Side effect profile and comparative tolerability of newer generation antidepressants in the acute treatment of major depressive disorder in children and adolescents: protocol for a systematic review and network meta-analysis

Abstract

Introduction

Major depressive disorder (MDD) is among the most common psychiatric disorders in children and adolescents. While previous meta-analyses have synthesised evidence on the efficacy and acceptability of newer-generation antidepressants in this population, specific adverse events (AEs) remain poorly characterised. This is of high clinical importance, as AEs are burdensome for patients, can reduce treatment adherence and lead to discontinuation. Here, we present a protocol for a network meta-analysis designed to evaluate the specific AE profile and comparative tolerability of newer-generation antidepressants in children and adolescents with MDD.

Methods and analysis

The planned study will include double-blind randomised controlled trials that compared one active drug with another and/or placebo for the acute treatment of MDD in children and adolescents. The following antidepressants will be considered: agomelatine, alaproclate, bupropion, citalopram, desvenlafaxine, duloxetine, edivoxetine, escitalopram, fluoxetine, fluvoxamine, levomilnacipran, milnacipran, mirtazapine, paroxetine, reboxetine, sertraline, venlafaxine, vilazodone and vortioxetine. The primary outcomes will include the number of patients experiencing at least one AE, specific non-serious AEs, serious AEs (eg, suicidal ideation) and AEs leading to treatment discontinuation. Published and unpublished studies will be retrieved through a systematic search in the following databases: PubMed, Embase, Cochrane Library (including the Cochrane Central Register of Controlled Trials), Web of Science Core Collection, PsycInfo and regulatory agencies’ registries. Study selection and data extraction will be performed independently by two reviewers. For each outcome, a network meta-analysis will be performed to synthesise all evidence. Consistency will be assessed through local and global methods, and the confidence in the evidence will be evaluated using the Confidence in Network Meta-Analysis web application. All analyses will be conducted in the R software.

Ethics and dissemination

The planned review does not require ethical approval. The findings will be published in a peer-reviewed journal and may be presented at international conferences.

PROSPERO registration number

CRD420251011399.

Correspondence to Dr Cagdas Türkmen; [email protected]

STRENGTHS AND LIMITATIONS OF THIS STUDY

• This will be the most comprehensive network meta-analysis on the safety of newer-generation antidepressants in children and adolescents with major depression, incorporating both published and unpublished data.

• Consistency will be assessed using both local and global methods, and the robustness of the results will be examined through network meta-regression.

• Antidepressant trials, particularly placebo-controlled trials, present a high risk of under-reporting and selection bias regarding adverse events.

• Study limitations will be addressed with the Cochrane Risk of Bias 2 tool, and the confidence in the evidence for network estimates of the main outcomes will be assessed with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework.

Introduction

Major depressive disorder (MDD) represents a common public health issue, with an estimated point prevalence of 8% among youth aged 10–19 years.¹ Early onset of depression is associated with multiple adverse outcomes in adulthood, including a threefold risk of depression, unemployment, elevated rates of anxiety and substance use disorders, lower educational attainment, and poorer physical health and social functioning.^{2 3} Depressive disorders are among the leading contributors to the global burden of disease among young people aged 10–24 years and their families,⁴ highlighting the need for safe, accessible and effective treatments.⁵

Newer-generation antidepressants, including selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs), are the primary pharmacologic treatment options. Current guidelines recommend SSRIs, either alone or in combination with cognitive behavioural therapy or interpersonal therapy, as the first-line treatment for moderate-to-severe MDD in youth.^{6 7} While previous network meta-analyses (NMAs) of randomised controlled trials (RCTs) suggest a modest benefit of SSRIs and SNRIs, clinicians must individually weigh both the benefits and risks of the respective antidepressant treatment.^8–10 Both SSRIs and SNRIs are associated with acute and late-emergent adverse events (AEs) in youth, which can reduce treatment adherence, increase discontinuation rates and lead to a problematic harm-benefit ratio.^8–11

While prior NMAs have identified variability in acceptability (all-cause discontinuation or discontinuation due to AEs) among antidepressants,^{8 9} specific AE profiles in youth with MDD remain poorly characterised. Among the AEs, most research has focused specifically on suicidal behaviour or ideation. Meta-analyses of both observational studies and RCTs have demonstrated an increased risk of suicide deaths, suicide attempts and suicidal ideation among adolescents and young adults taking antidepressants compared with those on placebo, whereas no increased risk has been observed in those aged 25 and older.^{12 13} Furthermore, recent network meta-analyses have consistently found an increased risk of suicidality in children and adolescents given venlafaxine, compared with those on placebo or other antidepressants.^8–10 The issue of antidepressant-related suicide risk is reflected in the US Food and Drug Administration’s boxed warning, which remains a matter of ongoing debate.^{14 15} A recent meta-analysis expanded the evidence base on AEs by examining insomnia as an AE in youth with MDD.¹⁶ This study found that SSRIs and SNRIs are associated with a modestly increased risk of treatment-emergent insomnia during acute treatment, with no difference between SSRIs and SNRIs.¹⁶ However, there was substantial variability regarding this risk among individual antidepressants.¹⁶ Expanding the line of research on specific AEs could provide a more precise understanding of the AE profile of individual antidepressants, which may improve clinical decision-making and contribute to treatment guidelines.¹⁷

