For Peer Review Only Journal: BMJ Open

BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

First- and second-generation antidepressants for depressive disorders in children, adolescents, and young adults: Study Protocol for a Systematic Review and Network Meta- Analysis

For peer review only Journal: BMJ Open

Manuscript ID: bmjopen-2015-007768

Article Type: Protocol

Date Submitted by the Author: 23-Jan-2015

Complete List of Authors: Zhou, Xinyu; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Bin, Qin; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Whittington, Craig; UCL, Clinical, Educational and Health Psychology Cohen, David; AP–HP, Hôpital Pitié–Salpétrière, Institut des Systèmes Intelligents et de Robotiques (ISIR), Centre National pour la Recherche Scientifique, Université Pierre et Marie Curie, Department of Child and Adolescent Psychiatry Liu, Yiyun; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Del Giovane, Cinzia; University of Modena and Reggio Emilia, Clinical and Diagnostic Medicine and Public Health http://bmjopen.bmj.com/ Michael, Kurt; Appalachian State University, Department of Psychology Zhang, Yuqing; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Xie, Peng; The First Affiliated Hospital of Chongqing Medical University, Department of Neuropsychiatry

Primary Subject Mental health Heading: on September 23, 2021 by guest. Protected copyright. Secondary Subject Heading: Paediatrics, Pharmacology and therapeutics

Child & adolescent psychiatry < PSYCHIATRY, Depression & mood Keywords: disorders < PSYCHIATRY, CLINICAL PHARMACOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 First- and second-generation antidepressants for depressive disorders in children, adolescents, 6 7 8 and young adults: Study Protocol for a Systematic Review and Network Meta-Analysis 9

10 1 1 2 3 1 4 5 11 Xinyu Zhou , Bin Qin , Craig Whittington , David Cohen , Yiyun Liu , Cinzia Del Giovane , Kurt D. Michael , Yuqing 12 13 Zhang1, and Peng Xie1 14 15 For peer review only 16 1Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China 17 18 19 2 Research Department of Clinical, Educational and Health Psychology, University College London, London, UK 20 21 22 3 Department of Child and Adolescent Psychiatry, AP–HP, Hôpital Pitié–Salpétrière, Institut des Systèmes Intelligents et 23 24 25 de Robotiques (ISIR), Centre National pour la Recherche Scientifique, Université Pierre et Marie Curie, Paris, France 26 27 28 4Department of Diagnostic, Clinical, and Public Health Medicine, University of Modena and Reggio, Emilia, Modena, 29 30 31 Italy 32 33 5

34 Department of Psychology, Appalachian State University, Boone, North Carolina, USA; http://bmjopen.bmj.com/ 35 36 37 Xinyu Zhou and Bin Qin contributed equally to the protocol. 38 39 40 Direct correspondence to: 41

42 on September 23, 2021 by guest. Protected copyright. Peng Xie, MD, Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi 43 44 45 Road, Yuzhong District, Chongqing 400016, China; Tel: +8602368485490; Fax: +8602368485111; 46 47 48 Email: [email protected] 49 50 51 Keywords: depression, child, adolescent, antidepressant, network metaanalysis 52 53 54 Word count: 2402. 55 56 57 58 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 ABSTRACT 6 7 8 Introduction: Depressive disorders are the most frequent psychiatric disorders in children, adolescents and young 9 10 11 adults, and have adverse effects on their psychosocial functioning. An increasing number of studies are aimed at its 12 13 antidepressants treatment. Questions concerning the efficacy and safety of antidepressants in the treatment of depression 14 15 For peer review only 16 in children and adolescents led us to integrate the available evidence by network metaanalysis to create hierarchies of 17 18 19 these drugs. 20 21 22 Methods and analysis: The databases with PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, 23 24 25 LiLACS, and PsycINFO will be searched from 1966 to January 5, 2015. There are no restrictions on language or type of 26 27 28 publication. Randomized clinical trials assessing first and secondgeneration antidepressants against another or placebo 29 30 31 as acute treatment for depressive disorder in patients with children, adolescent, and young adults (under 25 years of age) 32 33

34 will be included. The primary outcome for efficacy will be change scores in depressive symptoms, as measured by the http://bmjopen.bmj.com/ 35 36 37 mean change score of depression rating scales (self or assessorrated) from baseline to endpoint. The tolerability of 38 39 40 treatment will be defined as sideeffect discontinuation, as defined by the proportion of patients who discontinued 41

42 on September 23, 2021 by guest. Protected copyright. treatment due to adverse events during the study. We will also assess the secondary outcome for efficacy (response in 43 44 45 depressive symptoms), acceptability (allcause discontinuation), and suiciderelated outcomes. We will perform the 46 47 48 Bayesian network metaanalyses for all relative outcome measures. Subgroup analyses and sensitivity analyses will be 49 50 51 conducted to assess the robustness of the findings. 52 53 54 55 56 57 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Dissemination: The network metaanalysis will provide useful information on antidepressant treatment for child, 6 7 8 adolescent and young adult depression. The results will be disseminated through peerreviewed publication or 9 10 11 conference presentations. 12 13 Protocol registration: PROSPERO CRD42015016023. 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Strengths and limitations of this study 6 7 8 1. This Bayesian network metaanalysis can integrate direct evidence with indirect evidence from multiple treatment 9 10 11 comparisons to estimate the interrelations across all treatments. 12 13 2. We will comprehensively assess the efficacy, tolerability, acceptability, and suiciderelated outcomes of first and 14 15 For peer review only 16 secondgeneration antidepressants for depression in children, adolescents and young adults. 17 18 19 3. Several subgroup and sensitivity analyses will address some clinically relevant questions. 20 21 22 4. This method comprehensively synthesizes data to provide a clinically useful summary that can guide treatment 23 24 25 decisions and guideline development. 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 BACKGROUND 6 7 8 The depressive disorder in children and adolescents is a major public health problem, which showed as the fourth most 9

10 1 11 important disease in the estimation of disease burden. The prevalence of experiencing at least one episode of major 12 13 depression before adulthood is estimated to be approximately 1% to 2% for children (6–12 years old), 2% to 5% for 14 15 For peer review only 16 adolescents (13–18 years old), and 14% to 25% in young adults (19–25 years old).2 As with adults, the course of 17 18 19 depressive disorders in children and adolescents is often characterized by frequent recurrence, protracted episodes, 20 21 22 comorbidity with psychiatric disorders.3 The consequences of an untreated major depression in young people are likely 23 24 25 to be serious impairment in social functioning, e.g. poor school achievement; relational problems with family members 26 27 28 and peers.4 A report from the American Academy of Child and Adolescent Psychiatry (AACAP) indeed that depression 29 30 31 is responsible for over 500,000 suicide attempts by children and adolescents a year, and most of them diagnosed with 32 33 5,6

34 treatable forms of mental illness. Thus, early recognition, diagnosis, and treatment of depression in children and http://bmjopen.bmj.com/ 35 36 37 adolescents is an important strategy for curbing the rising rate of youth suicide seen in many developed and advanced 38

39 7 40 developing nations. 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 Since the late 1960 years, firstgeneration antidepressants, e.g., tricyclic antidepressant drugs (TCAs), have been first 46 47 48 used to treat depressive symptoms in young patients.8 In the U.S., the use of antidepressants in children and adolescents 49 50 51 grew 3 to 10fold between 1987 and 1996.9 Recently, the efficacy of TCAs has been investigated in at least 13 52 53 54 randomized placebocontrolled studies,10 which showed marginal evidence to support the use of TCAs in the treatment 55 56 57 of depression in only adolescents. However, methodological deficiencies in these studies, including small sample sizes 58 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 and diagnostic heterogeneity, as well as none of trials as being at low risk of bias, restrict statistical inference and 6 7 8 generalizability of the findings. At the same time, cardiovascular effects and overdoserelated mortality associated with 9

10 11,12 11 TCA use have greatly limited their use in clinical practice. Nevertheless, nortriptyline is still approved by the Food 12 13 and Drug Administration (FDA) for the treatment of depression in adolescents and adults.13 14 15 For peer review only 16 17 18 19 In recent decades, newgeneration antidepressants, e.g., selective serotonin reuptake inhibitors (SSRIs), have been 20 21 22 wildly used for the treatment of depression in children and adolescents.14The frequency of prescription of SSRIs in 23 24 25 children and adolescents has progressively increased.15 In European countries, there has been a doubling of SSRI use 26 27 28 over the 4year period.16 However, only fluoxetine, a SSRIs antidepressant, was approved by the U.S FDA for treating 29 30 31 depression in children and adolescents in January 2003.17 At the same year, concerns about the increased risk of suicide 32 33 18

34 and suicide attempts with SSRIs were first raised. In September 2004, the Food and Drug Administration cautioned http://bmjopen.bmj.com/ 35

36 19 37 practitioners in the use of antidepressant medications in children and adolescents. Similar warnings were issued by 38

39 2022 40 other health regulatory agencies. Thus, concerns about this issue have refocused attention on the question of how 41

42 on September 23, 2021 by guest. Protected copyright. effective antidepressants are with youth depression, the most prominent medication alternative. 43 44 45 46 47 48 The relative effect sizes between different antidepressant medications in these previous reviews were derived from 49 50 51 indirect comparisons between separate placebocontrolled trials,14,17,23,24 while very few from direct comparisons 52 53 54 between activecontrolled trials.25 Moreover, the problem of lumping together different antidepressant drugs would limit 55 56 57 the available effect size of each antidepressant. For these reasons, we employed a network metaanalysis –a 58 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 methodological approach that allows the simultaneous comparison of multiple psychotherapeutic interventions within a 6

7 26 8 single analysis, while preserving randomization. This approach is applied to integrate direct evidence (from studies 9 10 11 directly comparing interventions) with indirect evidence (information about two treatments derived via a common 12 13 comparator) from multiple treatment comparisons to estimate the interrelations across all treatments.27 We have 14 15 For peer review only 16 previously compared the efficacy and acceptability of psychotherapies for depression in children and adolescents28 and 17 18 19 the augmentation agents for treatmentresistant depression in adults29 in this way. The aim of the network metaanalysis 20 21 22 is of randomized controlled studies is to re analyze systematically the efficacy, tolerability, acceptability and suicide 23 24 25 risk of both first and secondgeneration antidepressants, either against another antidepressant or against a control 26 27 28 condition, in the treatment of child and adolescent depression. 29 30 31 32 METHODS 33

34 http://bmjopen.bmj.com/ 35 Criteria for included studies 36 37 38 Types of studies 39 40 41 Any prospective randomized controlled trials (RCTs) will be included. When trial with crossover design, the

42 on September 23, 2021 by guest. Protected copyright. 43 44 results will be only included from the first randomization period. However, quasirandomized trials (e.g., those 45 46 47 allocating using alternate days of the week) will be excluded. Trials with sample sizes smaller than 10 will be excluded 48 49 50 in the review. 51 52 53 Types of participants 54 55 Children, adolescents, and young adults (aged from 6 to 25 when they initially enrolled in the studies) with a primary 56 57 58 diagnosis of current major depressive disorder according to standardized diagnostic interviews, e.g., the Diagnostic and 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 3032 33,34 5 Statistical Manual of Mental Disorders (DSM) or the International Classification of Diseases (ICD) will be 6 7 8 included. Where patients with adults and youths, the data will be included if data on the depressed youths can be 9 10 11 extracted separately, or obtained from trial authors. We will not excluded trials in which patients with a secondary 12 13 diagnosis of comorbid medical or general psychiatric disorders. However, we will exclude studies in which those have a 14 15 For peer review only 16 diagnosis of resistant depression or psychotic depression. 17 18 19 Types of interventions 20 21 22 RCTs comparing any first and second generation antidepressant against another or either placebo for treatment of 23 24 25 depression in children and adolescents will be included. Trials comparing the same type of antidepressant but at 26 27 28 different therapeutic dose (fixed or flexible dose) and different treatment duration will be considered as the same node 29 30 31 in the network analysis. We will exclude studies involving combination therapy (i.e., combination of antidepressants, 32 33

34 combination of antidepressants with psychotherapy, or other nonpsychotherapeutic interventions); however, studies http://bmjopen.bmj.com/ 35 36 37 will be considered as eligible if the concomitant psychotherapy is not predefined in the study. 38 39 40 Types of outcome measures 41

42 on September 23, 2021 by guest. Protected copyright. The acute phase will be defined as from 4 to 16 week, and if a study presents data for more than one time period within 43 44 45 our predefined acute phase periods, we will take the 8week time point. Where depression symptoms are measured 46 47 48 using more than one depression scale in a trial, we will extract data from the depressive scales on the basis of a 49 50 51 hierarchy of rating scales. This hierarchy will be based on psychometric properties and appropriateness for use with 52 53 54 children and adolescents and for consistency of use across trials (the most commonly used tool)14 (see Table 1). The 55 56 57 Children’s Depression Rating Scale (CDRSR)35 is adapted for children and adolescents from the Hamilton Depression 58 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 36 37 5 Rating Scale (HAMD) , a tool validated and commonly used in adult populations. Both the CDRSR and HAMD 6

7 37 38 39 8 have good reliability and validity. The Beck Depression Inventory (BDI) /Children’s Depression Inventory (CDI) 9 10 11 are the most commonly used among depression symptom severityself rated scales and are ranked the second highest in 12 13 the hierarchy. 14 15 For peer review only 16 1. Overall efficacy 17 18 19 1.1 The primary outcome for efficacy will be change scores in depressive symptoms, as measured by the mean change 20 21 22 score of depression rating scales (self or assessorrated) from baseline to endpoint. 23 24 25 1.2 The secondary outcome for efficacy will be response (or remission) in depressive symptoms, as estimated by the 26 27 28 proportion of patients who achieved a decrease of a certain percentage or below the threshold in depression rating 29 30 31 score.40 32 33

34 2. Overall tolerability http://bmjopen.bmj.com/ 35 36 37 The tolerability of treatment will be defined as sideeffect discontinuation in this review, as defined by the proportion of 38 39 40 patients who discontinued treatment due to adverse events during the study. 41

42 on September 23, 2021 by guest. Protected copyright. 3. Overall acceptability 43 44 45 The acceptability of treatment will be defined as allcause discontinuation, as measured by the proportion of patients 46 47 48 who discontinued treatment (during the delivery of the intervention) up to the post intervention time point. 49 50 51 4. Suiciderelated outcomes 52 53 54 We will also assess the suiciderelated outcome, as estimated by the proportion of patients with one or more events of 55 56 57 definite suicidal behavior or ideation during the acute treatment phase.41 58 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Data Sources and Search Strategy 6 7 8 Seven electronic databases (PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, LiLACS, and 9 10 11 PsycINFO) will be searched from 1966 to January 5, 2015 with Medical Subject Headings (MeSH) and text 12 13 words： ”depress*” or “dysthymi*” or “mood disorder*” or “affective disorder*” and “selective serotonin reuptake 14 15 For peer review only 16 inhibitor*” or “SSRIs” or “serotonin norepinephrine reuptake inhibitor*” or “SNRIs” or “citalopram” or “fluoxetine” or 17 18 19 “paroxetine” or “sertraline” or “escitalopram[” or “fluvoxamine” or “venlafaxine” or “duloxetine” or “milnacipran” or 20 21 22 “reboxetine” or “bupropion” or “mirtazapine” or “tricyclic” or “amersergide” or “amineptine” or “amitriptyline” or 23 24 25 “amoxapine” or “butriptyline” or “chlorpoxiten” or “clomipramine” or “clorimipramine” or “demexiptiline” or 26 27 28 “desipramine” or “dibenzipin” or “dothiepin” or “doxepin” or “imipramine” or “lofepramine” or “melitracen” or 29 30 31 “metapramine” or “nortriptyline” or “noxiptiline” or “opipramol” or “protriptyline” or “quinupramine” or “tianeptine” 32 33

34 or “trimipramine” and “adolesc*” or “child*” or “boy*” or “girl*” or “juvenil*” or “minors” or “paediatri*” or http://bmjopen.bmj.com/ 35 36 37 “pediatri*” or “pubescen*” or “school*” or “student*” or “teen*” or “young” or “youth*”. Also, ClinicalTrials.gov, 38 39 40 World Health Organization’s trial portal and U.S. Food and Drug Administration (FDA) reports will be reviewed. There 41

42 on September 23, 2021 by guest. Protected copyright. are no restrictions on language or type of publication. Additional studies will be searched in the reference lists of all 43 44 45 identified publications including relevant metaanalyses and systematic reviews. All relevant authors and principal 46 47 48 manufacturers will be contacted to supplement incomplete reports of the original papers or to provide new data for 49 50 51 unpublished studies. 52 53 54 Study Selection 55 56 57 58 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Two reviewers (QB and LYY) will be independently scan citations at the title/abstract level identified from the search 6 7 8 strategies and then obtain potentially relevant studies in full text and determine whether to include them by the same 9 10 11 eligibility criteria. Besides, the references of relevant reviews and included trials will also be checked by the two 12 13 reviewers. The reasons for exclusion of trials will be reported in the characteristics of excluded studies tables. Any 14 15 For peer review only 16 disagreements will be resolved by a third review author (XYZ). 17 18 19 Data Extraction and Risk of bias assessment 20 21 22 Two independent reviewers (YYL, BQ) will independently extract the key study parameters using a standardized data 23 24 25 abstraction form and assess the risk of bias. The standardized data extraction forms will include the study characteristics 26 27 28 (e.g., first listed author, publication year, journal, country, institution, and sponsor), patient characteristics (e.g. 29 30 31 diagnostic criteria for depression, the type of patients, the number of patients, and patients’ baseline), intervention 32 33

34 details (e.g. antidepressant type, dose of antidepressant, the duration of treatment, treatment duration) and outcome http://bmjopen.bmj.com/ 35 36 37 measures (efficacy, tolerability, acceptability, and suiciderelated outcome). The risk of bias in trials will be used by the 38

39 42 40 risk of bias tool from the Cochrane Handbook. Trials attracting a rating of high risk of bias in one or more domains 41

42 on September 23, 2021 by guest. Protected copyright. will be considered as ‘high risk’ literature, low risk of bias in all domains as ‘low risk’ literature, and one or more 43 44 45 unclear risk of bias in each domain as ‘unclear risk’ literature.42 Any disagreements will be resolved by a third review 46 47 48 author (XYZ). Also, we will calculate the interrater reliability of the two raters. 49 50 51 Data Synthesis and Analysis 52 53 54 We will perform Bayesian network metaanalysis to compare the relative outcomes of different antidepressants and 55 56 57 placebo with each other from the median of the posterior distribution.26,27 The pooled estimates of standardised mean 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 difference (SMD) with 95% credible intervals (CrIs) will be calculated for the continuous outcomes; and odds ratios 6 7 8 (or) with 95% CrIs will be calculated for the categorical outcomes. The SMD is that the difference in means (MD) of 9 10 11 change scores between the two groups divided by the pooled standard deviation (SD) of the measurements, with a 12 13 negative SMD value indicates greater symptomatic relief. In the presence of minimally informative priors, CrIs can be 14 15 For peer review only 16 interpreted similarly to confidence intervals, and at conventional levels of statistical significance a twosided p <0.05 17 18 19 can be assumed if 95% CrIs do not include 0.42 If means and standard deviations (SDs) are not provided, we will 20 21 22 calculate them from the pvalue or other statistical indices as described elsewhere.42,43 Results from intentiontotreat 23 24 25 analysis (ITT) or modified ITT will be preferred over results from completer analyses, while we will also consider the 26 27 28 dataset for the means and SDs that are presented in the literature. 29 30 31 The pooled estimates will be obtained using the Markov Chains Monte Carlo method. Two Markov chains will be run 32 33

