whenever appropriate, involve partners in the discussion and review the decision and supporting evidence for it directly or through educational materials.

7.1.2. Screening for bipolar disorder during pregnancy and postpartum

All women with depressive symptoms should be screened for BD during pregnancy and the postpartum period., Standardized screening tools such as the Mood Disorder Questionnaire alone or in conjunction with the Edinburgh Postnatal Depression Scale are useful., , , Importantly, screening should be followed by a clinical interview to confirm or exclude a diagnosis of BD. Women should also be assessed for other psychiatric disorders that commonly co-occur with BD, such as anxiety disorders or obsessive compulsive disorder (OCD).

7.1.3. Pharmacological management of bipolar disorder during pregnancy

Given the complexity of the risks faced by women with BD in pregnancy and puerperium, it is good clinical practice to encourage liaison between the mental health and obstetrics/gynecology teams. A longitudinal study conducted in a tertiary care centre found a high risk of recurrence during pregnancy: 85% of pregnant women with BD who discontinued a mood stabilizer and 37% of those who were maintained on one or more mood stabilizers experienced a mood episode--predominately depressive or mixed--during pregnancy. For nearly half of the patients, recurrence occurred in the first trimester, with the median time for recurrence for those abruptly discontinuing treatment being 2 weeks, compared to 22 weeks for those who were gradually tapered off. However, studies from primary care, as well as obstetric centres, found low rates of relapse or hospitalizations in pregnancy.

The hierarchies presented throughout these guidelines should be followed for management of the various phases of BD, with consideration given to specific risks associated with the use of each medication using the most up-to-date information available from the US Food and Drug Administration (FDA) website ( http://www.fda.gov/ForConsumers/ByAudience/ForWomen/WomensHealthTopics/ucm117976.htm).

Table 21 includes a brief overview of medications commonly used in BD and the risk categories. This list should not be viewed as complete or comprehensive. Further, the FDA has replaced these risk categories in 2015 with Pregnancy and Lactation Labeling Final Rule (PLLR) with narrative sections and subsections. For instance, for pregnancy, the information for each medication is provided using the following subheadings: Pregnancy exposure registry, Risk summary, Clinical considerations, and Data. The last three subsections apply to medication risks during lactation. The FDA has not finalized and published the data for all medications as of completion date of these guidelines (February 2018), and new data appear to suggest that risks may have been overestimated for some medications such as lithium. Thus clinicians are strongly advised to use all the current data, including the FDA PLLR information if available, in collaboration with patient and family members to make final treatment decisions.

Wherever possible, psychosocial strategies should be preferred over medications in the first trimester as this period holds the highest risk for teratogenicity. When medications are deemed necessary, preference should be given to monotherapy using the lowest effective dose.

Each pregnancy should be closely monitored and appropriate screening tests (eg, foetal ultrasound if lithium is used in the first trimester) should be performed. Divalproex should be avoided during pregnancy due to elevated risk of neural tube defects (up to 5%), even higher incidences of other congenital abnormalities, and evidence of striking degrees of neurodevelopmental delay in children at 3 years of age and loss of an average of nine IQ points., , Because of changes in physiology in the second and early third trimesters, such as increased plasma volume, hepatic activity, and renal clearance, patients may require higher doses of medications towards the later part of the pregnancy. Prenatal vitamins, including high-dose (5 mg/day) folic acid, are also recommended, preferably even before conception and continuously through pregnancy; and preparations containing choline have recently been recommended as possibly preventive of the later development of schizophrenia. While it is important to note that folic acid may reduce the effectiveness of lamotrigine, the anti-teratogenic effects of folate may outweigh the potential for this loss of effectiveness. However, recent concerns have been raised regarding a potential association between very high plasma levels of maternal folate and risk of autism spectrum disorders.

7.1.4. Pharmacological management of bipolar disorder during the postpartum period

The postpartum period is a time of elevated risk for recurrence, with 66% of women who were medication free during pregnancy and 23% of those on treatment experiencing a mood episode following the delivery. The risk of postpartum relapse is highest in women who also experienced a mood episode during pregnancy and those who are not on prophylactic treatment. Despite the high prevalence of postpartum episodes, there is a dearth of studies investigating the efficacy of medications during this period. There is evidence of efficacy of benzodiazepines, antipsychotics, and lithium for postpartum mania, and quetiapine for postpartum bipolar depression (level 4). There are no studies of psychotherapy in the acute or preventative treatment of bipolar postpartum depression.

