canmat guidelines -2013 update

Guidelines Update

Canadian Network for Mood and AnxietyTreatments (CANMAT) and InternationalSociety for Bipolar Disorders (ISBD)collaborative update of CANMAT guidelinesfor the management of patients with bipolardisorder: update 2013

Yatham LN, Kennedy SH, Parikh SV, Schaer A, Beaulieu S, Alda M,ODonovan C, MacQueen G, McIntyre RS, Sharma V, Ravindran A,Young LT, Milev R, Bond DJ, Frey BN, Goldstein BI, Lafer B,Birmaher B, Ha K, Nolen WA, Berk M.Canadian Network for Mood and Anxiety Treatments (CANMAT)and International Society for Bipolar Disorders (ISBD) collaborativeupdate of CANMAT guidelines for the management of patientswith bipolar disorder: update 2013.Bipolar Disord 2013: 15: 144. 2012 John Wiley & Sons A S.Published by Blackwell Publishing Ltd.

The Canadian Network for Mood and Anxiety Treatments publishedguidelines for the management of bipolar disorder in 2005, withupdates in 2007 and 2009. This third update, in conjunction with theInternational Society for Bipolar Disorders, reviews new evidence andis designed to be used in conjunction with the previous publications.The recommendations for the management of acute mania remain

largely unchanged. Lithium, valproate, and several atypicalantipsychotic agents continue to be rst-line treatments for acutemania. Monotherapy with asenapine, paliperidone extended release(ER), and divalproex ER, as well as adjunctive asenapine, have beenadded as rst-line options.For the management of bipolar depression, lithium, lamotrigine, and

quetiapine monotherapy, as well as olanzapine plus selective serotoninreuptake inhibitor (SSRI), and lithium or divalproex plusSSRI bupropion remain rst-line options. Lurasidone monotherapyand the combination of lurasidone or lamotrigine plus lithium ordivalproex have been added as a second-line options. Ziprasidone aloneor as adjunctive therapy, and adjunctive levetiracetam have been addedas not-recommended options for the treatment of bipolar depression.Lithium, lamotrigine, valproate, olanzapine, quetiapine,

aripiprazole, risperidone long-acting injection, and adjunctiveziprasidone continue to be rst-line options for maintenance treatmentof bipolar disorder. Asenapine alone or as adjunctive therapy have beenadded as third-line options.

Lakshmi N Yathama, Sidney HKennedyb, Sagar V Parikhb, AyalSchafferb, Serge Beaulieuc, MartinAldad, Claire ODonovand, GlendaMacQueene, Roger S McIntyreb,Verinder Sharmaf, Arun Ravindranb,L Trevor Younga, Roumen Milevg,David J Bonda, Benicio N Freyh,Benjamin I Goldsteini, Beny Laferj,Boris Birmaherk, Kyooseob Hal,Willem A Nolenm andMichael Berkn,o

doi: 10.1111/bdi.12025

Key words: bipolar CANMAT depression

guidelines mania treatment

Received 1 April 2012, revised and accepted for

publication 30 September 2012

Corresponding author:

Lakshmi N. Yatham,MBBS,FRCPC,MRCPsych(UK)

Department of Psychiatry

University of British Columbia

2255 Wesbrook Mall

Vancouver, BC V6T 2A1

Canada

Fax: 604-822-7922

E-mail: [email protected]

Alilations for all authors are listed before the references.

Bipolar Disorders 2013: 15: 144 2012 John Wiley and Sons A/SPublished by Blackwell Publishing Ltd.

BIPOLAR DISORDERS

1

Section 1. Introduction

In 2005, the Canadian Network for Mood andAnxiety Treatments (CANMAT) published guide-lines for the management of bipolar disorder (BD)(1), followed by updates in early 2007 (2) and in2009 [in collaboration with the InternationalSociety for Bipolar Disorders (ISBD)] (3). Thisupdate includes data published in 2009 throughearly 2012, and is designed to be used in conjunc-tion with the 2005 CANMAT guidelines andprevious updates (13).The purpose of this update is to add previously

unpublished material to the guidelines. This updateis designed to be used with the previous iterations ofthe guidelines. As in the previous updates, the guide-lines are divided into eight sections (Table 1.0) andthe same numbering system has been used for thesections and tables in order to facilitate ease of use.New evidence is incorporated into the managementrecommendations, and changes to the recommen-dation tables have been clearly denoted with bolditalics and a footnote, and have been describedin the text. The objective is to ensure that theCANMAT guidelines for treatment of BD remaincurrent and useful for the practicing clinician.Central to this update are the tables showing

rst-line, second-line, third-line, and not-recom-mended treatment options. These tables may assistin the selection of treatment, while the text of thisupdate and the previous guideline iterationsprovide the details of the evidence that was usedto make the recommendations. Similarly, thetreatment algorithms condense key managementinformation into a decision-tree ow-chart; theclinician should begin by positioning the patient inthe decision tree, and then follow the arrows forsubsequent management suggestions.Search strategies and methods to assess evidence

were as described in the original guidelines (1).Evidence available only in abstract form was alsoconsidered in order to ensure that the recommen-

dations are as up to date as possible. The criteriafor rating strength of evidence and making aclinical recommendation are shown in Tables 1.1and 1.2.We caution the readers that the evidence-based

guidelines are limited by the data that are avail-able. For instance, drugs that have patents arelikely to have been more widely studied and theirdesign was likely inuenced by the goals of thesponsor to obtain approval. Generic drugs,although may be useful, may not have been widelystudied because of lack of sponsorship, thusaecting their placement in the treatment algo-rithm. Finally, it is important to understand thatthe lack of evidence for a particular drug does notimply inecacy or ecacy. Clinicians must exercisecaution and choose treatments based on a carefulriskbenet analysis for each situation.

Section 2. Foundations of management

Epidemiology

Prevalence. The World Mental Health SurveyInitiative, involving 61392 people in nine countriesin North and South America, Europe, and Asia,reported lifetime (and 12-month) prevalence esti-mates of 0.6% (0.4%) for BD I, 0.4% (0.3%) forBD II, and 1.4% (0.8%) for subthreshold BD (4).However, there were large cross-national dier-ences in rates, with the lifetime rates ranging from0 to 1% for BD I, 0 to 1.1% for BD II, and 0.1 to2.4% for subthreshold BD.In the Canadian Community Health Survey

Mental Health and Well-Being (CCHS 1.2), theprevalence of BD was signicantly lower among

Table 1.0. Overview of guideline sections

Section 1. IntroductionSection 2. Foundations of managementSection 3. Acute management of bipolar maniaSection 4. Acute management of bipolar depressionSection 5. Maintenance therapy for bipolar disorderSection 6. Special populationsSection 7. Acute and maintenance management of bipolar IIdisorder

Section 8. Safety and monitoringClosing statementDisclosuresReferences

Table 1.1. Evidence criteria

1 Meta-analysis or replicated double-blind (DB), randomizedcontrolled trial (RCT) that includes a placebo condition

2 At least one DB-RCT with placebo or active comparisoncondition

3 Prospective uncontrolled trial with at least ten or moresubjects

4 Anecdotal reports or expert opinion

Table 1.2. Treatment recommendation

First line Level 1 or level 2 evidence plus clinicalsupport for efficacy and safety

Second line Level 3 evidence or higher plus clinicalsupport for efficacy and safety

Third line Level 4 evidence or higher plus clinicalsupport for efficacy and safety

Not recommended Level 1 or level 2 evidence for lackof efficacy

Yatham et al.

2

immigrant, compared to non-immigrant, subjects,but immigrants with BD were signicantly lesslikely to report contact with mental health profes-sionals (5).

Impact. A meta-analysis of 15 studies identied ahigh prevalence of lifetime suicide attempts both inpatients with BD I (36.3%) and in those with BD II(32.4%) (6). In the Systematic Treatment Enhance-ment Program for Bipolar Disorder (STEP-BD)(n = 4360), the completed suicide rate was 0.014per 100 person-years (7). A large cohort studyfound that among men, the absolute risk of suicidewas highest with BD (7.8%) compared to anyother psychiatric condition, and among women,BD was associated with the second highest risk, at4.8%, just below schizophrenia at 4.9% (8). Claimsdatabase data demonstrate the signicant eco-nomic impact of suicide attempts, with one-yearhealthcare costs in the period post-attempt beingmore than double those in the year prior to anattempt (9).The large, two-year, prospective, observational

European Mania in Bipolar disorder Longitudi-nal Evaluation of Medication (EMBLEM) study(n = 2289) found high work impairment in 69%of patients at baseline and 41% at two years(10). Rapid cycling, high baseline work impair-ment, lower levels of education, recent admis-sions, mania symptom severity, and overallseverity all predicted higher work impairment,while living in a relationship and independenthousing predicted lower work impairment atfollow-up. Similarly, the Understanding PatientsNeeds, Interactions, Treatment, and Expectations(UNITE) global survey (n = 1300) revealed thatonly one-third of patients with BD were em-ployed full-time (11). In the UNITE survey,treatment of depression, weight gain, and qualityof life were identied by patients with BD asaspects of care most in need of improvement(11).A meta-analysis of data from 12 trials

(n = 1838) found that the self-esteem of patientswith remitted BD was signicantly lower than thatof controls but signicantly higher than that ofpatients with remitted major depressive disorder(MDD) (12). In addition, self-esteem may follow auctuating course during remission of BD.In a health claims database, risk of arrest was

associated with substance use, poor rell compli-ance, and prior arrest (13). Among patients treatedwith an atypical antipsychotic agent, there was alower risk of arrest in those who had frequentoutpatient visits (approximately monthly) com-pared to those who did not.

Course. In a sub-analysis of 771 patients in thetwo-year EMBLEM study, approximately one inthree presented with a mixed episode, which wasassociated with a lower likelihood of recovery andgreater use of antidepressant therapy compared toa pure manic state during follow-up (14). TheSystematic Treatment Optimization Program forEarly Mania (STOP-EM) project followed 53patients presenting with a rst episode of mania,and found that more than half experienced recur-rence of a mood episode during the one-yearfollow-up, with a mean time to event of7.9 months (15). The mean duration of moodepisodes in BD I, in a longitudinal analysis of 219patients followed for up to 25 years, was found tobe 13 weeks (16).