34 simultaneously with different arbitrarily chosen initial values. To ensure convergence, trace plots and the Brooks http://bmjopen.bmj.com/ 35

36 44 37 GelmanRubin statistic will be assessed. Convergence will be found to be adequate after running 50,000 samples for 38 39 40 both chains. These samples will be then discarded as “burnin”, and posterior summaries will be based on 100,000 41

42 on September 23, 2021 by guest. Protected copyright. subsequent simulations. The node splitting method will be used to calculate the inconsistency of the model, which 43 44 45 separate evidence on a particular comparison into direct and indirect evidence.45 Probability values will be summarized 46 47 48 and reported as surface under the cumulative ranking curve (SUCRA) and rankograms, a simple transformation of the 49 50 51 mean rank used to provide a hierarchy of the treatments and accounts for both the location and the variance of all 52 53 54 relative treatment effects.46 Network metaanalysis will be performed using the WinBUGS software package (version 55 56 57 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4

5 1.4.3, MRC Biostatistics Unit, Cambridge, UK) with random effects models for multiarm trials. The other analyses will 6 7 8 be performed and presented by the Stata 11.0 and R 2.11.1 software packages. 9 10 11 Subgroup analyses 12 13 The antidepressants will be coded according to clinical characteristics, risk of bias, and sample size. We will conduct 14 15 For peer review only 16 the subgroups analyses of data in primary outcome for efficacy. We will performed the following subgroups analyses by 17 18 19 using the metaregression model and calculated Somer’s D (a correlation coefficient for a dichotomous and an ordinal 20 21 22 variable)47: (i) sex ratio (maletofemale ratio> 1 vs. maletofemale ratio<1); (ii) age group (mean age of less than 13 23 24 25 vs. mean age of more than or equal 13); (iii) treatment duration (<8 weeks vs. ≥8 weeks); (iv) severity of depressive 26 27 28 symptom (mildtomoderate vs. moderatetosevere); (v) comorbid general psychiatric disorders ( with vs. without 29 30 31 comorbid); (vi) risk of bias (‘high risk’ literature vs. ‘unclear and low risk’ literature); (viii) sample size (≤100 patients 32 33

34 vs. >100 patients); (ix) company sponsor (with vs. without sponsor); and (x) the type of trials (published literature vs. http://bmjopen.bmj.com/ 35 36 37 unpublished literature). When the limitation by small number of comparisons for some potential modifiers in carrying 38 39 40 out subgroup analyses on these variables, we will perform the sensitivity analyses by omitting specific studies from the 41

42 on September 23, 2021 by guest. Protected copyright. overall analysis. 43 44 45 Other analyses 46 47 48 We will conduct an additional analysis in Bayesian network model based on including only secondgeneration 49 50 51 antidepressants in primary efficacy and tolerability outcomes. The funnel plot analyses will be performed to check for 52 53 54 publication bias. Moreover, we will carry out metaregression analyses to investigate the effect of sponsorship or year 55 56 57 published on outcome estimate. 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Ethics and dissemination 6 7 8 This network metaanalysis does not need ethical approval, as data used here are not individual or private. It will be 9 10 11 published in a peerreviewed journal. The results will provide a general overview and evidence of efficacy and safety of 12 13 antidepressants for depression in children, adolescents and young adults. They will also have implications for clinical 14 15 For peer review only 16 practice and further research. 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Contributors XZ and BQ conceived the study and drafted the protocol. XZ and PX wrote the first draft of the 6 7 8 manuscript. KDM, CW and DC assisted in protocol design and revision. YL and YZ participated in the search strategy 9 10 11 development. CDG and BQ participated in the design of data synthesis and analysis. All authors have approved the 12 13 publication of the protocol 14 15 For peer review only 16 Funding This work is supported by the National Basic Research Program of China (973 Program) (Grant No. 17 18 19 2009CB918300). The funders had no role in the protocol design; the writing of the protocol; or the decision to submit 20 21 22 the protocol for publication. 23 24 25 Competing interests None. 26 27 28 Provenance and peer review Not commissioned; externally peerreviewed. 29 30 31 Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution Non 32 33

34 Commercial (CC BYNC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non http://bmjopen.bmj.com/ 35 36 37 commercially, and license their derivative works on different terms, provided the original work is properly cited and the 38 39 40 use is noncommercial. See: See: http://creativecommons.org/licenses/bync/4.0/ 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 TABLES 6 7 Table 1 Hierarchy of depression symptom severity measurement scales 8 Hierarchy Depression symptom severity measurement scales Abbreviation 9 10 1 Children’s Depression Rating Scale CDRS 11 2 Hamilton Depression Rating Scale HAMD 12 13 3 Montgomery Asberg Depression Rating Scale MADRS 14 4 Beck Depression Inventory BDI 15 For peer review only 16 5 Children’s Depression Inventory CDI 17 6 Schedule for Affective Disorders and Schizophrenia for School KSADS 18 19 Aged Children 20 7 Mood and Feeling Questionnaire MFQ 21 8 Reynolds Adolescent Depression Scale RADS 22 23 9 Bellevue Index of Depression BID 24 10 Child Depression Scale CDS 25 26 11 Centre for Epidemiologic Studies Depression Scale CESD 27 12 Child Assessment Schedule CAS 28 29 13 Child Behavior ChecklistDepression CBCLD 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 References 6 7 8 1. Merikangas KR, Nakamura EF, Kessler RC. Epidemiology of mental disorders in children and adolescents. 9 10 11 Dialogues Clin Neurosci 2009;11:720. 12 13 2. Ryan ND. Treatment of depression in children and adolescents. Lancet 2005;366:93340. 14 15 For peer review only 16 3. Rao U, Ryan ND, Birmaher B, et al. Unipolar depression in adolescents: clinical outcome in adulthood. J Am Acad 17 18 19 Child Adolesc Psychiatry 1995;34:56678. 20 21 22 4. Jaffee SR, Moffitt TE, Caspi A, et al. Differences in early childhood risk factors for juvenileonset and adultonset 23 24 25 depression. Arch Gen Psychiatry 2002;59:21522. 26 27 28 5. Office of the Surgeon General. The Surgeon General's call to action to prevent suicide. Washington, DC: Department 29 30 31 of Health and Human Services, 1999. 32 33

34 6. Centers for Disease Control and Prevention (CDC). Suicide trends among youths and young adults aged 1024 years http://bmjopen.bmj.com/ 35 36 37 United States, 19902004. MMWR Morb Mortal Wkly Rep 2007,56:905908. 38 39 40 7. Pfeffer CR. Suicide among youth: Perspectives on risk and prevention. Washington, DC: American Psychiatric Press, 41

42 on September 23, 2021 by guest. Protected copyright. 43 1989;20325. 44 45 46 8. Lieberman JAI. History of the use of antidepressants in primary care. J Clin Psychiatry 2003;5:610. 47 48 9. Zito JM, Safer DJ, DosReis S, et al. Psychotropic practice patterns for youth: a 10year perspective. Arch Pediatr 49 50 51 Adolesc Med 2003;157:1725. 52 53 54 10. Hazell P, Mirzaie M. Tricyclic drugs for depression in children and adolescents. Cochrane Database Syst Rev 55 56 57 2013;6:CD002317. 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 11. Biederman J. Sudden death in children treated with a tricyclic antidepressant. J Am Acad Child Adolesc Psychiatry 6 7 8 1991;30:4958. 9 10 11 12. Varley CK, McClellan J. Case study: two additional sudden deaths with tricyclic antidepressants. J Am Acad Child 12 13 Adolesc Psychiatry 1997;36:3904. 14 15 For peer review only 16 13. Green WH. Child and Adolescent clinical psychopharmacology. 3 rd ed. Philadelphia: Lippincott Williams and 17 18 19 Wilkins, 2003:16972. 20 21 22 14. Hetrick SE, McKenzie JE, Cox GR, et al. Newer generation antidepressants for depressive disorders in children and 23 24 25 adolescents. Cochrane Database Syst Rev 2012;11:CD004851. 26 27 28 15. Bennett K, Teeling M, Feely J. Overprescribing antidepressants to children: pharmacoepidemiological study in 29 30 31 primary care. BMJ 2005;331:14512. 32 33

34 16. Fegert JM, Kölch M, Zito JM, et al. Antidepressant use in children and adolescents in Germany. J Child Adolesc http://bmjopen.bmj.com/ 35 36 37 Psychopharmacol 2006;16:197206. 38 39 40 17. Gentile S. Antidepressant use in children and adolescents diagnosed with major depressive disorder: what can we 41

42 on September 23, 2021 by guest. Protected copyright. 43 learn from published data? Rev Recent Clin Trials 2010;5;6375. 44 45 46 18. Healy D. Lines of evidence on the risks of suicide with selective serotonin reuptake inhibitors. Psychother 47 48 Psychosom 2003;72:719. 49 50 51 19. U.S. Food and Drug Administration (FDA). Suicidality in children and adolescents being treated with antidepressant 52 53 54 medications. 55 56 57 58 59 60 18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 http://www.fda.gov/drugs/drugsafety/postmarketdrugsafetyinformationforpatientsandproviders/ucm161679.htm 6 7 8 (accessed 11 Dec 2014). 9 10 11 20. Adverse Drug Reactions Advisory Committee. Use of SSRI antidepressants in children and adolescents. 12 13 https://www.tga.gov.au/usessriantidepressantschildrenandadolescentsjune2004 (accessed 1 Dec 2014). 14 15 For peer review only 16 21. Canadian Psychiatric Association Bulletin. Health Canada to examine SSRIs in children, April 2004. 17 18 19 http://ww1.cpaapc.org/Publications/Archives/Bulletin/2004/april/news32En.asp (accessed 18 Dec 2014). 20 21 22 22. National Institute of Mental Health. Antidepressant medications for children and adolescents: information for 23 24 25 parents and caregivers. http://www.nimh.nih.gov/health/topics/childandadolescentmentalhealth/antidepressant 26 27 28 medicationsforchildrenandadolescentsinformationforparentsandcaregivers.shtml (accessed 14 Dec 2014). 29 30 31 23. Papanikolaou K, Richardson C, Pehlivanidis A, et al. Efficacy of antidepressants in child and adolescent depression: 32 33

34 a metaanalytic study. J Neural Transm 2006;113:399415. http://bmjopen.bmj.com/ 35 36 37 24. Whittington CJ, Kendall T, Fonagy P, et al. Selective serotonin reuptake inhibitors in childhood depression: 38 39 40 systematic review of published versus unpublished data. Lancet 2004;363:13415. 41

42 on September 23, 2021 by guest. Protected copyright. 25. Qin B, Zhang Y, Zhou X, et al. Selective serotonin reuptake inhibitors versus tricyclic antidepressants in young 43 44 45 patients: a metaanalysis of efficacy and acceptability. Clin Ther 2014;36:108795. 46 47 48 26. Salanti G, Higgins JP, Ades AE, et al. Evaluation of networks of randomized trials. Stat Methods Med Res 49 50 51 2008;17:279301. 52 53 54 55 27. Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 56 57 58 2004;23:310524. 59 60 19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 21 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 28. Zhou X, Ravindran A, Qin B, et al. Comparative Efficacy, Acceptability and Tolerability of Augmentation Agents in 6 7 8 TreatmentResistant Depression: Systematic Review and Network MetaAnalysis. J Clin Psychiatry 2014; In press. 9 10 11 29. Qin B, Zhou X, Michael KD, et al. Psychotherapy for Depression in Children and Adolescents: Study Protocol for a 12 13 14 Systematic Review and Network MetaAnalysis. BMJ Open 2014; In press. 15 For peer review only 16 17 30. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSMIII). 3rd edition. 18 19 20 Washington, DC: American Psychiatric Association, 1980. 21 22 23 31. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSMIIIR). 3rd revised 24 25 26 edition. Washington, DC: American Psychiatric Association, 1987. 27 28 29 32. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th edn. Washington, DC: 30 31 32 American Psychiatric Association, 1994. 33

34 http://bmjopen.bmj.com/ 33. World Health Organization (WHO). The Ninth Revision of the International Classification of Diseases and Related 35 36 37 Health Problems (ICD9). Geneva: World Health Organization, 1978. 38 39 40 34. World Health Organization (WHO). The Tenth Revision of the International Classification of Diseases and Related 41

42 on September 23, 2021 by guest. Protected copyright. 43 Health Problems (ICD10). Geneva: World Health Organization, 1992. 44 45 46 35. Poznanski EO, Mokros HB. Children's depression rating scale, revised (CDRSR): manual. Western Psychological 47 48 49 Services, 1996. 50 51 52 36. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23:5662. 53 54 55 37. Brooks SJ, Kutcher S. Diagnosis and measurement of depression: a review of commonly utilized instruments. J 56 57 58 Child Adolesc Psychopharmacol 2001;11:34176. 59 60 20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 21 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 38. Beck AT, Steer R. Beck depression inventory: manual. San Antonio, Texas: Psychological Corporation, 1987. 6 7 8 39. Kovacs M. The Children's Depression, Inventory (CDI). Psychopharmacol Bull 1985;21:9958. 9 10 11 40. Riedel M, Möller HJ, Obermeier M, et al. Response and remission criteria in major depressiona validation of 12 13 current practice. J Psychiatr Res 2010;44:10638. 14 15 For peer review only 16 41. Hammad TA. Review and evaluation of clinical data. Washington, DC, US Food and Drug Administration. 17 18 19 http://www.fda.gov/ohrms/dockets/ac/04/briefing/20044065b110 tab08hammadsreview.pdf (accessed 9 Dec 2014). 20 21 22 42. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions version 5.1. 0 [updated March 23 24 25 2011]. The Cochrane Collaboration, 2011. 26 27 28 43. Follmann D, Elliott P, Suh I, et al. Variance imputation for overviews of clinical trials with continuous response. J 29 30 31 Clin Epidemiol 1992;45:76973. 32 33

34 http://bmjopen.bmj.com/ 44. Brooks SP, Gelman A. General Methods for Monitoring Convergence of Iterative Simulations. J Comput Graph Stat 35 36 37 1998;7:43455. 38 39 40 45. Lu GB, Ades AE. Assessing evidence inconsistency in mixed treatment comparisons. J Am Stat Assoc 41

42 on September 23, 2021 by guest. Protected copyright. 43 2006;101:44759. 44 45 46 46. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple 47 48 49 treatment metaanalysis: an overview and tutorial. J Clin Epidemiol 2011; 64:16371. 50 51 52 47. Dias S, Sutton AJ, Welton NJ, et al. Evidence synthesis for decision making 3: heterogeneitysubgroups, meta 53 54 55 regression, bias, and biasadjustment. Med Decis Making 2013;33:61840. 56 57 58 59 60 21 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

Comparative efficacy and tolerability of first- and newergeneration antidepressants for depressive disorders in children and adolescents: Study Protocol for a Systematic Review and Network Meta-Analysis

For peer review only Journal: BMJ Open

Manuscript ID: bmjopen-2015-007768.R1

Article Type: Protocol

Date Submitted by the Author: 28-Apr-2015

Primary Subject Mental health Heading: on September 23, 2021 by guest. Protected copyright. Secondary Subject Heading: Paediatrics, Pharmacology and therapeutics

Child & adolescent psychiatry < PSYCHIATRY, Depression & mood Keywords: disorders < PSYCHIATRY, CLINICAL PHARMACOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Comparative efficacy and tolerability of first- and newer-generation antidepressants for 6 7 8 depressive disorders in children and adolescents: Study Protocol for a Systematic Review and 9 10 11 Network Meta-Analysis 12 13 Xinyu Zhou1, Bin Qin1, Craig Whittington2, David Cohen3, Yiyun Liu1, Cinzia Del Giovane4, Kurt D. Michael5, Yuqing 14 15 For peer review only 16 Zhang1, and Peng Xie1 17 18 19 1Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China 20 21 22 2 Research Department of Clinical, Educational and Health Psychology, University College London, London, UK 23 24 25 3 Department of Child and Adolescent Psychiatry, AP–HP, Hôpital Pitié–Salpétrière, Institut des Systèmes Intelligents et 26 27 28 de Robotiques (ISIR), Centre National pour la Recherche Scientifique, Université Pierre et Marie Curie, Paris, France 29 30 31 4Department of Diagnostic, Clinical, and Public Health Medicine, University of Modena and Reggio, Emilia, Modena, 32 33

34 Italy http://bmjopen.bmj.com/ 35

36 5 37 Department of Psychology, Appalachian State University, Boone, North Carolina, USA; 38 39 40 Xinyu Zhou and Bin Qin contributed equally to the protocol. 41

42 on September 23, 2021 by guest. Protected copyright. Direct correspondence to: 43 44 45 Peng Xie, MD, Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi 46 47 48 Road, Yuzhong District, Chongqing 400016, China; Tel: +8602368485490; Fax: +8602368485111; 49 50 51 Email: [email protected] 52 53 54 Keywords: depression, child, adolescent, antidepressant, network metaanalysis 55 56 57 Word count: 2402. 58 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 ABSTRACT 6 7 8 Introduction: Depressive disorders are among the most common psychiatric disorders in children and adolescents, and 9 10 11 have adverse effects on their psychosocial functioning. Questions concerning the efficacy and safety of antidepressants 12 13 in the treatment of depression in children and adolescents led us to integrate the direct and indirect evidence using 14 15 For peer review only 16 network metaanalysis to create hierarchies of these drugs. 17 18 19 Methods and analysis: Seven databases with PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, 20 21 22 LiLACS, and PsycINFO will be searched from 1966 to December 2013 (updated to January 5, 2015). There are no 23 24 25 restrictions on language or type of publication. Randomized clinical trials assessing first and newergeneration 26 27 28 antidepressants against active comparator or placebo as acute treatment for depressive disorder in children and 29 30 31 adolescents (under 18 years of age) will be included. The primary outcome for efficacy will be mean improvement in 32 33

34 depressive symptoms, as measured by the mean change score of a depression rating scale from baseline to post http://bmjopen.bmj.com/ 35 36 37 treatment. The tolerability of treatment will be defined as sideeffect discontinuation, as defined by the proportion of 38 39 40 patients who discontinued treatment due to adverse events during the study. We will also assess the secondary outcome 41

42 on September 23, 2021 by guest. Protected copyright. for efficacy (response rate ), acceptability (allcause discontinuation), and suiciderelated outcomes. We will perform 43 44 45 the Bayesian network metaanalyses for all relative outcome measures. Subgroup analyses and sensitivity analyses will 46 47 48 be conducted to assess the robustness of the findings. 49 50 51 Dissemination: The network metaanalysis will provide useful information on antidepressant treatment for child and 52 53 54 adolescent depression. The results will be disseminated through peerreviewed publication or conference presentations. 55 56 57 Protocol registration: PROSPERO CRD42015016023 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Strengths and limitations of this study 6 7 8 1. This Bayesian network metaanalysis can integrate direct evidence with indirect evidence from multiple treatment 9 10 11 comparisons to estimate the interrelations across all treatments. 12 13 2. We will comprehensively assess the efficacy, tolerability, acceptability, and suiciderelated outcomes of first and 14 15 For peer review only 16 newergeneration antidepressants for depression in children and adolescents. 17 18 19 3. Several subgroup and sensitivity analyses will address some clinically relevant questions. 20 21 22 4. This method comprehensively synthesizes data to provide a clinically useful summary that can guide treatment 23 24 25 decisions and guideline development. 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 BACKGROUND 6 7 8 Depressive disorder in children and adolescents is a major public health problem, demonstrated by the disorders ranking 9

