Depressive disorders are prevalent among young people [1,2] and are frequently untreated or inadequately treated . The use of antidepressants in the management of depression in children and adolescents is contentious. Selective serotonin reuptake inhibitors (SSRIs) had been widely used in paediatric depression until warnings about their use were issued following concerns that children and adolescents who had participated in trials evaluating the effectiveness of SSRIs were at an increased risk of self-harm. This resulted in comprehensive reviews of the trials by government regulatory agencies which all concluded that only fluoxetine should be recommended for use in children and adolescents with depressive disorders [4–7].
Claims and counter claims about the effectiveness of SSRIs and risk related to their use have followed [8–13]. Concerns have been raised that if antidepressant use is limited young people could be deprived of effective treatments for a condition that carries with it considerable morbidity and mortality [8,9,14]. Recent research has suggested that pessimism about effective treatments has resulted in reluctance on the part of patients to seek help . Further, lower rates of depressive disorder diagnoses  may indicate that clinicians are reluctant to make this diagnosis, perhaps due to uncertainty about effective treatments.
The two key questions when considering the use of SSRIs in the treatment of children and adolescents with depressive disorder are, do they work and are they well tolerated? Answering these questions requires consideration of the evidence.
Do selective serotonin reuptake inhibitors work?
We undertook a Cochrane systematic review to evaluate the evidence about the effectiveness and risk of this class of medication for children and adolescents .
Twelve trials met the inclusion criteria for our review. The outcome ‘response’ was reported in all trials. This was a binary outcome created from a cut-point on a continuous depression scale. Both the scale and cut-point varied between trials. The pooled treatment effect of all SSRI compounds on the outcome ‘response’ demonstrated that a greater percentage of participants receiving SSRIs responded compared with those receiving placebo [relative risk (RR) 1.28; 95% confidence interval (CI) 1.17, 1.41] (Fig. 1).
There was evidence that effectiveness of SSRIs on ‘response’ varied by compound (with 81.3% of the variation across SSRI compound effect estimates due to heterogeneity rather than chance; P = 0.001). There was a statistically significant difference in the percentage of children and adolescents who ‘responded’ when receiving fluoxetine compared with placebo (RR 1.86; 95% CI 1.49, 2.32). The percentage of those ‘responding’ while on fluoxetine ranged between 41 and 61% compared with a range of 20–35% in those on placebo. Depressive symptom scores [Children's Depression Rating Scale-Revised scale (CDRS-R); range 17–113] were also statistically significantly lower for those receiving fluoxetine compared with placebo at the end of treatment [ranging between 8 and 12 weeks of treatment; treatment effect 5.63 (95% CI −7.38, −3.88)].
For paroxetine, there was no statistically significant difference in the percentage of children and adolescents who ‘responded’ to paroxetine compared with those who ‘responded’ to placebo (RR 1.09; 95% CI 0.95, 1.26). There was a lack of consistency for both sertraline and citalopram. There was no statistically significant difference in ‘response’ rate between the sertraline and control groups (RR 1.17; 95% CI 1.00, 1.36). However, depressive symptom scores were significantly lower in the sertraline group (treatment effect CDRS-R 3.56; 95% CI −6.69, −0.42). There was a statistically significant difference in the percentage of children and adolescents who ‘responded’ when receiving citalopram/escitalopram (RR 1.30; 95% CI 1.02, 1.67), compared with placebo but no significant difference in depressive symptom score (CDRS-R).
Functioning is perhaps the most relevant outcome for young people as it measures their ability to sustain school/work, interpersonal relations, and social functioning on a day-to-day basis . There was no evidence that SSRIs were effective in improving functioning.
What are the strengths and weaknesses of the evidence base?
It is important to be cautious when interpreting the results of our review. The interpretation of the outcome ‘response’, the selection of patient outcomes, the clinical significance of treatment effects, and other issues of internal and external validity all need consideration.