The aim of the planned study—Project SOTERIA (Safety Outcomes and Tolerability: Evaluating and Revisiting the Initiation of Antidepressants)—is to evaluate the specific AE profile and comparative tolerability of newer generation antidepressants in children and adolescents with MDD.

Methods and analysis

The information outlined in this protocol is reported in accordance with the guidelines of the Preferred Reporting Items for Systematic Review and Meta-Analysis Protocols (PRISMA-P).^{18 19} The completed PRISMA-P 2015 Checklist is available in ^{online supplemental file 1}. In the case of protocol amendments, each amendment will include the date, details of the change and the rationale. The study was initiated on 1 April 2025 and is expected to be completed by 1 April 2026. The study was prospectively registered on PROSPERO (CRD420251011399).

Types of studies

We will include RCTs comparing the antidepressants of interest (see Types of interventions) with one another and/or placebo in the acute phase of antidepressant treatment (min. 6 weeks and max. 12 weeks) in a double-blind, parallel-group fashion. Only monotherapy trials will be included (ie, those in which the antidepressant is used as an add-on treatment or where additional interventions are evaluated will be excluded). Quasi-randomised trials will be excluded, while cluster randomised trials will be included. Longer-term RCTs will be considered, provided that they report data for the acute treatment period (initial 6–12 weeks). Both fixed-dose and flexible-dose designs will be included. RCTs permitting the use of any rescue/pro re nata medications will be included.

Types of participants

The population of interest comprises children and adolescents (≤18 years of age), of any gender, with a primary diagnosis of MDD. RCTs employing standard operationalised diagnostic criteria such as the Diagnostic and Statistical Manual of Mental Disorders (DSM)-III, DSM-IV, DSM-5, or the International Classification of Diseases (ICD)-10 and ICD-11 will be included. Secondary psychiatric comorbidities will not be grounds for exclusion. RCTs in which MDD is not the primary focus or is a comorbidity of a primary psychiatric condition (eg, eating disorder or attention-deficit/hyperactivity disorder) will be excluded.

Types of interventions

The planned study will evaluate the following newer-generation antidepressants: agomelatine, alaproclate, bupropion, citalopram, desvenlafaxine, duloxetine, edivoxetine, escitalopram, fluoxetine, fluvoxamine, levomilnacipran, milnacipran, mirtazapine, paroxetine, reboxetine, sertraline, venlafaxine, vilazodone and vortioxetine. Data will be obtained from both head-to-head and placebo-controlled trials, in which antidepressants are administered within their licensed dose range. ^{Figure 1} illustrates the network of all possible pairwise comparisons between the interventions. We anticipate that any patient meeting the inclusion criteria could, in principle, be randomly assigned to any of the interventions within the synthesis comparator set (transitivity assumption).

Figure 1. Network of possible pairwise comparisons between the eligible interventions.

Outcomes and categorisation of adverse events (AEs)

The tolerability of the antidepressants will be evaluated using the following outcomes:

• Total number of patients experiencing serious AEs (eg, suicidal behaviour), with separate meta-analyses conducted for each AE.

• Total number of patients experiencing AEs leading to treatment discontinuation.

• Total number of patients experiencing specific non-serious AEs (eg, somnolence, nausea), with separate meta-analyses conducted for each AE.

• Total number of patients experiencing at least one AE.

Two independent reviewers will extract all AEs reported in the trials, ensuring that no events are double-counted. The reviewers will then use the preferred terms from the Medical Dictionary for Regulatory Activities (MedDRA) (https://www.meddra.org/) to categorise each AE. Details regarding MedDRA are described elsewhere.²⁰ If different MedDRA terms are used to describe similar AEs, these will be merged into broader categories based on clinical judgement and validated by another member of the review team. Any disagreements will be resolved through consensus within the review team.

As a definition for serious AEs, we will use the classification used by the US FDA (https://www.fda.gov/):

• Resulted in death.

• Life-threatening.

• Hospitalisation (initial or prolonged).

• Disability or permanent damage.

• Congenital anomaly/birth defect.

• Required intervention to prevent permanent impairment or damage (devices).