Patients should be encouraged to initiate or optimize maintenance treatment as soon after giving birth as possible, with preference given to medications that have previously been shown to be successful. Near delivery, close monitoring is essential for early detection and management of symptoms that might signal onset of a mood or psychotic episode. If an acute mood episode emerges in the postpartum period, the hierarchies for non-postpartum episodes should be followed, but, because most psychotropic medications are excreted in breast milk, treatment choice should take into consideration safety in breastfeeding when applicable.

The FDA website mentioned in the previous section as well as Table 21 also include information on lactation. The FDA PLLR should be consulted for further information about medication risks, as many are secreted in milk, if breastfeeding is being considered.

The potential risks and benefits of taking medications while breastfeeding should be discussed with the patient. Education about early recognition of drug toxicity and requirement for ongoing monitoring of infants is also critical. A recent systematic review suggested quetiapine and olanzapine as preferred choices for breastfeeding, considering their relatively lower infant dosages. The impact of medication on the infant can be reduced by scheduling medication administration after breastfeeding.

Replacing or supplementing breast milk with formula can also be considered. Although there are many benefits to breastfeeding, associated sleep disruption may increase the risk of mood episodes in women with BD. If possible, bottle feeding at night by the woman's partner or a support can be beneficial to allow the woman to maintain a better sleeping schedule. In women with postpartum psychosis or mania, breastfeeding may be more risky, and therefore may not be indicated, as the mother may be too disorganized to safely breastfeed.

As childbirth can be a trigger for first onset of hypomania/mania in women with MDD, antidepressants should be used cautiously, especially in women with a family history of BD. Women with first onset of depression in the postpartum period or those who have recurrence of depression during the early postpartum period, may also be at a high risk of switching to BD following treatment with antidepressants.

7.1.5. Impact of the menstrual cycle on symptoms

Despite the paucity of large, well-designed research studies examining the impact of the menstrual cycle on mood symptoms in BD, accumulating evidence suggests that hormonal changes can impact the course of illness. Several case reports and prospective studies suggest that women who experience premenstrual symptom exacerbation are more likely to have a highly symptomatic and relapse prone illness., One of the largest studies (n = 1099) found that women who met DSM-5 provisional criteria for premenstrual dysphoric disorder (PMDD) had an earlier illness onset, more comorbid Axis I disorders, a higher number of hypomanic/manic and depressive episodes, and higher rates of rapid cycling. In this study, there was a closer gap between BD onset and age of menarche in women with comorbid PMDD, which suggests that sensitivity to endogenous hormones may influence the onset and the clinical course of BD. Premenstrual syndrome (PMS) and PMDD also occur more frequently in women with BD., Importantly, an accurate diagnosis of comorbid PMDD in women with BD is made during euthymia, with a minimum of 2 months of prospective symptom charting.

7.1.6. Menopause

For many women, stress and hormonal changes associated with the transition to menopause may increase or trigger mood symptoms., , A post hoc analysis of the prospective Systematic Treatment Enhancement Program for Bipolar disorder (STEP-BD) study showed increased rates of depressive, but not manic episodes during menopause transition. However, due to the paucity of clinical trials in this area, more data are needed before treatment recommendations can be made.

7.2.1. Presentation and diagnosis

As the following section comprises only a brief overview of the epidemiology, phenomenology, and differential diagnosis of BD in children and adolescents, the reader is referred to more detailed reviews for further information., , , , ,

Between one-third (community samples) and two-thirds (clinical samples) of patients with BD experience their first mood episode during childhood or adolescence, with an earlier onset related to a more severe illness characterized by increased symptom burden and comorbidity., In contrast to the controversies of as recently as a decade ago, there is now far greater consensus in the field that, although there are developmental differences in the manner in which symptoms manifest themselves, the actual diagnosis of BD in children and adolescents should be made based on the same set of symptoms as applied to adults. When defined rigorously according to DSM-5 criteria, the course of illness in childhood and adolescence is characterized by eventual high rates of symptomatic recovery, but also high rates of recurrence, even in the context of naturalistic treatment. While the concepts of "over-diagnosis" and "over-treatment" in pediatric BD have received substantial attention, representative population studies demonstrate that adolescent BD is characterized by low rates of treatment, alongside high rates of suicidality and comorbidity., Risks of incorrectly diagnosing and treating BD in a child or adolescent should thus be carefully weighed against the risk of incorrectly or not diagnosing or treating, keeping in mind that the duration of treatment delay has been shown to be an independent risk factor for a poor outcome in adulthood.

