Show Summary Details
Page of

(p. 45) Treatment of Depression and Bipolar Disorder 

(p. 45) Treatment of Depression and Bipolar Disorder
(p. 45) Treatment of Depression and Bipolar Disorder

Moira Rynn

, Paul Crits-Christoph

, David Brent

, Robert Findling

, and Karen Dineen Wagner

Page of

PRINTED FROM OXFORD CLINICAL PSYCHOLOGY ONLINE ( © Oxford University Press, 2020. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a title in Oxford Clinical Psychology Online for personal use (for details see Privacy Policy and Legal Notice).

date: 25 October 2020

(p. 46) Although it is clear that adolescent mood disorders exist and lead to significant immediate and lifelong impairment for the child, as shown in Chapter 1, as compared to the adult treatment literature there exists limited treatment research in this special population. This has led clinicians to consult with the adult literature to provide guidance in their treatment approaches. In fact, the adult research studies have provided the template for the present interventions being explored in adolescent treatment studies. To appreciate how the adult literature informs the field of adolescent mood disorder treatments, the current status of the adult literature for both psychosocial and psychopharmacologic treatments will be reviewed. This will be followed by an appraisal of the same treatment areas for adolescents.

Psychosocial Treatments for Major Depressive Disorder

In 2014, an estimated 15.7 million adults aged 18 or older in the United States (6.7% of all adults) had one or more episodes of major depressive disorder in the past year (Center for Behavioral Health Statistics and Quality, 2015). There is now substantial evidence that major depressive disorder (MDD) can be treated successfully with certain targeted psychotherapies. This literature is briefly reviewed below.

Psychosocial Treatment of Major Depression in Adults

The strongest empirical evidence exists for three manual-based psychotherapies for the treatment of MDD (behavior therapy, cognitive-behavioral therapy, and interpersonal therapy) with less but still substantial evidence existing for two other forms of psychotherapy (brief psychodynamic therapy and problem-solving therapy). Initial evidence (not reviewed here) exists for humanistic-experiential therapy (Elliot et al., 2013).

Cognitive Therapy

The most widely studied psychotherapy for MDD is cognitive therapy, also known as cognitive-behavioral therapy (CBT) (Beck, Rush, Shaw, & Emery, 1979). This treatment is based on the model that the cognitions (conscious or readily accessible to consciousness) of depressed individuals are negatively biased. This negative bias is evident in negative beliefs about the self, the world, and the future. Such negative cognitions are one factor that plays a role in the initiation and maintenance of depressive symptoms. Cognitive therapy, typically consisting of 16 to 20 sessions over a period of 12 to 16 weeks, involves the application of both behavioral and cognitive techniques. The behavioral techniques serve to help patients engage in activities that give them pleasure, while cognitive techniques are used to help patients recognize negative cognitions and to evaluate the veracity of their beliefs.

The most recent meta-analytic review of CBT for depression covered a total of 115 studies (Cuijpers et al., 2013). This meta-analysis revealed a mean effect size (ES) for 94 comparisons from 75 studies of CBT and control groups of Hedges’ g = 0.71 (95% confidence interval [CI] 0.62–0.79) (number needed to treat = 2.6). However, there was no evidence CBT was more or less effective than other psychotherapies or pharmacotherapy, but combined CBT and medication was found to be significantly more effective than pharmacotherapy alone (g = 0.49). Another recent meta-analysis (N = 11 studies) of comparisons of CBT and second-generation antidepressants for MDD found no statistically significant difference in effectiveness between second-generation antidepressants and CBT for response (risk ratio 0.91, 95% CI 0.77–1.07), remission (0.98, 0.73–1.32), or change in the 17-item Hamilton Rating Scale for Depression score (weighted mean difference, −0.38, −2.87 to 2.10) (Amick et al., 2015).

Even among those with recurrent depression, and among those with moderate to severe MDD, cognitive therapy has been found to be similar in efficacy as antidepressant medication (Blackburn & Moore, 1997; DeRubeis et al., 2005). Moreover, a direct examination of the comparative effects of cognitive therapy and medication across four studies revealed no evidence of a difference among those with (p. 47) moderate to severe depression (DeRubeis, Gelfand, Tang, & Simons, 1999). However, a meta-analysis examining studies comparing cognitive therapy to treatment as usual concluded that cognitive therapy may not be superior to treatment as usual (Jakobsen, Hansen, Storebø, Simonsen, & Gluud, 2011a, 2011b).

Despite the lack of evidence that CBT is uniquely efficacious as compared to other psychotherapies in the treatment of MDD, the overall weight of the evidence is that CBT is an efficacious acute-phase treatment for MDD.

Behavior Therapy

The original behavioral model of MDD treatment was Lewinsohn’s (1974) approach. In this model, the primary goal is to increase the frequency of pleasant activities in the patient’s life. Derived from Lewinsohn’s model is a newer version that has been labeled behavioral activation (BA) treatment of depression (Lejuez et al., 2011; Martell, Dimidjian, & Herman-Dunn, 2013).

A meta-analysis of 17 efficacy studies reported that behavior therapy was superior to control (Cohen’s d effect size = 0.70), brief psychotherapy (d = 0.56), and supportive therapy (d = 0.75) and equal to standard CBT (d = −0.08) in the treatment of depression (Ekers et al., 2008). However, 12 of these 17 trials were done in the 1970s or 1980s, prior to the advent of current BA treatment methods; 12 of the 17 trials were very small (N’s per group of 15 or less); and many of the studies did not diagnose patients.

There have been two major efficacy trials of a modern form of BA for MDD completed to date. In the first study (Jacobson et al., 1996), there was no evidence that a full cognitive therapy produced better outcomes than BA at either termination or 6-month follow-up, despite excellent adherence to treatment protocols by the therapists and an allegiance of the therapists and supervisor to full cognitive therapy. In the second study (Dimidjian et al., 2006), patients (N = 241) were randomized to BA, standard CBT, antidepressant medication, and pill placebo. No significant differences between treatments were evident among low-severity patients. In the high-severity group, BA was significantly better than cognitive therapy. Furthermore, remission rates strongly favored BA (56%) compared to antidepressant medication (23%) and CBT (36%) in the more severe subgroup. The response rate for BA in the more severe group was 76%.

As with CBT, there appears to be evidence that behavior therapy is an efficacious, but not uniquely effective, acute treatment for MDD. The possibility that BA may be more efficacious than both CBT and pharmacotherapy among more severely depressed patients awaits replication.

Interpersonal Therapy for Depression

Klerman and Weissman’s (1989) interpersonal psychotherapy (IPT) for depression assumes that although depression is caused by a number of factors (genetic, biological, social) interacting in complex ways, it is usually triggered by problems in four interpersonal domains: role transition, grief, interpersonal deficits, and interpersonal disputes. In IPT, the interpersonal problem that triggered the current depressive episode is addressed and the person is helped to build communication and interaction skills to resolve it. The acute phase of IPT typically lasts for 16 to 20 sessions.

IPT is generally recognized as an evidence-based treatment for depression. A meta-analysis of IPT for depression reported an overall effect size (Cohen’s d) for 16 studies that compared IPT and a control group of 0.63 (95% CI 0.36–0.90), corresponding to a number needed to treat of 2.91 (Cuijpers, Geraedts, van Oppen, Markowitz, & van Straten, 2011). Ten studies that compared IPT and other psychological treatments showed a nonsignificant differential effect size of 0.04 (95% CI –0.14 to 0.21). However, based on nine studies (removing one outlier), pharmacotherapy was more effective than IPT (d = –0.19, 95% CI –0.38 to –0.01; number needed to treat = 9.43), but combination treatment was not more effective than IPT alone, although the small number of studies precluded drawing definite conclusions.

IPT has also been used to treat antepartum and postpartum depression. Treatment with IPT (p. 48) was superior to a parenting education program for women with antepartum depression in all measures of mood at termination (Spinelli & Endicott, 2003). Also, for symptomatic relief and social adjustment, IPT was superior to a waitlist control for women with postpartum depression (O’Hara, Stuart, Gorman, & Wenzel, 2000). In a different study, IPT for postpartum depression was found to be as effective as a mother–infant therapy group and superior to a waitlist control (Clark, Vittengl, Kraft, & Jarrett, 2003). Finally, in the only randomized clinical trial (RCT) of a Western psychotherapy adapted for Africa, group IPT was better than treatment as usual for depressed people in rural Uganda for depressive symptomatology and social functioning (Bolton et al., 2003).

Brief Dynamic Therapy

Psychodynamic psychotherapy comes in many forms. Brief versions of this treatment typically have a clear interpersonal or intrapsychic focus and use therapist interpretations as the key intervention designed to increase self-understanding about interpersonal or intrapsychic issues that might be contributing to or maintaining depressive symptoms.

The empirical status of dynamic therapy for MDD is supported by the results of a meta-analytic review of 54 studies (33 RCTs) (Driessen et al., 2015). Brief dynamic therapy was found to be significantly more effective than control groups at posttreatment on depression, general psychopathology, and quality-of-life measures (d = 0.49–0.69). No significant differences were evident between brief dynamic therapy and other psychotherapies at posttreatment (d = −0.14) or follow-up (d = −0.06). There was evidence, however, that brief dynamic therapy was significantly more efficacious than other psychotherapies on anxiety measures at both posttreatment (d = 0.35) and follow-up (d = 0.76).

Recently, two large-scale randomized noninferiority trials evaluated whether brief dynamic therapy was inferior to CBT in the treatment of MDD. Both studies, one conducted in the Netherlands (Driessen et al., 2013) and one conducted in a community mental health center in the United States (Connolly Gibbons et al., 2016), found that brief dynamic therapy was statistically not inferior to CBT in change in depressive symptoms.

