Article Text


Evidence-based treatment of adolescent major depression
  1. B Dubicka1,
  2. P Wilkinson2
  1. 1Honorary Lecturer, University of Manchester, Manchester, UK and Consultant, The Junction Adolescent Unit, Piccadilly, Lancaster, UK
  2. 2Clinical Lecturer, Section of Developmental Psychiatry, University of Cambridge, Cambridge, UK
  1. Dr B Dubicka, The Junction Adolescent Unit, Scotforth, Piccadilly, Lancaster LA1 4PW, UK; Bernadka.Dubicka{at}

Statistics from

Adolescent depression is a serious and debilitating disorder (see box). Up to one in 20 adolescents suffer from major depression at any point in time, and 20% of adolescents have at least one episode of clinical depression by the age of 18. Once depression is established it often becomes chronic. About a fifth of adolescents with major depression will continue to have a persistent disorder, and another third will recover but go on to have recurrent episodes.


Depression runs in families and children of depressed parents have an increased risk of becoming depressed, which is likely to be due to both the direct effect of genes as well as the adverse influence of depressed parenting. Recent research has highlighted the complex interaction of genetic and environmental influences, such as adverse life events, in depression.

The six Cs of depression

  • Common

  • Chronic

  • Causation complex

  • Comorbidity

  • Complications

  • Commit suicide


Pure depression is rare, and it is usually accompanied by other psychiatric disorders. A recent UK study found that 89% of cases had a comorbid disorder and the average number of additional disorders was three.1 Other complications include school refusal, academic failure, impaired peer relations, drug and alcohol abuse, and family relationship problems. However, the most important complication is suicide and depression is the most important risk factor for suicidality. Findings from 20-year follow-up data of depressed children and adolescents have shown that 2.5% had committed suicide and nearly half had attempted suicide.

Any treatment plan for adolescent depression therefore needs to take account of the chronic, relapsing nature of the disorder, consider aetiological factors such as parental depression, and address any concurrent psychiatric disorder and psychosocial complications, particularly suicidality.


Early studies of cognitive behavioural therapy (CBT) were promising in the prevention and treatment of depression and meta-analyses found large effect sizes. However, a recent meta-analysis, which included studies with active controls, reported a relatively small effect size of 0.3.2 The largest study, the Treatment of Adolescent Depression Study (TADS), which had more severely impaired participants than earlier studies, found that CBT alone did not differ from placebo on any of 15 outcomes.3 However, there is evidence from a sub-analysis that CBT may be effective in adolescents from higher income families4 and for high levels of cognitive disorders.

Recent findings in adult depression present an interesting line of evidence for the treatment of more severe depressive disorder. Dimidjian and colleagues reported that behavioural activation (using activity scheduling and graded tasks) without a cognitive component was comparable to antidepressant medication, and both significantly outperformed cognitive therapy.5 These results suggest that behavioural activation may be a useful treatment for more severe disorders and a viable alternative to antidepressant medication, although this needs to be tested in adolescents.

There have been three randomised controlled trials (RCTs) with positive results for interpersonal therapy (IPT). One trial was particularly encouraging as IPT was delivered by social workers with only basic training in therapeutic skills and proved to be superior to treatment as usual from school mental health teams.6 However, there are very few trained IPT therapists in the UK.

Although the role of parenting is thought to be important in both the genesis and maintenance of adolescent depression, the few studies of family intervention have generally yielded disappointing results. The addition of a parent group to CBT for depressed adolescents has not been shown to improve outcomes,7 and CBT significantly outperformed systemic family therapy in another study.8


There is little evidence that tricyclic antidepressants are efficacious in children and adolescents,9 and, in view of the side effects and risks of toxicity in overdose, they are not recommended in either UK or US guidelines.10 ,11

In recent years there has been much controversy surrounding the efficacy and safety of newer generation antidepressants (NGAs) in children and adolescents, in particular with regards to the issue of suicidality.

Efficacy of newer generation antidepressants

Interpretation and application of the evidence base from trial data are limited by numerous exclusion criteria; recruitment by advertising rather than clinical referral; differing methodologies and definitions of outcome; short-term data; the preponderance of US studies; and sponsorship of most studies by the pharmaceutical industry. Samples have varied in severity, comorbidity, suicidality and age, as well as placebo response. In particular, the most suicidal and complex cases have tended to be excluded, therefore findings may not be applicable to a typical clinic population and to the most severe cases where antidepressants would be considered.

Bridge and colleagues compared efficacy for the NGAs for different indications and found that, although there was a significant drug-placebo difference for all indications, this was in fact smaller for depression than for anxiety disorders.12 However, fluoxetine showed the highest pooled risk difference in major depression (20%), and is the only NGA to demonstrate efficacy on a primary outcome in two trials.3 ,13 Nefazadone also had a similar risk difference but this was based on a secondary outcome in one trial and the drug is not recommended due to concerns regarding hepatotoxicity.11 Only two other NGAs have demonstrated efficacy on a primary outcome: sertraline and citalopram. However, the sertraline trial consisted of two identical negative trials that had a positive outcome when combined as one trial, and citalopram has one negative, as well as one positive, trial; thus the evidence base is weaker than that for fluoxetine.1416