Distinguishing early-onset mania or hypomania from other psychiatric disorders is important as there is a high level of symptomatic overlap for multiple conditions including but not limited to ADHD, oppositional defiant disorder (ODD), disruptive mood dysregulation disorder (DMDD), substance abuse, personality disorders and generalized anxiety disorder, (Table 22). The discrete episodes of mania/hypomania and the non-overlapping symptoms can facilitate accurate diagnosis. When a comorbidity is present, such as ADHD, overlapping symptoms (eg, distractibility and hyperactivity) should only count towards a diagnosis of mania or hypomania if they intensify during intervals of elation or irritability. Notably, ADHD is an ongoing condition whereas BD is episodic, and decreased sleep, hypersexuality, hallucinations or delusions, and homicidal or suicidal thoughts and actions occur with childhood mania, but are rare or absent in uncomplicated ADHD.

It is important to note that, while chronic irritability with episodic behavioural outbursts or rages can be seen in multiple paediatric psychiatric disorders (including emerging personality disorders, substance abuse, ODD, pervasive developmental disorders, and major depressive episodes), such irritability and explosiveness are not sufficient to make a diagnosis, even when severe. The recent DSM-5 diagnosis DMDD--which includes chronic irritability as a defining feature--lists BD as an exclusion criterion. However, the DMDD phenotype is evident in about 25% of adolescents with episodic BD, and is associated with factors such as greater family conflict and ADHD comorbidity. Classical BD and chronic irritability are therefore not mutually exclusive, the nonspecific nature of the latter notwithstanding.

A significant minority of children or youth with MDD will eventually go on to develop BD, with an average rate of 28% being reported., Risk factors for switch to mania following a depressive episode include a family history of mood disorders, emotional and behavioural dysregulation, subthreshold manic symptoms, cyclothymia, atypical depression and psychosis. A recent meta-analysis suggested that the most potent predictors were family history, an earlier age of onset and the presence of psychotic symptoms. There is an increased prevalence of BD among offspring of parents with BD., , Although there is no uniform strategy for managing depression (or ADHD, anxiety, etc.) in the child of a parent with BD, increased caution is warranted when prescribing antidepressant or stimulant medications as these have the potential to precipitate mania/hypomania. Patients and their parents should be informed of the potential switch risk and close monitoring for treatment-emergent manic/hypomanic switch should be instituted.

Self-report and/or parent-reported questionnaires can be informative and can raise the index of suspicion for BD. However, scores on such questionnaires as the Child Behaviour Checklist (CBCL) "dysregulation phenotype", previously described as "BD phenotype", have poor capacity for differentiating BD from other complex and severe symptomatic presentations. Questionnaires can be used as screeners, but do not substitute for a thorough diagnostic evaluation. Longitudinal rating by parents may be most helpful in diagnosis and assessment of treatment response. An online program for weekly parental ratings (of depression, anxiety, ADHD, oppositional behaviour, and mania) of children aged 2-12 years is available at http://www.bipolarnews.org, click on Child Network.

7.2.2. Pharmacological management

7.3.1. Presentation and course

Because of the aging population in Canada and many countries around the world, knowledge of pertinent issues related to the management of older adults is becoming increasingly important. Approximately 6% of geriatric psychiatry outpatients and 10% of inpatients have BD, and proportionally this population is one of the highest users of psychiatric and physical health services. Approximately 25% of the patients with BD in the USA in 2005 were over the age of 60 years, and by 2030 >50% of patients with BD are expected to be aged >60 years.

The lifetime prevalence of late-life BD is about 1%-2% with a 1-year prevalence of 0.1%-0.7% in the general population. About 90%-95% of older adults with bipolar disorder have their initial episode prior to age 50 years, although there is a minority who will have a later onset., Late onset is often related to neurological or physical comorbidity, and may carry a negative prognosis, although this is not a consistent finding.