Two controlled studies have demonstrated the efficacy of dynamic psychotherapy in combination with antidepressant medication. Both de Jonghe, Kool, van Aalst, Dekker, and Peen (2001) and Burnand, Andreoli, Kolatte, Venturini, and Rosset (2002) found that combined dynamic psychotherapy and medication interventions were statistically and clinically superior to medication alone in the treatment of MDD. In addition, the Burnand et al. (2002) investigation implemented a supportive session with a nurse in the medication condition, and therefore the results of this study indicate that the between-group effects may be due to the specific interventions of the dynamic treatment rather than the nonspecific relationship effects of a purely supportive intervention.

The evidence base for brief dynamic therapy as an efficacious acute standalone treatment of MDD is now reasonably strong, though more studies comparing brief dynamic therapy to pharmacotherapy and credible control groups are needed.

Problem-Solving Therapy

Problem-solving therapy (PST) is a particular form of CBT for depression that has a specific focus on training in adaptive problem-solving attitudes and skills. In PST, the therapist attempts both to foster the adoption of a positive problem orientation and to facilitate the acquisition and real-life application of a rational problem-solving style. To instill a positive problem orientation, the therapist will encourage the patient to appraise problems as opportunities for benefit, to adopt the belief that problems are solvable and that he or she has the ability to solve problems effectively, and to recognize that effective problem solving takes time and effort.

A meta-analysis of 21 independent studies concluded that PST was equally effective as other psychosocial therapies and medication (p. 49) treatments and significantly more effective than no treatment and support/attention control groups in the treatment of depression (Bell & D’Zurilla, 2009). Effect sizes (Cohen’s d) were 0.45 for the comparison to supportive therapy/attention control groups and 2.38 for the comparison to no treatment (waitlist control groups).

PST also appears to be promising for the acute treatment of MDD. Limitations of the existing literature are that several studies focus on minor depression rather than MDD, and that follow-up data are more limited, with heterogeneous results across studies. In addition, larger studies are needed, as are more comparisons to other standard psychotherapies for MDD.

Prevention of Relapse and Recurrence

From the literature reviewed in the sections on CBT, BA, IPT, PST, and brief dynamic therapy, it is apparent that the efficacy of short-term acute-phase treatment of MDD is similar across these various psychotherapy modalities. However, CBT has a particular focus on the prevention of relapse and recurrence of depressive episodes. To address this important question, controlled relapse/recurrence prevention studies involving CBT have been conducted. A meta-analysis of 28 such studies documented that responders to CBT have a relapse or recurrence at a rate of 29% within 1 year and 54% within 2 years (Vittengl et al., 2007). These relapse/recurrence rates were similar to those associated with other depression-specific psychotherapies. Across seven studies, relapse/recurrence rates were lower for CBT (39%) compared with pharmacotherapy (68%).

In an effort to achieve even lower relapse/recurrence rates, investigators have examined continuation therapy (i.e., additional therapy sessions after an acute-phase response has been achieved). One form of CBT, mindfulness-based cognitive therapy (MBCT; Segal, Williams, & Teasdale, 2012), was developed specifically to prevent relapse and is delivered after a person has recovered from depression following acute treatment. MBCT is delivered in a group format over 3 months. The treatment encourages patients to process experience without judgment through mindfulness and meditation techniques. A meta-analysis of six MBCT studies found that MBCT has a 34% reduction in risk of relapse compared to controls, with an even higher reduction in relapse rates (43% reduction) among patients who have had three or more previous episodes of depression (Piet & Hougaard, 2011).

A recent meta-analysis of 29 studies of continuation treatments examined the recurrence of new episodes for CBT, MBCT, and IPT (Clarke et al., 2015). The reductions in relapse rates at 12 months were similar for CBT (25%), MBCT (21%), and IPT (22%). The effect was maintained at 24 months for CBT, but not for IPT (no 24-month MBCT studies were available). In relation to comparator groups, CBT, MBCT, and IPT reduced the risk of relapse compared to treatment as usual (relative risk [RR] = 0.79, 95% CI 0.70–0.91) and active comparators (RR = 0.77, 95% CI 0.68–0.87), but there was no significant difference among CBT, MBCT, and IPT.

Less research has been conducted on the prevention of relapse/recurrence of MDD following BA treatment. However, one study found BA to be at least as effective as CBT and continuation medication in the prevention of relapse of MDD (Dobson et al., 2008).


A relatively large and growing body of literature has substantiated the efficacy of targeted psychotherapies in the treatment of MDD in adults. Research evidence from controlled clinical trials in particular supports the efficacy of CBT, IPT, BA, and brief dynamic therapy. Acute-phase CBT and IPT reduce the risk of relapse or recurrence of MDD, and continuation treatment with CBT, IPT, or MBCT reduces such risks further. CBT, BA, and IPT generally have been found to be equally efficacious to medications, even with more severely depressed patients.

Although many of the psychotherapeutic interventions studied in adults have been adapted to the treatment of adolescents, including IPT, CBT, and BA, the adult literature is (p. 50) characterized by a much broader set of studies than in the child and adolescent MDD treatment literature. Even with the relatively large number of MDD studies conducted to date, numerous questions remain about psychotherapy for MDD in adults. Despite the success of certain psychotherapies in the treatment of adult MDD, it may be risky to assume such treatments are likely to be the best psychosocial treatments for childhood and adolescent MDD. The biological, developmental, cognitive, and experiential differences between children and adolescents and adults raise the question of whether wholly different intervention strategies may be most effective with children and adolescents (Mueller & Orvaschel, 1997). For example, psychodynamic psychotherapy, especially variants that rely heavily on symbolic interpretations, may not be appropriate for younger individuals who lack the cognitive maturity to understand such interventions. Treatments that have been little studied in adults, such as family therapy, may have much greater relevance among children and adolescents. New treatments that incorporate developmental issues may be needed. Although it may be hazardous to export treatment modalities developed for adults to children and adolescents, research on children and adolescents can benefit greatly from the methodological developments in the treatment of MDD in adults, particularly the study of prevention of relapse or recurrence. Ongoing dialog and interchange among investigators in the adult and child areas is likely to facilitate the continued development of literature on treatment of children with MDD.

Psychosocial Treatment of Child and Adolescent Major Depression

In this overview of psychosocial treatments for early-onset depression, we review the RCTs in children and adolescents with depressive disorders and symptomatology. Both CBT and IPT have been shown in several trials to be efficacious in the treatment and prevention of adolescent depression. In addition, other promising psychosocial treatments are described that have positive but unreplicated results. Finally, given the frequent interrelationship between depression and suicidal behavior, we review the published clinical trials that show some evidence with regard to the treatment of adolescent suicide attempters, suicidal ideation, and nonsuicidal self-injury (NSSI).

CBT of Youth Depression

CBT approaches for youth depression are quite heterogeneous, but all have two key components: a focus on identifying and modifying negative cognitive bias that leads to distress and depression, and the employment of behavioral activation, in which the depressed individual is encouraged to participate in activities that are likely to be associated with a sense of mastery or pleasure. Other common techniques used in CBT include problem solving, emotion-regulation strategies, and social skills to improve interpersonal effectiveness, since all of these domains can be impaired in depression (Kaslow & Thompson, 1998, Kazdin & Weisz, 1998, Weersing & Brent, 2006). CBT, when compared to alternative treatments for youth depression, shows consistently positive but modest effects (d = 0.35) (Weisz et al., 2006). This result was more recently validated in a network meta-analysis that found that CBT for depression was significantly more effective at posttreatment than control conditions, play therapy, and problem-solving therapy in children and adolescents (Zhou et al., 2015). Below we discuss the highest-quality studies conducted in settings most relevant to clinical practice.

Wood et al. (1996) compared the impact of a five- to eight-session CBT intervention with a comparable dose of relaxation training in the treatment of early- to middle-adolescent outpatients with depressive disorders, with 54% of the CBT group and 26% of the relaxation group remitting by the end of treatment. Similar results were obtained on self-report measures of depressive symptoms, self-esteem, and general psychosocial adjustment. At 6-month follow-up, the outcomes of the two treatment groups converged because of continued improvement in the relaxation group and relapse in the CBT group. Younger age (p. 51) of diagnosis and higher level of functioning at intake were associated with better outcome (Jayson et al., 1998). The addition of a median of six monthly booster CBT sessions after acute treatment resulted in a much lower relapse rate than acute treatment alone in a quasi-experimental study (20% vs. 50%) (Kroll et al., 1996).

Brent et al. (1997) tested a version of CBT adapted for adolescents from Beck et al. (1979) against systemic behavioral family therapy (SBFT) and a nondirective supportive therapy (NST), using a primarily clinically referred sample (2/3 vs. 1/3 from newspaper advertisements) of depressed adolescents. In comparison to the treatment used by Wood et al. (1996), these treatments consisted of more sessions (12–16 weekly sessions).

At posttreatment assessment, significantly fewer of the subjects receiving CBT (17%) than NST (42%) continued to have diagnosable MDD. Remission, as defined by the absence of MDD and at least three consecutive Beck Depression Inventory (BDI) scores of less than 9, was more common in the CBT group (60%) than in either SBFT (38%) or NST (39%). Reductions in suicidality and improvements in general psychosocial adjustment were not different across groups. CBT resulted in a greater change in cognitive distortions compared to either SBFT or NST, although changes in depressive symptoms were not mediated by changes in cognitive style (Kolko et al., 2000).