• Other serious (important medical events), for example, seizures that did not result in hospitalisation.

All serious AEs will be included in the meta-analysis. The following variables have been selected as the most important outcomes to be evaluated, given their inclusion in FDA’s boxed warning¹⁴ and/or their high clinical relevance, and will be subject to risk of bias and certainty of evidence assessments:

• Suicidal ideation.

• Suicidal behaviour (preparatory behaviours and/or suicide attempts).

• Discontinuation due to AEs.

To determine which specific non-serious AEs should be included in the statistical analyses, we will incorporate findings on core AE outcomes from the ongoing International Network for Research Outcomes in Adolescent Depression Studies (IN-ROADS) project (https://www.in-roads.org/).²¹ Importantly, the IN-ROADS project involves youth with lived experience and parents as active collaborators, ensuring that the results are co-constructed. The protocol will be updated accordingly once the IN-ROADS results on the prioritisation of outcomes are available. However, if the findings regarding core AE outcomes are not available at the time of analysis and/or if IN-ROADS project members are not able to provide such information, we will use an alternative approach that considers both the frequency of AEs in our meta-analyses and the perceived importance of AEs based on the findings of a recent international study that aimed to identify the most important non-serious AEs of antidepressants according to adult patients’ rankings.²² We will select up to 10 AEs which are both the most common in our meta-analyses and ranked as the most important AEs by adult patients.²² Last, we will review whether AEs were assessed through spontaneous reports or systematic assessments.

Search strategy and study selection

The planned study will adopt the same search strategy as in a recent meta-analysis¹⁶ and perform an update of the search. As the previous study focused on insomnia as an AE, sleep-related search terms will be removed from the search strings. Additionally, RCTs that were previously excluded solely due to the absence of reported data on insomnia as an AE or the lack of a placebo group will be reassessed for eligibility. The following databases will be systematically searched to identify both published and unpublished RCTs from 31 August 2023 (the search end date in the previous meta-analysis¹⁶) through the date of the search: PubMed, Embase, Cochrane Library (including the Cochrane Central Register of Controlled Trials for unpublished RCTs), Web of Science Core Collection and PsycInfo. No language restrictions will be applied. Regulatory agencies’ registries (eg, US FDA) will also be searched for relevant studies and/or data. The search will be supplemented by screening the reference lists of newly published studies that meet the eligibility criteria since the recent meta-analysis.¹⁶ The search strings are provided in ^{online supplemental file 2}.

Two reviewers will independently screen the titles and abstracts of potentially relevant studies identified in the previous meta-analysis (described in the Supplementary Materials of the previous meta-analysis¹⁶) and newly published studies since the date of the last search. Full-text reports of potentially relevant studies will be retrieved and assessed for eligibility. Disagreements regarding study selection will be resolved by consensus or, if necessary, with the involvement of another member of the review team. Inter-rater agreement will be reported in terms of percentage agreement and Cohen’s kappa.²³

Record management

Records that have been identified through the systematic search will first be imported into EndNote (Clarivate, V.21, 2023)²⁴ and deduplicated using the software’s built-in functions. After deduplication, the records will be exported from EndNote to an Excel spreadsheet for formal screening. At this stage, records will be coded as either clearly irrelevant (excluded prior to full-text review based on title and/or abstract) or potentially relevant (to be assessed at the full-text review stage). Following full-text review, potentially relevant records will be coded as either included or excluded, with reasons for exclusion documented for each record.

Data extraction

Two reviewers will independently extract data from all included RCTs for the following clinical and methodological variables: authors, year of publication, diagnostic criteria, name of antidepressant and dose range, treatment duration, number of participants, age range (and mean), proportion of young people who are females, country of recruitment site, patient setting (eg, outpatient vs inpatient), race/ethnicity (when available) and funder. This information will be summarised narratively in the text and/or presented in tabular form.

Categorical data (yes/no) for the outcomes of interest (AEs) will be extracted using a predefined structured extraction sheet. Efforts will be made to obtain unpublished data on AEs from trial registries and study summaries from drug company websites, as there is strong evidence indicating that a large amount of information on AEs remains unpublished and that both the number and scope of AEs are greater in unpublished versions compared with published versions of the same study.^{16 25 26} Pharmaceutical companies will be contacted to request additional data.

As discrepancies in AE data (eg, between publications and FDA reports) may occur, we will prioritise data sources based on the following hierarchy, where applicable:

• FDA/regulatory agency reports.

• Reanalyses by independent groups (eg, reanalysis of Study 329²⁶).

• Company trial documents.

• Trial registry entries.

• Peer-reviewed publications.