While symptoms of mania or hypomania are generally less prominent in older adults, depressive and cognitive symptoms are more often observed, and hyperactivity, aggression, insomnia, impulsivity, and self-neglect may pose a significant risk to the patient and others., Psychiatric comorbidity is also generally lower than in younger patients, with anxiety and substance use being the most common. Compared to younger patients, older adults are less likely to utilize inpatient, outpatient, and emergency room services and more likely to use case-management and conservator services.

Cognitive dysfunction is a significant concern for this population, with >30% showing significant deficits across all mood states, including euthymia. This cognitive dysfunction is relatively stable, related to the number of mood episodes earlier in life, and does not appear to exceed normal aging in 2-5-year follow-up., , Lithium use has been associated with lower rates of cognitive disorders in BD, and higher lithium levels in drinking water may be associated with lower dementia risks,, although prospective trials are required to definitively assess this. Standardized instruments, such as the Montreal Cognitive Assessment (MoCA), should be used to quantify cognitive dysfunction. Because of the link between cognition and functioning in BD,, the impact of medications (particularly those with a high anticholinergic burden) on cognition should be considered when making treatment decisions. Furthermore, improvement of modifiable risk factors such as diet, exercise, and mental stimulation should also be promoted in order to further diminish the risk of cognitive decline.

7.3.2. Medical comorbidity

Older adults with BD have an average of three to four medical comorbidities, with metabolic syndrome, hypertension, diabetes, cardiovascular disease, arthritis, and endocrine abnormalities being the most common., Together, these contribute to a reduction in life expectancy of 10-15 years in BD compared to non-psychiatric populations. Because of these high rates of comorbidities, assessment of an older adult with BD should include a thorough physical and neurological examination, including clinical laboratory tests. Neuroimaging should also be applied as indicated, particularly in the presence of focal neurological signs and symptoms or abrupt late onset, or if the presentation is different from prior episodes. Coordination with other health care providers is also imperative, as this can optimize physical health, as can smoking cessation.

7.3.3. Pharmacological treatment

The data supporting efficacy of medications in various mood states in this population are limited, with only a single RCT exclusively in geriatric patients completed to date, comparing lithium vs divalproex for the treatment of mania/hypomania. Despite this, open-label trials, naturalistic studies, and post hoc analyses of mixed aged RCTs suggest that medications efficacious in adults overall will also be effective in older adults, although additional considerations regarding medication tolerability and age-related changes in pharmacokinetics and pharmacodynamics must be taken into account. Because of the high number of medical comorbidities as well as physical changes related to the aging process, strict attention must be paid in these patients to potential pharmacokinetic issues, drug-drug interactions, side effects, and the need for ongoing monitoring (see Section 8).

Amongst other effects, lithium has been associated with adverse neurological effects and renal disease. Divalproex has been associated with motor side effects and metabolic effects (weight gain and diabetes mellitus). Carbamazepine induces cytochrome P450 enzymes and can reduce the levels of divalproex and other medications. Regarding antipsychotics, which are now very commonly used for BD in older adults, dose reduction may be beneficial in some aging patients to lower the risk of motor, sedation, metabolic syndrome, and cognitive effects. There is an association between mortality and antipsychotics in patients with dementia but it is unclear how this should be managed for patients with BD. Recently there have also been data linking antipsychotics with acute kidney injury.

In particular, when lithium is used in this population, lithium level and renal monitoring should occur at least every 3-6 months, as well as 5-7 days following a lithium dose adjustment or adjustment of non-steroidal anti-inflammatory drugs (NSAIDs), antiontensin II receptor blockers (ARBs), angiotensin-converting enzyme inhibitors (ACEIs), or thiazide diuretic dosing. Special consideration also needs to be given to dose selection. A post hoc analysis of the STEP-BD study found that while, on average, older patients required a similar number of medications to younger patients to achieve recovery, lower doses were used. In that sample, over twice as many older patients as younger ones recovered using lithium alone (42% vs 21%, respectively). In general, starting at a lower dose (eg, 150 mg nightly for lithium) is recommended, with gradual adjustments to reach the lowest end of the therapeutic range for adults, with subsequent titration based on tolerability and effectiveness; keeping in mind that some older patients will require similar blood levels as the general adult population in order to achieve remission. Further discussion, including clinical guidance and more detailed treatment recommendations, can be found in the ISBD Task Force report. In light of very limited international guidelines for maintenance treatment of older adults with BD, an ISBD Task Force is currently undertaking a Delphi survey of international experts, and clinicians are advised to consult this as results become available in the future (Shulman K, personal communication).