In the search for effective treatments, the ideal goal is a resolution of symptoms (i.e., remission) and a return to normal function. In the trials in our review, remission or functioning were inconsistently reported, not reported, or not collected. Most trials reported ‘response’ which equated to a reduction in symptoms that fell short of remission. Inconsistencies in the choice of measurement scale used between trials coupled with varying definitions of ‘response’ (e.g., [13,18,19]) make it difficult to interpret this outcome. It highlights a lack of consensus over both the most appropriate measurement scale to use in trials, and what is considered a clinically meaningful improvement on these scales. In addition, varying cut-points for dichotomizing a continuous measure may also be indicative of a form of within study selective reporting bias in which the most favourable cut-point is chosen on the basis of the results .
Statistical versus clinical significance
The size of the difference in improvements between groups is of questionable clinical significance. For example, for fluoxetine, the average CDRS-R score at follow-up was 5.63 lower for those in the treatment group compared with the placebo group, with both groups improving substantially. A difference of 25 points on the CDRS-R scale has been shown to differentiate clinical and nonclinical groups, whereas a difference of 19 points has been shown to differentiate clinical groups with and without depressive disorder .
There was generally limited information reported about the conduct of the trials, so it was difficult to assess their internal validity. However, two methodological components which were reported, attrition and blinding, provided cause for concern. These have empirically been shown to be associated with bias (, chapter 8, pp. 217–221). High attrition rates were observed in all trials; ranging from 17 to 46% in the control groups and 17–40% in the intervention groups. The bias associated with this attrition may be substantial. In the included trials, the clinicians were also the outcome assessors for the primary outcomes. It is possible that clinicians and participants may have been able to guess whether the participants were on active treatment or placebo from side effect profiles.
Generalizability of the included trials is limited. Most trials excluded placebo responders, those with complex comorbidity, and those at risk of suicide so that participants were not typical of those seen in clinical practice. It is unclear how effective these medications would be for such young people. In addition, the trials were of short duration, spanning only 7–12 weeks. Therefore, the effectiveness and risks of these medications over a 6–12 month time period, which is the typically recommended treatment duration [23,24], is unknown. It is possible, for example, that functioning may take a longer period of time to return to normal, but these trials have not evaluated this.
In conclusion, the effectiveness of SSRIs in patients typical of those presenting in clinical practice is unknown. For those young people with uncomplicated depression there may be a small advantage to using fluoxetine in the short-term. However, there is no evidence that treatment with fluoxetine will improve their function.
Are selective serotonin reuptake inhibitors well tolerated?
There is conflicting evidence about the link between SSRIs and suicidality. Consistent with US Food and Drug Administration (FDA) findings and other meta-analyses [25,26], our review showed an increased risk of suicidal ideation and behaviour in those receiving an SSRI compared with those receiving a placebo (RR 1.80; 95% CI 1.19, 2.72). However, compared with rates found in community studies [27–31], the rates of these phenomena were very low (Fig. 2). This is likely to be because of the exclusion of those most at risk. These phenomena were measured in various ways across trials, or not reported at all. The Treatment for Adolescents with Depression Study (TADs) trial was the only one to report suicidal ideation using a validated questionnaire (Suicide Ideation Questionnaire ), and in contrast to the increased risk of suicidal ideation behaviours observed across other trials, participants receiving fluoxetine had lower suicidal ideation at follow-up than at the start of treatment and compared with those in the placebo group . Although there were no reports of completed suicide in a total sample of 2240 children and adolescents included in our review; suicide is rare, the trials were of short duration, and were not powered to assess suicide risk from SSRIs. Although appropriate caution is required in interpreting the risk of suicidal ideation and behaviours due to inconsistency in reporting of this outcome across trials, the increased risk for those receiving SSRIs raises concern. Adding to this concern is the fact that the risk may be different, perhaps larger, if SSRIs are used in a population of patients who are typical of those seen in clinical practice.
In contradiction to the FDA findings  and other reviews of SSRIs [16,25], evidence from routinely collected data show a different story. In the decades prior to 2004, internationally, there was an increase in the prescribing rates of SSRIs. If SSRIs do increase suicidal behaviours, an increase in suicide rates might be expected. However, one study suggested that increased prescribing of antidepressants in those years was associated with reduced suicide rates in some countries . Further, since 2004 there has been a reduction in SSRI prescriptions for young people due to regulatory warnings, and over the same period of time an increase in the suicide rate in the Netherlands and the United States of America with claims that these two phenomena are related . The low overall rates of suicide make these claims questionable  with the link between antidepressant medication use and suicide rates contested .