The evidence base for other NGAs is still more limited. Paroxetine has demonstrated efficacy in secondary outcomes in one trial17 but not in others,18 ,19 but it has also been associated with increased agitation. Venlafaxine and escitalopram have only shown positive effects in sub-analyses in adolescents.20 ,21

How do antidepressants compare to psychological treatment? In TADS, fluoxetine was significantly superior to CBT on the majority of outcome measures.3 One other trial has compared antidepressants (sertraline) to CBT, and this found a superior effect of CBT, however, the study was underpowered.22


The recent meta-analysis by Bridge and colleagues also examined suicidality.12 Overall, there was an increased risk difference of suicidal ideation/suicide attempt across all trials and indications for drug versus placebo (0.7%, NNTH 143), but the risk differences within each indication were not statistically significant. There were no completed suicides. Interestingly, a recent FDA analysis found that this increased risk extends to the age of 24, after which the risk appears to be neutral and possibly protective, until the age of 65, when the risk reduces.23

However, these meta-analyses are limited by the fact that most primary studies excluded participants with pre-existing suicidality, and other study designs do not consistently support this increased risk. For example, autopsy studies of suicide in youths have not found evidence for an increased risk in suicide with NGAs;24 risk for attempts appears to be highest in the month prior to starting antidepressants;25 and longer-term treatment has been found to be associated with reduced attempts in adolescents.26 Moreover, recent data from USA and The Netherlands have found a fall in antidepressant prescribing of 22% in each country since the suicidality warnings were issued, and this has been accompanied by an alarming increase in suicide rates in youths in both countries (14% during 2003–2004 in the USA, and 49% during 2003–2005 in The Netherlands).27

Therefore the evidence for an increased risk in suicidality is not clear-cut, but any possible increased risk needs to be balanced against the well established risk of actual suicide in untreated depression.

Official recommendations on NGAs

Recommendations from regulatory agencies have varied significantly. The UK Committee on Safety in Medicines concluded that only fluoxetine showed a favourable risk:benefit ratio in the treatment of depression in young people and other NGAs have been contraindicated.28 Conversely, the US Food and Drug Administration (FDA) issued a warning for all NGAs, regardless of indication, but did not contraindicate them in order to allow access to these drugs for those who could benefit.29 The European Medicines Agency contraindicated all NGAs in children and adolescents except for their licensed indications.30

Clinical guidelines in the USA and UK also present conflicting advice. The UK National Institute for Health and Clinical Excellence (NICE) advises that antidepressants should only be given if psychological treatment is not effective or is not acceptable to the patient, and that fluoxetine should be the first line antidepressant in moderate to severe adolescent depression, with citalopram and sertraline as second line.10 In addition, antidepressants should only be given with a concurrent psychological treatment. In the USA, the Texas Children’s Medication Algorithm Project11 recommends any of these three antidepressants as a first line pharmacological treatment, once non-pharmacological interventions have been considered. All SSRIs are permitted as second line. Augmentation agents are suggested in partial responders, prior to switching to a third line drug (bupropion, mirtazapine, duloxetine or venlafaxine) if necessary.

As a result of these differing recommendations, clinicians are left uncertain as to which is the most appropriate treatment pathway.


TADS found a superior effect of fluoxetine combined with CBT when compared to either monotherapy (effect size 0.98 for depression scores);3 however, for more severe cases, there was no advantage of combined treatment over fluoxetine alone.4 The UK Adolescent Depression Antidepressant and Psychotherapy Trial (ADAPT),1 which had more severely depressed participants than TADS and higher levels of suicidality, failed to find an advantage of combined CBT and a selective serotonin reuptake inhibitor (SSRI) over an SSRI, in the context of usual care. Other studies have also not found an advantage of combined treatment over SSRIs.22 ,31 Results from TADS suggested that the addition of CBT to fluoxetine may reduce suicidality, although the ADAPT study failed to replicate this finding.1 Interestingly, ADAPT also reported that 20% of adolescents improved with a brief initial psychosocial intervention.


The evidence for CBT is conflicting, but it probably has a role in mild depression and for adolescents who refuse medication. There is some evidence to suggest that CBT may be more beneficial in high income families and for high levels of cognitive distortions. CBT is a scarce and expensive resource which therefore needs to be targeted at those who may benefit most. Interpersonal psychotherapy appears to be a promising treatment as is behavioural activation, but more training and research is needed in these areas. In moderate to severe depression, the greatest evidence for efficacy (compared with both placebo and CBT) has been demonstrated for fluoxetine. This should therefore be offered early in treatment, with an explanation of the evidence base and possible risk of suicidality with antidepressants, versus the known risk of suicide in depression, so families can make an informed decision regarding the possible risks and benefits of both medication and psychological treatment. However, ADAPT findings suggest that a non-specific brief psychosocial intervention may be helpful in some cases prior to initiating a more specialised treatment, and this requires further investigation. In light of recent evidence, the current UK NICE guidelines need revising.10 In the future we need improved trial methodology which reflects typical clinical populations, an improved understanding of phenomenology, and more effective targeted treatments, which take account of the heterogeneity found in adolescent depression.


View Abstract


  • Competing interests: BD has attended educational meetings sponsored by Lilly and Janssen.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • EBMH notebook
    Steven Reid