Pharmacological treatments have been evaluated using the criteria for strength of evidence for efficacy (Table 1) in older adults as well as safety and tolerability (Section 8). Unfortunately, there is a dearth of literature for efficacy of treatments in older adults. However, clinical experience supports the notion that treatments known to work in general adult populations are also effective in older adults. Tolerability may be different and this is an important consideration in treatment selection in older adults. General principles for management of acute episodes described in Sections 3 and 4 also apply to this population.

7.4.1. Comorbid psychiatric disorders

7.4.2. Comorbid metabolic disorders

7.4.3. Other comorbid medical conditions

Two studies from a random sample of 1 million people, taken from a large population-based retrospective cohort in Taiwan, demonstrated a reduced risk of both stroke and cancer associated with lithium treatment for BD. The lithium group was compared with propensity-matched controls.

The first study reported hazard ratio for stroke over 11 years of 0.39 (95% CI 0.22-0.68) for those prescribed lithium, even when adjusting for the risks associated with typical and atypical antipsychotics. The reduced risk was also correlated with a higher dose, longer treatment duration and a higher rate of exposure to lithium. In the second study, in a sample of 4729 patients diagnosed with BD, lithium exposure was associated with a reduced risk of cancer, compared to a group prescribed anticonvulsant medications. (Lithium with or without anticonvulsant HR = 0.735, 95% CI 0.55-0.97). The study also demonstrated a dose-response relationship. In a subsequent large BD cohort study (n = 9651) focusing on genitourinary cancer, however, lithium was not associated with any change in risk.

A recent meta-analysis suggests increased risk of dementia in BD. There is some evidence that lithium in drinking water reduces the risk of dementia in the general population as does the use of lithium in patients with BD.

8.3.1. Weight gain

As described in Section 6, despite normal weight at illness onset, it is common for patients with BD to become overweight or obese, and several medications used to treat the illness may also exacerbate this effect. The likelihood of medications to cause weight gain should be carefully considered, as this is one of the most frequent treatment-related factors of non-adherence; contributing to upwards of 60% of cases. The medications most commonly associated with weight gain are olanzapine, clozapine, risperidone, quetiapine, gabapentin, divalproex and lithium; with carbamazepine, lamotrigine, and ziprasidone being the safer or options associated with less weight gain. Recent reviews further suggest that asenapine and aripiprazole (longer term use) also may lead to weight gain, but the impact of lurasidone on weight is minimal. All patients should be regularly monitored for weight changes.

8.3.2. Gastrointestinal symptoms

Both lithium and divalproex are commonly associated with nausea, vomiting, and diarrhoea, with 35%-45% of patient experiencing these side effects., For lithium, this is particularly pronounced during treatment initiation, or rapid dose increases. Gradual dose titration, taking the medication at bedtime, taking medications with food, and slow release preparations may reduce nausea and other side effects.

8.3.3. Renal toxicity

Lithium has a well-recognized potential for renal toxicity, including nephrogenic diabetes insipidus (NDI), chronic tubulointerstitial nephropathy, and acute tubular necrosis, with NDI reported in 20%-40% of patients., , Upwards of 70% of patients on chronic lithium treatment will experience polyuria, which can cause impairment in work and daily functioning. This side effect is commonly underreported, unless it is directly inquired about. Long-term administration (ie, 10-20+ years) is further associated with decreased glomerular filtration rate and chronic kidney disease. While long-term lithium administration is probably an important risk factor for developing chronic kidney disease, factors that may increase susceptibility include higher plasma lithium levels, multiple daily lithium doses (vs once daily), concurrent medications (eg NSAIDs, ARBs, ACEIs and diuretics), somatic illnesses (eg, hypertension, diabetes mellitus and coronary artery disease) and older age., Instances of lithium toxicity will also greatly increase risk of renal dysfunction. Lithium use is associated with a two-fold higher risk of chronic kidney disease in older adults (>66 years). While the overall risk for progressive renal failure is low, plasma creatinine concentrations and ideally estimated glomerular filtration rate (eGFR) for these patients should be measured at least every 3-6 months. Since 37% of patients aged >70 years have an eGFR <60 mL/min per 1.73 m², a strict eGFR cut-off for lithium discontinuation is difficult. The UK National Institute for Healthcare and Excellence (NICE) guidelines for chronic kidney disease (CKD) recommend nephrologist consultation if there is rapidly declining eGFR (>5 mL/min per 1.73 m² in 1 year, or >10 mL/min per 1.73 m² within 5 years), if the eGFR falls below 45 in two consecutive readings, or if the clinician is concerned.,