Diagnostic assessment

The proposed fth edition of the Diagnostic andStatistical Manual of Mental Disorders (DSM-5) isscheduled to be completed by mid-2013. Revisionssuggested by the International Society for BipolarDisorders Diagnostic Guidelines Task Force (17)were summarized in the previous update to theseguidelines (3). DSM-5 will have separate chaptersfor bipolar and related conditions, and depressivedisorders. The condition BD not otherwise spec-ied (NOS) has been replaced with bipolarconditions not elsewhere classied. Substance-induced BD and BD associated with a generalmedical condition have been added.In the criteria for a manic episode, abnormally

and persistently increased activity or energy hasbeen added to criterion A, which previouslyreferred only to a distinct period of abnormallyand persistently elevated, expansive, or irritablemood. A manic episode emerging during antide-pressant treatment can qualify as a manic episodeof BD, provided that the symptoms persistedbeyond the physiological eects of treatment. Themixed episode diagnosis has been replaced with amixed features specier, requiring three symptomsof the opposite pole, which would apply to manic,hypomanic, and depressive episodes (18). In addi-tion, dimensional speciers for anxiety and suiciderisk have also been proposed.

Chronic disease management

BD is a chronic illness and patients require long-term multi-disciplinary management as describedin the 2005 guidelines (1). A small, cluster-ran-domized controlled trial (RCT) examined the eectof community mental health teams (n = 23) whoreceived enhanced training in relapse prevention

CANMAT guidelines for bipolar disorder

3

versus treatment-as-usual (TAU) in 96 patientswith BD (19). The median survival time of patientstreated by the trained teams was prolonged by8.5 weeks compared to those receiving TAU(42 weeks versus 33.5 weeks). A collaborative caremodel including clinician support through the useof simplied guidelines was found to result insignicantly greater guideline-concordant therapyover a three-year follow-up period compared toTAU in patients with BD (n = 306) (20).Data suggest that use of a symptom checklist can

substantially increase the recognition of earlywarning signs for depressive or manic relapse(21). There was a positive correlation between thefrequency of monitoring and social occupationalfunctioning.

Psychosocial interventions

When used as adjuncts to pharmacotherapy, psy-chosocial interventions such as group psychoedu-cation, cognitive behavior therapy (CBT), andinterpersonal and social rhythm therapy (IPSRT)have demonstrated signicant benets, both in thetreatment of acute depressive episodes and alsoas long-term maintenance treatment, includingdecreased relapse rates, mood uctuations, needfor medications, and hospitalizations, as well asincreased functioning and medication adherence(13). Therefore, providing psychological treat-ments and, in particular, brief psychoeducation,which has been demonstrated to be as eective asCBT at much lower cost (22) is an essentialaspect of managing patients with BD.A family-focused treatment approach designed to

help caregivers improve illness management skillsand their own self-care was shown to eectivelyreduce depressive symptoms and health-risk behav-ior among caregivers and family members, andreduce depressive symptoms in patients (23).The availability of internet-based strategies has

grown substantially, with demonstrated ecacy inreducing depressive symptoms and improving psy-chological quality of life (2427).

Section 3. Acute management of bipolar mania

Emergency management of acute mania

The acutely manic bipolar patient may present inan agitated state that acts as a barrier to therapy,interrupts the physicianpatient alliance, andcreates a disruptive, even hazardous, environ-ment. Whenever possible, oral therapy should beoered rst, as evidence suggests that oral agentscan be as eective as intramuscular agents (28,

29). Intramuscular injections oer an alternativewhen oral therapy cannot be reliably adminis-tered.Based on current data, the oral atypical anti-

psychotic agents, risperidone (level 2) (29, 30),olanzapine (level 2) (30), and quetiapine (level 3)(30, 31), should be considered rst in the treatmentof acute agitation. In patients who refuse oralmedications, intramuscular olanzapine (level 2)(3235), ziprasidone (level 2) (3538), and aripip-razole (level 2) (39) or a combination of intramus-cular haloperidol and a benzodiazepine shouldbe considered (level 2) (29, 35, 38, 40, 41). Ingeneral, benzodiazepines should not be used asmonotherapy, but are useful adjuncts to sedateacutely agitated patients (1).New data also support the use of intravenous

sodium valproate (level 3) (42) and oral divalproexER (level 3) (31) for rapid improvement of acutemania.

Pharmacological treatment of manic episodes

Pharmacological management of acute manic epi-sodes should follow the algorithm outlined inFigure 3.1 (13). New clinical trial data, and theavailability of several agents, justify some changesto the recommendations. Monotherapy with asen-apine, paliperidone ER, and divalproex ER, as wellas adjunctive asenapine, have been added as rst-line options (Table 3.3).

Step 1. Review general principles and assess medi-cation status: Recommendations from 2005 guide-lines remain unchanged.

Step 2. First-line therapies: A comprehensive meta-analysis of 68 trials supported the ecacy ofpharmacotherapy for the treatment of acute mania(43). Lithium, divalproex, risperidone ER, pali-peridone ER, olanzapine, quetiapine, aripiprazole,ziprasidone, and asenapine (rst line), carbamaze-pine, and haloperidol (second line), were signi-cantly more eective than placebo, whereasgabapentin, lamotrigine, and topiramate were not(not recommended) (43). Haloperidol was moreeective than a number of antimanic agents butnot olanzapine or risperidone, both of which weremore eective than valproate, ziprasidone, andlamotrigine. Two other recent meta-analyses alsosupport the ecacy of lithium divalproex andatypical antipsychotic agents for the treatment ofacute mania (44, 45).

Lithium divalproex. The ecacy of lithium anddivalproex in the management of acute mania is

Yatham et al.

4

well established (level 1) (13). Two large, 12-week,open, randomized trials comparing lithium todivalproex found comparable ecacy andtolerability of these agents for the treatment ofacute mania (46, 47).A large (n = 521), 12-week RCT compared

divalproex, olanzapine, and placebo in patientswith mild to moderate mania. At three weeks,improvements in mania scores were signicant witholanzapine versus placebo but not with divalproexversus olanzapine or placebo. After 12 weeks oftreatment, improvements in both active treatmentgroups were signicant versus placebo, but ola-nzapine was signicantly more ecacious thandivalproex (48).

The results of two three-week RCTs assessingthe ecacy of the ER formulation of divalproexfor the treatment of acute mania have now beenpublished (49, 50) One study demonstratedstatistically signicant improvements in manicsymptoms compared to placebo (level 2) (49),while the other did not (50). In the rst, Bowdenet al. (49) found signicantly greater improve-ment in manic symptoms and higher responserates (48% versus 34%, p = 0.012) with divalp-roex ER versus placebo, while Hirschfeld et al.(50) found no statistically signicant dierence inmania scores with divalproex ER versus placebo;however, discontinuation rates were over 80%and dosing may have been lower than optimal.

Table 3.3. Recommendations for pharmacological treatment of acute mania

First line Monotherapy: lithium, divalproex, divalproex ER a, olanzapineb, risperidone, quetiapine,quetiapine XR, aripiprazole, ziprasidone, asenapine a, paliperidone ER a

Adjunctive therapy with lithium or divalproex: risperidone, quetiapine, olanzapine, aripiprazole, asenapine a

Second line Monotherapy: carbamazepine, carbamazepine ER, ECT, haloperidol a

Combination therapy: lithium + divalproex

Third line Monotherapy: chlorpromazine, clozapine, oxcarbazepine, tamoxifen, cariprazine a

(not yet commercially available)Combination therapy: lithium or divalproex + haloperidol, lithium + carbamazepine, adjunctive tamoxifen

Not recommended Monotherapy: gabapentin, topiramate, lamotrigine, verapamil, tiagabineCombination therapy: risperidone + carbamazepine, olanzapine + carbamazepine

ECT = electroconvulsive therapy; XR or ER = extended release.aNew or change to recommendation.bGiven the metabolic side effects, use should be carefully monitored.

Assess safety/functioningEstablish treatment setting

D/C antidepressantsD/CRule out medical causes

D/C caffeine, alcohol, and illicit substances Behavioural strategies/rhythms, psychoeducation

Step 1Review general

principles &

assess medication On first-line agentNot on medicationstatus

Step 2Initiate/optimize

Initiate Li, DVP, AAP, or 2-drug AAP 2-drug combination

or first-line agent+

, check compliance

Step 3

combinationLithium or

DVPAAP

(Li or DVP + AAP)

Add or Add or switch to Replace one or both

No response

Add-on or switch therapy

switch to AAP Li or DVP agents with other first-line agents

Consider adding or Replace one or both Step 4No response

switching to second or third-line agent or ECT

pagents with other first-line agents


Step 5Add on novel or

No response

Consider adding novel i l-

experimental agentsor experimenta agent

Fig. 3.1. Treatment algorithm for acute mania. Novel experimental agents: zotepine, levetiracetam, phenytoin, mexiletine, omega-3-fatty acids, calcitonin, rapid tryptophan depletion, allopurinol, amisulpride, folic acid, memantine. D C = discontinue;Li = lithium; DVP = divalproex; AAP = atypical antipsychotic agent.


5

Given the level 1 evidence to support the imme-diate-release formulation of divalproex, as well asthe high discontinuation rate and dosing issues inthe negative trial, divalproex ER has been addedas a rst-line therapy, although, if prescribed,attention should be paid to dosing and serumlevels.