Statistical synthesis of data

We expect that a quantitative synthesis will be feasible; however, if it is deemed inappropriate, we will provide a systematic narrative synthesis summarising study characteristics and findings in text and/or table(s). The narrative synthesis will evaluate both within-study and between-study results.

Pairwise meta-analyses

For pairwise comparisons informed by ≥400 participants in at least one treatment arm, a random-effects meta-analysis model will be used to obtain ORs and 95% CIs, based on the assumption that treatment effects are similar but not identical across study settings.²⁷ This sample size threshold was chosen based on power calculations indicating sufficient statistical power to detect a small effect (standardised mean difference=0.20, corresponding to an OR=1.43) with α=0.05 and β=0.20. Trials will be grouped by the class of antidepressants (eg, SSRIs, SNRIs), rather than individual antidepressants, to improve feasibility. Based on previous research, we expect that some AEs, particularly suicide-related outcomes, will be rare (ie, low counts) in some studies.^{10 16} For rare AEs, we will use the Mantel-Haenszel method, which avoids continuity corrections that might bias results, as our primary model.^{28 29} We will compare the results of the inverse variance model (which assumes a common treatment effect) and the Mantel-Haenszel method. If there are considerable discrepancies, the inverse variance results will be discarded. In further sensitivity analysis, we will also use a Bayesian model with the exact Binomial likelihood. In the case of many double-zero studies, or when the total events count in one or both treatments is zero, we will use a Mantel-Haenszel model for risk difference.³⁰ Forest plots will be presented for all pairwise meta-analyses, with a 0.5 continuity correction for studies with zero events in one treatment arm. We will visually inspect the forest plots for any heterogeneity and report the I² statistic, along with its 95% CIs, for all analyses. As an additional way of assessing heterogeneity, we will present prediction intervals.^{31 32}

Assessment of the transitivity assumption of NMA

Transitivity is a key underlying assumption of NMA.^{33 34} To assess its validity, we will examine whether study-level characteristics that could affect relative treatment effects, such as age, sex, depressive severity at baseline assessment and dosing schedule. If significant discrepancies are found, we will restrict our NMA to studies with similar distributions. Furthermore, to ensure transitivity in our network, RCTs in which MDD is not the primary focus or is a comorbidity of a primary psychiatric condition (eg, eating disorder or attention-deficit/hyperactivity disorder) will be excluded.

Network meta-analyses

If the transitivity assumption holds, we will synthesise the evidence for individual antidepressants using NMA.^34–36 For outcomes that are not rare, a random-effects frequentist NMA model will be used, which assumes a common heterogeneity parameter across all comparisons. Results will be presented in a ‘league table’ including estimated treatment effects and corresponding 95% CIs. To assess heterogeneity, we will calculate prediction intervals for each outcome and comparison.^{31 32}

For rare outcomes (ie, with zero events in a treatment arm), we will use a fixed-effects Mantel-Haenszel NMA model without continuity correction³⁷ as the primary analysis and compare the results with the inverse variance NMA model (that includes continuity correction) as a sensitivity analysis. We will perform additional sensitivity analyses using a penalised likelihood NMA model.³⁸ In disconnected networks, we will perform NMAs in each of the corresponding subnetworks that include enough data to be meaningfully synthesised.

Assessment of inconsistency

Inconsistency refers to statistical disagreement between different sources of evidence in a network, which can challenge the validity of the transitivity assumption in NMA.^34–36 To assess inconsistency, we will use two methods, namely a global method (the design-by-treatment test)³⁹ and a local method (‘Separate Indirect from Direct Evidence’).⁴⁰ The global method tests the null hypothesis of overall consistency in the network, while the local method compares direct and indirect evidence for each treatment comparison to detect ‘hot spots’ of inconsistency in the network.

If inconsistency is detected, we will initially check for data extraction errors and reassess the plausibility of transitivity, particularly in the presence of hot spots. If the cause of inconsistency cannot be identified, we will interpret the NMA results with caution. It is important to note that tests for inconsistency may have low power in detecting violations of the transitivity assumption, particularly when outcomes are rare. Therefore, we will carefully assess the transitivity assumption, even in the absence of evidence for inconsistency.

Exploring heterogeneity and inconsistency and sensitivity analyses

Given the variety of study settings we intend to include, we anticipate at least a small degree of heterogeneity and inconsistency. For all outcomes, we will examine whether treatment effects remain consistent through subgroup analyses and network meta-regression if enough data become available, considering the following factors: (1) age (children vs adolescents), (2) sex, (3) depressive severity at baseline assessment and (4) dosing schedule. If few studies are available per comparison, we will instead group all antidepressants together and conduct pairwise meta-regressions comparing antidepressants with placebo.

To assess the robustness of our conclusions, we will conduct sensitivity analyses by evaluating (1) only studies with unpublished data (excluding those with only published data) and (2) only studies without a high risk of bias for the respective outcome.