10 1 11 as the third most important in the estimation of disease burden. The prevalence of experiencing at least one episode of 12 13 major depression before adulthood is estimated to be approximately 1% to 2% for children (6–12 years old), and 2% to 14 15 For peer review only 16 5% for adolescents (13–18 years old).2 The course of major depression in young people is often characterized by 17 18 19 frequent recurrence, protracted episodes, and comorbid psychiatric disorders.3 The consequences of an untreated 20 21 22 episode of major depression in young people are likely to be serious impairment in social functioning, e.g. poor school 23 24 25 achievement; relational problems with family members and peers.4 A report from the American Academy of Child and 26 27 28 Adolescent Psychiatry (AACAP) suggested that depression is responsible for over 500,000 suicide attempts by children 29 30 31 and adolescents a year, and most of them diagnosed with treatable forms of mental illness.5,6 Thus, early recognition, 32 33

34 diagnosis, and treatment of depression in children and adolescents is an important strategy for curbing the rising rate of http://bmjopen.bmj.com/ 35

36 7 37 youth suicide seen in many developed and advanced developing nations. 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. Since the late 1960s, firstgeneration antidepressants, e.g., tricyclic antidepressant drugs (TCAs), have been used to treat 43 44 45 depressive symptoms in young patients.8 In the U.S., the use of antidepressants in children and adolescents grew 3 to 46 47 48 10fold between 1987 and 1996.9 The efficacy of TCAs has been investigated in 13 randomized placebocontrolled 49 50 51 studies,10 which showed marginal evidence to support the use of TCAs in the treatment of depression in only 52 53 54 adolescents. However, methodological deficiencies in these studies, including small sample sizes and diagnostic 55 56 57 heterogeneity, restrict statistical inference and generalizability of the findings. At the same time, cardiovascular effects 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 11,12 5 and overdoserelated mortality associated with TCA use have greatly limited their utility in clinical practice. 6 7 8 Nevertheless, nortriptyline is still approved by the Food and Drug Administration (FDA) for the treatment of depression 9

10 13 11 in adolescents and adults. 12 13

14 15 For peer review only 16 In recent decades, newergeneration antidepressants, including second generation antidepressants (e.g., selective 17 18 19 serotonin reuptake inhibitors [SSRIs]) and third generation antidepressants (e.g., serotoninnorepinephrine reuptake 20 21 22 inhibitors [SNRIs]), have been widely used for the treatment of depression in children and adolescents.14 The frequency 23 24 25 of prescription of SSRIs and SNRIs in children and adolescents has progressively increased.15 In European countries, 26 27 28 there has been a doubling of SSRI use over the 4year period.16 However, only fluoxetinewas approved by the U.S FDA 29 30 31 for treating depression in children and adolescents in January 2003.17 In the same year, concerns about the increased 32 33 18

34 risk of suicide and suicide attempts with SSRIs were first raised. In September 2004, the FDA cautioned practitioners http://bmjopen.bmj.com/ 35

36 19 37 in the use of antidepressant medications in children and adolescents. Similar warnings were issued by other health 38

39 2022 40 regulatory agencies. Thus, concerns about this issue have refocused attention on the question of how effective 41

42 on September 23, 2021 by guest. Protected copyright. antidepressants are with youth depression. 43 44 45 Nonetheless, there is currently no published metaanalysis combining direct and indirect evidence for the use of 46 47 48 antidepressants on children and adolescents, and it is an important one to perform, given the conflicting results 49 50 51 regarding the efficacy and tolerability of various antidepressants in this age group, and lack of head to head trials of 52 53 54 such drugs.2325. For these reasons, we will employ a network metaanalysis –a methodological approach that allows the 55 56 57 simultaneous comparison of multiple psychotherapeutic interventions within a single analysis, while preserving 58 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 26 5 randomization. This approach is will be used to integrate direct evidence (from studies directly comparing 6 7 8 interventions) with indirect evidence (information about two treatments derived via a common comparator) from 9

10 27 11 multiple treatment comparisons to estimate the interrelations across all treatments. We have previously compared the 12 13 efficacy and acceptability of psychotherapies for depression in children and adolescents28 and the augmentation agents 14 15 For peer review only 16 for treatmentresistant depression in adults29 in this way. The aim of the network metaanalysis is of randomized 17 18 19 controlled studies is to reanalyze systematically the efficacy, tolerability, acceptability and suicide risk of both first 20 21 22 and newergeneration antidepressants, either against active comparator or control condition, in the treatment of child 23 24 25 and adolescent depression. 26 27 28 29 METHODS 30 31 32 Criteria for included studies 33

34 http://bmjopen.bmj.com/ 35 Types of studies 36 37 38 Any prospective randomized controlled trials (RCTs), including crossover design and clusterrandomized trials 39 40 41 will be included. However, quasirandomized trials (e.g., those allocating using alternate days of the week) will be

42 on September 23, 2021 by guest. Protected copyright. 43 44 excluded. Trials with sample sizes smaller than 10 will be excluded in this review. 45 46 47 Types of participants 48 49 50 Children and adolescents (aged from 6 to 18 when they initially enrolled in the studies) with a primary diagnosis of 51 52 53 current major depressive disorder according to standardized diagnostic interviews, e.g., the Diagnostic and Statistical 54 55 Manual of Mental Disorders (DSM)3032 or the International Classification of Diseases (ICD)33,34 will be included. 56 57 58 Where patients with adults and youths, the data will be included if data on the depressed youths can be extracted 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 separately, or obtained from trial authors. We will exclude studies focusing on child or adolescent bipolar disorder, but 6 7 8 will include trials involving patients with comorbid general psychiatric disorders, such as attentionhyperactivity 9 10 11 disorder (ADHD) and anxiety disorder. However, we will not exclude studies in which those have a diagnosis of 12 13 psychotic depression, and will examine those patients in a subgroup analysis. But we will exclude studies in which 14 15 For peer review only 16 those have a diagnosis of treatmentresistant depression, because they always involving augmentation or combination 17 18 19 therapy, 20 21 22 Types of interventions 23 24 25 RCTs comparing any first and newer generation antidepressant against active comparator or either placebo for 26 27 28 treatment of depression in children and adolescents will be included. Trials comparing the same type of antidepressant 29 30 31 but at different therapeutic dose (fixed or flexible dose) and different treatment duration will be considered as the same 32 33

34 node in the network analysis. We will exclude studies involving combination therapy (i.e., combination of http://bmjopen.bmj.com/ 35 36 37 antidepressants, combination of antidepressants with psychotherapy, or other nonpsychotherapeutic interventions); 38 39 40 however, studies will be considered as eligible if the concomitant psychotherapy is not predefined in the study. 41

42 on September 23, 2021 by guest. Protected copyright. Types of outcome measures 43 44 45 The acute phase will be defined as from 4 to 16 week, and if a study presents data beyond 16 weeks or for more 46 47 48 than one time period within our predefined acute phase periods, we will take the 8week or close to 8week time 49 50 51 point.35 We will exclude trials with treatment duration of less than 4 weeks. . Where depression symptoms are measured 52 53 54 using more than one depression scale in a trial, we will extract data from the depressive scales on the basis of a 55 56 57 hierarchy of rating scales. This hierarchy will be based on psychometric properties and appropriateness for use with 58 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 14 5 children and adolescents and for consistency of use across trials (referred from Hetrick SE study) (see Table 1). The 6

7 36 8 Children’s Depression Rating Scale (CDRSR) is adapted for children and adolescents from the Hamilton Depression 9

10 37 38 11 Rating Scale (HAMD) , a tool validated and commonly used in adult populations. Both the CDRSR and HAMD 12 13 have good reliability and validity.38 The Beck Depression Inventory (BDI)39/Children’s Depression Inventory (CDI)40 14 15 For peer review only 16 are the most commonly used among depression symptom severityself rated scales and are ranked the second highest in 17 18 19 the hierarchy. 20 21 22 1. Overall efficacy 23 24 25 1.1 The primary outcome for efficacy will be mean improvement in depressive symptoms, as measured by the mean 26 27 28 change score of depression rating scales (self or assessorrated) from baseline to endpoint. 29 30 31 1.2 The secondary outcome for efficacy will be response (or remission) in depressive symptoms, as estimated by the 32 33

34 proportion of patients who achieved a decrease of a certain percentage (e.g., a reduction of 50% or more) or below the http://bmjopen.bmj.com/ 35

36 41 37 published threshold in depression rating score (e.g., CDRSR ≤28). When ‘remission’ is not reported, we will use 38 39 40 ‘response’ if available. 41

42 on September 23, 2021 by guest. Protected copyright. 2. Overall tolerability 43 44 45 The tolerability of treatment will be defined as sideeffect discontinuation in this review, as defined by the proportion of 46 47 48 patients who discontinued treatment due to adverse events during the study. 49 50 51 3. Overall acceptability 52 53 54 The acceptability of treatment will be defined as allcause discontinuation, as measured by the proportion of patients 55 56 57 who discontinued treatment (during the delivery of the intervention) up to the post intervention time point. 58 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 4. Suiciderelated outcomes 6 7 8 We will also assess the suiciderelated outcome, as estimated by the proportion of patients with one or more events of 9

10 42 11 definite suicidal behavior or ideation during the acute treatment phase. 12 13 Data Sources and Search Strategy 14 15 For peer review only 16 Seven electronic databases (PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, LiLACS, and 17 18 19 PsycINFO) will be searched from 1966 to December 2013 (update to January, 2015) with Medical Subject Headings 20 21 22 (MeSH) and text words： ”depress*” or “dysthymi*” or “mood disorder*” or “affective disorder*” and “selective 23 24 25 serotonin reuptake inhibitor*” or “SSRIs” or “serotonin norepinephrine reuptake inhibitor*” or “SNRIs” or 26 27 28 “citalopram” or “fluoxetine” or “paroxetine” or “sertraline” or “escitalopram[” or “fluvoxamine” or “venlafaxine” or 29 30 31 “duloxetine” or “milnacipran” or “reboxetine” or “bupropion” or “mirtazapine” or “tricyclic” or “amersergide” or 32 33

34 “amineptine” or “amitriptyline” or “amoxapine” or “butriptyline” or “chlorpoxiten” or “clomipramine” or http://bmjopen.bmj.com/ 35 36 37 “clorimipramine” or “demexiptiline” or “desipramine” or “dibenzipin” or “dothiepin” or “doxepin” or “imipramine” or 38 39 40 “lofepramine” or “melitracen” or “metapramine” or “nortriptyline” or “noxiptiline” or “opipramol” or “protriptyline” or 41

42 on September 23, 2021 by guest. Protected copyright. “quinupramine” or “tianeptine” or “trimipramine” and “adolesc*” or “child*” or “boy*” or “girl*” or “juvenil*” or 43 44 45 “minors” or “paediatri*” or “pediatri*” or “pubescen*” or “school*” or “student*” or “teen*” or “young” or “youth*”. 46 47 48 Also, ClinicalTrials.gov, World Health Organization’s trial portal and U.S. Food and Drug Administration (FDA) 49 50 51 reports will be reviewed. There are no restrictions on language or type of publication. Additional studies will be 52 53 54 searched in the reference lists of all identified publications including relevant metaanalyses and systematic reviews. All 55 56 57 58 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 relevant authors and principal manufacturers will be contacted to supplement incomplete reports of the original papers 6 7 8 or to provide new data for unpublished studies. 9 10 11 Study Selection 12 13 Two reviewers (QB and LYY) will independently scan citations at the title/abstract level identified from the search 14 15 For peer review only 16 strategies and then obtain potentially relevant studies in full text and determine whether to include them by the same 17 18 19 eligibility criteria. Besides, the references of relevant reviews and included trials will also be checked by the two 20 21 22 reviewers. The reasons for exclusion of trials will be reported in the characteristics of excluded studies tables. Any 23 24 25 disagreements will be resolved by a third review author (XYZ). 26 27 28 Data Extraction and Risk of bias assessment 29 30 31 Two independent reviewers (YYL, BQ) will independently extract the key study parameters using a standardized data 32 33

34 abstraction form and assess the risk of bias. The standardized data extraction forms will include the study characteristics http://bmjopen.bmj.com/ 35 36 37 (e.g., first listed author, publication year, journal, country, institution, and sponsor), patient characteristics (e.g. 38 39 40 diagnostic criteria for depression, the type of patients, the number of patients, level of depressive symptoms, 41

42 on September 23, 2021 by guest. Protected copyright. comorbidities, the age of patients, and the gender of patients), intervention details (e.g. antidepressant type, dose of 43 44 45 antidepressant, and the duration of treatment) and outcome measures (efficacy, tolerability, acceptability, and suicide 46 47 48 related outcome). The risk of bias in trials will be assessed by the Cochrane risk of bias tool.42 Trials attracting a rating 49 50 51 of high risk of bias in one or more domains will be considered as ‘high risk’, low risk of bias in all domains as ‘low 52 53 54 risk’, and one or more unclear risk of bias in each domain as ‘unclear risk’.42 Any disagreements will be resolved by a 55 56 57 third review author (XYZ). In addition, we will calculate the interrater reliability of the two raters. 58 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Data Synthesis and Analysis 6 7 8 We will perform Bayesian network metaanalysis to compare the relative outcomes of different antidepressants and 9

10 26,27 11 placebo with each other from the median of the posterior distribution. The pooled estimates of standardized mean 12 13 difference (SMD) with 95% credible intervals (CrIs) will be calculated for the continuous outcomes; and odds ratios 14 15 For peer review only 16 (or) with 95% CrIs will be calculated for the categorical outcomes. The SMD is that the difference in means (MD) of 17 18 19 change scores between the two groups divided by the pooled standard deviation (SD) of the measurements, with a 20 21 22 negative SMD value indicates greater symptomatic relief. In the presence of minimally informative priors, CrIs can be 23 24 25 interpreted similarly to confidence intervals, and at conventional levels of statistical significance a twosided p <0.05 26 27 28 can be assumed if 95% CrIs do not include 0.42 If means and standard deviations (SDs) are not provided, we will 29 30 31 calculate them from the pvalue or other statistical indices as described elsewhere.43,44 Results from intentiontotreat 32 33

34 analysis (ITT) or modified ITT will be preferred over results from completer analyses, while we will also consider the http://bmjopen.bmj.com/ 35 36 37 dataset for the means and SDs that are presented in the literature. 38 39 40 The pooled estimates will be obtained using the Markov Chains Monte Carlo method. Two Markov chains will be run 41

42 on September 23, 2021 by guest. Protected copyright. simultaneously with different arbitrarily chosen initial values. To ensure convergence, trace plots and the Brooks 43 44 45 GelmanRubin statistic will be assessed.45 Convergence will be found to be adequate after running 50,000 samples for 46 47 48 both chains. These samples will be then discarded as “burnin”, and posterior summaries will be based on 100,000 49 50 51 subsequent simulations. The node splitting method will be used to calculate the inconsistency of the model, which 52 53 54 separate evidence on a particular comparison into direct and indirect evidence.46 Probability values will be summarized 55 56 57 and reported as surface under the cumulative ranking curve (SUCRA) and rankograms, a simple transformation of the 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 mean rank used to provide a hierarchy of the treatments and accounts for both the location and the variance of all 6

7 47 8 relative treatment effects. Network metaanalysis will be performed using the WinBUGS software package (version 9

10 11 1.4.3, MRC Biostatistics Unit, Cambridge, UK) with random effects models for multiarm trials. The other analyses will 12 13 be performed and presented by the Stata 11.0 and R 2.11.1 software packages. The quality of evidence for all outcomes 14 15 For peer review only 16 will be judged using the Grading of Recommendations Assessment, Development and Evaluation working group 17 18 19 methodology. The quality of evidence will be assessed across the domains of risk of bias, 20 21 22 consistency, directness, precision and publication bias. Additional domains may be considered where appropriate. 23 24 25 Quality will be adjudicated as high (further research is very unlikely to change our confidence in the estimate of effect), 26 27 28 moderate (further research is likely to have an important impact on ourconfidence in the estimate of effect and may 29 30 31 change the estimate), low (further research is very likely to have an important impact on our confidence in the estimate 32 33

34 of effect and is likely to change the estimate), or very low (very uncertain about the estimate of effect). http://bmjopen.bmj.com/ 35 36 37 Subgroup analyses 38 39 40 The antidepressants will be coded according to clinical characteristics, risk of bias, and sample size. We will conduct 41

42 on September 23, 2021 by guest. Protected copyright. the subgroups analyses of data in primary outcome for efficacy. We will perform the following subgroups analyses by 43 44 45 using the metaregression model and calculate Somer’s D (a correlation coefficient for a dichotomous and an ordinal 46 47 48 variable)48: (i) sex ratio (maletofemale ratio> 1 vs. maletofemale ratio<1); (ii) age group (mean age of less than 13 49 50 51 vs. mean age of more than or equal 13); (iii) treatment duration (<8 weeks vs. ≥8 weeks); (iv) severity of depressive 52 53 54 symptom (mildtomoderate vs. moderatetosevere); (v) comorbid general psychiatric disorders ( with vs. without 55 56 57 comorbid); (vi) risk of bias (‘high risk’ literature vs. ‘unclear and low risk’ literature); (viii) sample size (≤100 patients 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 vs. >100 patients); (ix) company sponsor (with vs. without sponsor); and (x) the type of trials (published literature vs. 6 7 8 unpublished literature). When the limitation by small number of comparisons for some potential modifiers in carrying 9 10 11 out subgroup analyses on these variables, we will perform the sensitivity analyses by omitting specific studies from the 12 13 overall analysis. 14 15 For peer review only 16 Other analyses 17 18 19 We will conduct an additional analysis in Bayesian network model based on including only newergeneration 20 21 22 antidepressants in primary efficacy and tolerability outcomes. The funnel plot analyses will be performed to check for 23 24 25 publication bias. Moreover, we will carry out metaregression analyses to investigate the effect of sponsorship or year 26 27 28 published on outcome estimate. 29 30 31 Ethics and dissemination 32 33

34 This network metaanalysis does not need ethical approval, as data used here are not individual or private. It will be http://bmjopen.bmj.com/ 35 36 37 published in a peerreviewed journal. The results will provide a general overview and evidence of efficacy and safety of 38 39 40 antidepressants for depression in children and adolescents. They will also have implications for clinical practice and 41

42 on September 23, 2021 by guest. Protected copyright. further research. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 References 6 7 8 1. Ferrari AJ, Charlson FJ, Norman RE, et al. Burden of depressive disorders by country, sex, age, and year: findings 9 10 11 from the global burden of disease study 2010. PLoS Med. 2013;10:e1001547. 12 13 2. Ryan ND. Treatment of depression in children and adolescents. Lancet 2005;366:93340. 14 15 For peer review only 16 3. Rao U, Ryan ND, Birmaher B, et al. Unipolar depression in adolescents: clinical outcome in adulthood. J Am Acad 17 18 19 Child Adolesc Psychiatry 1995;34:56678. 20 21 22 4. Jaffee SR, Moffitt TE, Caspi A, et al. Differences in early childhood risk factors for juvenileonset and adultonset 23 24 25 depression. Arch Gen Psychiatry 2002;59:21522. 26 27 28 5. Office of the Surgeon General. The Surgeon General's call to action to prevent suicide. Washington, DC: Department 29 30 31 of Health and Human Services, 1999. 32 33