Because of its narrow therapeutic window, acute lithium intoxication is also a complication, which, though reversible, can lead to reductions in glomerular filtration rate., Drugs that alter renal function and general medical conditions characterized by decreased circulating volume all contribute to increased risk.

8.3.4. Haematological effects

Carbamazepine may be a risk factor for leucopenia,, , although this finding is not robust. This side effect is generally reversible with dose reduction or discontinuation. There is also some concern about rapidly developing bone marrow suppression resulting from hypersensitivity, particularly in older patients.,

Clozapine carries the greatest risk for drug-induced changes in white blood cell counts, with approximately 0.18% of patients experiencing changes rated as probably or definitely drug induced. All patients started on clozapine should have a baseline haematological profile established and be enrolled in the clozapine monitoring programme which requires regular monitoring of haematological parameters: weekly at first and then every 2-4 weeks later in the course of treatment.

8.3.5. Cardiovascular effects

Lithium can increase the risk of abnormal QT prolongation or T-wave abnormalities, an impact more pronounced with age, as almost 60% of older patients on lithium maintenance therapy have ECG abnormalities. Several antipsychotics, including risperidone, olanzapine, ziprasidone and asenapine, are also associated with arrhythmias, QTc prolongation, and other cardiovascular adverse events. Clozapine may increase the risk of several rare but serious events such as dilated cardiomyopathy, myocarditis, and pericarditis. Of the antipsychotics, lurasidone and aripiprazole are considered safe from a cardiac perspective, although aripiprazole may increase the risk for hypotension.

8.3.6. Endocrine effects

There is also a strong link between lithium maintenance treatment and hypothyroidism, which is also associated with increased risk of affective episodes, rapid cycling, and more severe depression., Routine screening of thyroid function is therefore recommended for all patients on lithium treatment. Since lithium can also cause hyperparathyroidism, routine monitoring for serum calcium is recommended, and, if elevated, further investigations should be performed to evaluate for hyperparathyroidisim. Hypothyroidism is ordinarily not an indication for lithium cessation in a patient with a good response; rather, thyroid supplementation is recommended.

New onset oligomenorrhoea or hyperandronism is more common in divalproex users. While there are reports of an increased incidence of polycystic ovary syndrome (PCOS) in divalproex treatment, a recent meta-analysis did not support this. In those who develop PCOS on divalproex, there is evidence from a small sample that discontinuing divalproex results in remission of some of the aspects of PCOS.

Hyperprolactinaemia is common with some antipsychotics, and can have short-term and long-term adverse effects. Risperidone, amisulpride and paliperidone are more likely than other compounds to cause it. Hyperprolactinaemia can induce amenorrhoea, sexual dysfunction, and galactorrhoea, amongst other effects. In the long term, it can cause gynaecomastia and osteoporosis. When such effects are seen, it may advisable to reduce the dose or switch to a different medication.

8.3.7. Cognition

While many patients experience cognitive impairment, these deficits may be attributable to the disease itself, with more pronounced effects in those with more severe or chronic illness. While a small study has led to suggestions that medicated patients who are euthymic do perform similarly to those not receiving treatment, other naturalistic trials point towards the potential negative impact of several medications, with the effects of antipsychotics being the most significant. Lithium can also lead to impairment in processing speed and memory, which patients may find distressing, although recent randomized controlled data suggest lithium is superior to quetiapine in this regard. Indeed, the effects of lithium on neurocognition are complex and further research is needed to fully elucidate its neurocognitive impact. Anticonvulsants, except for lamotrigine, are also linked to subjective cognitive impairment. Given the importance of cognition on a patient's function and quality of life, further studies are needed in this area.