Assessment of reporting bias

Antidepressant trials, particularly placebo-controlled ones, present a high risk of reporting bias (bias due to missing evidence), especially in the reporting of AEs.^41–45 The risk of reporting bias and its impact on the estimated relative treatment effects in the NMA will be evaluated with the Risk Of Bias due to Missing Evidence in Network meta-analysis (ROB-MEN) framework, which involves several statistical and conceptual considerations.^{46 47} ROB-MEN offers a structured approach tailored to NMA and will be applied to the three most important outcomes: suicidal ideation, suicidal behaviour and discontinuation due to AEs. For the classic pairwise comparisons between antidepressant classes, we will use the Risk Of Bias due to Missing Evidence (ROB-ME) tool to assess bias due to missing evidence for the same outcomes.⁴⁸

Model implementation

All models will be fitted in the R software using the meta⁴⁹ package for the pairwise meta-analysis models, the netmeta⁵⁰ package for the frequentist NMAs, and the crossnma package for network meta-regression.⁵¹

Risk of bias

Two reviewers will independently assess the risk of bias for the three most important outcomes (suicidal ideation, suicidal behaviour and discontinuation due to AEs) in each study using the Cochrane Risk of Bias 2 (RoB 2) tool.⁵² The following domains will be assessed:

• Bias arising from the randomisation process.

• Bias due to deviations from the intended interventions.

• Bias due to missing outcome data.

• Bias in the measurement of the outcome.

• Bias in the selection of the reported result.

Each domain will be rated as ‘low risk of bias’, ‘some concerns’ or ‘high risk of bias’. The overall risk of bias for each outcome will be assigned according to the least favourable rating among the domains. The assessments will be managed using the RoB 2 Excel tool. Any disagreements in RoB 2 ratings will be resolved by consensus or, if necessary, with the involvement of another member of the review team. Inter-rater agreement will be reported in terms of percentage agreement and Cohen’s kappa.²³

Certainty of evidence of NMA

Two reviewers will independently assess the certainty of evidence for the three most important outcomes (suicidal ideation, suicidal behaviour and discontinuation due to AEs) using a validated framework.⁵³ We will apply a margin of equivalence of 0.95–1.05 for the ORs of the three most important outcomes (suicidal ideation, suicidal behaviour and discontinuation due to AEs).⁵⁴ This narrow margin was selected due to the clinical importance of even small differences in these outcomes. The Confidence in Network Meta-Analysis^{54 55} web application will be used to complete the assessments. The certainty in the body of evidence will be rated as high, moderate, low or very low. Justifications will be provided for decisions to downgrade or upgrade the certainty of the evidence, as well as for the importance rating of each outcome.

Discussion

AEs that emerge during the acute antidepressant treatment pose a significant clinical challenge; they may reduce treatment efficacy, worsen treatment adherence and increase the risk of discontinuation.^8–11 While prior studies have primarily assessed serious AEs, specifically suicidal behaviour or ideation,^{8 10 12 56} more precise understanding of other (non-serious) AEs is significantly lacking. Such an understanding may help clinicians better balance the risks and benefits of antidepressant treatments, ultimately improving patient care. This study aims to comprehensively evaluate AEs associated with newer-generation antidepressants, which may help tailor antidepressant treatment to the specific needs of children and adolescents with MDD. For example, in situations where multiple antidepressants demonstrate comparable efficacy, tolerability profiles may be used to guide treatment selection. Importantly, we are striving to integrate lived experience perspectives into this NMA by drawing on findings from the IN-ROADS project,²¹ which actively involves youth with lived experience and parents as partners in co-constructing core outcome priorities. Finally, by refining risk assessment and optimising prescribing, these findings have the potential to improve clinical decision-making and refine clinical guidelines for the treatment of children and adolescents with MDD.

Ethics and dissemination

The planned review does not require ethical approval. The findings will be published in a peer-reviewed journal and may be presented at international conferences.

Footnote

Contributors CT, the guarantor of the review, conceived the study and drafted the protocol. All other authors revised and edited the first draft of the protocol. All authors approved the final version of the protocol to be published.

Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

Competing interests JK has received a speaker fee from Janssen Pharmaceuticals. TAF reports personal fees from Boehringer-Ingelheim, Daiichi Sankyo, DT Axis, Micron, Shionogi, SONY and UpToDate, and a grant from DT Axis and Shionogi, outside the submitted work; in addition, TAF has a patent 7448125 and a pending patent 2022-082495, and has licensed intellectual properties for Kokoro-app to DT Axis. ACi has received research, educational and consultancy fees from INCiPiT (Italian Network for Paediatric Trials), CARIPLO Foundation, Lundbeck and Angelini Pharma. RAS has received consultancy or speakers fees from Janssen Pharmaceuticals, Clexio Biosciences, GH research, participated in industry sponsored clinical trials from Compass Pathways and Novartis, and received an investigator initiated grant for real world data project from J&J, all outside the submitted work. JRS has received research support from the National Institutes of Health, MindMed and AbbVie. He has also received material support from Myriad Genetics. Additionally, he receives royalties from Springer Publishing and Cambridge University Press, honoraria from the Neuroscience Education Institute, and serves as an author for UpToDate. JRS has consulted for MindMed, AbbVie (Cerevel), Alkermes, Otsuka, Supernus, Vistagen and Genomind. GJE receives research support from the American Foundation for Suicide Prevention, Janssen Research and Development, LLC, National Institutes of Health, Patient-Centered Research Outcomes Institute (PCORI), and the State of Texas. The other authors have no conflicts of interest to disclose.

Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ Shorey S, Ng ED, Wong CHJ. Global prevalence of depression and elevated depressive symptoms among adolescents: A systematic review and meta-analysis. Br J Clin Psychol 2022;61:287–305. doi:10.1111/bjc.12333

² Clayborne ZM, Varin M, Colman I. Systematic Review and Meta-Analysis: Adolescent Depression and Long-Term Psychosocial Outcomes. J Am Acad Child Adolesc Psychiatry 2019;58:72–9. doi:10.1016/j.jaac.2018.07.896

³ Schlechter P, Ford T, Neufeld SAS. Depressive symptom networks in the UK general adolescent population and in those looked after by local authorities. BMJ Ment Health 2023;26:e300707. doi:10.1136/bmjment-2023-300707

⁴ Martin F, Dahmash D, Wicker S, et al. Psychological well-being and needs of parents and carers of children and young people with mental health difficulties: a quantitative systematic review with meta-analyses. BMJ Ment Health 2024;27:e300971. doi:10.1136/bmjment-2023-300971

⁵ Cortese S, Moreno C. Advancing the evidence base for child and adolescent psychopharmacology. BMJ Ment Health 2025;28:e301634. doi:10.1136/bmjment-2025-301634

⁶ Jane Garland E, Kutcher S, Virani A, et al. Update on the Use of SSRIs and SNRIs with Children and Adolescents in Clinical Practice. J Can Acad Child Adolesc Psychiatry 2016;25:4–10.

⁷ Walter HJ, Abright AR, Bukstein OG, et al. Clinical Practice Guideline for the Assessment and Treatment of Children and Adolescents With Major and Persistent Depressive Disorders. J Am Acad Child Adolesc Psychiatry 2023;62:479–502. doi:10.1016/j.jaac.2022.10.001

⁸ Zhou X, Teng T, Zhang Y, et al. Comparative efficacy and acceptability of antidepressants, psychotherapies, and their combination for acute treatment of children and adolescents with depressive disorder: a systematic review and network meta-analysis. Lancet Psychiatry 2020;7:581–601. doi:10.1016/S2215-0366(20)30137-1

⁹ Cipriani A, Zhou X, Del Giovane C, et al. Comparative efficacy and tolerability of antidepressants for major depressive disorder in children and adolescents: a network meta-analysis. Lancet 2016;388:881–90. doi:10.1016/S0140-6736(16)30385-3

¹⁰ Hetrick SE, McKenzie JE, Bailey AP, et al. New generation antidepressants for depression in children and adolescents: a network meta-analysis. Cochrane Database Syst Rev 2021;5:Cd013674. doi:10.1002/14651858.CD013674.pub2

¹¹ Strawn JR, Mills JA, Poweleit EA, et al. Adverse Effects of Antidepressant Medications and their Management in Children and Adolescents. Pharmacotherapy 2023;43:675–90. doi:10.1002/phar.2767

¹² Stone M, Laughren T, Jones ML, et al. Risk of suicidality in clinical trials of antidepressants in adults: analysis of proprietary data submitted to US Food and Drug Administration. BMJ 2009;339:b2880. doi:10.1136/bmj.b2880

¹³ Barbui C, Esposito E, Cipriani A. Selective serotonin reuptake inhibitors and risk of suicide: a systematic review of observational studies. CMAJ 2009;180:291–7. doi:10.1503/cmaj.081514

¹⁴ Fornaro M, Anastasia A, Valchera A, et al. The FDA “Black Box” Warning on Antidepressant Suicide Risk in Young Adults: More Harm Than Benefits? Front Psychiatry 2019;10:294. doi:10.3389/fpsyt.2019.00294

¹⁵ Stone MB. In Search of a Pony: Sources, Methods, Outcomes, and Motivated Reasoning. Med Care 2018;56:375–81. doi:10.1097/MLR.0000000000000895