Moderators of a favorable response to CBT relative to the other two treatments were comorbid anxiety and the total number of adverse predictors to overall treatment response (Brent et al., 1998). On the other hand, a history of sexual abuse and current maternal depressive symptoms both eliminated the advantage that CBT had over alternative treatments (Barbe et al., 2004; Brent et al., 1998). At 2-year follow-up, differences between treatment groups were not significant (Birmaher et al., 2000). Recurrence of depression over the 2-year follow-up period was predicted by greater severity of depression symptoms at intake, higher levels of parent–child conflict, and a lifetime history of sexual abuse (Barbe et al., 2004; Birmaher et al., 2000).

CBT has been adapted to be delivered via the Internet. One brief (four or five sessions) intervention that primarily emphasized problem solving was not superior to a waitlist control (van der Zanden et al., 2012), but two RCTs with CBT have shown quite significant effects compared to waitlist control, including one for symptomatic youth excluded from mainstream education (Fleming et al., 2012; Hoek et al., 2012).

Weisz et al. (2009) conducted a community-based effectiveness trial comparing CBT to treatment as usual, and CBT was found to produce similar outcomes to community treatment, but in fewer sessions, with fewer additional services, lower costs, and great parent satisfaction.

There have been no treatment trials of CBT in preadolescent depressed patients in clinical settings, but one form of CBT, Positive and Secondary Control Enhancement Therapy (PASCET), has been shown to be superior to no intervention in reducing depressed symptoms in nonreferred, symptomatic youth, and has also been tested in clinically depressed youth with inflammatory bowel disease (Thompson et al., 2012; Weisz et al., 1997).

The Treatment of Adolescent Depression Study (TADS) compared 439 depressed adolescents randomized to placebo, CBT, fluoxetine, and a combination of CBT and fluoxetine (combination). At 12 weeks, the response rate for CBT alone was no different than placebo (43% vs. 35%) and inferior to fluoxetine (61%) and combination (71%) (March et al., 2004). By 18 weeks, the CBT-alone group had caught up with the other active treatment groups. Positive predictors of CBT response were high levels of cognitive distortions and higher family income (Curry et al., 2006). Combination treatment resulted in the highest rates of remission and fastest reduction in symptoms of suicidal ideation and depression (Kennard et al., 2006; March et al., 2004).

The Antidepressant and Psychotherapy trial (ADAPT) randomized 208 depressed adolescents to either fluoxetine alone or a combination of CBT and fluoxetine, and in contrast to TADS, did not find a benefit from the addition of CBT (Goodyer et al., 2007). The ADAPT sample was (p. 52) clinically severe, and these findings were consistent with TADS, in which combination treatment did not outpace fluoxetine alone for more severely ill youth (Curry et al., 2006).

Clarke et al. (2005) randomized 152 depressed youth treated in primary care to either medication treatment as usual, or medication treatment plus a structured, individual form of CBT based on the Coping With Depression for Adolescents (CWD-A) intervention. Although the combination was superior to medication alone in some secondary measures, there were no differences in the rate of diagnosable depression at the end of the trial, or upon follow-up.

The Treatment of SSRI-Resistant Depression in Adolescents (TORDIA) study examined four strategies for the management of 334 depressed youth who had not responded to an adequate trial with a selective serotonin reuptake inhibitor (SSRI): switch to another SSRI, switch to venlafaxine, switch to SSRI plus CBT, or switch to venlafaxine plus CBT. The combination of CBT and either medication was superior to medication monotherapy at 12 weeks (55% vs. 41%), although the outcomes among groups converged over longer-term follow-up (Brent et al., 2008). Similar to earlier studies of CBT alone (Brent et al., 1997, 1998), combination treatment was superior to medication monotherapy for those with comorbidity (most commonly attention-deficit/hyperactivity disorder [ADHD] or anxiety), whereas a history of abuse predicted a greater advantage of medication monotherapy over combination treatment (Asarnow et al., 2009).

Two meta-analyses have been conducted to see if CBT in addition to medication is superior to medication monotherapy for adolescent depression. The first, which focused on adolescents with primary depression, found that while combination treatment was superior to medication monotherapy for improving functional status, it was not superior to medication monotherapy in the reduction of depression or suicidal events (Dubicka et al., 2010). The second, which included studies of depression comorbid with substance abuse, for which CBT focused on substance abuse, found an advantage of combination treatment over medication alone for depression and with respect to the incidence of suicidal events (Cox, Callahan, Churchill, et al., 2012).

Interpersonal Psychotherapy for Adolescent Depression

Interpersonal psychotherapy for adolescents (IPT-A), an adaptation of IPT, is a time-limited, focused psychotherapy that addresses common adolescent developmental issues that are closely related to depression: separation from parents, authority and autonomy issues in the parent–teen relationship, development of dyadic interpersonal relationships, peer pressure, loss, and issues related to single-parent families (Mufson et al., 1994, 1999). Summarizing all available literature, a network meta-analysis found that IPT for depression was significantly more effective than control conditions, play therapy, and problem-solving therapy in adolescents (Zhou et al., 2015). Specific higher-quality studies are discussed below.

In a controlled, 12-week, clinical trial of IPT-A, 48 adolescents with major depression were randomly assigned to either weekly IPT-A or biweekly to monthly 30-minute sessions of clinical monitoring (Mufson et al., 1999). The sample was largely Hispanic and female. At termination, a much lower proportion of those in IPT-A still met criteria for major depression (12.5% vs. 41.6%) and a greater proportion were remitted (75% vs. 46%). Patients who received IPT-A reported a significant decrease in depressive symptoms and a greater improvement in overall social functioning, functioning with friends, and problem-solving skills.

Mufson et al. (2004) assessed the effectiveness of IPT-A in school-based mental health clinics in New York City: they randomized 63 depressed adolescents referred for a mental health intake visit to IPT-A or treatment as usual, both performed by the clinic staff. At termination, the adolescents in the IPT-A treatment experienced significantly greater symptomatic relief than the treatment-as-usual group. IPT has been shown to be especially effective compared to clinical management or usual care in depressed adolescents with poorer (p. 53) interpersonal functioning, high levels of conflict with parents, greater depressive severity, and comorbid anxiety (Gunlicks-Stoessel et al., 2010; Young et al., 2009).

Rossello and Bernal (1999) compared the efficacy of a 12-week, individually administered CBT program to IPT and waitlist control in adolescents with diagnosed MDD and/or dysthymia referred by school personnel. Although the study was limited by poor attendance in both treatment arms, 59% of adolescents in the CBT condition and 82% of those receiving IPT achieved clinically significant improvement in depression symptoms by posttreatment; data were not provided for the waitlist control condition. In a subsequent study, participants were randomized to group or individual IPT or CBT (four cells) versus treatment as usual. Counter to the previous study, either form of CBT resulted in greater functional improvement and reduction in depressive symptomatology than either form of IPT (Rossello et al., 2008).

Attachment-Based Family Therapy

Diamond, Reis, Diamond, Siqueland, and Isaacs (2002) conducted a randomized clinical trial comparing attachment-based family therapy (ABFT) with a waitlist control in 32 clinically referred, depressed adolescents in a largely (69%) African-American, poor, inner-city sample. On average, subjects received eight sessions of ABFT, which focused on strengthening family bonds, reducing conflict, improving trust, and communication. Those in the waitlist condition received 15 minutes of weekly telephone monitoring of their clinical condition as well as a face-to-face assessment at week 6, at which point those still meeting criteria for major depression were offered ABFT (the treatment data from these latter cases were not included in the primary outcome analyses). At posttreatment, 81% of those treated with ABFT no longer met criteria for depression, compared to 47% in the waitlist control group. A significantly greater number of those assigned to ABFT reported a BDI score of less than 9 (62% vs. 19%). A significant treatment-by-time interaction favoring ABFT was found for interview-rated depression, self-reported anxiety, and child-reported parent–child conflict, with nonsignificant trends favoring ABFT for attachment to mother, suicidal ideation, and hopelessness.

A replication of these effects in suicidal adolescents in discussed below (Diamond et al., 2010). One other study of adolescent self-harm, in which the participants were also depressed, showed a positive effect on depression using mentalization-based therapy and is described below (Rossouw and Fonagy, 2012).

Fine, Forth, Gilbert, and Haley (1991) compared two forms of short-term group therapy, either social skills training or therapeutic support, for depressed adolescents in an RCT. Subjects in the therapeutic support group showed significantly greater reductions in clinical depression and significant increases in self-concept, although there were no group differences by 9-month follow-up. Subjects in the support group maintained their gains, and those who attended the social skills group caught up.


A wellness-oriented CBT offered along with continuation medication for adolescents who responded to fluoxetine treatment was found to prevent depressive relapse compared to medication alone (15% vs. 37%, hazard ratio [HR] = 0.11; Kennard et al., 2008). This result was replicated in a subsequent larger trial (9% vs. 26.5%, HR = 0.31; Kennard et al., 2014). In a four-site study of the offspring of depressed parents, the majority of whom had had a previous depressive episode, a version of the CWD-A has been shown to prevent the onset or recurrence of depression compared to ordinary care at 8 and 33 months after the end of the intervention (21% vs. 33%, HR = 0.63 at 8 months) (Beardslee et al., 2013; Garber et al., 2009). A family group CBT has been shown to prevent the onset of depression in the offspring of depressed parents compared to usual care even 2 years after the intervention (odds ratio [OR] = 0.34; Compas et al., 2011), with effects mediated by improvements in youth coping (p. 54) and positive parenting. CBT, in combination with motivational interviewing, was more efficacious than CBT alone in preventing the incidence of depression in primary care (Saulsberry et al., 2013; Van Voorhees et al., 2009). Group IPT and individual IPT were found to be equally efficacious as group CBT and individual CBT for the prevention of depression in at-risk adolescents (Horowitz et al., 2007).