42 on September 23, 2021 by guest. Protected copyright. 43 1989;20325. 44 45 46 8. Lieberman JAI. History of the use of antidepressants in primary care. J Clin Psychiatry 2003;5:610. 47 48 9. Zito JM, Safer DJ, DosReis S, et al. Psychotropic practice patterns for youth: a 10year perspective. Arch Pediatr 49 50 51 Adolesc Med 2003;157:1725. 52 53 54 10. Hazell P, Mirzaie M. Tricyclic drugs for depression in children and adolescents. Cochrane Database Syst Rev 55 56 57 2013;6:CD002317. 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

42 on September 23, 2021 by guest. Protected copyright. 43 learn from published data? Rev Recent Clin Trials 2010;5;6375. 44 45 46 18. Healy D. Lines of evidence on the risks of suicide with selective serotonin reuptake inhibitors. Psychother 47 48 Psychosom 2003;72:719. 49 50 51 19. U.S. Food and Drug Administration (FDA). Suicidality in children and adolescents being treated with antidepressant 52 53 54 medications. 55 56 57 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

42 on September 23, 2021 by guest. Protected copyright. 25. Qin B, Zhang Y, Zhou X, et al. Selective serotonin reuptake inhibitors versus tricyclic antidepressants in young 43 44 45 patients: a metaanalysis of efficacy and acceptability. Clin Ther 2014;36:108795. 46 47 48 26. Salanti G, Higgins JP, Ades AE, et al. Evaluation of networks of randomized trials. Stat Methods Med Res 49 50 51 2008;17:279301. 52 53 54 55 27. Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 56 57 58 2004;23:310524. 59 60 18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

42 on September 23, 2021 by guest. Protected copyright. 43 Health Problems (ICD10). Geneva: World Health Organization, 1992. 44 45 46 35. Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 12 newgeneration 47 48 49 antidepressants: a multipletreatments metaanalysis. Lancet. 2009;373:74658.36. Poznanski EO, Mokros HB. 50 51 52 Children's depression rating scale, revised (CDRSR): manual. Western Psychological Services, 1996. 53 54 55 37. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23:5662. 56 57 58 59 60 19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 38. Brooks SJ, Kutcher S. Diagnosis and measurement of depression: a review of commonly utilized instruments. J 6 7 8 Child Adolesc Psychopharmacol 2001;11:34176. 9 10 11 39. Beck AT, Steer R. Beck depression inventory: manual. San Antonio, Texas: Psychological Corporation, 1987. 12 13 40. Kovacs M. The Children's Depression, Inventory (CDI). Psychopharmacol Bull 1985;21:9958. 14 15 For peer review only 16 41. Riedel M, Möller HJ, Obermeier M, et al. Response and remission criteria in major depressiona validation of 17 18 19 current practice. J Psychiatr Res 2010;44:10638. 20 21 22 42. Hammad TA. Review and evaluation of clinical data. Washington, DC, US Food and Drug Administration. 23 24 25 http://www.fda.gov/ohrms/dockets/ac/04/briefing/20044065b110 tab08hammadsreview.pdf (accessed 9 Dec 2014). 26 27 28 43. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions version 5.1. 0 [updated March 29 30 31 2011]. The Cochrane Collaboration, 2011. 32 33

34 44. Follmann D, Elliott P, Suh I, et al. Variance imputation for overviews of clinical trials with continuous response. J http://bmjopen.bmj.com/ 35 36 37 Clin Epidemiol 1992;45:76973. 38 39 40 45. Brooks SP, Gelman A. General Methods for Monitoring Convergence of Iterative Simulations. J Comput Graph Stat 41

42 on September 23, 2021 by guest. Protected copyright. 43 1998;7:43455. 44 45 46 46. Lu GB, Ades AE. Assessing evidence inconsistency in mixed treatment comparisons. J Am Stat Assoc 47 48 49 2006;101:44759. 50 51 52 47. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple 53 54 55 treatment metaanalysis: an overview and tutorial. J Clin Epidemiol 2011; 64:16371. 56 57 58 59 60 20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 48. Dias S, Sutton AJ, Welton NJ, et al. Evidence synthesis for decision making 3: heterogeneitysubgroups, meta 6 7 8 regression, bias, and biasadjustment. Med Decis Making 2013;33:61840. 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

Page 22 of 35 Checklist item Checklist (registration number CRD42015016023). (registration number

Psychology, University College London, London, UK UK London, London, College Psychology, University 2 Research Department of Clinical, Educational and and Health Educational Clinical, Department 2 of Research

Chongqing Medical University, Chongqing, China Chongqing, University, Medical Chongqing 1 Department of Neurology, The First Affiliated Hospital Hospital of First Affiliated The 1 Neurology, of Department Author Affiliations Author Affiliations

[email protected] [email protected] *Corresponding author: Peng Xie Peng author: *Corresponding

updated on 28 updated 2015 28 April, on Systematic Reviews (PROSPERO) on 19 January 2015 and was and last was 2015 on 19 January (PROSPERO) Systematic Reviews was registered with the International Prospective Register of Register Prospective the International waswith registered In accordance with the guidelines, our systematic review protocol review protocol systematic our guidelines, In with the accordance

MetaAnalysis MetaAnalysis Study Protocol for a Systematic Review and Network and Network Systematic Review a Protocol for Study antidepressants for depressive disorders in children and adolescents: and adolescents: in children for disorders antidepressants depressive Comparative efficacy and tolerability of first and newergeneration and newergeneration first of and tolerability Comparative efficacy

BMJ Open

2 2 3a 3a 1a 1a 1b 1b http://bmjopen.bmj.com/ Item No No Item on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

For peer review only

Contact Contact Authors: Authors: Registration Registration Update Identification Identification Title: Administrative information Administrative Section and topic topic and Section 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

obotiques obotiques design of data synthesis and analysis. All authors have approved the approved the authors have All analysis. and design synthesis of data the search strategy development. CDG and BQ participated in in the participated BQ development. and CDG the strategy search assisted in protocol design and revision. YL and YZ participated and participated in YZ and revision. YL assisted design in protocol PX wrote the first draft of the manuscript. KDM, CW and and CWDC KDM, manuscript. first of the PX wrotethe draft XZ and BQ conceived the study and drafted the protocol. XZ and XZ the protocol. and study the and XZ conceived drafted BQ and Peng Xie [email protected] and Peng Xie [email protected] [email protected] Yuqing Zhang [email protected] [email protected] Zhang Yuqing [email protected] [email protected] Kurt D. Michael Michael D. Kurt [email protected] [email protected] Cinzia Del Giovane Del Cinzia Giovane [email protected] David Cohen [email protected]hopparis.fr Yiyun Liu Liu Yiyun David Cohen [email protected]hopparis.fr Email: Xinyu Zhou [email protected] Bin Qin Bin Qin Email: [email protected] Zhou Xinyu

5 Department of Psychology, Appalachian State University, Boone, University, Boone, State Appalachian 5 of Psychology, Department

University of Modena and Reggio, Emilia, Modena, Modena, Italy Emilia, and ofModena University Reggio, 4 Department of Diagnostic, Clinical, and Public Health Medicine, Medicine, Health Public and 4 Clinical, of Diagnostic, Department

Pierre et Marie Curie, Paris, France Curie, France Pierre Paris, et Marie (ISIR), Centre National pour la Recherche Scientifique, Université Université Scientifique, Recherche la pour (ISIR),Centre National Pitié–Salpétrière, Institut des Systèmes Intelligents et de et R de Intelligents Systèmes des Pitié–Salpétrière, Institut 3 Department of Child and Adolescent Psychiatry, AP–HP, Hôpital Hôpital AP–HP, Psychiatry, and Adolescent 3 of Child Department

publication of the the publication protocol of [email protected] Craig Whittington [email protected] [email protected] Craig Whittington [email protected] North Carolina, USA North USA Carolina, BMJ Open 3b 3b http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Contributions Contributions Page 23 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from Page 24 of 35 funded this work . . work this funded

Studies will be selected according to the criteria outlined below. criteria below. outlined the according to selected Studiesbe will Eligibility Eligibility criteria condition, in the treatment of child and adolescent depression. adolescent child and depression. condition, the of treatment in antidepressants, either against active comparator or control control or comparator active either against antidepressants, acceptability and suicide risk of both first and newergeneration firstand newergeneration and risk both of suicide acceptability studies is to reanalyze systematically the efficacy, tolerability, tolerability, efficacy, the studies systematically to reanalyze is The aim of the network metaanalysis is of randomized controlled controlled randomized is metaanalysis of The of network aim the and Network MetaAnalysis] and MetaAnalysis] Network children and adolescents: Study Protocol for a Systematic Review Review a Systematic for Protocol adolescents: Study and children newergeneration antidepressants for depressive disorders in disorders in depressive for antidepressants newergeneration [Comparative efficacy and tolerability of first and and first of and tolerability efficacy [Comparative decision to submit the protocol for publication. for publication. the decision protocol to submit no role in the protocol design; the writing of the protocol; or the the or protocol; of the the writing protocol no in the design; role China (973 Program) (Grant No. 2009CB918300). The funders had had The funders 2009CB918300). (Grant No. China Program) (973 This work is funded by the National Basic Research Program Program of Basic Research National funded the is Thisby work The National Basic Research Program of China Program The Basic Research National China (973 Program) (Grant No. 2009CB918300). 2009CB918300). (Grant No. China Program) (973 This work is funded by the National Basic Research Program Program of Basic Research National funded the is Thisby work

BMJ Open

8 8 7 7 6 6 4 4 5c 5c 5a 5a 5b 5b http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only

Eligibility criteria criteria Eligibility Methods Objectives Objectives Rationale Rationale Introduction Role of sponsor or funder orfunder Role of sponsor Sponsor Sponsor Sources Sources Support: Support: Amendments Amendments 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from always involving augmentation or combination therapy, combination therapy, augmentation involving alwaysor have a diagnosis of treatmentresistant depression, because they because depression, haveof treatmentresistant diagnosis a in a subgroup analysis. But we will exclude studies in which those those exclude in which studies will But we analysis. ina subgroup diagnosis of psychotic depression, and will examine those patients those examine will depression, and diagnosis psychotic of However, we will not exclude studies in which those have a have those which in studies exclude However, not will we attentionhyperactivity disorder (ADHD) and anxiety anxiety disorder. and disorder (ADHD) attentionhyperactivity child or adolescent bipolar disorder, but will include trials involving trials involving but include will disorder, child bipolar adolescent or included if data on the depressed youths can be extracted separately, separately, extracted be can depressed youths the included data if on focusing on studies exclude will We authors. or from trial obtained included. Where patients with adults and youths, the data will be data be will and the with youths, adults included. patients Where (DSM) or the International Classification of Diseases be will (ICD) Diseases of Classification or the (DSM) International e.g., the Diagnostic and Statistical Manual of Mental Disorders Disorders of Mental Statistical Manual and e.g., the Diagnostic depressive disorder according to standardized diagnostic interviews, interviews, diagnostic to according standardized depressive disorder enrolled in the studies) with a primary diagnosis of current major major current of diagnosis a primary in with enrolled the studies) Children and adolescents (aged from 6 to 18 when they initially initially when they to 18 from 6 (aged and Children adolescents

Participants Participants

sizes smaller than 10 will be excluded in this review. this review. in excluded 10 sizes be will smaller than alternate days of the week) will be excluded. Trials with sample with sample Trials excluded. be will week) alternate the days of However, quasirandomized trials (e.g., those allocating using allocating those (e.g., trials However, quasirandomized crossover design and clusterrandomized trials will be included. included. be will trials clusterrandomized and crossover design Any prospective randomized controlled trials (RCTs), including including (RCTs), trials controlled randomized Any prospective

Study designs designs Study patients with comorbid general psychiatric disorders, such as as such psychiatric disorders, comorbid patients general with BMJ Open http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Page 25 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from Page 26 of 35 a hierarchy of rating scales. This hierarchy will be based on based on be will scales. This hierarchy a of rating hierarchy trial, we will extract data from the depressive scales on the basis basis of on the scales from the depressive extract trial, data will we symptoms are measured using more than one depression scale in a a scale in more one using depression than measured symptoms are treatment duration of less than 4 weeks. Where depression Where weeks. depression 4 less than treatment of duration or close to 8week time point.35 We will exclude trials with trials exclude will We or time point.35 to close 8week within our predefined acute phase periods, we will take the 8week the 8week will we take acute phase periods, within predefined our The acute phase will be defined as from 4 to 16 week, and if a a and study if 16 fromweek, to as 4 defined be phase will The acute

Outcomes

concomitant psychotherapy is not predefined in the study. study. in the is not psychotherapy predefined concomitant interventions); however, studies will be considered as eligible if the as eligible if considered be studies will interventions); however, antidepressants with psychotherapy, or other nonpsychotherapeutic nonpsychotherapeutic or other antidepressants psychotherapy, with (i.e., combination of antidepressants, combination of combination of antidepressants, (i.e., of combination analysis. We will exclude studies involving combination therapy therapy combination involving studies exclude analysis. will We duration will be considered as the same node in the network network the node in as the same duration considered be will therapeutic dose (fixed or flexible dose) and different treatment treatment and different flexible dose) or therapeutic dose (fixed comparing the same type of antidepressant but at different different but at antidepressant the comparingtype same of depression in children and adolescents will be included. Trials included. adolescents be will and children depression in against active comparator or either placebo for treatment of for treatment or either placebo comparator against active RCTs comparing any first and newer generation antidepressant antidepressant and newer generation any first RCTs comparing

Interventions

psychometric properties and appropriateness for use with children children for with use and appropriateness psychometric properties presents data beyond 16 weeks or for more than one time period time period than weeks one for more presentsor beyond 16 data BMJ Open http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from R and HAMD R and HAMD 3. Overall acceptability acceptability 3. Overall the study. the study. discontinuation in this review, as defined by the proportion of the proportion as by in this discontinuation defined review, The tolerability of treatment will be defined as sideeffect as sideeffect defined be will treatment The of tolerability 2. Overall 2. Overall tolerability not reported, we will use ‘response’ if available. if ‘response’ use notwe will reported, depression rating score (e.g., CDRSR ≤28).41 When ‘remission’ is ‘remission’ When CDRSR ≤28).41 depression (e.g., score rating reduction of 50% or more) or below the published threshold in threshold the published below or more) reductionor of 50% of patients who achieved a decrease of a certain percentage (e.g., a percentage a (e.g., certain a decrease of achieved of who patients remission) in depressive symptoms, as estimated by the proportion the by estimated proportion as symptoms, remission) depressive in 1.2 The secondary outcome for efficacy will be response (or be response efficacy will for 1.2 outcome The secondary endpoint. endpoint. depression rating scales (self or assessorrated) from baseline to from assessorrated) baseline or depression(self scales rating depressive symptoms, as measured by the mean change score of change score the measured mean by as depressive symptoms, 1.1 The primary outcome for efficacy will be mean improvement in improvement mean efficacy be for will 1.1outcome primary The 1. Overall efficacy 1. Overall scales and are ranked the second highest in the hierarchy. in the hierarchy. the scalesare highest second ranked and commonly used among depression symptom severityself rated severityself symptom among depression commonly used (BDI)/Children’s Depression Inventory (CDI) are the the most are (CDI) Inventory (BDI)/Children’s Depression have good reliability and validity. The Beck Depression Inventory Inventory validity. Depression Beck and The have reliability good commonly used in adult populations. Both CDRS Both the populations. adult commonly in used Hamilton Depression Rating Scale (HAMD), a tool validated and a validated tool Scale (HAMD), Rating Hamilton Depression (CDRSR) is adapted for children and adolescents from the from adolescents and children is for (CDRSR) adapted from Hetrick SE study). The Children’s Depression Rating Scale Rating Scale The Children’s Depression from study). Hetrick SE and adolescents and for consistency of use across trials across trials (referred consistency of use for and and adolescents patients who discontinued treatment due to adverse events during during events adverse to due treatment discontinued patients who BMJ Open http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Page 27 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from Page 28 of 35

language or type of publication. publication. of language or type (FDA) reports will be reviewed. There are no restrictions are no on restrictions There reviewed. (FDA) be will reports Organization’s trial portal and U.S. Food and Drug Administration and Drug Food U.S. and portal trial Organization’s in children and adolescents. Also, ClinicalTrials.gov, World Health World Health ClinicalTrials.gov, Also, adolescents. inand children first and newergeneration antidepressants for depressive disorders disorders for depressive antidepressants first and newergeneration with Medical Subject Headings (MeSH) and text words related related to words and text (MeSH) Headings Subject with Medical Library, Web of Science, CINAHL, LiLACS, and PsycINFO) will will PsycINFO) and CINAHL, LiLACS, Web Library, Science, of Seven electronic databases (PubMed, Embase, the Cochrane Cochrane the Embase, Seven (PubMed, electronic databases There will be no restrictions classification of language. classification of language. restrictions There no be will

Language Language

There will be no restrictions by type of setting. setting. of by restrictions type There no be will

Setting Setting

We will also assess the suiciderelated outcome, as estimated by the by the estimated as outcome, the Wesuiciderelated assess also will 4. Suiciderelated 4.outcomes Suiciderelated to the post intervention time point. point. time tointervention post the discontinued treatment (during the delivery of the intervention) up up of the intervention) the delivery discontinued (during treatment discontinuation, as measured by the proportion of patients who who patients of the proportion by as discontinuation, measured The acceptability of treatment will be defined as as allcause defined be will of treatment The acceptability be searched from 1966 to December 2013 (update to January, 2015) to 2015) (update January, 2013 to December from be searched 1966 behavior or ideation during the acute treatment phase. phase. treatment the acute behavior ideation during or proportion of patients with one or more events of definite suicidal suicidal of definite events or more one proportion with patients of BMJ Open 9 9 http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Information Information sources 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from tricyclic drugs) or amersergide or amineptine or amitriptyline or or amitriptyline amineptine or amersergide or or drugs) tricyclic citalopram or fluoxetine or paroxetine or sertraline or sertraline or or fluoxetine paroxetine or citalopram (( (