¹⁶ Türkmen C, Machunze N, Lee AM, et al. Systematic Review and Meta-Analysis: The Association Between Newer-Generation Antidepressants and Insomnia in Children and Adolescents With Major Depressive Disorder. J Am Acad Child Adolesc Psychiatry 2025. doi:10.1016/j.jaac.2025.01.006

¹⁷ Stringaris A, Burman C, Delpech R, et al. Comparing apples and oranges in youth depression treatments? A quantitative critique of the evidence base and guidelines. BMJ Ment Health 2025;28:e301162. doi:10.1136/bmjment-2024-301162

¹⁸ Moher D, Shamseer L, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Syst Rev 2015;4:1. doi:10.1186/2046-4053-4-1

¹⁹ Shamseer L, Moher D, Clarke M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015: elaboration and explanation. BMJ 2015;350:g7647. doi:10.1136/bmj.g7647

²⁰ Tomlinson A, Efthimiou O, Boaden K, et al. Side effect profile and comparative tolerability of 21 antidepressants in the acute treatment of major depression in adults: protocol for a network meta-analysis. Evid Based Mental Health 2019;22:61–6. doi:10.1136/ebmental-2019-300087

²¹ Monga S, Monsour A, Stallwood E, et al. Core Outcome Set Development for Adolescent Major Depressive Disorder Clinical Trials: A Registered Report. J Am Acad Child Adolesc Psychiatry 2020;59:1297–8. doi:10.1016/j.jaac.2020.07.905

²² Chevance A, Tomlinson A, Ravaud P, et al. Important adverse events to be evaluated in antidepressant trials and meta-analyses in depression: a large international preference study including patients and healthcare professionals. Evid Based Ment Health 2022;25:e41–8. doi:10.1136/ebmental-2021-300418

²³ Cohen J. A Coefficient of Agreement for Nominal Scales. Educ Psychol Meas 1960;20:37–46. doi:10.1177/001316446002000104

²⁴ EndNote [program]. Clarivate; 2023.

²⁵ Golder S, Loke YK, Wright K, et al. Reporting of Adverse Events in Published and Unpublished Studies of Health Care Interventions: A Systematic Review. PLoS Med 2016;13:e1002127. doi:10.1371/journal.pmed.1002127

²⁶ Le Noury J, Nardo JM, Healy D, et al. Restoring Study 329: efficacy and harms of paroxetine and imipramine in treatment of major depression in adolescence. BMJ 2015;351:h4320. doi:10.1136/bmj.h4320

²⁷ Borenstein M, Hedges LV, Higgins JPT, et al. A basic introduction to fixed-effect and random-effects models for meta-analysis. Res Synth Methods 2010;1:97–111. doi:10.1002/jrsm.12

²⁸ MANTEL N, HAENSZEL W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719–48.

²⁹ Efthimiou O. Practical guide to the meta-analysis of rare events. Evid Based Ment Health 2018;21:72–6. doi:10.1136/eb-2018-102911

³⁰ Xu C, Furuya-Kanamori L, Zorzela L, et al. A proposed framework to guide evidence synthesis practice for meta-analysis with zero-events studies. J Clin Epidemiol 2021;135:70–8. doi:10.1016/j.jclinepi.2021.02.012

³¹ Borenstein M, Higgins JP, Hedges LV, et al. Basics of meta-analysis: I(2) is not an absolute measure of heterogeneity. Res Synth Methods 2017;8:5–18. doi:10.1002/jrsm.1230

³² IntHout J, Ioannidis JPA, Rovers MM, et al. Plea for routinely presenting prediction intervals in meta-analysis. BMJ Open 2016;6:e010247. doi:10.1136/bmjopen-2015-010247

³³ Efthimiou O, Debray TPA, van Valkenhoef G, et al. GetReal in network meta-analysis: a review of the methodology. Res Synth Methods 2016;7:236–63. doi:10.1002/jrsm.1195

³⁴ Mavridis D, Giannatsi M, Cipriani A, et al. A primer on network meta-analysis with emphasis on mental health. Evid Based Ment Health 2015;18:40–6. doi:10.1136/eb-2015-102088

³⁵ Cipriani A, Higgins JPT, Geddes JR, et al. Conceptual and technical challenges in network meta-analysis. Ann Intern Med 2013;159:130–7. doi:10.7326/0003-4819-159-2-201307160-00008

³⁶ Salanti G. Indirect and mixed-treatment comparison, network, or multiple-treatments meta-analysis: many names, many benefits, many concerns for the next generation evidence synthesis tool. Res Synth Methods 2012;3:80–97. doi:10.1002/jrsm.1037