Suicide Attempters, Suicidal Ideation, and Self-Harm

Depression is the most significant psychiatric risk factor for adolescent suicide (Brent et al., 1993, 1999; Shaffer et al., 1996). We first review the interventions that focus on depression that also examined the impact of treatment on suicidal events (attempters or emergency referral for suicidal ideation), and then turn to studies that have examined impact on suicidal ideation, suicide attempts, or self-harm (including suicidal behavior and NSSI). NSSI is an important treatment target because of evidence that nonsuicidal self-harm is a stronger predictor of a suicide attempt than a previous attempt in some studies (Asarnow et al., 2011a; L.J. Cox et al., 2012; Wilkinson et al., 2011) and is also a predictor of completed suicide (Hawton et al., 2012). In the studies reviewed below, the term “self-harm” refers to behavior that is self-injurious regardless of suicidal intent.

In the TADS study, it was reported that the combination of medication and CBT was protective against suicidal events relative to medication alone (March et al., 2004, 2007), but this was not found in two other studies of combination therapy in adolescent depression (Goodyer et al., 2007), or in an overall meta-analysis of treatment of primary depression (Dubicka et al., 2010). However, in a meta-analysis that included studies of adolescents with depression comorbid with substance abuse, and including CBT treatments aimed at substance abuse, combination treatment did seem to protect against suicidal events relative to medication monotherapy (G.R. Cox et al., 2012).

Suicidal Ideation

A meta-analysis (N = 13 studies) of CBT found that CBT was significantly better than control groups in reducing suicidal ideation (Hedges’ g = 0.40, 95% CI 0.30–0.49, p < .001) (Labelle et al., 2015).

Diamond et al. (2010) found that ABFT was four to six times more efficacious than clinical management in reducing suicidal ideation to below a clinical cutpoint in 66 suicidal adolescents; among adolescents in the study with significant depression, similar effects were found on reduction in depressive symptomatology. Similar to the first study, one weakness is the imbalance in duration and the intensity of the treatments (10 sessions for ABFT vs. three sessions for clinical management). The effects appeared to be mediated by improvement in family climate (Shpigel et al., 2012).

King et al. (2006, 2009) conducted two clinical trials in suicidal adolescents recruited from inpatient units using the Youth-Nominated Support Team intervention (YST). This brief intervention helped the youth to identify possible supportive adults and coach those adults on how to respond to the adolescent’s request for support. Although there were no main effects on reattempts or on suicidal ideation, subgroups in each of these two studies (girls in the first, and multiple attempters in the second) showed greater reductions in suicidal ideation than usual care. The studies had relatively low participation of eligible subjects, and, at least in the second study, the intervention may have been overpowered by the intensity of treatment as usual, which in both intervention arms was over 30 sessions.

Tang et al. (2009) randomized 73 adolescents with high suicidal ideation, hopelessness, depression, or anxiety to either an intensive modification of IPT (IPT-A-IN) or supportive counseling in a school-based intervention. Participants in IPT-A-IN received 12 sessions over 6 weeks plus a weekly 30-minute follow-up call, whereas the counseling group received 30 to 60 minutes of support once or twice a week. The focus on IPT-A-IN was on interpersonal problems that precipitated suicidality or urges for self-harm, and included dyadic sessions with (p. 55) teachers or peers where relationships were problematic. The active treatment resulted in much greater reductions in ideation, hopelessness, anxiety, and depression, although its impact on suicidal behavior or self-harm was not reported. Limitations include the imbalance in attention between groups, lack of report on suicidal or self-harm activity, and exclusion of adolescents with a recent attempt from the study.

Suicide Attempts

Esposito-Smythers et al. (2011) compared integrated CBT (iCBT) to treatment as usual (consisting of “usual CBT” and medication management in 36 suicidal, alcohol- or substance-abusing adolescents). iCBT, in addition to the components of standard CBT, included motivational interviewing for substance abuse, and intensive family work to improve monitoring, supervision, and positive parenting. Despite its small size, this is the only study that has reported a difference in suicide attempts upon 18-month follow-up, which was paralleled by a decrease in alcohol/substance abuse, hospitalizations, and arrests. In the Labelle et al. (2015) meta-analysis, there was no significant effect for CBT compared to control groups for suicide attempts, possibly because of a low baseline prevalence of suicide attempts in most studies.

Huey et al. (2004) reported that Multi-Systemic Therapy (MST) compared to hospitalization resulted in a decreased incidence of suicide attempts. However, the treatment groups were not distinct, as 44% of those in the MST group were also hospitalized, and the rate of reattempt was the same in both groups, but the rate of previous attempts was higher in the MST group, making the results difficult to interpret.

Self-Harm (Includes Both Attempts and NSSI)

The Labelle et al. (2015) meta-analysis reported that, across eight studies, CBT was significantly better than control groups in reducing self-harm (Hedges’ g = 0.27, 95% CI 0.17–0.38).

Rossouw and Fonagy (2012) compared mentalization-based therapy (MBT) to usual care in 80 mostly depressed and borderline personality–disordered adolescents who had engaged in self-harm. MBT, consisting of around 20 sessions (vs. 17 sessions in usual care) had weekly individual and monthly family sessions and aimed at helping the patient and family to understand action in terms of thoughts and feelings. MBT compared to usual care showed a dramatic reduction in the frequency of self-harm and also a very strong effect on depression and borderline personality symptoms. The effect of the treatment on self-harm was mediated by a decrease in avoidant attachment and improved self-reported ability to mentalize.

Pineda and Dadds (2013) randomized 48 adolescents with self-harm to either the Resourceful Adolescent Parenting Program (RAP-P) plus treatment as usual versus treatment as usual alone. RAP-P consisted of four 2-hour, biweekly, didactic sessions aimed at augmenting parental strengths. RAP-P was associated with a much greater reduction in self-harm and symptomatic improvement, and these effects were mediated by improvement in parent-reported family functioning.

A number of large trials for adolescent self-harm have been negative, including treatments that used a brief home-based family therapy and a group skills psychotherapy intervention, termed developmental group therapy (DGP) (Green et al., 2011; Harrington et al., 1998; Hazell et al., 2009). These three studies may have had too low a dose of treatment, especially relative to the intensity of treatment as usual. For example, in the Green et al. study, participants in the experimental condition received about nine DGP group sessions plus 10 treatment-as-usual sessions, versus 10 treatment-as-usual sessions in the comparison treatment. There are no definitive findings that support any specific intervention for suicidal or self-harming youth, but we review those that had some positive findings.

Treatment Engagement

Asarnow et al. (2011b) found that a one-session emergency department–based session for adolescent suicide attempters was successful in (p. 56) increasing attendance at subsequent sessions but not in reducing the incidence of suicidal events or attempts. Similarly, Ougrin et al. developed a brief 30-minute intervention for adolescents engaging in self-harm and found that it increased engagement, but this increased attendance to treatment did not translate into a reduction in recurrent self-harm (Ougrin et al., 2011, 2012, 2013).


Both CBT and IPT-A are effective in reducing depressive symptoms compared to comparator treatments as well as in preventing the onset of depression in at-risk youth. CBT can be successfully delivered over the Internet for either treatment or prevention. In addition, there is evidence that both interventions can be delivered in community settings with good results. Although IPT has not been studied in combination with medication, CBT when added to medication improves the outcome for treatment-resistant depression over medication monotherapy, and functional outcomes across studies of depressed youth. CBT does particularly well in patients with comorbidity (e.g., anxiety, ADHD, and conduct disorder), whereas IPT-A does well in patients with greater severity, interpersonal dysfunction, and high levels of parent–child discord. Although not studied in IPT-A, abuse and parental depression appear to be negative moderators for CBT. The combination of CBT and fluoxetine continuation treatment may be superior to fluoxetine continuation alone for the prevention of depressive recurrence. Both CBT and IPT may be promising interventions for preadolescents, but further work is needed. Both CBT and IPT have been shown to be efficacious in preventing the onset of depression in high-risk youth.

In addition to CBT and IPT-A, ABFT appears to be a promising treatment for depression, and perhaps MBT as well. With regard to suicidal risk in depressed adolescents, CBT alone is effective in reducing suicidal ideation and self-harm behaviors. The addition of CBT to medication, however, does not appear to convey additional protection for primary depression. There are some promising, albeit unreplicated other treatments for suicidal ideation (ABFT, IPT-A-IN; perhaps YST), self-harm (MBT, RAP-P), and suicide attempt (iCBT).

Pharmacologic Treatment of Adult Major Depression


At this time the first-line medications for treating adult MDD are the SSRIs, which include fluoxetine (Prozac), sertraline (Zoloft), paroxetine (Paxil), fluvoxamine (Luvox), citalopram (Celexa), and escitalopram (Lexapro). They have a greater affinity for the serotonin transporter than for the noradrenergic transporter, and each compound selectively inhibits 5-hydroxytryptamine (5-HT) reuptake and has unique secondary binding properties. This class of drug is rarely associated with fatalities and given its safety profile provides an easy treatment option for the clinician (Farvolden, Kennedy, & Lam, 2003).