8.3.8. Sedation

Sedation is a concern for many, and is reported by over half of patients as a reason for treatment non-adherence. Divalproex and atypical antipsychotics are most likely to lead to these effects, with 30%-50% of patients on atypical antipsychotics experiencing sedation, compared to 8%-13% with placebo, , , , and 21%-29% of patients on divalproex., This is not a concern with all antipsychotics, however; quetiapine, clozapine, and olanzapine will generally have higher rates of sedation compared to ziprasidone, risperidone, and aripiprazole. Lamotrigine and lithium have both been found to be less likely to cause sedation than divalproex.,

8.3.9. Neurological effects, including EPS

Tremor can be a significant cause of frustration for many patients, and is experienced by up to 10% of those treated with lithium or divalproex., , New onset neurological symptoms in patients on divalproex should raise suspicion of hyperammonaemic encephalopathy, which, while rare, can be potentially fatal, and hence early detection and discontinuation of divalproex is critical to prevent morbidity and mortality Sustained release formulations and dose reductions may limit symptoms., , While conventional antipsychotics such as haloperidol are often associated with EPS (including pseudoparkinsonism, akathisia, acute dystonia, and tardive dyskinesia), this risk is either absent or low with atypical antipsychotic agents., Among the atypical agents, risperidone, aripiprazole, cariprazine, ziprasidone and lurasidone are more likely to cause EPS, particularly at higher doses. In older patients, impaired swallowing function and dysphagia have also been linked to atypical agents, particularly at higher doses.,

Although rare, neuroleptic malignant syndrome (NMS) is a potentially life-threatening adverse event associated with antipsychotic agents. The risk with atypical agents was considered negligible initially; however, while the risk is very low, several atypical agents have nevertheless been associated with NMS. While generally unpredictable, the risk is greatest during the initial phase of treatment or change of dosage, with intravenous or intramuscular administration, with high dosages or polypharmacy, when the patient is physically restrained or dehydrated, in high ambient temperatures, in older patients and in patients with medical or psychiatric comorbidities. Patients with a previous history of NMS and/or a personal or family history of catatonia are also at higher risk. Antipsychotics may also impact thermoregulation, with case studies indicating the potential for both heat-related illnesses and hypothermia; thus, patients should be made aware and monitored for these risks during periods of extreme temperatures.

8.3.10. Dermatological reactions

Approximately 10% of patients being treated with lamotrigine will experience a non-serious rash, with 0.3%-1% developing a serious rash such as toxic epidermal necrolysis and SJS, although for those initiated on a dose of 25 mg with a gradual titration increasing the dose by 25 mg biweekly, the risk of developing serious rash may be as low as 0.02% or 1 in 5000. In some cases, an even lower rate of titration may be used (ie 12.5 mg/day and then gradually increase as per instructions. Carbamazepine is also associated with increased risk of rash and SJS, especially in the first 8 weeks of therapy, although the baseline risk is extremely low. Similarly, while these can also occur with divalproex, the risk is extremely low. Nevertheless, it is important that patients be informed of these risks and told to report any rash immediately, and these treatments should be discontinued if a serious rash is suspected.

Lithium is also linked with a variety of potentially distressing skin conditions, including acne, psoriasis, eczema, hair loss, hidradenitis suppurativa, nail dystrophy and mucosal lesions, with overall estimates ranging from 3% to 45% depending on the criteria applied. Most cases can be managed without treatment discontinuation.

8.3.11. Metabolic syndrome, hyperglycaemia, type 2 diabetes and dyslipidaemia

As described in Section 6, patients with BD are already at elevated risk for these physical illnesses and this risk is further exacerbated by some atypical antipsychotic agents and mood stabilizers. Clozapine and olanzapine are associated with the greatest level of risk, followed by quetiapine (particularly in higher doses) and risperidone, with a more minimal impact of aripiprazole, ziprasidone, asenapine, and lurasidone. Lithium and divalproex are also associated with weight gain. All patients on atypical antipsychotics should be monitored for changes in blood glucose and lipid profiles as indicated previously in this section, and if disturbances are detected, the atypical antipsychotic should be discontinued if possible and appropriate treatment initiated if necessary.

8.3.12. Fracture risk and osteoporosis

Some anticonvulsants, antidepressants, and antipsychotics may decrease bone mineral density and increase the risk of fracture in high-risk patients., This risk is increased by the presence of mood disorders, as well as known risks for mood disorders such as physical inactivity, smoking and poor diet quality. Thus, screening for this population may be indicated.