Atypical antipsychotic monotherapy. SubstantialRCT data support the ecacy of atypical anti-psychotic monotherapy with olanzapine, risperi-done, quetiapine, ziprasidone, and aripiprazolefor the rst-line treatment of acute mania (level 1)(13).As reviewed earlier, a large (n = 521), 12-week

RCT comparing divalproex, olanzapine, andplacebo in patients with mild to moderate maniafound that improvements in mania scores with ola-nzapine were signicantly greater than with pla-cebo after three weeks, and greater than with bothdivalproex and placebo after 12 weeks (48).In a meta-analysis of six aripiprazole monother-

apy RCTs in acute mania involving 2303 patients,the eect size was 0.34 versus placebo at weekthree, with a response generally being seen at daythree (level 1) (51). A 12-week RCT of aripiprazolemonotherapy in acute mania found signicantlygreater improvements in the Young Mania RatingScale (YMRS) scores at week three with aripip-razole ()12.0, p < 0.05) or haloperidol ()12.8,p < 0.01) compared to placebo ()9.7), which weremaintained to week 12; haloperidol was includedas an active control and was not statisticallycompared to aripiprazole (52).Additional data are also available comparing

ziprasidone to placebo and haloperidol in a12-week RCT. Improvements in mania scores andresponse rates at week three were signicantlygreater than with placebo for both active treat-ments, but haloperidol was signicantly moreeective than ziprasidone. During the nine-weekextension phase, responses were maintained for themajority of patients receiving active treatments.Ziprasidone showed a superior tolerability proleand lower discontinuation rates during the exten-sion phase (53).Two three-week, double-blind RCTs demon-

strating the ecacy of paliperidone ER in patientswith manic or mixed episodes, which were previ-ously cited in abstract form, have now beenpublished (level 1) (54, 55). Paliperidone ER hasbeen upgraded to a rst-line option.Two three-week, double-blind RCTs demon-

strating the ecacy of asenapine as monotherapyfor acute mania (56, 57), as well as nine-week and40-week extension phase results demonstrating the

maintenance of benets (58, 59), which werepreviously cited in abstract form, have now beenpublished (level 1). Asenapine has been upgradedto a rst-line option.

Atypical antipsychotic combination therapy. Aspreviously reported, a six-week, placebo-controlledRCT showed that adding aripiprazole to lithium ordivalproex in 384 patients with an inadequateresponse was signicantly more eective thanplacebo from week 1 onward (60). A 46-weekopen-label extension of this study found thataripiprazole as an adjunct to lithium or divalproexprovided continued improvement in mania but notdepression (61).A 12-week RCT demonstrating signicant

improvements in mania symptoms with adjunctiveasenapine added to lithium divalproex comparedto placebo that was previously cited in abstractform has still not been published, but 40-weekextension results have now been reported (level 2)(62). Of the original 318 patients, 71 completed the40-week extension; there were additional improve-ments in mania scores at 52 weeks in both theasenapine and placebo groups. Adjunctive asena-pine has been moved to a rst-line option.

Step 3. Add-on or switch therapy (alternate rst-linetherapies): No changes from 2005 guidelines.

Step 4. Add-on or switch therapy (second- and third-line therapies):

Second-line options. A small RCT in 44 patientswith manic, mixed, or depressive episodes foundthat among patients who were on, or thought tobenet from, carbamazepine, there were no dier-ences in mood ratings or in the total level ofadverse events with immediate-release versus ERcarbamazepine (63, 64). However, there weresignicantly fewer autonomic and gastrointestinaladverse events with carbamazepine ER (64).While electroconvulsive therapy (ECT) can be

an eective option, research studies have not beenrigorous and therefore it continues to be recom-mended as a second-line therapy (level 3) (1). Inan RCT of ECT as adjunct to antipsychotictherapy, bilateral, twice-weekly ECT delivered atstimulus intensities just above seizure thresholdwas as eective and safe as ECT administeredat stimulus intensities 2.5 times the seizurethreshold in rapidly resolving the symptoms ofacute mania (65).In a meta-analysis of 13 haloperidol-controlled

trials, the drug was signicantly more eective thanlithium, divalproex, quetiapine, aripiprazole,

Yatham et al.

6

ziprasidone, carbamazepine, asenapine, and lamo-trigine (43). Given the strong data for ecacy,haloperidol has been upgraded to a second-lineoption. However, haloperidol should only be usedon a short-term basis to treat acute mania ascontinuation of haloperidol may increase the riskof a depressive episode (43).

Third-line options. A small (n = 60), 12-week RCTcomparing oxcarbazepine to divalproex in patientswith acute mania found no signicant dierencesin improvements in mania scores or remission ratesbetween the two treatments, but divalproex wasassociated with more adverse events (66). Inanother small (n = 52) RCT, adjunctive ox-carbazepine was more eective than carbamaze-pine as add-on therapy in patients who hadpreviously been inadequate responders to lithium,although both agents improved manic and depres-sive scores versus baseline (67). As described in theprevious updates to these guidelines, there areother small positive trials, but there is also anegative placebo-controlled RCT, and these newtrials are small and do not include a placebocontrol arm; therefore, oxcarbazepine remains as athird-line option.Cariprazine, a new dopamine D3 D2 receptor

antagonist, appears promising for the treatment ofacute mania, but has not yet been approved byCanadian or US regulatory agencies. The resultsof a three-week, phase 2, RCT, presented inabstract form, reported signicant reductions inmania scores with cariprazine compared to placebo(level 2) (68).

Step 5. Add-on novel or experimental agents:Zotepine is an antipsychotic agent that has beenapproved in some European countries and inJapan for the treatment of schizophrenia. In afour-week, single-blind trial, adjunctive zotepineadded to lithium or divalproex therapy in 45inpatients with moderate-to-severe mania was aseective as adjunctive haloperidol in improvingmania scores (69).Two RCTs have now demonstrated the ecacy

of adjunctive allopurinol for the treatment of acutemania (level 1) (70, 71). In an eight-week RCT,allopurinol as adjunct to lithium plus haloperidolwas found to be signicantly more eective thanplacebo in 82 patients hospitalized with acutemania (70). In the second RCT (n = 180), com-paring the addition of allopurinol, dipyridamole,or placebo to lithium for four weeks, allopurinolled to signicantly greater improvements in maniascores and remission rates versus placebo (71).Although there is level 1 evidence for the use of

allopurinol, given that it can cause hepatomegalyas well as hypersensitivity reactions such asStevenJohnson syndrome and toxic epidermalnecrolysis, it is recommended only for thosepatients that are refractory to other rst-, second,and third-line treatments.Preliminary evidence previously suggested

antimanic ecacy associated with tamoxifen (1).A six-week, placebo-controlled RCT has nowdemonstrated signicantly greater improvementsin mania scores with tamoxifen as an adjunct tolithium in 40 inpatients with acute mania com-pared to lithium alone (level 2) (72).In a three-week RCT in 88 acutely manic

patients on divalproex, adjunctive folic acid wassignicantly better than placebo in improvingmania scores (level 2) (73).A three-week, open-label, pilot trial in 33 patients

with manic or mixed episode BD I found that 3050% of patients responded to doses of memantineranging from 20 mg to 40 mg (level 3) (74).Given the limited data, at this time, these agents

can only be recommended as add-on therapiesafter failure of standard therapies.

Adjunctive therapies with negative data requiringfurther study: A six-week RCT that found nosignicant improvements in manic symptoms withadjunctive exible-dose paliperidone in patientswith manic or mixed episodes who had notresponded to lithium or divalproex, was previouslycited in abstract form and has now been published(level 2, negative) (75). Given that paliperidonemonotherapy is eective, and that lithium orvalproate does not aect the metabolism of pali-peridone, the lack of ecacy of combinationtherapy is surprising. In spite of the fact that itwas a exible-dose trial, paliperidone ER may havebeen under-dosed, since the monotherapy studiessuggest that 12 mg day is the most eective doseand the mean dose used was 8.1 (3.30) mg day,with 45% of the patients receiving a nal dose of6 mg day in this trial. In addition, a post-hocsubgroup analysis found that adjunctive paliperi-done ER was superior to lithium or divalproexmonotherapy for patients diagnosed with a manicepisode (p = 0.020).A three-week RCT in over 600 patients with BD

mania mixed episodes found no signicantbenets with adjunctive ziprasidone at either high(120160 mg day) or low (4080 mg day) dosescompared to placebo (76). The trial has not yetbeen published, but the results are available athttp://www.clinicaltrials.gov.In light of these negative trials, adjunctive use of

paliperidone ER or ziprasidone, at the dosages


7

used by the above-noted studies, is not recom-mended.

Mania with psychotic features

A meta-analysis of four RCTs of aripiprazolesupports its antipsychotic eects, as measuredby the Positive and Negative Syndrome Scale(PANSS) score, during the acute manic andmaintenance phases of BD (77). The eect sizesfor aripiprazole versus placebo were highest for thePANSSpositive subscale (0.28) and the PANSShostility subscale (0.24).

Mixed states

A six-week RCT (n = 202) evaluating adjunctiveolanzapine compared to adjunctive placebo dem-onstrated signicantly greater and earlier reduc-tions in manic and depressive symptoms in patientswith mixed episodes inadequately controlled withdivalproex (78). Post-hoc analysis of this studyfound that response (Clinical Global ImpressionSeverity decrease 1) at day two was predictive ofmixed symptom remission (79).A post-hoc analysis of two asenapine RCTs in

patients with manic or mixed episodes demon-strated statistically signicant decreases in depres-

sion scores with asenapine versus placebo inpatients with severe baseline depressive symptoms(n = 604); dierences between the active compar-ator olanzapine and placebo were not signicant,which makes the interpretation of these resultsmore dicult as it raises the possibility of anegative study (80).