Also approved in the United States for the treatment of adult MDD are the selective serotonin and noradrenaline reuptake inhibitors (SNRIs), which include venlafaxine (Effexor), desvenlafaxine (Pristiq), duloxetine (Cymbalta), and levomilnacipran (Fetzima). Another SNRI, milnacipran (Savella), is not approved in the United States but is approved in other countries for the treatment of MDD. It appears that venlafaxine possesses a selective high affinity for the noradrenergic and serotonergic reuptake sites; however, it is only at higher doses (150–225 mg) that its noradrenergic reuptake becomes activated. This dual reuptake inhibition may contribute to the high rates of remission of depressive symptoms compared to those with SSRIs (Thase, Entsuah, & Rudolph, 2001). Although venlafaxine’s side-effect profile is similar to that of SSRIs, at 200 mg there is a 5.5% clinically significant elevation of blood pressure and at 300 mg the incidence of hypertension reaches 13%. Duloxetine reportedly has dual reuptake inhibition that is equal at clinical doses.

(p. 57) Reboxetine is a selective norepinephrine reuptake inhibitor not approved in the United States for the treatment of major depression. Mirtazapine, which belongs to the piperazine–azepine group of compounds, has not been investigated in the treatment of pediatric depression.

A more recent (2013) approved agent for adult MDD is vortioxetine (Brintellix), a serotonin modulator and stimulator. A meta-analysis of studies evaluating vortioxetine concluded that although it was more effective than placebo for acute treatment of MDD, it was potentially less effective than SNRIs (Meeker et al., 2015). The most common adverse events with vortioxetine, relative to placebo, are nausea and vomiting (Meeker et al., 2015).

Another compound used for the treatment of adult depression is bupropion, which works by blocking noradrenergic and dopamine reuptake (Dong & Blier, 2001). It has the side effects of insomnia, nausea, increased anxiety, and restlessness. Another potential side effect is an increased incidence of seizures, which occur at a rate of 0.4% with daily doses below 450 mg and 2.4% with daily doses between 450 and 600 mg (Johnston et al., 1991).

Older classes of antidepressants efficacious in treating adult MDD, such as the tricyclic and heterocyclic antidepressants (amitriptyline, amoxapine, clomipramine, desipramine, doxepin, imipramine, maprotiline, nortriptyline, and protriptyline), inhibit different combinations of serotonin, noradrenergic, and dopamine receptors. Their antagonism at other receptor types causes difficult side effects such as dry mouth, confusion, orthostatic hypotension, tachycardia, weight gain, urinary retention, and constipation. In addition, doses outside of the therapeutic range may be lethal because of conduction abnormalities. The monoamine oxidase inhibitors (MAOIs; phenelzine and tranylcypromine) irreversibly inhibit the MAO isozymes A and B. It is thought that the blockade of the isozyme A lends these compounds their clinical efficacy (Mann et al., 1989). One major issue with this class of antidepressants is the required tyramine-restricted diet; if the diet is not adhered to, dangerous and possibly fatal elevation of the blood pressure can occur. Also, drugs that increase synaptic monoamines must be avoided, such as over-the-counter cold medications, tricyclic antidepressants, SSRIs, stimulants, and cocaine. There is also a risk of lethal, rare hyperthermic reactions that occur with meperidine and other opiates. Given these issues of tolerability and potentially serious adverse events, tricyclics and MAOIs have been more often reserved as second- and third-line agents for patients who have failed to respond to treatment with one of the newer antidepressant classes.

Pharmacologic Combination and Augmentation Therapy

Combination and augmentation strategies have been used when there is an inadequate response to an initial antidepressant and for treatment-resistant depression. The definition of treatment-resistant depression is the failure to clinically respond to one or two antidepressant trials with an adequate amount of time and dosage. There are several strategies for treating adults with an inadequate response or treatment-resistant depression: maximize the dose and duration of treatment; switch to another antidepressant within a class or another class; use a combination of antidepressants; and augment with other compounds.

By far the most ambitious attempt to examine the effects of switching and augmentation was the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) trial (Rush et al., 2004). In STAR*D, patients with MDD (N = 3,671) were initially treated with citalopram. Those who did not achieve remission or could not tolerate citalopram were then randomly assigned using an equipoise stratified randomized design (allowing patients to decline treatment options) to four switch treatment options (sustained-release bupropion, cognitive therapy, sertraline, or extended-release venlafaxine) and three augmentation options (citalopram plus bupropion, buspirone, or cognitive therapy). Those who did not achieve remission at this second step were offered a third step consisting of two medication switch strategies (mirtazapine or (p. 58) nortriptyline) or two medication augmentation strategies (lithium or triiodothyronine [T3, 25 mg]). A fourth step consisted of a single randomization to either tranylcypromine or extended-release venlafaxine plus mirtazapine. Remission rates were 36.8%, 30.6%, 13.7%, and 13.0% for the first, second, third, and fourth acute treatment steps, respectively, with an overall cumulative remission rate of 67% (Rush et al., 2006). Switching and augmentation appear to achieve similar remission rates (Gaynes et al., 2012). Thus, an initial one or two attempts at augmentation/switching appear useful for achieving remission, but additional (three or more) treatment changes do not yield meaningful further improvements.

Earlier studies supported the efficacy of lithium and triiodothyronine augmentation with patients who did not respond to SSRIs (one study with citalopram and one with fluoxetine) and newer antidepressants (Baumann, 1996; Katona et al., 1993). Other evidence supports the use of lithium and triiodothyronine as augmentation treatment for patients who do not respond to tricyclics (Aronson, Offman, Joffee, & Naylor, 1996; Freemantle, Anderson, & Young, 2000).

Several approaches are now approved by the U.S. Food and Drug Administration (FDA) for adjunctive (augmentation) therapy in the treatment of MDD. These include the atypical antipsychotics aripiprazole, quetiapine, brexpiprazole, and olanzapine (the latter in combination with fluoxetine). A meta-analysis of 17 trials of augmentation of MDD treatment with atypical antipsychotics found that the remission rate and overall response rate of adjunctive treatment were significantly higher than placebo treatment (remission: 32.6% for adjunctive groups vs. 18.2% for placebo; response: 43.8% for adjunctive groups vs. 28.7% for placebo), although there was a higher discontinuation rate due to adverse effects with the atypical antipsychotics (Wen et al., 2014).

Although effective, the approval of atypical antipsychotics for augmentation therapy has generated concern about the metabolic effects of these agents. The concept of metabolic syndrome is used to evaluate these risks. Metabolic syndrome has been defined by a combination of high blood pressure, low levels of high-density lipoprotein cholesterol, elevated triglycerides, and central obesity (Expert Panel, 2001). Depression itself is believed to be an independent risk factor for cardiovascular disease (Niranjan et al., 2012), so it is important to separate out the effects of the disorder from the influence of treatment on metabolic syndrome. This was done in a meta-analysis of studies that assessed metabolic parameters in MDD patients and found that use versus non-use of atypical antipsychotics was significantly associated with metabolic syndrome (Vancampfort et al., 2014).

Prevention of Relapse and Recurrence

Although multiple antidepressant agents are effective for short-term treatment of acute episodes of MDD in adults, if treatment is not continued relapse/recurrent rates have been found to range from 1.56% to 17.8% per month across 45 studies (Baldessarini et al., 2015). Given high rates of relapse/recurrence, longer-term treatment (beyond 1 year) has been evaluated to reduce such rates. In a meta-analysis of 72 trials, antidepressants were found to be more effective than placebos in preventing relapses (relative response rates [RR] = 1.90, number needed to treat = 4.4; p < .0001) and recurrences (RR = 2.03, number needed to treat = 3.8; p < .0001), with only minor differences evident among drug types (Sim et al., 2015).


SSRIs remain a first-choice option for pharmacologic therapy for adult MDD, with SNRIs also widely used. Longer-term treatment (>1 year) with SSRIs/SNRIs has been successful in reducing relapse/recurrence rates. For those who do not achieve remission with (or fail to tolerate) a first agent, switching or augmenting, particularly with an approved atypical antipsychotic, is a viable option for improving the likelihood of achieving remission of symptoms, at least through a second switch/augmentation trial. However, the possibility that atypical (p. 59) antipsychotics may contribute to metabolic syndrome raises concerns about longer-term use.

Pharmacologic Treatment of Adolescent Major Depression

The only antidepressants that have demonstrated efficacy in double-blind placebo-controlled trials for children and adolescents with major depression are the SSRIs. Two SSRIs have FDA approval for acute and maintenance treatment of major depression in children and adolescents: fluoxetine (>8 years old) and escitalopram (>12 years old). Three double-blind placebo-controlled trials demonstrated the efficacy and safety of fluoxetine in the treatment of children and adolescents with major depression (Emslie et al., 1997, 2002; Treatment for Adolescents with Depression Study [TADS] Team 2004). Fluoxetine doses ranged from 10 to 40 mg in these studies. Clinical Global Impressions-Improvement (CGI-I) scores of less than 2 (much or very much improved) were 56%, 52%, and 61% for the fluoxetine groups compared to 33%, 37%, and 43% respectively for the placebo groups. The efficacy and safety of escitalopram for the treatment of major depression in adolescents has been demonstrated in two double-blind placebo-controlled multicenter trials. In the first (Emslie et al., 2009), 157 adolescents were randomly assigned to escitalopram (dose range 10–20 mg/day) or placebo for an 8-week trial. A statistically significant greater improvement in depression (CDRS-R scores) was found for escitalopram compared to the placebo: 64% of patients treated with escitalopram were much or very much improved versus 53% of patients treated with placebo. In the second trial (Findling et al., 2013), improvement in CDRS-R scores was significantly greater for escitalopram than for placebo, with response rates also significantly different (63.6% for escitalopram; 47.1% for placebo).