6. limit 5 to (human and clinical clinical trial) and 6. to limit (human 5 5. 1 and 2 and 3 and 4 4 and 5.and and 3 2 1 or metapramine or nortriptyline or noxiptiline or opipramol or or opipramol or noxiptiline or or nortriptyline metapramine dothiepin or doxepin or dothiepin or doxepin melitracen or lofepramine or imipramine clorimipramine or demexiptiline or desipramine or dibenzipin or or or dibenzipin or desipramine demexiptiline clorimipramine or amoxapine or butriptyline or chlorpoxiten or clomipramine or or clomipramine chlorpoxiten or amoxapine butriptyline or 4. milnacipran or reboxetine or bupropion or or bupropion or milnacipran or reboxetine mirtazapine).ti,ab. escitalopram or fluvoxamine or venlafaxine or duloxetine or duloxetine venlafaxine or or escitalopram or fluvoxamine 3. teen$ or young or youth$).ti,ab. or youth$).ti,ab. teen$ young or 2. (adolesc$ or child$ or boy$ or girl$ or juvenil$ or minors or or or minors juvenil$ boy$ or girl$ or child$ or 2. or (adolesc$ disorder$)).ti,ab. disorder$)).ti,ab. 1. (depress$ or dysthymi$ or (mood disorder$) or (affective (affective or disorder$) (mood 1.or or dysthymi$ (depress$ Draft EMBASE searchOvid interface DraftsearchOvid EMBASE Appendix1 Appendix1 to supplement incomplete reports of the original papers or or to papers the original of reports to incomplete supplement All relevant authors and principal manufacturers will be contacted contacted will be manufacturers principal authors and All relevant systematic reviews. systematic reviews. identified publications including relevant metaanalyses and metaanalyses relevant including identified publications Additional studies will be searched in the reference lists of all all lists the of reference searched in be Additional will studies protriptyline or quinupramine or tianeptine or trimipramine).ti,ab. trimipramine).ti,ab. tianeptine or quinupramine protriptyline or or paediatri$ or pediatri$ or pubescen$ or school$ or student$ or student$ or pubescen$ or school$ or pediatri$ paediatri$ or or provide new data for unpublished studies. studies. unpublished for provide data new BMJ Open

10 10 http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Study records: Study Search strategy Search strategy Page 29 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from Page 30 of 35 the the Endnote 7.0 and 7.0 Endnote to to ndently extract the key study study the key extract ndently imported

comorbidities, the age of patients, and the gender of patients), patients), of the patients, and gender comorbidities, the of age characteristics (e.g. diagnostic criteria for depression, the of type the forcriteria depression, characteristics diagnostic (e.g. We will extract study characteristics (e.g., first listed author, author, first listed (e.g., characteristics study We extract will reliability of the two raters. two raters. reliability the of third review author. In addition, we will calculate the interrater interrater the calculate we will In addition, author. third review suiciderelated outcome). Any disagreements will be resolved by a by a resolved be will disagreements Any suiciderelated outcome). outcome measures (efficacy, tolerability, acceptability, and and acceptability, (efficacy, tolerability, outcome measures characteristics, patient characteristics, intervention details and details intervention characteristics, characteristics, patient standardized data extraction forms will include the study study the include will forms extraction standardized data Two independent reviewers will indepe will Two reviewers independent will be resolved by a third review author. author. a review third will resolved by be two reviewers. The reasons for exclusion of trials will be reported will in reported be forof trials exclusion two reasons The reviewers. disagreements Any tables. studies excluded the of characteristics of relevant reviews and included trials will also be checked by the the by checked be also trials will and of reviews included relevant include them by the same eligibility criteria. Besides, the references references the criteria. Besides, eligibility same the include them by level identified from the search strategies and then and obtain then strategies search level from the identified Two reviewers will independently scan citations at the title/abstract title/abstract at the citations scan independently Two will reviewers criteria. criteria. will screening citations based on the inclusion and exclusion exclusion and the citations inclusion based on will screening duplicates will be removed during the study selection process. We We process. selection the study during duplicates removed be will Literature search results will be Literature results search patients, the number of patients, level of depressive symptoms, symptoms, of depressive level patients, patients, of number the publication year, journal, country, institution, and sponsor), patient sponsor), and institution, country, journal, publication year, parameters using a standardized data abstraction form. The form. abstraction data a parameters standardized using potentially relevant studies in full text and determine whether to whether and full determine text in studies potentially relevant BMJ Open 12 12 11c 11c 11a 11a 11b 11b http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Data items Data items Data collection process collection process Data Selection process process Selection Data management Data management 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from 4. Suiciderelated 4.outcomes Suiciderelated to the post intervention time point. point. time tointervention post the discontinued treatment (during the delivery of the intervention) up up of the intervention) the delivery discontinued (during treatment discontinuation, as measured by the proportion of patients who who patients of the proportion by as discontinuation, measured The acceptability of treatment will be defined as as allcause defined be will of treatment The acceptability 3. Overall acceptability acceptability 3. Overall the study. the study. discontinuation in this review, as defined by the proportion of the proportion as by in this discontinuation defined review, The tolerability of treatment will be defined as sideeffect as sideeffect defined be will treatment The of tolerability 2. Overall 2. Overall tolerability not reported, we will use ‘response’ if available. if ‘response’ use notwe will reported, depression rating score (e.g., CDRSR ≤28). When ‘remission’ is is When ‘remission’ CDRSR ≤28). depression (e.g., score rating reduction of 50% or more) or below the published threshold in threshold the published below or more) reductionor of 50% of patients who achieved a decrease of a certain percentage (e.g., a percentage a (e.g., certain a decrease of achieved of who patients remission) in depressive symptoms, as estimated by the proportion the by estimated proportion as symptoms, remission) depressive in 1.2 The secondary outcome for efficacy will be response (or be response efficacy will for 1.2 outcome The secondary endpoint. endpoint. depression rating scales (self or assessorrated) from baseline to from assessorrated) baseline or depression(self scales rating depressive symptoms, as measured by the mean change score of change score the measured mean by as depressive symptoms, 1.1 The primary outcome for efficacy will be mean improvement in improvement mean efficacy be for will 1.1outcome primary The 1. Overall efficacy 1. Overall tolerability, acceptability, and suiciderelated outcome). outcome). and suiciderelated acceptability, tolerability, and the duration of treatment) and outcome measures (efficacy, (efficacy, measures and outcome treatment) and the of duration intervention details (e.g. antidepressant type, dose of antidepressant, antidepressant, dose of type, antidepressant intervention (e.g. details patients who discontinued treatment due to adverse events during during events adverse to due treatment discontinued patients who BMJ Open 13 13 http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Outcomes and prioritization and Outcomesprioritization Page 31 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from Page 32 of 35 SDs that are presented in the literature. literature. are in the SDs that presented analyses, while we will also consider the dataset for the means and and the means for the dataset consider also analyses,we will while or modified ITT will be preferred over results from completer from completer over results preferred be or ITT will modified CrIs do not include 0. Results from intentiontotreat analysis (ITT) analysis (ITT) intentiontotreat from Results 0. CrIs not include do statistical significance a twosided p <0.05 can be assumed if 95% if assumed 95% can be p a <0.05 twosided statistical significance similarly to confidence intervals, and at conventional levels of levels at conventional and confidence similarly intervals, to negative SMD value indicates greater symptomatic relief. In the relief. symptomatic negative greater indicates value SMD (MD) of change scores between the two groups divided by the the by divided two groups the scores between of change (MD) the categorical outcomes. The SMD is that the difference in means means in the is difference that SMD The the outcomes. categorical outcomes; and odds ratios (or) with 95% CrIs will be calculated for for calculated will CrIs be andratios (or) odds with 95% outcomes; 95% credible intervals (CrIs) will be calculated for the continuous continuous for the calculated be will 95% intervals (CrIs) credible The pooled estimates of standardized mean difference (SMD) with (SMD) difference mean standardized The estimates of pooled relative outcomes of different antidepressants and placebo with each each with placebo and antidepressants of different relative outcomes of distribution. posterior other the the median from We will perform Bayesian network metaanalysis to compare the the to compare metaanalysis network Bayesian We perform will each domain as ‘unclear risk’. risk’. as ‘unclear each domain all domains as ‘low risk’, and one or more unclear risk of bias in in risk bias of unclear more and one risk’, as or all ‘low domains more domains will be considered as ‘high risk’, low risk of bias in in risk bias of low as risk’, ‘high considered more be will domains The risk of bias in trials will be assessed by the Cochrane risk risk of the Cochrane assessed by trials will in The of be risk bias We will also assess the suiciderelated outcome, as estimated by the by the estimated as outcome, the Wesuiciderelated assess also will presence of minimally informative priors, CrIs can be interpreted interpreted be can CrIs informative priors, presence minimally of pooled standard deviation (SD) of the measurements, with a a with measurements, of the (SD) pooled deviation standard bias tool. Trials attracting a rating of high risk of bias in one or bias in one of aattracting high risk rating biasof tool. Trials behavior or ideation during the acute treatment phase. phase. treatment the acute behavior ideation during or proportion of patients with one or more events of definite suicidal suicidal of definite events or more one proportion with patients of BMJ Open 14 14 15a 15a 15b 15b http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Data synthesis Data synthesis Risk of bias in individual studies studies individual ofRisk in bias 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from sponsorship or year published on outcome estimate. estimate. on outcome published sponsorship or year carry out metaregression analyses to investigate the effect of effect of the investigate analyses to carry metaregression out will be performed to check for publication bias. Moreover, we will will we Moreover, bias. for publication to check will performed be We will conduct an additional analysis in Bayesian network model network analysis in Bayesian an additional We conduct will

specific studies from the overall analysis. analysis. overall from the specific studies variables, we will perform the sensitivity analyses by omitting omitting by analyses sensitivity perform the variables, will we When the limitation by small number of comparisons for some for of comparisons small number When by limitation the the type of trials (published literature vs. unpublished literature). literature). unpublished vs. literature trials the of (published type low risk’ literature); (viii) sample size (≤100 patients vs. >100 >100 patients vs. (≤100 sample size low literature); (viii) risk’ comorbid); (vi) risk of bias (‘high risk’ literature vs. ‘unclear and vs. ‘unclear literature risk’ (‘high risk bias (vi) comorbid); of comorbid general psychiatric disorders ( with vs. without ( vs. with comorbid disorders general psychiatric depressive symptom (mildtomoderate vs. moderatetosevere); (v) (v) vs. moderatetosevere); (mildtomoderate depressive symptom (iii) treatment duration (<8 weeks vs. ≥8 weeks); (iv) of severity (iv) weeks); weeks vs. ≥8 (<8 (iii) duration treatment (mean age of less than 13 vs. mean age of more than or equal 13); equal 13); or than of more age mean 13 vs. (mean than age of less (maletofemale ratio> 1 vs. maletofemale ratio<1); (ii) age group group age (ii) ratio<1); vs. 1 ratio> (maletofemale maletofemale coefficient for a dichotomous and an ordinal variable): (i) sex ratio ratio sex (i) variable): and an for coefficientordinal a dichotomous metaregression model and calculate Somer’s D (a correlation Somer’s (a calculate D and metaregression model will perform the following subgroups analyses by using the using analyses by subgroups following willthe perform subgroups analyses of data in primary outcome for efficacy. We efficacy. for outcome analyses in primary of data subgroups characteristics, risk of bias, and sample size. We will conduct the the conduct will We and size. sample bias, characteristics, risk of The antidepressants will be coded according to clinical to clinical coded be will according The antidepressants primary efficacy and tolerability outcomes. The funnel plot analyses analyses plot funnel The outcomes. and primary tolerability efficacy based on including only newergeneration antidepressants in in antidepressants newergeneration basedonly on including potential modifiers in carrying out subgroup analyses these on out analyses carrying in subgroup potential modifiers patients); (ix) company sponsor (with vs. without sponsor); and (x) sponsor); (x) and without vs. (with company sponsor patients); (ix) BMJ Open 15c 15c http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Page 33 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from Page 34 of 35

of of effect). change the estimate), or very low (very uncertain about the estimate estimate the about uncertain low very (very or change estimate), the impact on our confidence in the estimate of effect and is likely and to likely is of effect estimate in the impact confidence our on estimate), low (further research is very likely to have an important an important have to likely very is low research estimate), (further ourconfidence in the estimate of effect and may change the change the may and of effect estimate ourconfidence the in moderate (further research is likely to have an important impact on impact important an have to likely is research moderate (further to change our confidence in the estimate of effect), effect), estimate of in the confidence tochange our Quality will be adjudicated as high (further research is very unlikely unlikely very is research (further as high adjudicated Quality be will domains may be considered where appropriate. appropriate. where domains considered be may consistency, directness, precision and publication bias. Additional bias. Additional and publication precision consistency, directness, will be assessed across the domains of risk of of bias, of domains assessed risk across the will be Evaluation working group methodology. The quality of evidence evidence of quality The group methodology. Evaluation working Grading of Recommendations Assessment, Development and Development Assessment, of Grading Recommendations The quality of evidence for all outcomes will be judged using the the using judged be outcomes for will all of evidence The quality reporting bias will be further explored by funnel plots. plots. funnel by explored further be reporting bias will outcomes is present (outcome reporting bias). The potential for potential The reporting bias). outcomes present is (outcome Administration. We will evaluate whether selective reporting of of reporting selective evaluate whether will We Administration. World Health Organization’s trial portal and U.S. Food and Drug and Food Drug and U.S. portal trial World Organization’s Health recruitment of patients of the study was started. For studies studies For started. study was of the recruitment patients of determine whether the protocol of the RCT was published before before published RCT was the of protocol determine the whether In order to determine whether reporting bias is present, we will will we present, is bias reporting whether In to determine order characteristics and findings of the included of included studies. the findings and characteristics A systematic narrative synthesis will be provided with information information with provided be will A synthesis narrative systematic published after January 2015, we will screen the ClinicalTrials.gov, ClinicalTrials.gov, the screen will we 2015, published January after presented in the text and tables to summarise and explain the explain and tables and to summarise presented text in the BMJ Open 17 17 16 16 15d 15d http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only Confidence in cumulative evidence evidence cumulative Confidence in Metabias(es) Metabias(es) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from BMJ Open http://bmjopen.bmj.com/ on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml For peer review only

Page 35 of 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

Comparative efficacy and tolerability of first- and newergeneration antidepressant medications for depressive disorders in children and adolescents: Study Protocol for a Systematic Review and Network Meta-Analysis For peer review only

Journal: BMJ Open

Manuscript ID: bmjopen-2015-007768.R2

Article Type: Protocol

Date Submitted by the Author: 29-May-2015

Complete List of Authors: Zhou, Xinyu; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Bin, Qin; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Whittington, Craig; UCL, Clinical, Educational and Health Psychology Cohen, David; AP–HP, Hôpital Pitié–Salpétrière, Institut des Systèmes Intelligents et de Robotiques (ISIR), Centre National pour la Recherche Scientifique, Université Pierre et Marie Curie, Department of Child and Adolescent Psychiatry Liu, Yiyun; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Del Giovane, Cinzia; University of Modena and Reggio Emilia, Clinical and http://bmjopen.bmj.com/ Diagnostic Medicine and Public Health Michael, Kurt; Appalachian State University, Department of Psychology Zhang, Yuqing; The First Affiliated Hospital of Chongqing Medical University, Department of Neurology Xie, Peng; The First Affiliated Hospital of Chongqing Medical University, Department of Neuropsychiatry

Primary Subject Mental health Heading: on September 23, 2021 by guest. Protected copyright.

Secondary Subject Heading: Paediatrics, Pharmacology and therapeutics

Child & adolescent psychiatry < PSYCHIATRY, Depression & mood Keywords: disorders < PSYCHIATRY, CLINICAL PHARMACOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Comparative efficacy and tolerability of first- and newer-generation antidepressant 6 7 8 medications for depressive disorders in children and adolescents: Study Protocol for a 9 10 11 Systematic Review and Network Meta-Analysis 12 13 Xinyu Zhou1, Bin Qin1, Craig Whittington2, David Cohen3, Yiyun Liu1, Cinzia Del Giovane4, Kurt D. Michael5, Yuqing 14 15 For peer review only 16 Zhang1, and Peng Xie1 17 18 19 1Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China 20 21 22 2 Research Department of Clinical, Educational and Health Psychology, University College London, London, UK 23 24 25 3 Department of Child and Adolescent Psychiatry, AP–HP, Hôpital Pitié–Salpétrière, Institut des Systèmes Intelligents et 26 27 28 de Robotiques (ISIR), Centre National pour la Recherche Scientifique, Université Pierre et Marie Curie, Paris, France 29 30 31 4Department of Diagnostic, Clinical, and Public Health Medicine, University of Modena and Reggio, Emilia, Modena, 32 33

34 Italy http://bmjopen.bmj.com/ 35

36 5 37 Department of Psychology, Appalachian State University, Boone, North Carolina, USA; 38 39 40 Xinyu Zhou and Bin Qin contributed equally to the protocol. 41

42 on September 23, 2021 by guest. Protected copyright. Direct correspondence to: 43 44 45 Peng Xie, MD, Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi 46 47 48 Road, Yuzhong District, Chongqing 400016, China; Tel: +86-023-68485490; Fax: +86-023-68485111; 49 50 51 E-mail: [email protected] 52 53 54 Keywords: depression, child, adolescent, antidepressant, network meta-analysis 55 56 57 Word count: 2402. 58 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 ABSTRACT 6 7 8 Introduction: Depressive disorders are among the most common psychiatric disorders in children and adolescents, and 9 10 11 have adverse effects on their psychosocial functioning. Questions concerning the efficacy and safety of antidepressant 12 13 medications in the treatment of depression in children and adolescents led us to integrate the direct and indirect 14 15 For peer review only 16 evidence using network meta-analysis to create hierarchies of these drugs. 17 18 19 Methods and analysis: Seven databases with PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, 20 21 22 LiLACS, and PsycINFO will be searched from 1966 to December 2013 (updated to May, 2015). There are no 23 24 25 restrictions on language or type of publication. Randomized clinical trials assessing first- and newer-generation 26 27 28 antidepressant medications against active comparator or placebo as acute treatment for depressive disorder in children 29 30 31 and adolescents (under 18 years of age) will be included. The primary outcome for efficacy will be mean improvement 32 33

34 in depressive symptoms, as measured by the mean change score of a depression rating scale from baseline to post- http://bmjopen.bmj.com/ 35 36 37 treatment. The tolerability of treatment will be defined as side-effect discontinuation, as defined by the proportion of 38 39 40 patients who discontinued treatment due to adverse events during the trial. We will also assess the secondary outcome 41

42 on September 23, 2021 by guest. Protected copyright. for efficacy (response rate), acceptability (all-cause discontinuation), and suicide-related outcomes. We will perform the 43 44 45 Bayesian network meta-analyses for all relative outcome measures. Subgroup analyses and sensitivity analyses will be 46 47 48 conducted to assess the robustness of the findings. 49 50 51 Dissemination: The network meta-analysis will provide useful information on antidepressant treatment for child and 52 53 54 adolescent depression. The results will be disseminated through peer-reviewed publication or conference presentations. 55 56 57 Protocol registration: PROSPERO CRD42015016023 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Strengths and limitations of this study 6 7 8 1. This Bayesian network meta-analysis can integrate direct evidence with indirect evidence from multiple treatment 9 10 11 comparisons to estimate the interrelations across all treatments. 12 13 2. We will comprehensively assess the efficacy, tolerability, acceptability, and suicide-related outcomes of first- and 14 15 For peer review only 16 newer-generation antidepressant medications for depression in children and adolescents. 17 18 19 3. Several subgroup and sensitivity analyses will address some clinically relevant questions. 20 21 22 4. This method comprehensively synthesizes data to provide a clinically useful summary that can guide treatment 23 24 25 decisions and guideline development. 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 BACKGROUND 6 7 8 Depressive disorder in children and adolescents is a major public health problem, demonstrated by the disorders ranking 9

34 diagnosis, and treatment of depression in children and adolescents is an important strategy for curbing the rising rate of http://bmjopen.bmj.com/ 35