Section 4. Acute management of bipolar depression

Pharmacological treatment of depressive episodes

Pharmacological management of acute bipolardepressive episodes should follow the algorithmoutlined in Figure 4.1 (13). The recommendationsfor rst- and second-line therapies are largelyunchanged, except for the addition of lurasidonemonotherapy and lurasidone or lamotrigine pluslithium or divalproex as second-line options. Basedon negative data, ziprasidone alone or as adjunc-tive therapy, and adjunctive levetiracetam havebeen added as not-recommended options for thetreatment of bipolar depression (Table 4.3).Several meta-analyses have assessed the ecacy

of atypical antipsychotic agents and other medica-tions for the treatment of bipolar depression(81, 82). A meta-analysis of atypical antipsychoticagents for bipolar depression included ve trials(two monotherapy trials with each of quetiapine

Assess safety/functioningBehavioural strategies/rhythms

Psychoeducation

Step 1Review general

principles&

assess medication status

+

On first-line agent

On DVP On OLZ, RIS, ARI, or ZIP

Not on medication

Step 2Initiate/optimize, check compliance

Li OLZ+SSRIa

Li + DVPAdd SSRIa/BUP or add/switch to

Li , LAM or QUE

Li or DVP+SSRIa/ BUP

LAM QUEAdd SSRIa, Li

or LAM or switch to Li, LAM or QUE

Add SSRIa/BUP

Step 3Add-on or

Add SSRIa/BUP or

Switch to QUE, QUE+SSRIa Li

Add/switchto Li

No response

Switch Li or DVP to QUE or

Add SSRI, Li or LAM or switch

or switchLi or DVP to LAM or QUE


or add/switch to LAM or QUE

, Li, Li + SSRI /BUP a

or LAMb

to Li or QUE

No response

OLZ or switch SSRIa/BUP to

LAMc

to Li, LAM or OLZ + SSRIa

Replace one or both agents with alternate first- or second-line agents

Step 4Add-on or

switch therapy

No

Step 5Add-on or

switch therapy

Consider ECT, third-line agents and novel or experimental options

response

Fig. 4.1. Treatment algorithm for the management of bipolar I depression. Novel experimental agents: adjunctive pramipexole,eicosapentaenoic acid (EPA), riluzole, topiramate, N-acetyl cysteine (NAC), ketamine, armodanil, and chronotherapy.DVP = divalproex; OLZ = olanzapine; RIS = risperidone; ARI = aripiprazole; ZIP = ziprasidone; SSRI = selective serotoninreuptake inhibitor; BUP = bupropion; Li = lithium; LAM = lamotrigine; QUE = quetiapine; ECT = electroconvulsive therapy.aExcept paroxetine. bOr switch the SSRI to another SSRI. cOr switch the SSRI or BUP to another SSRI or BUP.

Yatham et al.

8

and aripiprazole, and one combination trial witholanzapine) and found signicantly greaterimprovement compared to placebo for weeks 16but not for weeks seven and eight (primarilyaccounted for by a tapering of eect in aripiprazolestudies) (81). This suggests that the ecacy of theseagents is not a class eect and that individualagents may show dierential benets, and, as such,generalizations on the role of atypical antipsy-chotic agents for depressive symptoms cannot bemade. Another meta-analysis included 19 trialsassessing mainly quetiapine (ve trials) and lamo-trigine (six trials), but also paroxetine, lithium,olanzapine, aripiprazole, phenelzine, and divalp-roex for the treatment of bipolar depression (82).This analysis found the highest reductions inMontgomeryAsberg Depression Rating Scale(MADRS) scores with the olanzapine plus uoxe-tine combination and quetiapine monotherapycompared to placebo. In this analysis, lamotrigine,paroxetine, aripiprazole, and lithium were notsignicantly dierent from placebo in improvingdepression scores. However, as cited in previousiterations, a meta-analysis of individual patientdata supported the ecacy of lamotrigine mono-therapy (83).

Step 1. Review general principles and assess medi-cation status: Recommendations from 2005 guide-lines remain unchanged.

Step 2. Initiate or optimize therapy and checkadherence (rst-line therapies): Lithium, lamotrigine,quetiapine, and quetiapine extended release (XR)

monotherapies, as well as lithium or divalproexplus selective serotonin reuptake inhibitor (SSRI),olanzapine plus SSRI, lithium plus divalproex, andlithium or divalproex plus bupropion all continueto be recommended as rst-line choices for bipolardepression.Data suggest that the absence of early improve-

ment (23 weeks) may be a highly reliable predic-tor of eventual non-response, suggesting that thesepatients may benet from a change in therapy(84, 85).

Lithium. The early results of the National Insti-tute of Mental Health (NIMH) lithium treatmentmoderate dose use study have been presented(86). This pragmatic study randomized 283patients with BD I or BD II to receive sixmonths of open-label moderate dose(600 mg day) lithium plus optimized treatment[per Texas Medication Algorithms (87)] versusoptimized treatment alone and found no signif-icant dierences between treatment groups. How-ever, given that this was an open-label study, andin the absence of further study details, recom-mendations for adjunctive lithium use remainunchanged.

Quetiapine monotherapy. The four large publishedRCTs demonstrating the ecacy of quetiapinemonotherapy in bipolar depression, which werecited in previous iterations of these guidelines, havenow all been published: BipOLar DEpRession(BOLDER) I (88) and II (89) and Ecacy ofMonotherapy SEROQUEL in BipOLar DEpres-

Table 4.3. Recommendations for pharmacological treatment of acute bipolar I depressiona

First line Monotherapy: lithium, lamotrigine, quetiapine, quetiapine XRCombination therapy: lithium or divalproex + SSRIb, olanzapine + SSRIb, lithium + divalproex, lithium ordivalproex + bupropion

Second line Monotherapy: divalproex, lurasidonec

Combination therapy: quetiapine + SSRIb, adjunctive modafinil, lithium or divalproex + lamotriginec, lithiumor divalproex + lurasidonec

Third line Monotherapy: carbamazepine, olanzapine, ECT d

Combination therapy: lithium + carbamazepine, lithium + pramipexole, lithium or divalproex + venlafaxine,lithium + MAOI, lithium or divalproex or AAP + TCA, lithium or divalproex or carbamazepine + SSRIb +lamotrigine, quetiapine + lamotriginec

Not recommended Monotherapy: gabapentin, aripiprazole, ziprasidone c

Combination therapy: adjunctive ziprasidonec, adjunctive levetiracetamc

AAP = atypical antipsychotic agent; ECT = electroconvulsive therapy; MAOI = monoamine oxidase inhibitor; TCA = tricyclic antide-pressant; SSRI = selective serotonin reuptake inhibitor; XR = extended release.aThe management of a bipolar depressive episode with antidepressants remains complex. The clinician must balance the desired effectof remission with the undesired effect of switching. See detailed discussion in Clinical questions and controversies section.bExcept paroxetine.cNew or change to recommendation.dCould be used as first- or second-line treatment in certain situations (see text).


9

sioN (EMBOLDEN) I (90) and EMBOLDEN II(91) (level 1).The eight-week RCT demonstrating signicantly

greater improvement in depressive symptoms withquetiapine XR monotherapy in patients with BD Ior BD II depression, which was previously cited,has now been published (92).

Olanzapine + uoxetine. There are level 1 datademonstrating the ecacy of olanzapineuoxe-tine combination (OFC) therapy for the treatmentof BD I depression (13). Follow-up results from apreviously described RCT [seven-week outcomedata (93)] found signicantly greater improvementsin depressive and manic symptoms with OFCversus lamotrigine in 410 patients with BD I atstudy end (94). However, OFC treatment wasassociated with a signicantly increased risk oftreatment-emergent hypercholesterolemia andweight gain. In addition, a post-hoc analysis of apreviously cited combination study (95) found thatboth OFC and olanzapine monotherapy were moreecacious than placebo in patients with BD Imixed depression (i.e., syndromal depression andsubsyndromal mania hypomania).

Step 3. Add-on or switch therapy (alternate rst- orsecond-line therapies):

Second-line options

Divalproex monotherapy. Four small RCTs haveassessed the ecacy of divalproex or divalproex ERfor the treatment of BD I or BD II depression (level1) (96, 97). Two meta-analyses of these trials (totaln = 142), by separate groups, concluded thatdivalproex was more eective than placebo for thetreatment of bipolar depression, but the strength ofthe conclusions was limited by sample size (96, 97).Therefore, given the limited evidence, divalproexcontinues to be recommended as a second-line option.

Lurasidone. Two six-week RCTs have demon-strated the ecacy of lurasidone as monotherapy(98) or as an adjunct (99) in patients with bipolardepression. Lurasidone monotherapy signicantlyreduced depressive symptoms in patients with BD Idepression as early as week two compared toplacebo (level 2) (98). Similarly, when used as anadjunct to lithium or divalproex, lurasidonesignicantly reduced depressive symptoms, andimproved functioning and quality of life comparedto placebo in patients with BD I depression whohad an inadequate response to lithium or divalp-roex alone (level 2) (99). These data look verypromising and if clinical experience supports e-

cacy, lurasidone will be upgraded to one of therst-line treatments in the next revision.

Lamotrigine + lithium or divalproex. In an eight-weekRCT, the acute eect of lamotriginewas greaterthan that of placebo as an add-on to lithium for BD Ior BD II depression (n = 124) (100). Non-respond-ers in this trial entered a second phase in whichparoxetinewas added; this addition showedbenet innon-responders to lithium + placebo, but not innon-responders to lithium + lamotrigine (101).Given the slow titration required for lamotrigine,this treatment is recommended either in monother-apy or as an add-on therapy primarily for those withmild-to-moderate bipolar depression, and in partic-ular for those with depression recurrences, given itsecacy in preventing depressive relapses.

Step 4. Add-on or switch therapy (alternate rst- orsecond-line therapies): No changes from 2005guideline (1).

Step 5. Add-on or switch therapy (third-line agentsand novel experimental therapies):

Third-line options

Olanzapine monotherapy. There are now two largeRCTs demonstrating the ecacy of olanzapinemonotherapy for the treatment of bipolar depres-sion (level 1) (95, 102). In the earlier of these twotrials, olanzapine monotherapy demonstrateda statistically signicant, but clinically modestantidepressant eect in a large (n = 833), eight-week RCT in patients with bipolar depression (95),and was recommended as a third-line option (1).In the subsequent large RCT, available in

abstract form, 514 patients with bipolar depres-sion achieved signicantly greater improvement indepressive symptoms with olanzapine comparedto placebo (MADRS )13.8 versus )11.7, p =0.018) over six weeks of treatment (102). How-ever, olanzapine was also associated with signif-icantly greater rates of metabolic changes (102). Asmall (n = 20), open-label study provided addi-tional support for the ecacy of olanzapinemonotherapy in patients with BD I or BD IIdepression (103).Although there is level 1 evidence, the magnitude

of benet of olanzapine monotherapy was onlymodestly greater than that of placebo (95, 102). Inthe earlier trial (95), the increased ecacy ofolanzapine relative to placebo was mainlyaccounted for by changes in sleep, appetite, andinner tension, which are not the core symptoms ofdepression (81). In addition, as adverse events were

Yatham et al.