Although not FDA approved for pediatric depression, two other SSRIs have demonstrated efficacy for the treatment of major depression in youth. In a double-blind placebo-controlled 8-week trial, 174 youths ages 7 to 17 years were randomized to citalopram (dose range 20–40 mg/day) for an 8-week trial (Wagner et al., 2004). The citalopram-treated group showed statistically significant greater improvement in depression (CDRS-R scores) than did the placebo group. The efficacy of sertraline for the treatment of children and adolescents with major depression was demonstrated in two a priori combined pooled analyses of identical double-blind placebo-controlled multicenter studies (Wagner et al., 2003). Three hundred seventy-six youth were randomized to sertraline (dose range 50–200 mg/day) or placebo for 10 weeks. The sertraline group showed a statistically significant greater improvement in depression (CDRS-R scores) compared with the placebo group. Negative double-blind placebo-controlled SSRI trials include citalopram for adolescent depression (von Knorring et al., 2006), paroxetine (Berard et al., 2006; Emslie et al., 2006; Keller et al., 2001), and an escitalopram study that included children and adolescents (Wagner et al., 2006).

Adverse events among children and adolescents treated with SSRIs are a particular concern. One study reported that 74% of children/adolescents experienced an adverse event to an SSRI over the course of their treatment (Wilens et al., 2004). Some differences in adverse events are evident in relation to age. Using adverse events from all double-blind placebo-controlled trials of SSRIs in children and adolescents that separated findings by age group, Safer and Zito (2006) found that activation and vomiting were twofold to threefold more prevalent in children than in adolescents, but rates of these events were low in adults. Furthermore, while somnolence was not common in children, rates of this adverse event increased with advancing age. Both insomnia and nausea were common adverse events across the full age span from children to adults. The most frequent reason for discontinuation from SSRI clinical trials in children was activation; in contrast, somnolence, nausea, and insomnia were the most common reasons for discontinuations in adults (Safer & Zito, 2006).

Of particular concern is the possibility that SSRIs increase suicidality in children and adolescents. Following up on reports of suicidal (p. 60) behavior in case reports and clinical trials of SSRI treatment in children and adolescents, the FDA conducted an analysis of 23 placebo-controlled trials of antidepressants in children/adolescents. Results indicated an overall statistically significant (p < .05) relative risk increase in suicide-related adverse events of 1.66 in MDD trials, and 1.95 when all trials were pooled, for SSRIs compared to placebo (Hammad, Laughren, & Racoosin, 2006). The data were interpreted as indicating that for every 100 treated child/adolescent patients treated with an SSRI, one to three patients will have an increase in suicidality beyond the risk that occurs with depression itself. There were no completed suicides reported in the data reviewed by the FDA (Hammad, Laughren, & Racoosin, 2006).

The FDA’s response to these results was to implement a “black box” label warning in 2004 for SSRI use in children and adolescents. The warning states that SSRIs may increase the risk of suicidal thinking and behavior in children and adolescents with MDD and other psychiatric disorders. A recommendation is given in the black box warning that children and adolescents taking SSRI medications should be closely monitored for any worsening in depression, emergence of suicidal thinking or behavior, or unusual changes in behavior. During the first 4 weeks of treatment monthly monitoring is stated to be especially important. The warning also adds that adults should be similarly closely monitored during the initial few months of drug therapy.

The data on suicidality risk with SSRIs in children and adolescents need to be weighed against the benefits of treating MDD with SSRIs. Despite the rare suicide-related adverse events that occur, studies indicated that treatment with SSRIs overall decreases suicidal ideation and suicide attempts in children/adolescents (Kutcher & Gardner, 2008; March et al., 2004). Moreover, population studies have found that regions of the United States with increased antidepressant use have shown a decrease in suicides in children and adolescents (Gibbons, Hur, Bhaumik & Mann, 2006; Olfson, Shaffer, Marcus & Greenberg, 2003). Consistent with such studies, postmortem studies have failed to find a link between SSRI use and youth suicide (Isacsson, Holmgren, & Ahlner, 2005; Leon et al., 2006). Thus, in general, SSRI use appears to decrease suicide rates rather than increase them. In 2007, the FDA modified the black box warning to include young adults ages 18 to 24, but also to include a statement that depression itself was associated with an increased risk of suicide, thereby highlighting the balance between the small risk associated with antidepressant treatment and the proven efficacy of such treatments. Initiation of antidepressants at high therapeutic doses has been associated with an increased risk of self-harm behavior in patients ages 10 to 24 years (Miller et al., 2014).

Other classes of antidepressants have not demonstrated superiority of medication to placebo in RCTs for the treatment of children and adolescents with major depression, including desvenlafaxine ( Identifier:NCT01372150), duloxetine (Atkinson et al., 2014; Emslie et al., 2014), mirtazapine (FDA, 2004), nefazodone (Rynn et al., 2002; FDA, 2004), venlafaxine (Emslie et al., 2007), and selegiline transdermal system (DelBello et al., 2011). Bupropion has not been evaluated in controlled studies for major depression in children and adolescents.

Overall, rates of response for depressed youth treated with antidepressants are 61% and for placebo are 50% in treatment trials (Bridge et al., 2007), Remission rates in trials of acute treatment with antidepressants ranged from 30% to 40%. Clinical response by week 12 of treatment increased the likelihood of clinical remission (Curry et al., 2011; Emslie et al., 2010). The number needed to treat to benefit is 10, whereas the number needed to harm is 112. Therefore, the benefits of the antidepressants outweigh the potential harm from suicidal ideation or attempt.

Consensus guidelines recommend that antidepressants be continued for 6 to 12 months after symptom remission (Birmaher et al., 2007; Hughes, 2007). There is one maintenance study of antidepressant treatment for adolescent depression (Cheung et al., 2008). In this study, 93 adolescents with major depression received sertraline for 12 weeks, and responders were (p. 61) enrolled in a 24-week continuation study. At the end of the continuation phase, responders were randomized to sertraline or placebo for a 52-week maintenance phase. Thirty-eight percent of the adolescents in the sertraline group maintained response (i.e., no recurrence), whereas none of the placebo-treated patients maintained response. This study suggests that treatment of depression in adolescents should continue for a minimum of 1 year following 9 months of adequate treatment response.

Since approximately 40% of youths do not respond to initial treatment with an SSRI, the clinician is faced with a decision regarding selection of another antidepressant. The findings from the TORDIA trial provide guidance about the next antidepressant choice (Brent et al., 2008), as described earlier. Of note, adverse events of increased diastolic blood pressure and pulse and skin problems were more frequent in the venlafaxine group compared to the SSRI group. On the basis of these findings, approximately 50% of adolescents who do not respond to initial treatment with an SSRI will respond to treatment with an alternate antidepressant. Since adverse events were lower in the SSRI-treated patients, it is recommended that adolescents who fail to respond to an initial SSRI switch to a different SSRI. If a depressed youth fails to respond to two SSRI trials, then it would be reasonable to switch to an alternate class of antidepressants such as venlafaxine, duloxetine, bupropion, or mirtazapine.

Some depressed youth have a partial response to an antidepressant. Unfortunately, there are no controlled data to guide treatment-augmentation strategies for depressed children and adolescents. In a small case series, augmentation with an atypical antipsychotic (quetiapine) improved treatment response for depressed adolescents (Pathak et al., 2005). Based upon extrapolation from adult data, augmentation of SSRIs with lithium, bupropion, or mirtazapine has been recommended (Hughes et al., 2007).

Repetitive transcranial magnetic stimulation (rTMS) has been studied in a small open-label trial for treatment-resistant depression. Nine adolescents received rTMS for 14 days (10 Hz, 2-second trains given 20 minutes per day). One third of the adolescents responded to treatment (response defined as a >30% reduction in CDRS-R score). In a 3-year follow-up of rTMS for eight adolescents, improvement in depression was maintained.


Pharmacologic treatment of MDD in children and adolescents requires a careful consideration of potential risks and benefits. Unlike in adults, only fluoxetine and escitalopram have shown efficacy and are approved for acute and maintenance treatment in the United States for pediatric patients. The data on rare suicide-related adverse events and the black box warning regarding SSRI use in children and adolescents led to a subsequent decrease in the use of such drugs in the pediatric population (Friedman, 2014). This was likely an overreaction, given that the studies that emerged demonstrated that, on balance, antidepressant use was associated with reduced suicidality among youth and that no actual suicides were present in the clinical trials database involving SSRI use in children/adolescents. Nevertheless, close monitoring for clinical worsening and any suicidal ideation or behaviors is critical when SSRIs are prescribed in both youth and young adults.

Bipolar Disorder

Mood disorders by definition disrupt functioning in several areas of an individual’s life, including school, family, and peer relationships. Practice guidelines for the treatment of bipolar disorder in adults recognize both pharmacotherapy and psychotherapy as essential components of optimal treatment (American Psychiatric Association, 1994). From a developmental standpoint, research supports the notion that early psychosocial impairment tends to promote later impairment, as the individual arrives at each progressive stage of development with inadequate resources available to meet the challenges unique to the ensuing developmental period (Cicchetti, Rogosch, & Toth, 1998). Thus, it is crucial that treatment be (p. 62) provided promptly and effectively to maintain a normal developmental trajectory as much as possible, and minimize the effects that symptoms have on functioning.

Given the recognition of bipolar disorder in childhood, there is a growing body of evidence supporting treatment strategies, both pharmacologic and psychotherapeutic, for this population. Potentially promising treatment approaches for pediatric bipolar disorder build upon the literature for treatment of adult bipolar disorder. In the next section we review the literature on the psychosocial treatment of adult bipolar disorder.

