Optimizing Pharmacological Treatment in Children and Adolescents With Major Depressive Disorder: An Individual...

The specific influence of individual characteristics on antidepressant efficacy and tolerability in pediatric major depressive disorder (MDD) remains unclear. This study aimed to characterize subgroup-specific treatment efficacy and to establish a preliminary scientific foundation for precision psychiatry in pediatric MDD management based on available data.

An individual participant data (IPD) analysis of double-blind randomized trials (RCTs) was conducted to compare antidepressants with placebo for the acute treatment of MDD in children and adolescents. Subgroup analyses identified treatment outcome differences by participant characteristics, with mixed-effects models used to examine effect modifiers. Primary outcomes were efficacy (depressive symptom change, mean difference [MD]) and tolerability (adverse event-related withdrawals, odds ratio [OR]). Secondary outcomes included deterioration and suicide-related outcomes. The protocol is registered with PROSPERO (CRD42016051657).

IPD from 10 of 40 eligible RCTs (2,584 participants) on 5 antidepressants (fluoxetine [n = 475], imipramine [n = 95], paroxetine [n = 413], desvenlafaxine [n = 358], duloxetine [n = 341]), and placebo [n = 902] were analyzed. Most trials had "some concerns" for risk of bias. There was also substantial missing data for key covariates (eg, 53% for family history of psychiatric disorders). Inconsistencies in the reporting and collection of suicide-related outcomes (ideation and attempts) limited our analyses. In the individual participant meta-analysis, subgroup analysis showed that fluoxetine was the only agent significantly more effective than placebo (MD = -2.43). Imipramine (OR = 5.43) and paroxetine (OR = 1.85) showed poor tolerability. In terms of subgroup analyses based on individual participant characteristics, fluoxetine demonstrated superior efficacy in children, female participants, White participants, healthy weight participants, and those with severe symptoms (MDs = -3.90 to -2.60). Fluoxetine also showed less deterioration in female participants (OR = 0.55, 95% CI = 0.37 to 0.83). No significant differences in suicide-related outcomes were observed across subgroups based on individual characteristics.

These provide a preliminary understanding of one possible precision medicine framework for pediatric MDD, indicating that individual characteristics (age, sex, ethnicity, body mass index [BMI], symptom severity, family history) may influence treatment outcomes. Exploratory subgroup analyses suggest fluoxetine's efficacy may be more pronounced in certain populations: children, female participants, White individuals, healthy weight participants, and those with severe symptoms. These findings should be interpreted with caution as IPD was only available from a quarter of eligible clinical trials, introducing data availability bias. These findings also require further validation in more diverse cohorts. Universal suicide-related outcome assessment across all subgroups (regardless of individual characteristics) remains essential. Broader data sharing is needed to improve representativeness in future IPD analyses.

Comparative efficacy and tolerability of new-generation antidepressants for major depressive disorder in children and adolescents: protocol of an individual patient data meta-analysis; https://www.crd.york.ac.uk/PROSPERO/view/CRD42016051657.