10

marked in the recent trial (102), this strategycontinues to be recommended as a third-lineoption.

Quetiapine + lamotrigine. A small, open trial in 39patients with BD I and BD II found that thecombination of lamotrigine plus quetiapine wasbenecial in treatment-resistant bipolar depression(level 3) (104).

Carbamazepine. There is additional evidence tosupport the use of carbamazepine (level 2), as a smallRCT in a mixed population of 44 patients with BDfound that ER carbamazepine was as eective as theimmediate-release form, with fewer autonomic andgastrointestinal adverse events (63, 64).

ECT. As stated in previous iterations of the guide-lines, the use of ECT should be considered earlier inpatients who have psychotic bipolar depression, inthose at high risk for suicide, and in those withsignicant medical complications due to not drink-ing and eating. Clinical experience and open-labeldata continue to accumulate and support the ecacyof ECT. In an open trial, similar rates of responseand remissionwere observed in patientswith bipolardepression (70% and 26%, respectively) and thosewith mixed states (66% and 30%, respectively)(105). A retrospective analysis of 201 patients withBD receiving ECT concluded that those receivingconcomitant anticonvulsants achieved comparablesymptomatic improvement to those not on anticon-vulsants; however, they required a signicantlygreater number of ECT sessions to achieve this(106). Two RCTs comparing dierent ECT proto-cols found no dierence in response rates in patientswith bipolar or unipolar depression (107, 108).

Novel or experimental agents

Data were previously described demonstrating thebenets of adjunctive use of the following agents:pramipexole (level 2), eicosapentaenoic acid (EPA)(level 2), riluzole (level 3), topiramate (level 3), andN-acetyl cysteine (NAC) (level 2) (13).Additional open-label data support the use of

adjunctive riluzole and adjunctive NAC (109, 110).Patients in a small, open-label, imaging studyreported improvements in depressive symptomswith riluzole (level 3) (109), while data from a large(n = 149), eight-week, open-label trial foundsignicant improvement in depressive symptomswith adjunctive NAC in patients with BD I, II, orNOS depression (110).Preliminary data are also available to support

other novel treatments not previously investigated

in patients with bipolar depression, includingadjunctive ketamine, armodanil, and chronother-apy. A two-week, crossover RCT assessing adjunc-tive ketamine infusion in patients with treatment-resistant bipolar depression identied a robust earlyantidepressant eect (within 40 min), with improve-ment remaining signicant versus placebo throughday three (level 2) (111). An eight-week RCT in 257patients with BD I depression reported a trendtoward greater improvement in depressive symp-toms with adjunctive armodanil versus placebo onsome, but not all depression symptom scales (level 2)(112). Adjunctive combined chronotherapy (sleepdeprivation, exposure to bright light, and sleep-phase advance) demonstrated a more rapid andsustained antidepressant response compared tomedication alone (lithium + antidepressant) in aseven-week RCT in 49 patients with bipolar depres-sion (level 3) (113). Patients receiving adjunctivechronotherapy experienced a signicantly greaterreduction in depressive symptoms within 48 hours,which was sustained throughout the seven weeks.

Not recommended for the treatment of acute bipolardepression:

Ziprasidone monotherapy. Data are now availablefrom two negative RCTs of ziprasidone monother-apy in BD I depression (level 1, negative) (114,115). The trials have not yet been published, butresults are available at http://www.clinicaltrials.gov(114, 115). Both were large trials (n = 381 andn = 504, respectively) in patients with BD Idepression, and both demonstrated no signicantimprovements in depression scores compared toplacebo. While ziprasidone is not recommendedfor bipolar depression, patients who are usingziprasidone with benet (initiated during mania)do not need to have it discontinued.

Adjunctive therapies with negative data requiringfurther study:

Adjunctive ziprasidone. A large RCT assessingadjunctive ziprasidone (mean dose 90 mg) addedto therapy with lithium, divalproex, or lamotriginein 298 patients with BD I depression found nosignicant dierence between active treatment andplacebo (level 2, negative) (116).

Adjunctive aripiprazole. While open-label trialssuggest a benet of adjunctive aripiprazole forthe treatment of bipolar depression (level 3) (117,118), a small RCT did not nd a signicant eectcompared to placebo (level 2, negative) (119). In asix-week RCT, 23 inpatients with bipolar depres-


11

sion on lithium or divalproex were given open-labelcitalopram, and randomized to adjunctive aripip-razole or placebo. Depressive symptoms improved,with no signicant dierences between treatmentgroups (level 2 negative) (119).In a 16-week, prospective, open-label trial,

aripiprazole (add-on or monotherapy) was associ-ated with a signicant decline in depressive symp-toms over 16 weeks among 85 patients with bipolardepression who were unresponsive to other med-ications (lithium, anticonvulsants, or antipsychoticagents) (level 3) (117). There was also a signicantreduction in self-rated anhedonia among patientswith BD I depression treated with aripiprazolein the same study (120). Another small (n = 20),six-week, open-label trial demonstrated improve-ment in depressive symptoms, with a 44% responserate with aripiprazole as add-on or monotherapyfor BD I, II, or NOS depression (level 3) (118).

Adjunctive levetiracetam. A six-week RCT in 32patients with bipolar depression found no signi-cant dierences in the change in depression scoreswith adjunctive levetiracetam versus placebo (level2, negative) (121).

Clinical questions and controversies

What is the role of antidepressants in patients withbipolar depression?

The role of antidepressants in patients withbipolar depression remains one of the most con-troversial areas in psychiatry. Antidepressants arethe most commonly used treatments for bipolardepression (122, 123) as clinicians continue tobelieve that, based on their clinical experience,these are eective for bipolar depression. However,the limited, but growing body of clinical trial datahas not been consistent in supporting their role.For instance, OFC was shown to be more eectivethan placebo or olanzapine monotherapy (93, 95),but the combination of paroxetine or bupropionwith a mood stabilizer was not more eective than amood stabilizer plus placebo (124). However, thisstudy had severe methodological limitations; mostpatients were also participating in a psychotherapytrial, and an unknown proportion of patientscontinued to use the previous antidepressant theyhad been on at baseline. In another study, paroxe-tine monotherapy (20 mg day) was not superior toplacebo in improving bipolar depressive symptoms(125). It is unknown if higher doses of paroxetinewould have been more eective.Although individual studies are contradictory,

the most recent meta-analysis, which included 15

RCTs (126), found a strong trend for superiority ofantidepressants over placebo for the acute treatmentof bipolar depression (p = 0.06). Antidepressantswere not associated with a signicantly increasedrisk of manic switch (126). Most negative studies ofantidepressants for bipolar depression to date haveemployed paroxetine as the antidepressant (91, 125,127). A meta-analysis of the ecacy of antidepres-sants in unipolar depression (128) suggested thatclinically important dierences exist between vari-ous antidepressants in terms of ecacy and accept-ability. Interestingly, paroxetine was inferior to anumber of other antidepressants in this meta-anal-ysis. The risk of manic hypomanic switch does notappear to be a major concern with modern antide-pressants when used in conjunction with a moodstabilizer or an atypical antipsychotic agent, at leastduring short-term treatment; therefore, safety doesnot appear to be a signicant issue during the acutetreatment of bipolar depression. An importantcaveat is that the current denition of switchrequires threshold mania; milder switches, whichare common, are not captured by the defaultdenition (129). Similarly, the metrics of cycleacceleration are not captured in current denitionsor trial designs (130).Given the above, we believe that the following

conclusions and recommendations are warrantedregarding the use of antidepressants for bipolardepression: (i) SSRIs (other than paroxetine) andbupropion could be used as rst-line treatments inconjunction with a mood stabilizer for acute short-term treatment of bipolar depression, with theobjective of tapering and discontinuing antidepres-sants 68 weeks after full remission of depression;(ii) avoid the use of tricyclic antidepressants andvenlafaxine (131, 132) as they are associated withan increased risk of manic switch; (iii) antidepres-sants should not be used to treat a current mixedepisode or in patients with a history of rapidcycling; (iv) monotherapy with antidepressants isnot recommended for bipolar depression.

Section 5. Maintenance therapy for bipolar disorder

Adherence

New data provide further insight into adherence tomaintenance therapy in patients with BD. Inseveral analyses, adherence was positively associ-ated with higher satisfaction with medication,monotherapy, a college degree, and fear of relapse,and was negatively associated with illness factors(substance use, previous hospitalization, psychoticsymptoms, reduced insight into illness), medicationfactors (side eects, no perceived daily benet,

Yatham et al.

12

diculties with medication routines), and patientattitudes (belief that medications are unnecessary,negative attitudes toward medications, perceivedchange in appearance, perceived interference withlife goals) (133139). Under-dosing can also lead tohigher discontinuation rates; patients receivinglower doses of ziprasidone had signicantly higherdiscontinuation rates than those receiving mediumor high doses (140).Non-adherence has been linked to a high

frequency of episodes (particularly depressiveepisodes), a higher risk of hospitalization andemergency room visits, as well as higher employeecosts of absenteeism, short-term disability, andworkers compensation (137, 141144). In ananalysis involving UK data, the direct costs ofcare were two to three times higher in patients whorelapsed compared to those who did not over the612-month follow-up (145).

Predictors of recurrence

In observational studies, predictors of symptom-atic remission and recovery during 12 years offollow-up in patients with manic episodes included:Caucasian ethnicity, a previous manic episode,good social functioning (no work or social impair-ment, living independently or with family), outpa-tient treatment, and being neither satised nordissatised with life (146, 147).In patients with rapid cycling treated with lithium

or divalproex, increased risk for non-stabilizationwas associatedwith a history of recent substance usedisorder (SUD), early-life verbal abuse, femalegender, and late onset of rst depressive episode(148). Among responders to long-term lithiumtherapy, the risk of recurrence was higher in thosewith atypical features (mainly mood-incongruentpsychotic symptoms), inter-episodic residual symp-tomatology, and rapid cycling (149).