Psychosocial Treatment of Adult Bipolar Disorder

Several different manualized psychosocial treatments have been applied to the treatment of bipolar disorder in adults; although they are based on different theoretical orientations, they share the goal of diminishing relapse to ultimately improve quality of life. Common areas of treatment focus include increasing treatment compliance, enhancing protective factors (e.g., support, self-care routines), and decreasing risk factors associated with relapse (e.g., stress, substance use).


CBT conceptualizes mood swings as a function of negative thought and behavioral patterns. These maladaptive patterns are then targeted in the therapy. Three RCTs with bipolar adults (for a review see Craighead, Miklowitz, Frank, & Vajk, 2002) suggest that adjunctive CBT leads to increased medication compliance, fewer hospitalizations, and improved social and occupational functioning.

IPT and Social Rhythm Therapy

Interpersonal therapy, as discussed earlier, is a short-term, present-oriented, problem-focused individual therapy developed and supported for the alleviation of symptoms of major depression (Klerman et al., 1984). In IPT, the onset of the depressive episode is placed in the context of interpersonal relationships, and current interpersonal difficulties are addressed. With the knowledge that circadian rhythm disturbances are linked to bipolar disorder (Ehlers, Frank, & Kupfer, 1988), Frank et al. (1997) supplemented IPT with social rhythm therapy (SRT) to create IPSRT for bipolar disorder. The focus of IPSRT is on the regularization of both social and circadian rhythms to control mood cycling. Results of a controlled trial with bipolar adults indicate that IPSRT is most effective in controlling the depressive symptoms of bipolar disorder, and also affects greater stabilization of sleep–wake cycles than case management treatment.

Family-Based Therapies

Several RCTs of family-based treatments have been documented in the literature. Inpatient family intervention (Clarkin et al., 1990), a nine-session intervention focused on psychoeducation, aims to modify negative family patterns and increase coping skills. Results indicate that female patients had better global and symptomatic functioning than those receiving standard hospital treatment (Clarkin et al., 1990).

Miklowitz and Goldstein (1990) developed family-focused therapy (FFT), a 9-month treatment incorporating psychoeducation, communication skills training, and problem-solving skills training. In a randomized trial, patients receiving FFT experienced fewer depressive symptoms, showed increased compliance with medication regimens, had fewer hospitalizations, and experienced a longer period to mood relapse than patients in a case management condition (Miklowitz et al., 2000).

FFT has been adapted and studied in adolescents in several studies by Miklowitz et al. (2004, 2006, 2008). The results suggest this form of therapy helps by shortening the duration of symptoms but must be used as adjunctive treatment to pharmacotherapy. We know that familial climate is related to relapse in bipolar adults (Miklowitz, Goldstein, Nuechterlein, Snyder, & Mintz, 1988). Therefore, family-based interventions among the families of bipolar children (p. 63) and adolescents should assist with improved long-term outcome.

Group Psychotherapy

Among adults with bipolar disorder, several different group approaches have been successful. For example, Colom et al. (2003) demonstrated increased time to relapse of mood symptoms, as well as lower rates of hospital readmission among remitted bipolar adults attending a psychoeducational group, compared with those receiving standard treatment (medications alone). Another structured group approach, The Life Goals Program (Bauer, McBride, Chase, Sachs, & Shea, 1998), consists of psychoeducation, behavioral skills, and individually tailored goals; at present, no data are available on outcome.

Psychosocial Treatment of Adolescent Bipolar Disorder

In adapting existing adult psychosocial treatments used with bipolar disorder for children and adolescents, several considerations should be made. For example, modifications should take into account the developmental level of the patient. Thus, information provided within an age-appropriate context may be more understandable and more widely accepted by the patient and family members; this may include the use of age-appropriate language rather than medical terminology. Psychoeducation conducted against the backdrop of a normal developmental trajectory may help distinguish normal childhood tantrums and adolescent moodiness from bipolar disorder (Fristad et al., 2009). Given the high rate of comorbidity in pediatric bipolar disorder (Findling et al., 2001; Goldstein, 2012), information about comorbid conditions may be included as part of psychoeducation. Comparative risks and benefits of psychotropic medications used with this population should be included. Psychoeducation may also focus on the manner in which early-onset bipolar disorder differs from adult bipolar disorder—childhood bipolar disorder often manifests in less discrete episodes and more frequent mood episodes (Findling et al., 2001; West & Pavuluri, 2009). Additionally, the manner in which symptoms manifest may need to be considered within a developmental framework, as children exhibit several affective symptoms differently than adults (Goldstein, 2012). Furthermore, age-specific issues may be targeted, including substance use, suicide prevention, family conflict, teasing, and academic concerns that may not be relevant in an adult population. Finally, because bipolar disorder is highly familial, it is likely that at least one of the parents of a child with pediatric bipolar disorder has either unipolar or bipolar mood disorder (Findling et al., 2001). The course of a parent’s mood disorder is likely to be intimately related to the child’s presentation, compliance, and management. Given that parental depression may interfere with a mood-disordered child’s response to treatment (Brent et al., 1998), it is particularly important that the parent’s own affective illness be addressed when treating a bipolar child or adolescent.


The empirical literature on psychosocial treatment of adult and adolescent bipolar disorder suggests that several forms of therapy, including FFT-A, CBT, IPT supplemented with SRT, family-based treatments, and group therapy, may be adjunctive treatments (Weinstein et al., 2013; West & Pavuluri, 2009). These treatments have been reported to produce better medication compliance, fewer hospitalizations, and improved social and occupational functioning.

Pharmacologic Treatment of Adolescent Bipolar Disorder

Compared with what is known about the pharmacotherapy of bipolar disorder in adults, relatively little is known about the medication management of bipolar disorder in young people. This is unfortunate, as there has been an increased appreciation over the past few years that bipolar disorder is a chronic, debilitating condition when it occurs in children and (p. 64) adolescents. Until recently, most prospective pharmacotherapy studies in pediatric bipolar disorder were uncontrolled and methodologically limited. However, over the past several years, more definitive treatment studies have been conducted.

The manic and mixed phases of bipolar disorder are more prevalent in teenagers compared to the depressed phase of illness. This may explain why more methodologically stringent work has been done in patients with manic and mixed states compared to adolescents with bipolarity who are depressed. It should be noted that most recently conducted studies are acute treatment studies, often only a few weeks in duration. As such, there is still a substantive need for longer-term controlled trials.

Fortunately, there has been a substantial increase in the number of methodologically stringent medication studies within this patient population over the past few years (Table 2.1). These data can provide practicing clinicians practical information on rational treatment approaches to the pharmacotherapy of this illness.

Table 2.1 Selected Double-Blind, Placebo-Controlled Acute Monotherapy Studies in Adolescent Bipolar Disorder



Mood State(s)

Age (years)

Study Length (wks)/Sample Size


Wagner et al., 2006


Manic or mixed



Oxcarbazepine not superior to placebo

Tohen et al., 2007


Manic or mixed



Olanzapine superior to placebo (p < .001); mean weight gain with olanzapine = 3.7 kg

Findling et al., 2009


Manic or mixed



Aripiprazole superior to placebo (p < .0001);

MCA = extrapyramidal disorder, sedation

Haas et al., 2009


Manic or mixed



Risperidone superior to placebo (p < .001); MCA = somnolence, headache

Wagner et al., 2009

Extended-release divalproex

Manic or mixed



Extended-release divalproex not superior to placebo (p = .604)

Findling et al., 2013


Manic or mixed



Ziprasidone superior to placebo (p = .0005); MCA = sedation, somnolence, headache

Pathak et al., 2013





Quetiapine superior to placebo (p < .001); MCA = somnolence, sedation

Findling et al., 2015


Manic or mixed



Lithium superior to placebo (p = .03); MCA = vomiting, nausea

MCA = most common adverse event


Despite lithium’s role as a benchmark treatment in adults (Severus et al., 2014), there is a limited research base in regard to adolescents. Some studies have focused on characterizing lithium’s pharmacokinetics (Findling et al., 2010) and dosing for youths (Findling et al., (p. 65) 2011). One large randomized trial evaluated monotherapies in adolescent bipolar disorder. The Treatment of Early Age Mania study targeted 6- to 15-year-old children and adolescents (N = 279) with manic or mixed-phase bipolar disorder and randomized them to receive lithium, divalproex sodium, or risperidone (Geller et al., 2012). In this study, risperidone was found to be significantly more efficacious than lithium or divalproex sodium for the initial treatment of childhood mania but was associated with potentially serious metabolic effects.

Only one placebo-controlled trial has been published. This recent study randomized youth ages 7 to 17 years with bipolar I manic or mixed episodes to lithium or placebo for up to 8 weeks of treatment (Findling et al., 2015). Results indicated that lithium was associated with significantly greater manic symptom improvement compared to placebo, but was not associated with weight gain. Response rates were 32% for lithium and 21% for placebo.

Carbamazepine has historically been given as a treatment to young people with a variety of neuropsychiatric conditions. Most reports describing the use of this drug in young people with neuropsychiatric conditions lack methodological rigor. However, one prospective open-label treatment study examined the effectiveness of an extended-release formulation of carbamazepine (CBZ-ERC) in 157 youths between the ages of 10 and 17 years. The CBZ-ERC was administered in divided doses with flexible total daily dosing ranging between 200 and 1.200 mg/day. The results of that prospective clinical trial provide preliminary data to suggest that CBZ-ERC may be effective in this patient population (Findling & Ginsberg, 2012). However, more definitive studies are needed to ascertain whether the efficacy that CBZ-ERC has in adults also is present in pediatric patients.