36 7 37 youth suicide seen in many developed and advanced developing nations. 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. Since the late 1960s, first-generation antidepressants, e.g., tricyclic antidepressant drugs (TCAs), have been used to treat 43 44 45 depressive symptoms in young patients.8 In the U.S., the use of antidepressant medication in children and adolescents 46 47 48 grew 3- to 10-fold between 1987 and 1996.9 The efficacy of TCAs has been investigated in 13 randomized placebo- 49 50 51 controlled trials,10 which showed marginal evidence to support the use of TCAs in the treatment of depression in only 52 53 54 adolescents. However, methodological deficiencies in these trials, including small sample sizes and diagnostic 55 56 57 heterogeneity, restrict statistical inference and generalizability of the findings. At the same time, cardiovascular effects 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 11,12 5 and overdose-related mortality associated with TCA use have greatly limited their utility in clinical practice. 6 7 8 Nevertheless, the TCA nortriptyline is still approved by the Food and Drug Administration (FDA) for the treatment of 9

10 13 11 depression in adolescents and adults. 12 13

14 15 For peer review only 16 In recent decades, newer-generation antidepressants, including second generation antidepressants (e.g., selective 17 18 19 serotonin reuptake inhibitors [SSRIs]) and third generation antidepressants (e.g., serotonin-norepinephrine reuptake 20 21 22 inhibitors [SNRIs]), have been widely used for the treatment of depression in children and adolescents.14 The frequency 23 24 25 of prescription of SSRIs and SNRIs in children and adolescents has progressively increased.15 In European countries, 26 27 28 there has been a doubling of SSRI use over the 4-year period.16 However, only fluoxetinewas approved by the U.S FDA 29 30 31 for treating depression in children and adolescents in January 2003.17 In the same year, concerns about the increased 32 33 18

34 risk of suicide and suicide attempts with SSRIs were first raised. In September 2004, the FDA cautioned practitioners http://bmjopen.bmj.com/ 35

36 19 37 in the use of antidepressant medications in children and adolescents. Similar warnings were issued by other health 38

39 20-22 40 regulatory agencies. Thus, concerns about this issue have refocused attention on the question of how effective 41

42 on September 23, 2021 by guest. Protected copyright. antidepressant medications are with youth depression. 43 44 45 46 47 48 Nonetheless, currently, no published meta-analysis combined direct and indirect evidence for the use of antidepressant 49 50 51 medications on children and adolescents, while it is an important one to perform, given the conflicting results regarding 52 53 54 the efficacy and tolerability of various antidepressant medications in this age group, and lack of head to head trials of 55 56 57 such drugs.23-25. For these reasons, we will employ a network meta-analysis –a methodological approach that allows the 58 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 simultaneous comparison of multiple psychotherapeutic interventions within a single analysis, while preserving 6

7 26 8 randomization. This approach is will be used to integrate direct evidence (from studies directly comparing 9 10 11 interventions) with indirect evidence (information about two treatments derived via a common comparator) from 12 13 multiple treatment comparisons to estimate the interrelations across all treatments.27 We have previously compared the 14 15 For peer review only 16 efficacy and acceptability of psychotherapies for depression in children and adolescents28 and the augmentation agents 17 18 19 for treatment-resistant depression in adults29 in this way. The aim of the network meta-analysis is of randomized 20 21 22 controlled trials is to re-analyze systematically the efficacy, tolerability, acceptability and suicide risk of both first- and 23 24 25 newer-generation antidepressant medications, either against active comparator or control condition, in the treatment of 26 27 28 child and adolescent depression. 29 30 31 32 METHODS 33

34 http://bmjopen.bmj.com/ 35 Criteria for included studies 36 37 38 Types of studies 39 40 41 Any prospective randomized controlled trials (RCTs), including cross-over design and cluster-randomized trials

42 on September 23, 2021 by guest. Protected copyright. 43 44 will be included. However, quasi-randomized trials (e.g., those allocating using alternate days of the week) will be 45 46 47 excluded. Trials with sample sizes smaller than 10 will be excluded in this review. 48 49 50 Types of participants 51 52 53 Children and adolescents (aged from 6 to 18 when they initially enrolled in the trials) with a primary diagnosis of 54 55 current major depressive disorder according to standardized diagnostic interviews, e.g., the Diagnostic and Statistical 56 57 58 30-32 33,34 Manual of Mental Disorders (DSM) or the International Classification of Diseases (ICD) will be included. 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Where a trial contains a portion of participants who are over 18, we will contact trial authors in order to obtain data 6 7 8 from only those participants within our age range. We will exclude trials focusing on child or adolescent bipolar 9 10 11 disorder, but will include trials involving patients with comorbid general psychiatric disorders, such as attention- 12 13 hyperactivity disorder (ADHD), anxiety disorder and substance-related disorder. However, we will not exclude trials in 14 15 For peer review only 16 which participants have a diagnosis of psychotic depression, and these participants will be considered within a separate 17 18 19 subgroup analysis. But we will exclude trials in which those have a diagnosis of treatment-resistant depression, 20 21 22 because those resistance patients always have different treatment response compared with those patients with non- 23 24 25 resistant depression, 26 27 28 Types of interventions 29 30 31 RCTs comparing any first- and newergeneration antidepressant drug against active comparator or either placebo for 32 33

34 treatment of depression in children and adolescents will be included. Trials comparing the same type of antidepressant http://bmjopen.bmj.com/ 35 36 37 but at different therapeutic dose (fixed or flexible dose) and different treatment duration will be considered as the same 38 39 40 node in the network analysis. We will exclude trials involving combination therapy (i.e., combination of antidepressant 41

42 on September 23, 2021 by guest. Protected copyright. medications, combination of antidepressant medication with psychotherapy, or other non-psychotherapeutic 43 44 45 interventions); however, trials will be considered as eligible if the concomitant psychotherapy is not predefined in the 46 47 48 study. 49 50 51 Types of outcome measures 52 53 54 The acute phase will be defined as from 4 to 16 weeks, and if a trial presents data beyond 16 weeks or for more than 55 56 57 one time period within our pre-defined acute phase periods, we will take the 8-week or close to 8-week time point.35 We 58 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 will exclude trials with treatment duration of less than 4 weeks. Where depression symptoms are measured using more 6 7 8 than one depression scale in a trial, we will extract data from the depressive scales on the basis of a hierarchy of rating 9 10 11 scales. This hierarchy will be based on psychometric properties and appropriateness for use with children and 12 13 adolescents and for consistency of use across trials (referred from Hetrick et al. study)14 (see Table 1). The Children’s 14 15 For peer review only 16 Depression Rating Scale (CDRS-R)36 is adapted for children and adolescents from the Hamilton Depression Rating 17 18 19 Scale (HAMD)37, a tool validated and commonly used in adult populations.38 Both the CDRS-R and HAMD have good 20 21 22 reliability and validity.38 The Beck Depression Inventory (BDI)39/Children’s Depression Inventory (CDI)40 are the most 23 24 25 commonly used among depression symptom severity-self rated scales and are ranked the second highest in the hierarchy. 26 27 28 1. Overall efficacy 29 30 31 1.1 The primary outcome for efficacy will be mean improvement in depressive symptoms, as measured by the mean 32 33

34 change score of depression rating scales (self- or assessor-rated) from baseline to endpoint. http://bmjopen.bmj.com/ 35 36 37 1.2 The secondary outcome for efficacy will be response in depressive symptoms, as estimated by the proportion of 38

39 41 40 patients who achieved a decrease of a certain percentage (e.g., a reduction of 50% or more) in depression rating score. 41

42 on September 23, 2021 by guest. Protected copyright. When ‘response’ is not reported, we will use ‘remission’ if available. The remission will be defined as the proportion of 43 44 45 patients who achieved a below the published threshold in depression rating score (e.g., CDRS-R ≤28).41 46 47 48 2. Overall tolerability 49 50 51 The tolerability of treatment will be defined as side-effect discontinuation in this review, as defined by the proportion of 52 53 54 patients who discontinued treatment due to adverse events during the study. 55 56 57 3. Overall acceptability 58 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 The acceptability of treatment will be defined as all-cause discontinuation, as measured by the proportion of patients 6 7 8 who discontinued treatment (during the delivery of the intervention) up to the post intervention time point. 9 10 11 4. Suicide-related outcomes 12 13 Suicide-related outcomes of dichotomous and continuous will be measured. If data are available, we will extracted the 14 15 For peer review only 16 number of participants with suicidal behavior/ideation during the acute treatment, as measured on a standardised, 17 18 19 validated and reliable rating scales, or reported cases of suicidality.42 In addition, we will also collect data on suicidal 20 21 22 ideation as a continuous outcome where a standardised, validated and reliable rating scale, such as the Suicidal Ideation 23 24 25 Questionnaire-Junior High School version (SIQ-JR),43 has been used. 26 27 28 29 30 31 Data Sources and Search Strategy 32 33

34 Seven electronic databases (PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, LiLACS, and http://bmjopen.bmj.com/ 35 36 37 PsycINFO) will be searched from 1966 to December 2013 (update to May, 2015) with Medical Subject Headings 38 39 ： 40 (MeSH) and text words ”depress*” or “dysthymi*” or “mood disorder*” or “affective disorder*” and “selective 41

42 on September 23, 2021 by guest. Protected copyright. serotonin reuptake inhibitor*” or “SSRIs” or “serotonin norepinephrine reuptake inhibitor*” or “SNRIs” or 43 44 45 “noradrenergic and specific serotonergic antidepressants” or “NaSSA” or “citalopram” or “fluoxetine” or “paroxetine” 46 47 48 or “sertraline” or “escitalopram[” or “fluvoxamine” or “venlafaxine” or “duloxetine” or “milnacipran” or “reboxetine” 49 50 51 or “bupropion” or “mirtazapine” or “tricyclic” or “amersergide” or “amineptine” or “amitriptyline” or “amoxapine” or 52 53 54 “butriptyline” or “chlorpoxiten” or “clomipramine” or “clorimipramine” or “demexiptiline” or “desipramine” or 55 56 57 “dibenzipin” or “dothiepin” or “doxepin” or “imipramine” or “lofepramine” or “melitracen” or “metapramine” or 58 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 “nortriptyline” or “noxiptiline” or “opipramol” or “protriptyline” or “quinupramine” or “tianeptine” or “trimipramine” 6 7 8 and “adolesc*” or “child*” or “boy*” or “girl*” or “juvenil*” or “minors” or “paediatri*” or “pediatri*” or “pubescen*” 9 10 11 or “school*” or “student*” or “teen*” or “young” or “youth*”. Also, ClinicalTrials.gov, World Health Organization’s 12 13 trial portal and U.S. Food and Drug Administration (FDA) reports will be reviewed. There are no restrictions on 14 15 For peer review only 16 language or type of publication. Additional studies will be searched in the reference lists of all identified publications 17 18 19 including relevant meta-analyses and systematic reviews. All relevant authors and principal manufacturers will be 20 21 22 contacted to supplement incomplete reports of the original papers or to provide new data for unpublished trials. 23 24 25 Study Selection 26 27 28 Two reviewers (QB and LYY) will independently scan citations at the title/abstract level identified from the search 29 30 31 strategies and then obtain potentially relevant studies in full text and determine whether to include them by the same 32 33

34 eligibility criteria. Besides, the references of relevant reviews and included trials will also be checked by the two http://bmjopen.bmj.com/ 35 36 37 reviewers. The reasons for exclusion of trials will be reported in the characteristics of excluded studies tables. Any 38 39 40 disagreements will be resolved by a third review author (XYZ). 41

42 on September 23, 2021 by guest. Protected copyright. Data Extraction and Risk of bias assessment 43 44 45 Two independent reviewers (YYL, BQ) will independently extract the key trial parameters using a standardized data 46 47 48 abstraction form and assess the risk of bias. The standardized data extraction forms will include the trial characteristics 49 50 51 (e.g., first listed author, publication year, journal, country, institution, and sponsor), patient characteristics (e.g. 52 53 54 diagnostic criteria for depression, the type of patients, the number of patients, level of depressive symptoms, 55 56 57 comorbidities, the age of patients, and the gender of patients), intervention details (e.g. antidepressant type, dose of 58 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 antidepressant, and the duration of treatment) and outcome measures (efficacy, tolerability, acceptability, and suicide- 6

7 42 8 related outcome). The risk of bias in trials will be assessed by the Cochrane risk of bias tool. Trials attracting a rating 9 10 11 of high risk of bias in one or more domains will be considered as ‘high risk’, low risk of bias in all domains as ‘low 12 13 risk’, and one or more unclear risk of bias in each domain as ‘unclear risk’.42 Any disagreements will be resolved by a 14 15 For peer review only 16 third review author (XYZ). In addition, we will calculate the inter-rater reliability of the two raters. 17 18 19 Data Synthesis and Analysis 20 21 22 We will perform Bayesian network meta-analysis to compare the relative outcomes of different antidepressant 23 24 25 medications and placebo with each other from the median of the posterior distribution.26,27 The pooled estimates of 26 27 28 standardized mean difference (SMD) with 95% credible intervals (CrIs) will be calculated for the continuous outcomes; 29 30 31 and odds ratios (or) with 95% CrIs will be calculated for the categorical outcomes. The SMD is that the difference in 32 33

34 means (MD) of change scores between the two groups divided by the pooled standard deviation (SD) of the http://bmjopen.bmj.com/ 35 36 37 measurements, with a negative SMD value indicates greater symptomatic relief. In the presence of minimally 38 39 40 informative priors, CrIs can be interpreted similarly to confidence intervals, and at conventional levels of statistical 41

42 on September 23, 2021 by guest. Protected copyright. significance a two-sided p <0.05 can be assumed if 95% CrIs do not include 0. If means and standard deviations (SDs) 43 44 45 are not provided, we will calculate them from the p-value or other statistical indices as described elsewhere.44,45 Results 46 47 48 from intention-to-treat analysis (ITT) or modified ITT will be preferred over results from completer analyses, while we 49 50 51 will also consider the dataset for the means and SDs that are presented in the literature. 52 53 54 The pooled estimates will be obtained using the Markov Chains Monte Carlo method. Two Markov chains will be run 55 56 57 simultaneously with different arbitrarily chosen initial values. To ensure convergence, trace plots and the Brooks- 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 46 5 Gelman-Rubin statistic will be assessed. Convergence will be found to be adequate after running 50,000 samples for 6 7 8 both chains. These samples will be then discarded as “burn-in”, and posterior summaries will be based on 100,000 9 10 11 subsequent simulations. The node splitting method will be used to calculate the inconsistency of the model, which 12 13 separate evidence on a particular comparison into direct and indirect evidence.47 Probability values will be summarized 14 15 For peer review only 16 and reported as surface under the cumulative ranking curve (SUCRA) and rankograms, a simple transformation of the 17 18 19 mean rank used to provide a hierarchy of the treatments and accounts for both the location and the variance of all 20 21 22 relative treatment effects.48 Network meta-analysis will be performed using the WinBUGS software package (version 23 24 25 1.4.3, MRC Biostatistics Unit, Cambridge, UK) with random effects models for multi-arm trials. The other analyses will 26 27 28 be performed and presented by the Stata 11.0 and R 2.11.1 software packages. 29 30 31 Subgroup analyses 32 33

34 The antidepressant medications will be coded according to clinical characteristics, risk of bias, and sample size. We will http://bmjopen.bmj.com/ 35 36 37 conduct the subgroups analyses of data in primary outcome for efficacy. We will perform the following subgroups 38 39 40 analyses by using the meta-regression model and calculate Somer’s D (a correlation coefficient for a dichotomous and 41

42 on September 23, 2021 by guest. Protected copyright. an ordinal variable)49: (i) sex ratio (male-to-female ratio> 1 vs. male-to-female ratio<1); (ii) age group; (iii) treatment 43 44 45 duration; (iv) severity of depressive symptom (mild-to-moderate vs. moderate-to-severe); (v) comorbid general 46 47 48 psychiatric disorders ( with vs. without comorbid); (vi) risk of bias (‘high risk’ literature vs. ‘unclear and low risk’ 49 50 51 literature); (vii) sample size; (viii) company sponsor (with vs. without sponsor); and (ix) the type of trials (published 52 53 54 literature vs. unpublished literature). When the limitation by small number of comparisons for some potential modifiers 55 56 57 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 in carrying out subgroup analyses on these variables, we will perform the sensitivity analyses by omitting specific trials 6 7 8 from the overall analysis. 9 10 11 Other analyses 12 13 We will conduct an additional analysis in Bayesian network model based on including only newer-generation 14 15 For peer review only 16 antidepressant medications in primary efficacy and tolerability outcomes. The funnel plot analyses will be performed to 17 18 19 check for publication bias. Moreover, we will carry out meta-regression analyses to investigate the effect of sponsorship 20 21 22 or year published on outcome estimate. 23 24 25 Ethics and dissemination 26 27 28 This network meta-analysis does not need ethical approval, as data used here are not individual or private. It will be 29 30 31 published in a peer-reviewed journal. The results will provide a general overview and evidence of efficacy and safety of 32 33

34 antidepressant medications for depression in children and adolescents. They will also have implications for clinical http://bmjopen.bmj.com/ 35 36 37 practice and further research. 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Contributors XZ and BQ conceived the study and drafted the protocol. XZ and PX wrote the first draft of the 6 7 8 manuscript. KDM, CW and DC assisted in protocol design and revision. YL and YZ participated in the search strategy 9 10 11 development. CDG and BQ participated in the design of data synthesis and analysis. All authors have approved the 12 13 publication of the protocol 14 15 For peer review only 16 Funding This work is supported by the National Basic Research Program of China (973 Program) (Grant No. 17 18 19 2009CB918300). The funders had no role in the protocol design; the writing of the protocol; or the decision to submit 20 21 22 the protocol for publication. 23 24 25 Competing interests None. 26 27 28 Provenance and peer review Not commissioned; externally peer-reviewed. 29 30 31 Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution- Non 32 33

34 Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non- http://bmjopen.bmj.com/ 35 36 37 commercially, and license their derivative works on different terms, provided the original work is properly cited and the 38 39 40 use is non-commercial. See: See: http://creativecommons.org/licenses/by-nc/4.0/ 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 TABLES 6 7 Table 1 Hierarchy of depression symptom severity measurement scales 8 Hierarchy Depression symptom severity measurement scales Abbreviation 9 10 1 Children’s Depression Rating Scale CDRS 11 2 Hamilton Depression Rating Scale HAMD 12 13 3 Montgomery Asberg Depression Rating Scale MADRS 14 4 Beck Depression Inventory BDI 15 For peer review only 16 5 Children’s Depression Inventory CDI 17 6 Schedule for Affective Disorders and Schizophrenia for School K-SADS 18 19 Aged Children 20 7 Mood and Feeling Questionnaire MFQ 21 8 Reynolds Adolescent Depression Scale RADS 22 23 9 Bellevue Index of Depression BID 24 10 Child Depression Scale CDS 25 26 11 Centre for Epidemiologic Studies Depression Scale CESD 27 12 Child Assessment Schedule CAS 28 29 13 Child Behavior Checklist-Depression CBCL-D 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 References 6 7 8 1. Ferrari AJ, Charlson FJ, Norman RE, et al. Burden of depressive disorders by country, sex, age, and year: findings 9 10 11 from the global burden of disease study 2010. PLoS Med. 2013;10:e1001547. 12 13 2. Ryan ND. Treatment of depression in children and adolescents. Lancet 2005;366:933-40. 14 15 For peer review only 16 3. Rao U, Ryan ND, Birmaher B, et al. Unipolar depression in adolescents: clinical outcome in adulthood. J Am Acad 17 18 19 Child Adolesc Psychiatry 1995;34:566-78. 20 21 22 4. Jaffee SR, Moffitt TE, Caspi A, et al. Differences in early childhood risk factors for juvenile-onset and adult-onset 23 24 25 depression. Arch Gen Psychiatry 2002;59:215-22. 26 27 28 5. Office of the Surgeon General. The Surgeon General's call to action to prevent suicide. Washington, DC: Department 29 30 31 of Health and Human Services, 1999. 32 33