Our meta-analysis  showed that there was limited evidence of efficacy of SSRIs. Fluoxetine was the only SSRI for which there was consistent evidence of its effectiveness in improving ‘response’ and depressive disorder symptoms in children and adolescents compared with placebo. Most children and adolescents who participated in the trials were not typical of those presenting for treatment; they had low rates of comorbidity and were at low risk of suicide or self-harm. Those who responded to placebo had been screened out, maximizing the chance of showing a difference between placebo and medication. Despite this, even for fluoxetine the improvement is modest. There was evidence of an increased risk of suicidal ideation and behaviour for those on an SSRI, although rates were low. However, one included trial showed a reduction in suicidal ideation for those prescribed fluoxetine and evidence from studies using routinely collected data have shown an inverse relationship between SSRI prescription and suicide.
When deciding treatment options, all interventions should be considered including psychotherapies and other antidepressants.
When beginning treatment, interventions of least risk should be considered first. Standard care, typical of most child and adolescent mental health services, may be sufficient for some to recover. Consideration to psychotherapies can also be given. There is some evidence that a range of psychotherapeutic interventions may be effective, at least in the short-term .
Antidepressants potentially pose greater risks. Our systematic review examined the evidence of SSRIs . However, a recent review included trials comparing all classes of antidepressant medications with placebo . Within this review, the authors undertook a meta-regression to examine whether a drug class was predictive of treatment effect for the outcome ‘response’ and concluded that there was no statistical evidence to support the hypothesis that SSRIs were more effective than tricyclic antidepressants.
It is possible that medications are effective for some individuals including those with more severe and complex presentations. However, only a few trials (e.g. ) have included these participants, so there is limited evidence about the effectiveness and risks of treatments in these subgroups.
Decisions on treatment options should always be made in consultation with the young person and their families. Given the equivocal evidence base about the effectiveness and risks of the treatments, there is a particular need to provide accurate information about the limits of the evidence base and elicit treatment preferences of the young people and their family; allowing them to be actively involved in their treatment decision [16,41–45].
The results of our systematic review, and others (e.g. ), have identified the need for further research. This could include a systematic review that allows for direct and indirect comparisons of various treatment options; including antidepressant medication and psychotherapeutic interventions. This may necessitate the need for a large head-to-head trial testing those interventions demonstrated to be the most effective in the systematic review. Such a trial should include young people similar to those seen in clinical practice and should be of at least 6–12 months duration. Consideration should be given to adequately powering the trial to test whether the effectiveness of the treatments differs by severity of depressive disorder.
Finally, it is evident from the variation in measurement instruments used between trials included in our review that there is a need to establish a standard set of outcome measures. This has been successfully done in rheumatology . Additional research may need to be undertaken to establish what outcomes are important to children and adolescents.
Untreated depressive disorder is associated with morbidity and mortality; however, the argument that SSRIs should be used to reduce the morbidity and mortality associated with depressive disorders is not well founded [16,47]. Our systematic review, along with others [16,39], has demonstrated the limited effectiveness of SSRIs and, furthermore, the limited information about their effectiveness in young people typically seen in clinical practice. Evidence regarding the risk of self-harm on SSRIs is conflicting. The search for effective treatments for depressive disorders in young people is not over .
In the meantime, it is essential that optimism and an active approach are maintained in the treatment of depressive disorders in children and adolescents. Depressive disorders in young people need to be detected, and actively managed, first utilizing interventions of least risk. Many will respond to simple remedies . There is considerable heterogeneity in young people who present with depressive disorder and SSRIs may be effective for subgroups of these young people, which are not yet defined. For young people who have not responded to other interventions, or those with a particularly severe or disabling disorder, a trial of medication could be considered after discussing the risks and limited evidence of effectiveness with the young person and their family. Close monitoring of symptoms and side effects is always crucial, but particularly once medication is utilized. Guidelines, although varying, consistently recommend such an approach, including more judicious use of medication [10,23,24,48]. We would support this.