Psychosocial interventions for maintenance therapy

As reported in previous iterations of these guide-lines, data have supported the benets of adjunc-tive psychoeducation, CBT, family therapy, andIPSRT in reducing recurrences and improvingsymptoms in patients with BD (13). However,recent meta-analyses assessing the ecacy of psy-chotherapies for patients with BD have reacheddiscordant conclusions. One meta-analysis of fourRCTs concluded that CBT had a small eect sizefor depressive symptoms compared to treatment asusual or wait-list controls (150). A second analysisincluded 12 trials and found low to medium eectsizes associated with adjunctive CBT at the end of

treatment and at follow-up (151). Anothermeta-analysis concluded that CBT was likely notan eective treatment strategy for the prevention ofrelapse in BD (152).Given the transient benet of CBT in the rst

major studyof this strategy inpatientswithBD(153)and the negative results of a large RCT (154), thetrue benet of CBT is unclear, beyond its commoncore element of psychoeducation (155). However,two new RCTs provide some promise. In one studycomparing a group CBT intervention to treatmentas usual in 50 patients, group CBT was associatedwith a longer median time to relapse, but nodierences in time to recurrence or number ofepisodes (156). In a 14-week study comparing theCBT (n = 27) and control pharmacotherapy(n = 14) groups, there was a slight, although non-signicant, reduction in depressive symptoms in theCBT group (157). Both of these small studiesdemonstrated the feasibility of group CBT ratherthan its ecacy, and proper replication is needed. Inaddition, a previously reported RCT (n = 204)(22), which has now been published, compared sixsessions of group psychoeducation to 20 sessions ofindividual CBT. Both treatments demonstrated asignicant benet in terms of mood stability andreduction of recurrence, but there were no dier-ences between the two treatments. Since thepsychoeducation treatment was designed to bedelivered by psychiatric nurses and was documen-ted to be much less expensive than individualCBT, the study suggested that group psychoedu-cation should be prioritized as a rst universalpsychosocial treatment for BD.RCTs of adjunctive group psychoeducation

programs demonstrated a longer time to recur-rence, fewer recurrences of any type, less timeacutely ill, and fewer days of hospitalization during15 years of follow-up (158, 159). A 12-weekdyadic (patientcompanion) group-based psycho-education program demonstrated signicantly low-er relapse rates and a longer time to relapsecompared to treatment as usual during a 60-weekfollow-up (160).

Pharmacological treatments for maintenance therapy

As discussed in previous guideline iterations, almostall modern maintenance studies have used anenriched design. The only exceptions are some ofthe older maintenance studies with lithium, whichshowed the ecacy of lithium for maintenancetreatment in non-enriched samples. Given the e-cacy of various treatments with enriched designstudies, it makes intuitive sense that, in general, thetreatment that worked during the acute phase is


13

likely to be eective in themaintenance phase [pleasesee the original 2005 guidelines: Section 5. Mainte-nance therapy for bipolar disorder: General principles(1)]. Based on new evidence, asenapine alone and asadjunctive therapy have been added as third-lineoptions (Table 5.5) (13).

First-line options: Lithium, divalproex, olanzapine,and quetiapine (for both depression and mania), aswell as lamotrigine (primarily for preventing depres-sion), risperidone long-acting injection (LAI) andziprasidone (primarily for preventing mania) con-tinue to be rst-line monotherapy options formaintenance treatment of BD (13). Quetiapine,risperidone LAI (mania), aripiprazole (mania), andziprasidone (mania) are also recommended asadjunctive therapy with lithium or divalproex.A systematic review of pharmacological inter-

ventions for the prevention of relapse in BDincluded 34 RCTs and quasi-RCTs, and concludedthat lithium, olanzapine, and aripiprazole hadsignicant eects in the prevention of manicrelapses, as did divalproex, lamotrigine, and imip-ramine in the prevention of depressive symptoms(161). A meta-analysis of 20 RCTs (n = 5364)assessing the relative risk for relapse in patientswith BD in remission conrmed the ecacy oflithium, divalproex, lamotrigine, and a number ofatypical antipsychotic agents in preventing relapseto any episode versus placebo (162).

Lithium divalproex. The Bipolar Aective disorderLithium ANti-Convulsant Evaluation (BAL-ANCE) study randomized 330 patients with BDto open-label lithium monotherapy, divalproexmonotherapy, or the combination after an activerun-in period on the combination (163). Both the

combination and lithium monotherapy weresignicantly more eective than divalproex mono-therapy in preventing relapse during up to twoyears of follow-up. Combination therapy was notsignicantly more eective than lithium alone. Thisstudy, however, had a number of methodologicallimitations, including an open design; hence, thesendings need to be conrmed in double-blind trialsbefore rm conclusions can be drawn.

Lamotrigine. During a one-year extension phase ofan RCT, median time to relapse or recurrence waslonger among responders receiving lamotriginecompared to those receiving placebo as an add-on to lithium paroxetine (ten versus 3.5months)(164).An open, randomized trial in patients with BD I

found no dierences in maintenance eectivenessbetween lithium (n = 78) and lamotrigine (n = 77)(165). Among patients followed for at least veyears, practically no patients were maintainedsuccessfully on monotherapy with either drug.

Olanzapine.Ameta-analysis ofveRCTs found thatolanzapine as monotherapy or an adjunct to lithiumor divalproex was more eective than adjunctplacebo in preventing a manic, but not any type ofdepressive, episode (166). The analysis concludedthat olanzapine may prevent manic episodes only inpatients who have responded to olanzapine for anacute episode and who have not previously had asatisfactory response to lithium or valproate.In two recent RCTs, olanzapine was included as

an active control arm, and in the continuationphase demonstrated a signicantly longer time torecurrence than either risperidone LAI (167) orpaliperidone ER (168).

Table 5.5. Recommendations for maintenance pharmacotherapy of bipolar disorder

First line Monotherapy: lithium, lamotrigine (limited efficacy in preventing mania), divalproex, olanzapinea,quetiapine, risperidone LAIb, aripiprazoleb

Adjunctive therapy with lithium or divalproex: quetiapine, risperidone LAIb, aripiprazoleb, ziprasidoneb

Second line Monotherapy: carbamazepine, palideridone ERc

Combination therapy: lithium + divalproex, lithium + carbamazepine, lithium or divalproex + olanzapine,lithium + risperidone, lithium + lamotrigine, olanzapine + fluoxetine

Third line Monotherapy: asenapinec

Adjunctive therapy: phenytoin, clozapine, ECT, topiramate, omega-3-fatty acids, oxcarbazepine,gabapentin, asenapinec

Not recommended Monotherapy: gabapentin, topiramate, or antidepressantsAdjunctive therapy: flupenthixol

LAI = long-acting injection; ER = extended release; ECT = electroconvulsive therapy.aGiven the metabolic side effects, use should be carefully monitored.bMainly for the prevention of mania.cNew or change to recommendation.

Yatham et al.

14

A large, observational study (EMBLEM) in-cluded 1076 patients in a comparison of olanzapinemonotherapy or as an adjunct and found nosignicant dierence in rates of improvement,remission, or recovery, but signicantly lowerrelapse rates with olanzapine alone compared toadjunctive olanzapine (p = 0.01) over the two-year follow-up (169).

Quetiapine. Three of the ve large RCTs describedin the 2009 guidelines (3), which demonstrated theecacy of quetiapine alone or in combination withlithium divalproex for maintenance therapy inBD, have now been published (level 1) (170172).The eight-week acute-phase results of the EM-BOLDEN I (90) and II (91) trials have also beenpublished but the long-term data remain availableonly in abstract form.

Risperidone LAI. Risperidone LAI monotherapy(level 2) (173) and adjunct (level 2) (174) werepreviously recommended as rst-line maintenancetherapies (3), based on RCTs presented in abstractform that have now been published. An additional18-month continuation study has now providedlevel 1 evidence for maintenance risperidone LAI.Time to recurrence of any mood episode wassignicantly longer with risperidone LAI comparedto placebo (level 1) (168). However, risperidoneLAI was less eective in preventing relapsecompared to olanzapine.In addition, a small (n = 29), long-term, open

trial demonstrated improvements in treatmentadherence, reductions in any relapse rates, andreductions in re-hospitalization rates with adjunc-tive risperidone LAI during a mean two-yearfollow-up (175).

Aripiprazole. Aripiprazole monotherapy has dem-onstrated ecacy for the prevention of manicepisodes in the maintenance treatment of patientswith BD I and was included in the 2009 guidelineas a rst-line maintenance therapy for the treat-ment and prevention of mania (level 1) (3).A 52-week, relapse prevention RCT demon-

strated the ecacy of aripiprazole as an adjunct tolithium or divalproex in patients with manic mixedepisodes and an inadequate response to lithium ordivalproex (176). Patients (n = 337) in remissionfor 12 weeks who were randomized to continueadjunctive aripiprazole had a lower rate of relapseto manic (5% versus 15%, p = 0.013) but notdepressive (10% versus 13%, p = 0.384) episodes.Since adjunctive aripiprazole demonstrated ecacyfor the prevention of any mood episode or manicepisodes, but not depressive episodes, it has been

added as a rst-line maintenance therapy for theprevention of manic episodes (level 2).Another 52-week relapse prevention study in 351

BD I patients with a manic mixed episode, avail-able in abstract form, showed a non-signicanttrend to lower rates of manic mixed relapse witharipiprazole versus placebo added to lamotrigine(11% versus 23%, p = 0.058) (177). There was anon-signicant trend to lower rates of any relapse,and no eect on depressive relapse rates.

Ziprasidone. An RCT demonstrating the ecacy ofadjunctive ziprasidone for maintenance treatmentof BD, previously available in abstract form, hasnow been published (level 2) (178). Adjunctiveziprasidone (80160 mg day) demonstrated e-cacy for the prevention of manic, but not depres-sive, episodes.

Second-line options:

Carbamazepine. A meta-analysis of four RCTs,including 464 patients, supports previous conclu-sions that maintenance treatment with carbamaz-epine has a similar ecacy to lithium for rates ofrelapses, with the caveat that there were fewerwithdrawals due to adverse eects with lithium(level 2) (179). Given the signicant tolerabilityissues with carbamazepine and the diculty incombining this agent with other psychotropicmedications because of its hepatic microsomalenzyme induction properties, carbamazepine con-tinues to be recommended as a second-line option.