At present, one published placebo-controlled study has examined the efficacy of extended-release divalproex sodium (Depakote-ER) in the treatment of pediatric mania; it was not found to be superior to placebo (Wagner et al., 2009). However, in an unpublished 8-week study that compared divalproex sodium (Depakote) to placebo in the treatment of pediatric manic or mixed states, divalproex was found to be superior to placebo (Kowatch et al., 2007). The specific reason for the divergent results between these two studies is not known.

Topiramate does not appear to have efficacy in the treatment of mania in adults, but preliminary data suggest that it might be of benefit for bipolar youths (DelBello et al., 2005). However, a double-blind, randomized trial of 120 Iranian youths with manic or mixed-episode bipolar disorder (ages 12–18 years) found that sodium valproate was superior to topiramate (Hebrani, Behdani, & Manteghi, 2009).

Another anticonvulsant medication that has been studied in this patient population is oxcarbazepine. In a multisite, placebo-controlled study, oxcarbazepine was shown not to be effective in the treatment of youth with mania (Wagner et al., 2006).

Several other agents have been examined in the context of acute prospective placebo-controlled monotherapy treatment trials (see Table 2.1). Most of these published placebo-controlled trials have focused on the atypical antipsychotics. These larger-scale studies have examined risperidone, olanzapine, quetiapine, aripiprazole, and ziprasidone. In addition, results from an 8-week open-label study of 15 patients provide preliminary data that paliperidone monotherapy might be of benefit for pediatric patients with bipolar illness (Joshi et al., 2013). For youth in a manic or mixed state, risperidone, olanzapine, quetiapine, and aripiprazole are FDA approved for treatment. As with MDD, there are ongoing concerns about risks for metabolic syndrome with antitypical antipsychotic use for those with bipolar disorder. The overall rate of metabolic syndrome in those with bipolar disorder was estimated in a meta-analysis of 37 studies to be about 37% (Vancampfort et al., 2013). The rate was 45.3% for those taking antipsychotics compared to 32.4% among those not taking antipsychotics (Vancampfort et al., 2013).

Asenapine has also been shown to be effective in the treatment of manic or mixed states in adults. No articles on the use of asenapine in the treatment of bipolar adolescents have been published.

(p. 66) As mentioned previously, most medication monotherapy trials in pediatric bipolar disorder have focused on patients either in the manic or mixed phase of illness. At present, only one published placebo-controlled study has examined the efficacy of medication monotherapy in youths with bipolar disorder who are currently in a depressive episode. In that clinical trial, 32 depressed youths between the ages of 12 and 18 who also met diagnostic symptom criteria for bipolar I disorder were administered either quetiapine or placebo for up to 8 weeks in a double-blind fashion (Delbello et al., 2009). The patients who were treated with active medication received between 300 and 600 mg/day of quetiapine. Quetiapine was found to have no greater efficacy in reducing depressive symptoms than placebo. More than two thirds of both placebo- and quetiapine-treated patients responded during study participation.

The atypical antipsychotic lurasidone has been shown to have efficacy in the treatment of the depressed phase of bipolar illness in adults. At present, there are no data regarding lurasidone’s use in adolescents with bipolar illness.

Psychopharmacologic Combinations

A substantial number of patients do not appear to derive optimal clinical response from acute treatment with a single drug. As a result, investigators have begun to explore whether pharmacotherapy with more than one agent may be useful in the acute treatment of symptomatic bipolar youth. The majority of these studies have explored intervention with (1) an atypical antipsychotic combined with a traditional mood stabilizer or (2) combined treatment with lithium and divalproex sodium.

Data show that the combination of olanzapine and fluoxetine (Symbyax) is effective in the treatment of the depressed phase of bipolar illness in teenagers (Detke et al., 2012). In that study 291 depressed youths between the ages of 10 and 17 were randomized to receive placebo or flexible doses of olanzapine/fluoxetine. Reductions in depressive symptomatology with active treatment were superior to those seen with placebo. The most common side effects noted in the patients receiving olanzapine/fluoxetine were weight gain and increased appetite. Olanzapine/fluoxetine is FDA approved for the treatment of depressed pediatric patients with bipolar I disorder.

Data from chart reviews suggest that clozapine may serve as a useful adjunct to traditional mood stabilizers for treatment-resistant patients (Emslie, Kennard, & Kowatch, 1995). Preliminary reports have noted that the antipsychotics haloperidol and risperidone might be helpful when combined with a traditional mood stabilizer in the treatment of teenagers with mania, particularly those with psychotic symptoms (Frazier et al., 1999; Kafantaris, Coletti, Dicker, Padula, & Kane, 2001; Pavuluri et al., 2004, 2006). Similarly, a case series reported that adjunctive olanzapine may provide benefit to pediatric patients who do not receive satisfactory benefit from lithium or anticonvulsants (Emiroglu et al., 2006).

Data from case series have described the potential usefulness of topiramate as an adjunct to antipsychotics and/or traditional mood stabilizers in this population (Barzman et al., 2005). In addition, preliminary data from a prospective trial suggest that adjunctive topiramate might be an effective way to reduce the weight gain associated with olanzapine use in this population (Wozniak et al., 2009).

One published study rigorously examined quetiapine treatment as an adjunct in patients ages 12 to 18 with bipolar I disorder (DelBello et al., 2002). In that 6-week trial, 30 hospitalized adolescents with mania or mixed presentations were treated with divalproex sodium at an initial dose of 20 mg/kg per day. Half of the subjects were randomized to receive adjunctive quetiapine and half were randomized to receive adjunctive placebo. The quetiapine was titrated to a total daily dose of 450 mg. Results showed greater reductions in manic symptomatology in the youths receiving combination therapy, but sedation was also more common in these youths.

Another prospective trial that examined combination pharmacotherapy in pediatric bipolar disorder described the response of 90 outpatients (mean age ~11 years) who were treated (p. 67) with combination lithium and divalproex sodium therapy for up to 20 weeks (Findling et al., 2003). Response rates were larger than those seen in other prospective trials in this population. Response rates were higher than those described in adults using a similar combination treatment paradigm. Combination treatment was well tolerated. Of particular interest was the observation that residual depressive symptomatology was not manifest after combination lithium plus divalproex treatment in this population. This is in direct contrast to adults, in whom depression was often seen as a problematic residual mood state.


On the basis of available evidence, it appears that combination pharmacotherapy with more than one mood-stabilizing agent may be a rational approach for some youth who have manic, hypomanic, or mixed states. Whether treatment should begin with drug monotherapy or combination pharmacotherapy should be a topic of further study.

Maintenance Trials

The first data to suggest that lithium maintenance therapy might be useful in the treatment of adolescents with bipolar disorder were published in 1990. That report described the results of a naturalistic, prospective, 18-month follow-up study of 37 youth with bipolar I disorder who had responded to lithium treatment (Strober et al., 1990). In the 13 youth who discontinued lithium maintenance therapy, the relapse rate was 2.5 times greater than in those who continued lithium treatment. Despite methodological limitations inherent in this study design, these data suggest that maintenance lithium treatment in young people who respond to this compound may be beneficial.

In one prospective trial of maintenance pharmacotherapy in pediatric bipolar disorder, 60 youths who had responded to acute treatment with combination therapy using both lithium and divalproex sodium were randomized to receive monotherapy treatment with either lithium or divalproex for up to 76 weeks in a double-blind fashion (Findling et al., 2005). The overall median survival time in the study was approximately 100 days. Lithium and divalproex sodium had similar effectiveness as monotherapy.

In another study, Findling et al. (2013) conducted a double-blind continuation study in a group of youths who had completed a 4-week, double-blind, placebo-controlled trial of aripiprazole. At the time of initial randomization, patients were between the ages of 10 and 17 years and received 10 mg/day or 30 mg/day of active treatment or placebo. Subsequently, 210 patients continued on their same blinded treatment assignment for up to 26 more weeks. Continued pharmacotherapy with aripiprazole was found to be more effective during this 26-week period than treatment with placebo. However, the rates of study completion were low: 45.3% for aripiprazole 10 mg/day; 31.0% for aripiprazole 30 mg/day; 18.8% for placebo.

Finally, lamotrigine has been shown to be an effective maintenance treatment in adults with bipolar illness. Currently, there are no published data about whether lamotrigine has efficacy either as a maintenance therapy or as an acute treatment for pediatric patients.


There are only limited data from methodologically rigorous trials in pediatric bipolar disorder. Evidence suggests that monotherapy with lithium, carbamazepine, divalproex sodium, or olanzapine may be useful in the treatment of young patients with mania and related mood states. Risperidone was found in one large trial to be superior to lithium and divalproex sodium for the initial treatment of childhood mania, but metabolic adverse effects were apparent. A consistent theme from the literature is that a substantial number of patients do not respond to monotherapy with these agents. For this reason, investigators have begun to explore combination pharmacotherapy. It seems that simultaneous treatment with more than one agent may be a rational form (p. 68) of intervention for some patients. The combination of divalproex sodium and quetiapine has shown efficacy compared to divalproex plus placebo, though sedation was a common adverse event with quetiapine. Beyond this study, circumstances in which combination drug therapy might be most rationally employed have not been identified. Because pediatric bipolar disorder is a chronic condition, it appears that young people with this illness will need long-term treatment. Although maintenance pharmacotherapy data are lacking, patients who respond to a given acute pharmacotherapy regimen may continue to benefit from ongoing treatment with the drug(s) that led to symptom amelioration. Concerns about long-term treatment with atypical antipsychotics suggest that other agents may have a potentially better risk/benefit balance for maintenance treatment.

In summary, pediatric bipolar disorder is a chronic condition associated with substantial dysfunction and suffering. There are few methodologically sound pharmacologic treatment studies; thus, more research on this topic is sorely needed.