1 Birmaher B, Ryan ND, Williamson DE, et al
. Childhood and adolescent depression: a review of the past 10 years: part 1. J Am Acad Child Adolesc Psychiatry 1996; 35:1427–1439.
2 Lewinsohn PM, Hops H, Roberts RE, et al
. Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-III-R disorders in high school students. J Abnorm Psychol 1993; 102:133–144.
3 Andrews G, Sanderson K, Corry J, Lapsley HM. Using epidemiological data to model efficiency in reducing the burden of depression. J Ment Health Policy Econ 2000; 3:175–186.
8 Brent DA. Antidepressants and pediatric depression: the risk of doing nothing. N Engl J Med 2004; 351:1598–1601.
9 Brent DA, Birmaher B. British warning on SSRIs questioned. J Am Acad Child Adolesc Psychiatry 2004; 43:379–380.
10 Cheung AH, Zuckerbrot RA, Jensen PS, et al
. Guidelines for adolescent depression in primary care (GLAD-PC): II treatment and ongoing management. Pediatrics 2007; 120:e1313–e1326.
11 Wagner KD. Pharmacotherapy for major depression in children and adolescents. Prog Neuropsychopharmacol Biol Psychiatry 2005; 29:819–826.
12 Whittington CJ, Kendall T, Fonagy P, et al
. Selective serotonin reuptake inhibitors in childhood depression: systematic review of published versus unpublished data. Lancet 2004; 363:341–345.
13 Jureidini JN, Doecke CJ, Mansfield PR, et al
. Efficacy and safety of antidepressants for children and adolescents. BMJ 2004; 328:879–883.
14 Findling R, Greenhill L. Message to AACAP members. Letter sent by e-mail to members of the AACAP. President of the American Academy of Child and Adolescent Psychiatrists; 2004.
15 Libby AM, Brent DA, Morrato EH, et al
. Decline in treatment of pediatric depression after FDA advisory on risk of suicidality with SSRIs. Am J Hum Genet 2007; 164:884–891.
16 Hetrick SE, Merry S, McKenzie J, et al
. Selective serotonin reuptake inhibitors (SSRIs) for depressive disorders in children and adolescents. Cochrane Database Syst Rev:CD004851. doi: 10.1002/14651858.CD004851.pub2.
17 Zimmerman M, McGlinchey JB, Posternak MA, et al
. Discordance between self-reported symptom severity and psychosocial functioning ratings in depressed outpatients: implications for how remission from depression should be defined. Psychiatry Res 2006; 141:185–191.
18 Emslie GJ, Rush R, Weinberg WA, et al
. A double blind, randomized, placebo-controlled trial of fluoxetine in children and adolescents with depression. Arch Gen Psychiatry 1997; 54:1031–1037.
19 Emslie GJ, Heiligenstein JH, Wagner KD, et al
. Fluoxetine for acute treatment of depression in children and adolescents: a placebo-controlled randomized clinical trial. J Am Acad Child Adolesc Psychiatry 2002; 41:1205–2015.
20 Williamson PR, Gamble C, Altman D, Hutton JL. Outcome selection bias in meta-analysis. Stat Methods Med Res 2005; 14:515–524.
21 Poznanski EO, Mokros HB. Children's Depression Rating Scale Revised. LA: WPS; 1996.
22 Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions version 5.0.2 [updated September 2009]. The Cochrane Collaboration; 2008. Available from www.cochrane-handbook.org
23 National Institute for Health and Clinical Excellence NICE. Depression in children and young people: identification and management in primary, community and secondary care. Leicester, UK: The British Psychological Society; 2005.
24 American Academy of Child and Adolescent Psychiatry (AACAP). Practice parameters for the assessment and treatment of children and adolescents with depressive disorders. J Am Acad Child Adolesc Psychiatry 2007; 46:1503–1526.
25 Dubicka B, Hadley S, Roberts C. Suicidal behaviour in youths with depression treated with new-generation antidepressants: meta-analysis. Br J Psychiatry 2006; 189:393–398.
26 Hammad TA, Laugren T, Racoosin J. Suicidality in pediatric patients treated with antidepressant drugs. Arch Gen Psychiatry 2006; 63:332–339.