34 6. Centers for Disease Control and Prevention (CDC). Suicide trends among youths and young adults aged 10-24 years- http://bmjopen.bmj.com/ 35 36 37 -United States, 1990-2004. MMWR Morb Mortal Wkly Rep 2007,56:905-908. 38 39 40 7. Pfeffer CR. Suicide among youth: Perspectives on risk and prevention. Washington, DC: American Psychiatric Press, 41

42 on September 23, 2021 by guest. Protected copyright. 43 1989;203-25. 44 45 46 8. Lieberman JAI. History of the use of antidepressants in primary care. J Clin Psychiatry 2003;5:6-10. 47 48 9. Zito JM, Safer DJ, DosReis S, et al. Psychotropic practice patterns for youth: a 10-year perspective. Arch Pediatr 49 50 51 Adolesc Med 2003;157:17-25. 52 53 54 10. Hazell P, Mirzaie M. Tricyclic drugs for depression in children and adolescents. Cochrane Database Syst Rev 55 56 57 2013;6:CD002317. 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 11. Biederman J. Sudden death in children treated with a tricyclic antidepressant. J Am Acad Child Adolesc Psychiatry 6 7 8 1991;30:495-8. 9 10 11 12. Varley CK, McClellan J. Case study: two additional sudden deaths with tricyclic antidepressants. J Am Acad Child 12 13 Adolesc Psychiatry 1997;36:390-4. 14 15 For peer review only 16 13. Green WH. Child and Adolescent clinical psychopharmacology. 3 rd ed. Philadelphia: Lippincott Williams and 17 18 19 Wilkins, 2003:169-72. 20 21 22 14. Hetrick SE, McKenzie JE, Cox GR, et al. Newer generation antidepressants for depressive disorders in children and 23 24 25 adolescents. Cochrane Database Syst Rev 2012;11:CD004851. 26 27 28 15. Bennett K, Teeling M, Feely J. Overprescribing antidepressants to children: pharmacoepidemiological study in 29 30 31 primary care. BMJ 2005;331:1451-2. 32 33

34 16. Fegert JM, Kölch M, Zito JM, et al. Antidepressant use in children and adolescents in Germany. J Child Adolesc http://bmjopen.bmj.com/ 35 36 37 Psychopharmacol 2006;16:197-206. 38 39 40 17. Gentile S. Antidepressant use in children and adolescents diagnosed with major depressive disorder: what can we 41

42 on September 23, 2021 by guest. Protected copyright. 43 learn from published data? Rev Recent Clin Trials 2010;5;63-75. 44 45 46 18. Healy D. Lines of evidence on the risks of suicide with selective serotonin reuptake inhibitors. Psychother 47 48 Psychosom 2003;72:71-9. 49 50 51 19. U.S. Food and Drug Administration (FDA). Suicidality in children and adolescents being treated with antidepressant 52 53 54 medications. 55 56 57 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 http://www.fda.gov/drugs/drugsafety/postmarketdrugsafetyinformationforpatientsandproviders/ucm161679.htm 6 7 8 (accessed 11 Dec 2014). 9 10 11 20. Adverse Drug Reactions Advisory Committee. Use of SSRI antidepressants in children and adolescents. 12 13 https://www.tga.gov.au/use-ssri-antidepressants-children-and-adolescents-june-2004 (accessed 1 Dec 2014). 14 15 For peer review only 16 21. Canadian Psychiatric Association Bulletin. Health Canada to examine SSRIs in children, April 2004. 17 18 19 http://ww1.cpa-apc.org/Publications/Archives/Bulletin/2004/april/news32En.asp (accessed 18 Dec 2014). 20 21 22 22. National Institute of Mental Health. Antidepressant medications for children and adolescents: information for 23 24 25 parents and caregivers. http://www.nimh.nih.gov/health/topics/child-and-adolescent-mental-health/antidepressant- 26 27 28 medications-for-children-and-adolescents-information-for-parents-and-caregivers.shtml (accessed 14 Dec 2014). 29 30 31 23. Papanikolaou K, Richardson C, Pehlivanidis A, et al. Efficacy of antidepressants in child and adolescent depression: 32 33

34 a meta-analytic study. J Neural Transm 2006;113:399-415. http://bmjopen.bmj.com/ 35 36 37 24. Whittington CJ, Kendall T, Fonagy P, et al. Selective serotonin reuptake inhibitors in childhood depression: 38 39 40 systematic review of published versus unpublished data. Lancet 2004;363:1341-5. 41

42 on September 23, 2021 by guest. Protected copyright. 25. Qin B, Zhang Y, Zhou X, et al. Selective serotonin reuptake inhibitors versus tricyclic antidepressants in young 43 44 45 patients: a meta-analysis of efficacy and acceptability. Clin Ther 2014;36:1087-95. 46 47 48 26. Salanti G, Higgins JP, Ades AE, et al. Evaluation of networks of randomized trials. Stat Methods Med Res 49 50 51 2008;17:279-301. 52 53 54 55 27. Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 56 57 58 2004;23:3105-24. 59 60 18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 28. Zhou X, Ravindran A, Qin B, et al. Comparative Efficacy, Acceptability and Tolerability of Augmentation Agents in 6 7 8 Treatment-Resistant Depression: Systematic Review and Network Meta-Analysis. J Clin Psychiatry 2014; In press. 9 10 11 29. Qin B, Zhou X, Michael KD, et al. Psychotherapy for Depression in Children and Adolescents: Study Protocol for a 12 13 14 Systematic Review and Network Meta-Analysis. BMJ Open 2014; In press. 15 For peer review only 16 17 30. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-III). 3rd edition. 18 19 20 Washington, DC: American Psychiatric Association, 1980. 21 22 23 31. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R). 3rd revised 24 25 26 edition. Washington, DC: American Psychiatric Association, 1987. 27 28 29 32. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th edn. Washington, DC: 30 31 32 American Psychiatric Association, 1994. 33

34 http://bmjopen.bmj.com/ 33. World Health Organization (WHO). The Ninth Revision of the International Classification of Diseases and Related 35 36 37 Health Problems (ICD-9). Geneva: World Health Organization, 1978. 38 39 40 34. World Health Organization (WHO). The Tenth Revision of the International Classification of Diseases and Related 41

42 on September 23, 2021 by guest. Protected copyright. 43 Health Problems (ICD-10). Geneva: World Health Organization, 1992. 44 45 46 35. Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 12 new-generation 47 48 49 antidepressants: a multiple-treatments meta-analysis. Lancet. 2009;373:746-58. 50 51 52 36. Poznanski EO, Mokros HB. Children's depression rating scale, revised (CDRS-R): manual. Western Psychological 53 54 55 Services, 1996. 56 57 58 37. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23:56-62. 59 60 19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 35 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 38. Brooks SJ, Kutcher S. Diagnosis and measurement of depression: a review of commonly utilized instruments. J 6 7 8 Child Adolesc Psychopharmacol 2001;11:341-76. 9 10 11 39. Beck AT, Steer R. Beck depression inventory: manual. San Antonio, Texas: Psychological Corporation, 1987. 12 13 40. Kovacs M. The Children's Depression, Inventory (CDI). Psychopharmacol Bull 1985;21:995-8. 14 15 For peer review only 16 41. Riedel M, Möller HJ, Obermeier M, et al. Response and remission criteria in major depression--a validation of 17 18 19 current practice. J Psychiatr Res 2010;44:1063-8. 20 21 22 42. Hammad TA. Review and evaluation of clinical data. Washington, DC, US Food and Drug Administration. 23 24 25 http://www.fda.gov/ohrms/dockets/ac/04/briefing/2004-4065b1-10- tab08-hammads-review.pdf (accessed 9 Dec 2014). 26 27 28 43. Reynolds WM. Suicidal Ideation Questionnaire-Junior. Odessa, FL: Psychological Assessment Resources, 1987. 29 30 31 44. Higgins JPT, Green S. Cochrane handbook for systematic reviews of interventions version 5.1. 0 [updated March 32 33

34 2011]. The Cochrane Collaboration, 2011. http://bmjopen.bmj.com/ 35 36 37 45. Follmann D, Elliott P, Suh I, et al. Variance imputation for overviews of clinical trials with continuous response. J 38 39 40 Clin Epidemiol 1992;45:769-73. 41

42 on September 23, 2021 by guest. Protected copyright. 43 46. Brooks SP, Gelman A. General Methods for Monitoring Convergence of Iterative Simulations. J Comput Graph Stat 44 45 46 1998;7:434-55. 47 48 49 47. Lu GB, Ades AE. Assessing evidence inconsistency in mixed treatment comparisons. J Am Stat Assoc 50 51 52 2006;101:447-59. 53 54 55 48. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple- 56 57 58 treatment meta-analysis: an overview and tutorial. J Clin Epidemiol 2011; 64:163-71. 59 60 20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 35 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 49. Dias S, Sutton AJ, Welton NJ, et al. Evidence synthesis for decision making 3: heterogeneity-subgroups, meta- 6 7 8 regression, bias, and bias-adjustment. Med Decis Making 2013;33:618-40. 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

Page 22 of 35 Checklist item Checklist (registration number CRD42015016023). (registration number

[email protected] [email protected] *Corresponding author: Peng Xie Peng author: *Corresponding

BMJ Open

2 2 3a 3a 1a 1a 1b 1b http://bmjopen.bmj.com/ Item No No Item on September 23, 2021 by guest. Protected copyright. For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml

For peer review only

5 Department of Psychology, Appalachian State University, Boone, University, Boone, State Appalachian 5 of Psychology, Department

BMJ Open

Participants Participants

Outcomes

Interventions

Language Language

There will be no restrictions by type of setting. setting. of by restrictions type There no be will

Setting Setting

For peer review only Journal: BMJ Open

Manuscript ID: bmjopen-2015-007768.R3

Article Type: Protocol

Date Submitted by the Author: 06-Aug-2015

Primary Subject Mental health Heading: on September 23, 2021 by guest. Protected copyright. Secondary Subject Heading: Paediatrics, Pharmacology and therapeutics

Child & adolescent psychiatry < PSYCHIATRY, Depression & mood Keywords: disorders < PSYCHIATRY, CLINICAL PHARMACOLOGY

For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 1 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

34 Italy http://bmjopen.bmj.com/ 35

36 5 37 Department of Psychology, Appalachian State University, Boone, North Carolina, USA; 38 39 40 Xinyu Zhou and Bin Qin contributed equally to the protocol. 41

42 on September 23, 2021 by guest. Protected copyright. Direct correspondence to: 43 44 45 Peng Xie, MD, Department of Neurology, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi 46 47 48 Road, Yuzhong District, Chongqing 400016, China; Tel: +86-023-68485490; Fax: +86-023-68485111; 49 50 51 E-mail: [email protected] 52 53 54 Keywords: depression, child, adolescent, antidepressant, network meta-analysis 55 56 57 Word count: 2585. 58 59 60 1 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 2 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 ABSTRACT 6 7 8 Introduction: Depressive disorders are among the most common psychiatric disorders in children and adolescents, and 9 10 11 have adverse effects on their psychosocial functioning. Questions concerning the efficacy and safety of antidepressant 12 13 medications in the treatment of depression in children and adolescents led us to integrate the direct and indirect 14 15 For peer review only 16 evidence using network meta-analysis to create hierarchies of these drugs. 17 18 19 Methods and analysis: Seven databases with PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, 20 21 22 LiLACS, and PsycINFO will be searched from 1966 to December 2013 (updated to May, 2015). There are no 23 24 25 restrictions on language or type of publication. Randomized clinical trials assessing first- and newer-generation 26 27 28 antidepressant medications against active comparator or placebo as acute treatment for depressive disorder in children 29 30 31 and adolescents (under 18 years of age) will be included. The primary outcome for efficacy will be mean improvement 32 33

42 on September 23, 2021 by guest. Protected copyright. for efficacy (response rate), acceptability (all-cause discontinuation), and suicide-related outcomes. We will perform the 43 44 45 Bayesian network meta-analyses for all relative outcome measures. Subgroup analyses and sensitivity analyses will be 46 47 48 conducted to assess the robustness of the findings. 49 50 51 Dissemination: The network meta-analysis will provide useful information on antidepressant treatment for child and 52 53 54 adolescent depression. The results will be disseminated through peer-reviewed publication or conference presentations. 55 56 57 Protocol registration: PROSPERO CRD42015016023 58 59 60 2 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 3 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 3 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 4 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 BACKGROUND 6 7 8 Depressive disorder in children and adolescents is a major public health problem, demonstrated by the disorders ranking 9

34 diagnosis, and treatment of depression in children and adolescents is an important strategy for curbing the rising rate of http://bmjopen.bmj.com/ 35

36 7 37 youth suicide seen in many developed and advanced developing nations. 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. Since the late 1960s, first-generation antidepressants, e.g., tricyclic antidepressant drugs (TCAs), have been used to treat 43 44 45 depressive symptoms in young patients.8 In the U.S., the use of antidepressant medication in children and adolescents 46 47 48 grew 3- to 10-fold between 1987 and 1996.9 The efficacy of TCAs has been investigated in 13 randomized placebo- 49 50 51 controlled trials,10 which showed marginal evidence to support the use of TCAs in the treatment of depression in only 52 53 54 adolescents. However, methodological deficiencies in these trials, including small sample sizes and diagnostic 55 56 57 heterogeneity, restrict statistical inference and generalizability of the findings. At the same time, cardiovascular effects 58 59 60 4 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 5 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

10 13 11 depression in adolescents and adults. 12 13

34 risk of suicide and suicide attempts with SSRIs were first raised. In September 2004, the FDA cautioned practitioners http://bmjopen.bmj.com/ 35

36 19 37 in the use of antidepressant medications in children and adolescents. Similar warnings were issued by other health 38

39 20-22 40 regulatory agencies. Thus, concerns about this issue have refocused attention on the question of how effective 41

42 on September 23, 2021 by guest. Protected copyright. antidepressant medications are with youth depression. 43 44 45 46 47 48 Nonetheless, currently, no published meta-analysis has combined direct and indirect evidence for the use of 49 50 51 antidepressant medications on children and adolescents, while it is an important one to perform, given the conflicting 52 53 54 results regarding the efficacy and tolerability of various antidepressant medications in this age group, and lack of head 55 56 57 to head trials of such drugs.23-25. For these reasons, we will employ a network meta-analysis –a methodological 58 59 60 5 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 6 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 approach that allows the simultaneous comparison of multiple psychotherapeutic interventions within a single analysis, 6

7 26 8 while preserving randomization. This approach is will be used to integrate direct evidence (from studies directly 9 10 11 comparing interventions) with indirect evidence (information about two treatments derived via a common comparator) 12 13 from multiple treatment comparisons to estimate the interrelations across all treatments.27 We have previously compared 14 15 For peer review only 16 the efficacy and acceptability of psychotherapies for depression in children and adolescents28 and the augmentation 17 18 19 agents for treatment-resistant depression in adults29 in this way. The aim of the network meta-analysis is of randomized 20 21 22 controlled trials is to re-analyze systematically the efficacy, tolerability, acceptability and suicide risk of both first- and 23 24 25 newer-generation antidepressant medications, either against active comparator or control condition, in the treatment of 26 27 28 child and adolescent depression. 29 30 31 32 METHODS 33

42 on September 23, 2021 by guest. Protected copyright. 43 44 included. However, quasi-randomized trials (e.g., those allocating using alternate days of the week) will be excluded. 45 46 47 Trials with sample sizes smaller than 10 will be excluded in this review. 48 49 50 51 52 53 Types of participants 54 55 Children and adolescents (aged from 6 to 18 when they initially enrolled in the trials) with a primary diagnosis of 56 57 58 current major depressive disorder according to standardized diagnostic interviews, e.g., the Diagnostic and Statistical 59 60 6 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 7 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 30-32 33,34 5 Manual of Mental Disorders (DSM) or the International Classification of Diseases (ICD) will be included. 6 7 8 Where a trial contains a portion of participants who are over 18, we will contact trial authors in order to obtain data 9 10 11 from only those participants within our age range. We will exclude trials focusing on child or adolescent bipolar 12 13 disorder, but will include trials involving patients with comorbid general psychiatric disorders, such as attention- 14 15 For peer review only 16 hyperactivity disorder (ADHD), anxiety disorder and substance-related disorder. However, we will not exclude trials in 17 18 19 which participants have a diagnosis of psychotic depression, and these participants will be considered within a separate 20 21 22 subgroup analysis. But we will exclude trials in which those have a diagnosis of treatment-resistant depression, 23 24 25 because those resistance patients tend to have different treatment response compared with those patients with non- 26 27 28 resistant depression. 29 30 31 32 33

34 Types of interventions http://bmjopen.bmj.com/ 35 36 37 RCTs comparing any first- and newergeneration antidepressant drug against active comparator or either placebo for 38 39 40 treatment of depression in children and adolescents will be included. Trials comparing the same type of antidepressant 41

42 on September 23, 2021 by guest. Protected copyright. but at different therapeutic dose (fixed or flexible dose) and different treatment duration will be considered as the same 43 44 45 node in the network analysis. We will exclude trials involving combination therapy (i.e., combination of antidepressant 46 47 48 medications, combination of antidepressant medication with psychotherapy, or other non-psychotherapeutic 49 50 51 interventions); however, trials will be considered as eligible if the concomitant psychotherapy is not predefined in the 52 53 54 study. 55 56 57 58 59 60 7 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 8 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Types of outcome measures 6 7 8 The acute phase will be defined as from 4 to 16 weeks, and if a trial presents data beyond 16 weeks or for more than 9

10 35 11 one time period within our pre-defined acute phase periods, we will take the 8-week or close to 8-week time point. We 12 13 will exclude trials with treatment duration of less than 4 weeks. Where depression symptoms are measured using more 14 15 For peer review only 16 than one depression scale in a trial, we will extract data from the depressive scales on the basis of a hierarchy of rating 17 18 19 scales. This hierarchy will be based on psychometric properties and appropriateness for use with children and 20 21 22 adolescents and for consistency of use across trials (referred from Hetrick et al. study)14 (see Table 1). The Children’s 23 24 25 Depression Rating Scale (CDRS-R)36 is adapted for children and adolescents from the Hamilton Depression Rating 26 27 28 Scale (HAMD)37, a tool validated and commonly used in adult populations.38 Both the CDRS-R and HAMD have good 29 30 31 reliability and validity.38 The Beck Depression Inventory (BDI)39/Children’s Depression Inventory (CDI)40 are the most 32 33

34 commonly used among depression symptom severity-self rated scales and are ranked the second highest in the hierarchy. http://bmjopen.bmj.com/ 35 36 37 1. Overall efficacy 38 39 40 1.1 The primary outcome for efficacy will be mean improvement in depressive symptoms, as measured by the mean 41