OFC. A six-month, continuation, RCT comparingOFC and lamotrigine monotherapy in patientswith bipolar depression that was previously cited inabstract form has now been published (94). OFCwas associated with a signicantly greater improve-ment in depressive and manic symptoms, but therewere no dierences in relapse rates of bipolardepression among responders to acute treatment.In the observational EMBLEM study in 1076patients with BD mania, adjunctive olanzapine wasless eective than olanzapine monotherapy inpreventing any relapse (169). By contrast, anotherstudy of open-label continuation treatment, in 114patients with bipolar depression who were re-sponders to OFC found that signicantly morepatients maintained their response to OFC com-pared to olanzapine monotherapy (180).

Paliperidone ER. In a three-month continuationstudy, continued paliperidone ER was signicantlymore eective than placebo in preventing relapse inprior paliperidone ER responders (level 2) (168).


15

However, paliperidone ER was less eective inpreventing relapse compared to the olanzapineactive control group.

Third-line options:

Asenapine. Nine-week and 40-week extensionphase results from two pooled three-week RCTsdemonstrated the maintenance of benets withasenapine monotherapy and olanzapine in pa-tients with BD mania (YMRS reduction )24.4and )23.9 at week 12, )28.6 and )28.2 at week52, respectively) (level 2) (58, 59). At one year, aworsening of mania was reported in 2.6% ofasenapine patients and 1.9% of olanzapinepatients, while a switch to a depressive episodeoccurred in 0% of asenapine and 3.0% ofolanzapine patients (59). Time to response wassignicantly longer with asenapine compared toolanzapine (p = 0.0127). In this trial, patients inthe placebo groups of the original RCTs wereblindly switched to asenapine.In addition, 40-week extension phase results

have now been reported in abstract form fromthe trial of adjunctive asenapine added to lith-ium divalproex compared to placebo (62). Of theoriginal 318 patients, 71 completed the 40-weekextension; improvements in mania ()17.2 versus)19.7) and depression ()3.3 versus )3.9) scores at52 weeks were seen in both the asenapine andplacebo groups, but the extension was not poweredfor statistical comparisons.Based on evidence of maintenance of benets,

but in the absence of relapse prevention data andclinical experience, asenapine is recommended as athird-line option for maintenance therapy.

Rapid cycling

In a six-monthRCT, patientswith BD (recentmanicepisode), SUD, and rapid cycling, who wereresponders to lithium plus divalproex, were ran-domized to continue combination therapyor lithiumalone (181). Of 149 patients enrolled into the open-label acute stabilization phase, 31 were assigned tomaintenance treatment and 55% relapsed. Therewere no signicant dierences between combinationand monotherapy in the rate of relapse or time torelapse. However, given the small sample size, thestudy was likely underpowered to detect this.In the STEP-BD study, among patients who

were responders to adjunctive antidepressants andcontinued this treatment, those with a rapid-cycling course had three times more depressiveepisodes compared to those without rapid cycling(1.29 versus 0.42 episodes year, p = 0.04) (182).

Mixed states

Several open-label and post-hoc analyses provideadditional insight into the role of atypical antipsy-chotic agents in the management of patients withmixed episodes or psychotic symptoms. In a post-hoc analysis of patients with mixed episodes(n = 121), where responders to olanzapine wererandomized to continue olanzapine or switch toplacebo, there were signicant reductions in relapserates with olanzapine compared to placebo (59.2%versus 91.1%, p < 0.001) (183). During 52 weeksof exible-dosed ziprasidone (40160 mg day)open-label extension treatment (n = 65), patientswith manic or mixed episodes, with or withoutpsychotic symptoms, showed comparable improve-ments in mania and overall subtypes acrosssubgroups (184). In a 24-week, open-label trial ofadjunctive risperidone in 114 patients with mixedor manic episodes, signicant reductions frombaseline in manic, depressive, and overall symptomscores were observed with combination therapy inboth the manic and mixed groups (185).

Section 6. Special populations

Issues in the management of BD in women

The management of BD in women can presentadditional challenges associated with the repro-ductive cycle.

Premenstrual syndrome (PMS) premenstrual dys-phoric disorder(PMDD): In the longitudinal STEP-BD study, among women with BD (n = 293),those with premenstrual exacerbation had moreepisodes (primarily depressive), more depressiveand manic symptoms overall, a shorter time torelapse, and greater symptom severity (186). Pre-menstrual exacerbation may predict a more symp-tomatic and relapse-prone phenotype in womenwith BD.A study in 61 women with BD I or BD II and

122 healthy women found that moderate-to-severePMS PMDD occurred signicantly more fre-quently in patients with BD II (51.6%) comparedto the healthy women (19.7%) (187). Similarly, in astudy of 92 women with BD I or BD II, patientswith PMDD were more likely to have BD II andcyclothymia than were patients without PMDD(188).Among patients seen at a specialty gynecology

clinic for chronic pelvic pain, those with endometri-osis (n = 27) were more likely to have BD (44.4%)and a poorer quality of life than women with non-endometriosis pelvic pain (n = 12) (0%) (189).

Yatham et al.

16

Pre-conception: Providing appropriate educationand guidance to patients considering pregnancy orwho may have recently become pregnant isan important component of BD management.Pre-conception counseling should include a carefulreview of risks and benets and a treatmentplan for ongoing monitoring. An analysis of alarge claims database including 16385 women ofchild-bearing age with BD or MDD revealed 1308women who were receiving a category D (12%) orcategory X (1%) medication during pregnancy(190). The most frequently used psychotropicdrugs were paroxetine, alprazolam, lorazepam,divalproex, lithium, and temazepam.

Pregnancy: The management of pregnant womenwith BD should incorporate careful planning.Updated recommendations on the use of psychiat-ric medications during pregnancy and lactation areavailable from the American Congress of Obste-tricians and Gynecologists (ACOG), and the read-er is referred to this Practice Bulletin for moreinformation (191). In addition, please visit theCanadian Hospital for Sick Children Motheriskwebsite (http://www.Motherisk.org). Therefore, inthe following section, we will provide only a briefupdate of some of the new data in patients withBD.The risk of teratogenicity associated with use

of psychotropic medications (please see ACOGrecommendations in Table 6.2) during the rsttrimester should be carefully weighed against therisks to the mother and the fetus of an untreatedmood episode. Psychotropic medications can beused in the second and third trimester if necessary.If lithium is used during the second and thirdtrimester, the serum lithium levels should bemonitored closely because of changes in bloodvolume during pregnancy, and the dose should beadjusted accordingly to maintain levels in thetherapeutic range.Cohort studies in various patient populations

conrm the teratogenic risk associated withdivalproex (192194) and carbamazepine (195)during pregnancy. In a case-controlled series of52 pregnancies, topiramate was associated withreduced birth weight but no decrease in gestationalage and no increase in structural defects (196).Several cohort studies in various patient popu-

lations found that, during pregnancy, antidepres-sants did not confer an increased risk of majorcongenital anomalies compared to unexposedcontrols (197, 198). However, in one analysisantidepressant use was associated with increasedrates of pregnancy complications, including in-duced delivery, caesarean section, and preterm

birth, as well as increased risk of persistentpulmonary hypertension of the newborn (199).Some analyses suggest increased risks with

individual antidepressants (198200), while othersdo not (197). In one analysis, there were associa-tions of uoxetine with ventricular septal defects,paroxetine with right ventricular outow tractdefects, and citalopram with neural tube defects,although the absolute risk for these specic eectswas small (198). The rate of major anomalies(primarily cardiovascular) in birth outcome amongpregnant patients (n = 314) with rst trimesterexposure was 4.7% with uoxetine and 5.2%with paroxetine, compared to 2.5% in controlpatients (n = 1467), in a multicentre, prospective,controlled study (200). Cardiovascular defects havealso been associated with paroxetine (199, 201) andtricyclic antidepressant exposure (199).A Canadian neonatal record analysis (n =

119547 live births) concluded that prenatal expo-sure to combination therapy with SSRI andbenzodiazepines conferred a higher incidence ofcongenital heart disease when compared to noexposure (202). SSRI monotherapy was not asso-ciated with an increased risk for major congenitalanomalies, but was associated with an increasedincidence of atrial septal defects.A retrospective chart review including 30092

total deliveries identied one major malformationamong 16 of the mothers who were treated withatypical antipsychotic agents during their preg-nancy (203). The Food and Drug Administration(FDA) issued a safety alert regarding the risks tonewborns associated with prenatal exposure totypical or atypical antipsychotic drugs (204). Thenew drug labels now contain information about thepotential risk for abnormal muscle movements andwithdrawal symptoms including agitation, abnor-mal muscle tone, tremor, sleepiness, breathing, andfeeding diculties in newborns.In an observational, prospective study in 14

women with BD undergoing maintenance treat-ment with lithium during pregnancy, the lithiumconcentration ratio from infant to mother was 0.96and congenital malformations were greater in thosereceiving higher doses versus lower doses; however,no signicant dierences in neonatal outcomes(gestational age weight, Apgar scores, or hospitalstay) were noted (205).