27 Kovacs M, Feinberg TL, Crouse-Novak MA, et al
. Depressive disorders in childhood I: a longitudinal prospective study of characteristics and recovery. Arch Gen Psychiatry 1884; 41:229–237.
28 Lewinsohn PM, Rohde P, Seeley JR. Adolescent suicidal ideation and attempts: prevalence, risk factors and clinical implications. Clin Psychol Sci and Pract 1996; 3:25–46.
29 Adolescent Health Research Group. A health profile of New Zealand youth who attend secondary school. N Z Med J 2003; 116:3–12.
30 Grunbaum J, Kann L, Kinchen SA, et al
. Morbidity and mortality weekly report. New York: Centres for Disease Control; 2002.
31 Andrews JA, Lewinsohn PM. Suicide attempts among older adolescents: prevalence and co-occurance with psychiatric disorder. J Am Acad Child Adolesc Psychiatry 1992; 31:655–662.
32 Reynolds WM. Professional manual for the suicidal ideation questionnaire. Odessa, Florida: Psychological Assessment Resources, Inc.; 1987.
33 March J, Silva S, Petrycki S, et al
. Fluoxetine, cognitive-behavioral therapy, and their combination for adolescents with depression: Treatment for Adolescents with Depression Study (TADS) randomized controlled trial. JAMA 2004; 292:807–820.
34 Isacsson G. Suicide prevention: a medical breakthrough? Acta Psychiatr Scand 2000; 102:113–117.
35 Gunnell D, Ashby D. Antidepressants and suicide: what is the balance of benefit and harm? BMJ 2004; 329:34–38.
36 Sheldon T. Dutch academics criticise suicide claims made in American psychiatry journal. BMJ 2008; 336:112.
37 Hetrick S, McKenzie J, Merry S. Response to commentary on Cochrane review of SSRIs in children and adolescents by Dr Bernadka Dubicka and Dr Paul Wilkinson. Evid-Based Child Health 2008; 3: 897–900. doi: 10.1002/ebch.256.
38 Watanabe N, Hunot V, Omori IM, et al
. Psychotherapy for depression among children and adolescents: a systematic review. Acta Psychiat Scand 2007; 116:84–95.
39 Tsapakis EM, Soldani F, Tondo L, Baldessarini RJ. Efficacy of antidepressants in juvenile depression: meta-analysis. Br J Psychiatry 2008; 193:10–17.
40 Goodyer I, Dubicka B, Wilkinson P, et al
. Selective serotonin reuptake inhibitors (SSRIs) and routine specialist care with and without cognitive behaviour therapy in adolescents with major depression: randomised controlled trial. Br J Psychiatry 2007; 335:142.
41 Ilic D, Egberts K, McKenzie JE, et al
. Informing men about prostate cancer screening: a randomized controlled trial of patient education materials. J Gen Intern Med 2008; 23:466–471.
42 O'Connor AM, Legare F, Stacey D. Risk communication in practice: the contribution of decision aids. BMJ 2003; 327:736–740.
43 Wennberg JE. Unwarranted variations in healthcare delivery: implications for academic medical centres. BMJ 2002; 325:961–964.
44 Whitney SN, McGuire AL, McCullough LB. A typology of shared decision making, informed consent, and simple consent. Ann Intern Med 2004; 140:54–59.
45 Hetrick SE, Simmons MB, Merry SN. SSRIs and depression in children and adolescents: the imperative for shared decision making. Australas Psychiatry. 2008; 16:354–358.
46 Tugwell P, Boers M, Brooks P, et al
. OMERACT: an international initiative to improve outcome measurement in rheumatology. Trials 2007; 8:38.
47 Jensen P. After TADS, can we measure up, catch up, and ante up? J Am Academy Child Adolesc Psychiatry 2006; 45:1456–1460.
48 Zuckerbrot RA, Cheung AH, Jensen PS, et al
, and the GLAD-PC Steering Group. Guidelines for adolescent depression in primary care (GLAD-PC): I identification, assessment and initial management. Pediatrics 2007; 120:e1313–e1326.