42 on September 23, 2021 by guest. Protected copyright. change score of depression rating scales (self- or assessor-rated) from baseline to endpoint. 43 44 45 1.2 The secondary outcome for efficacy will be response in depressive symptoms, as estimated by the proportion of 46 47 48 patients who achieved a decrease of a certain percentage (e.g., a reduction of 50% or more) in depression rating score.41 49 50 51 When ‘response’ is not reported, we will use ‘remission’ if available. The remission will be defined as the proportion of 52 53 54 patients who achieved a below the published threshold in depression rating score (e.g., CDRS-R ≤28).41 55 56 57 2. Overall tolerability 58 59 60 8 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 9 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 The tolerability of treatment will be defined as side-effect discontinuation in this review, as defined by the proportion of 6 7 8 patients who discontinued treatment due to adverse events during the study. 9 10 11 3. Overall acceptability 12 13 The acceptability of treatment will be defined as all-cause discontinuation, as measured by the proportion of patients 14 15 For peer review only 16 who discontinued treatment (during the delivery of the intervention) up to the post intervention time point. 17 18 19 4. Suicide-related outcomes 20 21 22 Suicide-related dichotomous and continuous outcomes will be measured. If data are available, we will extracted the 23 24 25 number of participants with suicide-related events (combined suicidal ideation and suicidal behavior) during the acute 26 27 28 treatment, as measured on a standardised, validated and reliable rating scales, or reported cases of suicidality.42 In 29 30 31 addition, we will also collect data on suicidal ideation as a continuous outcome where a standardised, validated and 32 33 43

34 reliable rating scale, such as the Suicidal Ideation Questionnaire-Junior High School version (SIQ-JR), has been used. http://bmjopen.bmj.com/ 35 36 37 38 39 40 Data Sources and Search Strategy 41

42 on September 23, 2021 by guest. Protected copyright. Seven electronic databases (PubMed, Embase, the Cochrane Library, Web of Science, CINAHL, LiLACS, and 43 44 45 PsycINFO) will be searched from 1966 to December 2013 (update to May, 2015) with Medical Subject Headings 46 47 48 (MeSH) and text words： ”depress*” or “dysthymi*” or “mood disorder*” or “affective disorder*” and “selective 49 50 51 serotonin reuptake inhibitor*” or “SSRIs” or “serotonin norepinephrine reuptake inhibitor*” or “SNRIs” or 52 53 54 “noradrenergic and specific serotonergic antidepressants” or “NaSSA” or “citalopram” or “fluoxetine” or “paroxetine” 55 56 57 or “sertraline” or “escitalopram[” or “fluvoxamine” or “venlafaxine” or “duloxetine” or “milnacipran” or “reboxetine” 58 59 60 9 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 10 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 or “bupropion” or “mirtazapine” or “tricyclic” or “amersergide” or “amineptine” or “amitriptyline” or “amoxapine” or 6 7 8 “butriptyline” or “chlorpoxiten” or “clomipramine” or “clorimipramine” or “demexiptiline” or “desipramine” or 9 10 11 “dibenzipin” or “dothiepin” or “doxepin” or “imipramine” or “lofepramine” or “melitracen” or “metapramine” or 12 13 “nortriptyline” or “noxiptiline” or “opipramol” or “protriptyline” or “quinupramine” or “tianeptine” or “trimipramine” 14 15 For peer review only 16 and “adolesc*” or “child*” or “boy*” or “girl*” or “juvenil*” or “minors” or “paediatri*” or “pediatri*” or “pubescen*” 17 18 19 or “school*” or “student*” or “teen*” or “young” or “youth*”. Also, ClinicalTrials.gov, World Health Organization’s 20 21 22 trial portal and U.S. Food and Drug Administration (FDA) reports will be reviewed. There are no restrictions on 23 24 25 language or type of publication. Additional studies will be searched in the reference lists of all identified publications 26 27 28 including relevant meta-analyses and systematic reviews. All relevant authors and principal manufacturers will be 29 30 31 contacted to supplement incomplete reports of the original papers or to provide new data for unpublished trials. 32 33

34 http://bmjopen.bmj.com/ 35 36 37 Study Selection 38 39 40 Two reviewers (QB and LYY) will independently scan citations at the title/abstract level identified from the search 41

42 on September 23, 2021 by guest. Protected copyright. strategies and then obtain potentially relevant studies in full text and determine whether to include them by the same 43 44 45 eligibility criteria. Besides, the references of relevant reviews and included trials will also be checked by the two 46 47 48 reviewers. The reasons for exclusion of trials will be reported in the characteristics of excluded studies tables. Any 49 50 51 disagreements will be resolved by a third review author (XYZ). 52 53 54 55 56 57 Data Extraction and Risk of bias assessment 58 59 60 10 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 11 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Two independent reviewers (YYL, BQ) will independently extract the key trial parameters using a standardized data 6 7 8 abstraction form and assess the risk of bias. The standardized data extraction forms will include the trial characteristics 9 10 11 (e.g., first listed author, publication year, journal, country, institution, and sponsor), patient characteristics (e.g. 12 13 diagnostic criteria for depression, the type of patients, the number of patients, level of depressive symptoms, 14 15 For peer review only 16 comorbidities, the age of patients, and the gender of patients), intervention details (e.g. antidepressant type, dose of 17 18 19 antidepressant, and the duration of treatment) and outcome measures (efficacy, tolerability, acceptability, and suicide- 20 21 22 related outcome). The risk of bias in trials will be assessed by the Cochrane risk of bias tool.42 Trials attracting a rating 23 24 25 of high risk of bias in one or more domains will be considered as ‘high risk’, low risk of bias in all domains as ‘low 26 27 28 risk’, and one or more unclear risk of bias in each domain as ‘unclear risk’.42 Any disagreements will be resolved by a 29 30 31 third review author (XYZ). In addition, we will calculate the inter-rater reliability of the two raters. 32 33

34 http://bmjopen.bmj.com/ 35 36 37 Data Synthesis and Analysis 38 39 40 We will perform Bayesian network meta-analysis to compare the relative outcomes of different antidepressant 41

42 on September 23, 2021 by guest. Protected copyright. medications and placebo with each other from the median of the posterior distribution.26,27 The pooled estimates of 43 44 45 standardized mean difference (SMD) with 95% credible intervals (CrIs) will be calculated for the continuous outcomes; 46 47 48 and odds ratios (or) with 95% CrIs will be calculated for the categorical outcomes. The SMD is that the difference in 49 50 51 means (MD) of change scores between the two groups divided by the pooled standard deviation (SD) of the 52 53 54 measurements, with a negative SMD value indicates greater symptomatic relief. In the presence of minimally 55 56 57 informative priors, CrIs can be interpreted similarly to confidence intervals, and at conventional levels of statistical 58 59 60 11 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 12 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 significance a two-sided p <0.05 can be assumed if 95% CrIs do not include 0. If means and standard deviations (SDs) 6

7 44,45 8 are not provided, we will calculate them from the p-value or other statistical indices as described elsewhere. Results 9 10 11 from intention-to-treat analysis (ITT) or modified ITT will be preferred over results from completer analyses, while we 12 13 will also consider the dataset for the means and SDs that are presented in the literature. 14 15 For peer review only 16 The pooled estimates will be obtained using the Markov Chains Monte Carlo method. Two Markov chains will be run 17 18 19 simultaneously with different arbitrarily chosen initial values. To ensure convergence, trace plots and the Brooks- 20 21 22 Gelman-Rubin statistic will be assessed.46 Convergence will be found to be adequate after running 50,000 samples for 23 24 25 both chains. These samples will be then discarded as “burn-in”, and posterior summaries will be based on 100,000 26 27 28 subsequent simulations. The node splitting method will be used to calculate the inconsistency of the model, which 29 30 31 separate evidence on a particular comparison into direct and indirect evidence.47 Probability values will be summarized 32 33

34 and reported as surface under the cumulative ranking curve (SUCRA) and rankograms, a simple transformation of the http://bmjopen.bmj.com/ 35 36 37 mean rank used to provide a hierarchy of the treatments and accounts for both the location and the variance of all 38

39 48 40 relative treatment effects. Network meta-analysis will be performed using the WinBUGS software package (version 41

42 on September 23, 2021 by guest. Protected copyright. 1.4.3, MRC Biostatistics Unit, Cambridge, UK) with random effects models for multi-arm trials. The other analyses will 43 44 45 be performed and presented by the Stata 11.0 and R 2.11.1 software packages. 46 47 48 49 50 51 Subgroup analyses 52 53 54 The antidepressant medications will be coded according to clinical characteristics, risk of bias, and sample size. We will 55 56 57 conduct the subgroups analyses of data in primary outcome for efficacy. We will perform the following subgroups 58 59 60 12 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 13 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 analyses by using the meta-regression model and calculate Somer’s D (a correlation coefficient for a dichotomous and 6

7 49 8 an ordinal variable) : (i) sex ratio (male-to-female ratio> 1 vs. male-to-female ratio<1); (ii) age group; (iii) treatment 9 10 11 duration; (iv) severity of depressive symptom (mild-to-moderate vs. moderate-to-severe); (v) comorbid general 12 13 psychiatric disorders ( with vs. without comorbid); (vi) risk of bias (‘high risk’ literature vs. ‘unclear and low risk’ 14 15 For peer review only 16 literature); (vii) sample size; (viii) company sponsor (with vs. without sponsor); and (ix) the type of trials (published 17 18 19 literature vs. unpublished literature). When the limitation by small number of comparisons for some potential modifiers 20 21 22 in carrying out subgroup analyses on these variables, we will perform the sensitivity analyses by omitting specific trials 23 24 25 from the overall analysis. 26 27 28 29 30 31 Other analyses 32 33

34 The funnel plot analyses will be performed to check for publication bias. Moreover, we will carry out meta-regression http://bmjopen.bmj.com/ 35 36 37 analyses to investigate the effect of sponsorship or year published on outcome estimate. 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. Ethics and dissemination 43 44 45 This network meta-analysis does not need ethical approval, as data used here are not individual or private. It will be 46 47 48 published in a peer-reviewed journal. The results will provide a general overview and evidence of efficacy and safety of 49 50 51 antidepressant medications for depression in children and adolescents. They will also have implications for clinical 52 53 54 practice and further research. 55 56 57 58 59 60 13 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 14 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 Contributors XZ and BQ conceived the study and drafted the protocol. XZ and PX wrote the first draft of the 6 7 8 manuscript. KDM, CW and DC assisted in protocol design and revision. YL and YZ participated in the search strategy 9 10 11 development. CDG and BQ participated in the design of data synthesis and analysis. All authors have approved the 12 13 publication of the protocol 14 15 For peer review only 16 Funding This work is supported by the National Basic Research Program of China (973 Program) (Grant No. 17 18 19 2009CB918300). The funders had no role in the protocol design; the writing of the protocol; or the decision to submit 20 21 22 the protocol for publication. 23 24 25 Competing interests None. 26 27 28 Provenance and peer review Not commissioned; externally peer-reviewed. 29 30 31 Open Access This is an Open Access article distributed in accordance with the Creative Commons Attribution- Non 32 33

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 14 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 15 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 TABLES 6 7 Table 1 Hierarchy of depression symptom severity measurement scales 8 Hierarchy Depression symptom severity measurement scales Abbreviation 9 10 1 Children’s Depression Rating Scale CDRS 11 2 Hamilton Depression Rating Scale HAMD 12 13 3 Montgomery Asberg Depression Rating Scale MADRS 14 4 Beck Depression Inventory BDI 15 For peer review only 16 5 Children’s Depression Inventory CDI 17 6 Schedule for Affective Disorders and Schizophrenia for School K-SADS 18 19 Aged Children 20 7 Mood and Feeling Questionnaire MFQ 21 8 Reynolds Adolescent Depression Scale RADS 22 23 9 Bellevue Index of Depression BID 24 10 Child Depression Scale CDS 25 26 11 Centre for Epidemiologic Studies Depression Scale CESD 27 12 Child Assessment Schedule CAS 28 29 13 Child Behavior Checklist-Depression CBCL-D 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 15 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 16 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 References 6 7 8 1. Ferrari AJ, Charlson FJ, Norman RE, et al. Burden of depressive disorders by country, sex, age, and year: findings 9 10 11 from the global burden of disease study 2010. PLoS Med. 2013;10:e1001547. 12 13 2. Ryan ND. Treatment of depression in children and adolescents. Lancet 2005;366:933-40. 14 15 For peer review only 16 3. Rao U, Ryan ND, Birmaher B, et al. Unipolar depression in adolescents: clinical outcome in adulthood. J Am Acad 17 18 19 Child Adolesc Psychiatry 1995;34:566-78. 20 21 22 4. Jaffee SR, Moffitt TE, Caspi A, et al. Differences in early childhood risk factors for juvenile-onset and adult-onset 23 24 25 depression. Arch Gen Psychiatry 2002;59:215-22. 26 27 28 5. Office of the Surgeon General. The Surgeon General's call to action to prevent suicide. Washington, DC: Department 29 30 31 of Health and Human Services, 1999. 32 33

34 6. Centers for Disease Control and Prevention (CDC). Suicide trends among youths and young adults aged 10-24 years- http://bmjopen.bmj.com/ 35 36 37 United States, 1990-2004. MMWR Morb Mortal Wkly Rep 2007,56:905-908. 38 39 40 7. Pfeffer CR. Suicide among youth: Perspectives on risk and prevention. Washington, DC: American Psychiatric Press, 41

42 on September 23, 2021 by guest. Protected copyright. 43 1989;203-25. 44 45 46 8. Lieberman JAI. History of the use of antidepressants in primary care. J Clin Psychiatry 2003;5:6-10. 47 48 9. Zito JM, Safer DJ, DosReis S, et al. Psychotropic practice patterns for youth: a 10-year perspective. Arch Pediatr 49 50 51 Adolesc Med 2003;157:17-25. 52 53 54 10. Hazell P, Mirzaie M. Tricyclic drugs for depression in children and adolescents. Cochrane Database Syst Rev 55 56 57 2013;6:CD002317. 58 59 60 16 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 17 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

34 16. Fegert JM, Kölch M, Zito JM, et al. Antidepressant use in children and adolescents in Germany. J Child Adolesc http://bmjopen.bmj.com/ 35 36 37 Psychopharmacol 2006;16:197-206. 38 39 40 17. Gentile S. Antidepressant use in children and adolescents diagnosed with major depressive disorder: what can we 41

42 on September 23, 2021 by guest. Protected copyright. 43 learn from published data? Rev Recent Clin Trials 2010;5;63-75. 44 45 46 18. Healy D. Lines of evidence on the risks of suicide with selective serotonin reuptake inhibitors. Psychother 47 48 Psychosom 2003;72:71-9. 49 50 51 19. U.S. Food and Drug Administration (FDA). Suicidality in children and adolescents being treated with antidepressant 52 53 54 medications. 55 56 57 58 59 60 17 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 18 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 http://www.fda.gov/drugs/drugsafety/postmarketdrugsafetyinformationforpatientsandproviders/ucm161679.htm 6 7 8 (accessed 11 Dec 2014). 9 10 11 20. Adverse Drug Reactions Advisory Committee. Use of SSRI antidepressants in children and adolescents. 12 13 https://www.tga.gov.au/use-ssri-antidepressants-children-and-adolescents-june-2004 (accessed 1 Dec 2014). 14 15 For peer review only 16 21. Canadian Psychiatric Association Bulletin. Health Canada to examine SSRIs in children, April 2004. 17 18 19 http://ww1.cpa-apc.org/Publications/Archives/Bulletin/2004/april/news32En.asp (accessed 18 Dec 2014). 20 21 22 22. National Institute of Mental Health. Antidepressant medications for children and adolescents: information for 23 24 25 parents and caregivers. http://www.nimh.nih.gov/health/topics/child-and-adolescent-mental-health/antidepressant- 26 27 28 medications-for-children-and-adolescents-information-for-parents-and-caregivers.shtml (accessed 14 Dec 2014). 29 30 31 23. Papanikolaou K, Richardson C, Pehlivanidis A, et al. Efficacy of antidepressants in child and adolescent depression: 32 33

42 on September 23, 2021 by guest. Protected copyright. 25. Qin B, Zhang Y, Zhou X, et al. Selective serotonin reuptake inhibitors versus tricyclic antidepressants in young 43 44 45 patients: a meta-analysis of efficacy and acceptability. Clin Ther 2014;36:1087-95. 46 47 48 26. Salanti G, Higgins JP, Ades AE, et al. Evaluation of networks of randomized trials. Stat Methods Med Res 49 50 51 2008;17:279-301. 52 53 54 55 27. Lu G, Ades AE. Combination of direct and indirect evidence in mixed treatment comparisons. Stat Med 56 57 58 2004;23:3105-24. 59 60 18 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 19 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 28. Zhou X, Ravindran A, Qin B, et al. Comparative Efficacy, Acceptability and Tolerability of Augmentation Agents in 6 7 8 Treatment-Resistant Depression: Systematic Review and Network Meta-Analysis. J Clin Psychiatry 2014; In press. 9 10 11 29. Qin B, Zhou X, Michael KD, et al. Psychotherapy for Depression in Children and Adolescents: Study Protocol for a 12 13 14 Systematic Review and Network Meta-Analysis. BMJ Open 2015; 5:e005918 15 For peer review only 16 17 30. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-III). 3rd edition. 18 19 20 Washington, DC: American Psychiatric Association, 1980. 21 22 23 31. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders (DSM-III-R). 3rd revised 24 25 26 edition. Washington, DC: American Psychiatric Association, 1987. 27 28 29 32. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th edn. Washington, 30 31 32 DC: American Psychiatric Association, 1994. 33

42 on September 23, 2021 by guest. Protected copyright. 43 Health Problems (ICD-10). Geneva: World Health Organization, 1992. 44 45 46 35. Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 12 new-generation 47 48 49 antidepressants: a multiple-treatments meta-analysis. Lancet. 2009;373:746-58. 50 51 52 36. Poznanski EO, Mokros HB. Children's depression rating scale, revised (CDRS-R): manual. Western Psychological 53 54 55 Services, 1996. 56 57 58 37. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23:56-62. 59 60 19 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 20 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

42 on September 23, 2021 by guest. Protected copyright. 43 46. Brooks SP, Gelman A. General Methods for Monitoring Convergence of Iterative Simulations. J Comput Graph Stat 44 45 46 1998;7:434-55. 47 48 49 47. Lu GB, Ades AE. Assessing evidence inconsistency in mixed treatment comparisons. J Am Stat Assoc 50 51 52 2006;101:447-59. 53 54 55 48. Salanti G, Ades AE, Ioannidis JP. Graphical methods and numerical summaries for presenting results from multiple- 56 57 58 treatment meta-analysis: an overview and tutorial. J Clin Epidemiol 2011; 64:163-71. 59 60 20 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml Page 21 of 22 BMJ Open BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 21 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml BMJ Open Page 22 of 22 BMJ Open: first published as 10.1136/bmjopen-2015-007768 on 9 September 2015. Downloaded from

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 For peer review only 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

34 http://bmjopen.bmj.com/ 35 36 37 38 39 40 41

42 on September 23, 2021 by guest. Protected copyright. 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 For peer review only - http://bmjopen.bmj.com/site/about/guidelines.xhtml