Postpartum period: Distinguishing bipolar depres-sion from MDD can be challenging in thepostpartum period because of a lack of screeninginstruments designed specically for use during thisperiod (206). A Mood Disorder Questionnaire(MDQ) validation study concluded that, with


17

Table 6.2. Psychiatric medications in pregnancy and lactation

AgentPregnancy riskcategorya

American Academy ofPediatrics rating

Lactation riskcategoryb

Anxiolytic medications

BenzodiazepinesAlprazolam D Unknown, of concern L3Chlordiazepoxide D N A L3Clonazepam D N A L3Clorazepate D N A L3Diazepam D Unknown, of concern L3, L4 if used chronicallyLorazepam D Unknown, of concern L3Oxazepam D N A L3

Benzodiazepines for insomniaEstazolam X N A L3Flurazepam X N A L3Quazepam X Unknown, of concern L2Temazepam X Unknown, of concern L3Triazolam X N A L3

Non-benzodiazepine anxiolytics and hypnoticsBuspirone B N A L3Chloral hydrate C Compatible L3Eszoplicone C N A N AZaleplon C Unknown, of concern L2Zolpidem B N A L3

Antiepileptic and mood-stabilizing medications

Lithium carbonate D Contraindicated L4Valproic acid D Compatible L2Carbamazepine D Compatible L2Lamotrigine C Unknown L3

Antidepressants

Tricyclic and heterocyclic antidepressantsAmitriptyline C Unknown, of concern L2Amoxapine C Unknown, of concern L2Clomipramine C Unknown, of concern L2Desipramine C Unknown, of concern L2Doxepin C Unknown, of concern L5Imipramine C Unknown, of concern L2Maprotiline B N A L3Nortriptyline C Unknown, of concern L2Protriptyline C N A N A

Selective serotonin reuptake inhibitorsCitalopram C N A L3Escitalopram C N A L3 in older infantsFluoxetine C Unknown, of concern L2 in older infants, L3 if used in

neonatal periodFluvoxamine C Unknown, of concern L2Paroxetine D Unknown, of concern L2Sertraline C Unknown, of concern L2

Other antidepressantsBupropion B Unknown, of concern L3Duloxetine C N A N AMirtazapine C N A L3Nefazodone C N A L4Trazodone C Unknown, of concern L2Venlafaxine C N A L3

Antipsychotic medications

Typical antipsychotic agentsChlorpromazine C Unknown, of concern L3Fluphenazine C N A L3Haloperidol C Unknown, of concern L2Loxapine C N A L4Perphenazine C Unknown, of concern N APimozide C N A L4

Yatham et al.

18

alternate scoring, the MDQ may be a usefulscreening instrument for BD in the postpartumperiod (206, 207). The optimal cut-o score waseight or more endorsed symptoms without thesupplementary questions (sensitivity 88% andspecicity 85%) (208).In a well-designed longitudinal study involving

344 pregnant women, there was an 8.4-foldincrease in hypomanic symptoms in the earlypostpartum period (11.7%) compared to duringthe rst trimester (1.4%) and to eight weekspostpartum (4.9%) (209). Although hypomanicsymptoms are common in the early puerperium,they are often overlooked, leading to a misdiag-nosis of MDD (210). This may be due in part to theabsence of hypomania from the postpartum-onsetspecier in the Diagnostic and Statistical Manualof Mental Disorders, 4th editiontext revised(DSM-IV-TR) (211). However, a study fromSpain suggested reconsideration of the DSMpostpartum-onset specier, as it did not appearto inuence prognosis or functioning in womenwith BD (212).According to a Danish population-based study

of new mothers, the period of highest risk forpsychiatric readmission was 10 to 19 days post-partum [relative risk (RR) = 2.7], and the periodof lowest risk was during pregnancy (RR = 0.5).Previous diagnosis of BD was the strongestpredictor of readmissions 1019 days postpartum(RR = 37.2), with 27% of women with BDbeing readmitted within the rst year (213).Similarly, psychotic illness has been shown topeak immediately following a rst childbirth(214).ACOG recommendations for the use of psychi-

atric medications during lactation are also shownin Table 6.2 (191).

Menopause: In a STEP-BD analysis of 164 patientswith BD followed for an average of 30 months,menopausal transition was associated with signif-icantly more visits due to depressive symptoms andfewer euthymic visits compared to a comparisongroup of non-menopausal women and men (215).

Issues in the management of BD in children andadolescents

Complete treatment recommendations for pediat-ric BD are beyond the scope of these guidelines; thereader is referred to specic guidelines for themanagement of children and adolescents with BD,such as those developed by the American Academyof Child and Adolescent Psychiatry (AACAP)(216). Therefore, in the following section, we willprovide only a brief overview of some of the issuesin this population without oering recommenda-tions for levels of treatment.

Presentation and diagnosis:Across the world, BD isthe fourth leading cause of disability amongadolescents (1519 years) (217). The presentationand diagnosis of BD in children and adolescentsremains controversial. Increasing billing and dis-charge diagnoses of pediatric BD contradict astable epidemiologic prevalence, suggesting thatdiagnostic criteria for BD may not be systemati-cally applied in some clinical settings (218, 219).However, much of the controversy regardingpediatric BD has focused on the group of youthwith severe, chronic, non-episode irritability, whichhas led to the proposed DSM-5 diagnosis ofdisruptive mood dysregulation disorder (DMDD)(220). The reader is referred to recent publicationsfocusing specically on dierential diagnosis anddevelopmental considerations in ascertaining

Table 6.2. (Continued).

AgentPregnancy riskcategorya

American Academy ofPediatrics rating

Lactation riskcategoryb

Thioridazine C N A L4Thiothixene C N A L4Trifluoperazine C Unknown, of concern N A

Atypical antipsychotic agentsAripiprazole C N A L3Clozapine B Unknown, of concern L3Olanzapine C N A L2Quetiapine C Unknown, of concern L4Risperidone C N A L3Ziprasidone C Unknown, of concern L4

[Copyright (2008) Wolters Kluwer Health; reprinted with permission from (191)]. N A = not available.aThe US Food and Drug Administration classifies drug safety using the following categories: A = controlled studies show no risk; B = noevidence of risk in humans; C = risk cannot be ruled out; D = positive evidence of risk; X = contraindicated in pregnancy.bLactation risk categories are listed as follows: L1 = safest; L2 = safer; L3 = moderately safe; L4 = possibly hazardous; L5 = contra-indicated.


19

manic symptoms (221, 222), and regarding con-cerns about the DMDD diagnosis (223). Prospec-tive studies of children and adolescents withrigorously dened BD demonstrate that this is anepisodic illness that continues into young adult-hood and is characterized by substantial impair-ment and morbidity (224226).Based on a meta-analysis of 12 epidemiological

studies in patients between the ages of seven and21 years (n = 16222), the overall prevalence of BDwas 1.8% (219). Of note, rates of BD in theseepidemiologic studies did not increase over timeand did not dier for studies within versus outsideof the USA. Among Canadian adolescents andyoung adults (1524 years), the CCHS 1.2 surveyrecorded a lifetime prevalence of BD of 3.0%(2.1% in those aged 1518 years; 3.8% in thoseaged 1924 years) (227). In a claims database, theone-year rate of a new diagnosis of BD amongpatients 17 years of age was 0.23%. Misdiagnosiswas common, with 47% being diagnosed withdepressive disorder and 37% with disruptivebehavior disorder in the previous year (228). Inthe Course and Outcome of Bipolar Youth(COBY) study (n = 364), rst episodes in patients 12 years were generally depressive, while those inpatients

Pharmacological management

Atypical antipsychotics. The AACAP published apractice parameter on the use of psychotropicmedication in children and adolescents, and thereader is referred to those guidelines for moredetails (245). Provided below is an overview ofcurrent data of the ecacy of atypical antipsy-chotic agents for the treatment of BD in youngerpatients.The US FDA has now approved quetiapine for

the rst-line treatment of acute manic mixedepisodes in pediatric patients. In light of safety tol-erability concerns, olanzapine (weight gain andmetabolic disturbances) and ziprasidone (QT pro-longation) were approved as second-line treat-ments only (246).A meta-analysis of nine RCTs included 1609

pediatric patients with acute BD mania and foundsignicantly greater improvements in YMRSscores in the atypical antipsychotic agent groupand mood stabilizer group relative to placebo. Theeect sizes were greater for the atypical antipsy-chotic agent group compared to a mood stabilizergroup (eect size 0.65 versus 0.20). However, themood stabilizer group included studies on topira-mate and oxcarbazepine, neither of which demon-strated ecacy as mood stabilizers. Further,medication-associated weight gain was greater withatypical antipsychotic agents than with moodstabilizers (eect size 0.53 versus 0.10) (44).In a small eight-week RCT in 32 adolescents (age

1218 years), quetiapine monotherapy was notsignicantly better than placebo for the treatmentof bipolar depression (247). This negative trialrequires replication in light of the small samplesize, high placebo response rates, and robustevidence of quetiapine ecacy in bipolar depres-sion in adults group (13). Quetiapine monotherapyhas also demonstrated ecacy as acute and main-tenance treatment in small, open-label studies (248,249).A pooled analysis of four olanzapine trials

(two RCTs, two open-label) in adolescents (age1317 years) with BD or schizophrenia revealedsignicantly more weight gain compared to adultpatients (7.4 kg versus 3.2 kg) in up to 32 weeksof treatment (250). Adolescents also experiencedsignicant changes in fasting glucose, total choles-terol, triglycerides, alanine aminotransferase, andprolactin. In an eight-week, open-label study theaddition of topiramate to olanzapine therapy in 40pediatric patients with BD resulted in signicantlyless weight gain than with olanzapine monothera-py, with no dierences in mania symptom scores(251).

In addition to the previously cited RCT data,further open-label data support the ecacy andtolerability of ziprasidone therapy in youth withBD (252, 253).A three-week RCT in 169 pediatric patients with

BD mania mixed episodes demonstrated greaterreductions in mania scores with risperidone versusplacebo (254). In a six-week RCT in 66 youth,risperidone resulted in more rapid improvement inmanic symptoms and a signicantly greater rate ofremission (63% versus 33%) compared to divalp-roex (255). Primary outcomes from the eight-weekTreatment of Early-Age Mania (TEAM) study of279 medication-nave youth (age 615 years)with BD I mania mixed episodes were recentlypublished (256). TEAM was a controlled, random-ized, no-patient-choice study comparing lithium,divalproex, and risperidone. The response rate forrisperidone (68.5%) was signicantly greater thanfor lithium (35.6%) and divalproex (24.0%), whichdid not dier signicantly from each other. How-ever, increasedweight gain, bodymass index (BMI),and serum prolactin levels were also signicantlygreater with risperidone than with the other medi-cations. Similar results were seen in a small cohortstudy in which risperidone resulted in a faster andgreater reduction of symptom scores versus divalp-roex (257). Hence, although it is premature toconclude that atypical antipsychotic agents havegreater ecacy than mood stabilizers in pediatricmania, a convergence of data suggest that this maybe the case.An RCT demonstrating the ecacy of aripip-

razole for BD I manic mixed episodes in pediatricpatients that was previously cited in abstract formhas now been published (258). Another six-weekRCT in 43 ped

canmat guidelines -2013 update

Documents