³⁷ Efthimiou O, Rücker G, Schwarzer G, et al. Network meta-analysis of rare events using the Mantel-Haenszel method. Stat Med 2019;38:2992–3012. doi:10.1002/sim.8158

³⁸ Evrenoglou T, White IR, Afach S, et al. Network meta-analysis of rare events using penalized likelihood regression. Stat Med 2022;41:5203–19. doi:10.1002/sim.9562

³⁹ Higgins JPT, Jackson D, Barrett JK, et al. Consistency and inconsistency in network meta‐analysis: concepts and models for multi‐arm studies. Res Synth Methods 2012;3:98–110. doi:10.1002/jrsm.1044

⁴⁰ Dias S, Welton NJ, Caldwell DM, et al. Checking consistency in mixed treatment comparison meta-analysis. Stat Med 2010;29:932–44. doi:10.1002/sim.3767

⁴¹ Cipriani A, Furukawa TA, Salanti G, et al. Comparative efficacy and acceptability of 21 antidepressant drugs for the acute treatment of adults with major depressive disorder: a systematic review and network meta-analysis. The Lancet 2018;391:1357–66. doi:10.1016/S0140-6736(17)32802-7

⁴² Lancee M, Lemmens CMC, Kahn RS, et al. Outcome reporting bias in randomized-controlled trials investigating antipsychotic drugs. Transl Psychiatry 2017;7:e1232–e32. doi:10.1038/tp.2017.203

⁴³ Schroll JB, Penninga EI, Gøtzsche PC. Assessment of Adverse Events in Protocols, Clinical Study Reports, and Published Papers of Trials of Orlistat: A Document Analysis. PLoS Med 2016;13:e1002101. doi:10.1371/journal.pmed.1002101

⁴⁴ Turner EH, Matthews AM, Linardatos E, et al. Selective publication of antidepressant trials and its influence on apparent efficacy. N Engl J Med 2008;358:252–60. doi:10.1056/NEJMsa065779

⁴⁵ Rolland P, Jutel A, Douget L, et al. Incomplete reporting of adverse events in duloxetine trials: a meta-research survey of randomized controlled trials vs placebo. J Clin Epidemiol 2025;180:111677. doi:10.1016/j.jclinepi.2025.111677

⁴⁶ Chiocchia V, Nikolakopoulou A, Higgins JPT, et al. ROB-MEN: a tool to assess risk of bias due to missing evidence in network meta-analysis. BMC Med 2021;19:304. doi:10.1186/s12916-021-02166-3

⁴⁷ Chiocchia V, Holloway A, Salanti G. Semi-automated assessment of the risk of bias due to missing evidence in network meta-analysis: a guidance paper for the ROB-MEN web-application. BMC Med Res Methodol 2023;23:223. doi:10.1186/s12874-023-02038-9

⁴⁸ Page MJ, Sterne JAC, Boutron I, et al. ROB-ME: a tool for assessing risk of bias due to missing evidence in systematic reviews with meta-analysis. BMJ 2023;383:e076754. doi:10.1136/bmj-2023-076754

⁴⁹ Schwarzer G. meta: An R package for meta-analysis, 7.2007:40–5.

⁵⁰ Balduzzi S, Rücker G, Nikolakopoulou A, et al. netmeta: An R Package for Network Meta-Analysis Using Frequentist Methods. J Stat Softw 2023;106:40. doi:10.18637/jss.v106.i02

⁵¹ Hamza T, Schwarzer G, Salanti G. crossnma: An R package to synthesize cross-design evidence and cross-format data using network meta-analysis and network meta-regression. BMC Med Res Methodol 2024;24:169. doi:10.1186/s12874-023-02130-0

⁵² Sterne JAC, Savović J, Page MJ, et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ 2019;366:l4898. doi:10.1136/bmj.l4898

⁵³ Salanti G, Del Giovane C, Chaimani A, et al. Evaluating the quality of evidence from a network meta-analysis. PLoS One 2014;9:e99682. doi:10.1371/journal.pone.0099682

⁵⁴ Nikolakopoulou A, Higgins JPT, Papakonstantinou T, et al. CINeMA: An approach for assessing confidence in the results of a network meta-analysis. PLoS Med 2020;17:e1003082. doi:10.1371/journal.pmed.1003082

⁵⁵ Papakonstantinou T, Nikolakopoulou A, Higgins JPT, et al. CINeMA: Software for semiautomated assessment of the confidence in the results of network meta-analysis. Campbell Syst Rev 2020;16:e1080. doi:10.1002/cl2.1080

⁵⁶ Hammad TA, Laughren T, Racoosin J. Suicidality in pediatric patients treated with antidepressant drugs. Arch Gen Psychiatry 2006;63:332–9. doi:10.1001/archpsyc.63.3.332