The exponential growth of human interactions on social media via the internet has revolutionized global communication, but it has also emerged as a critical factor in mental health linked to suicidal ideation and mental disorders. This systematic review and meta-analysis aimed to synthesize evidence on the most prevalent disorders in adolescents and young adults associated with social media use based on previous research, highlighting risk factors and key findings. Publications from 2020 to 2024 in highly relevant databases were reviewed following the PRISMA declaration guidelines. The meta-analysis (conducted in R software) of the included documents (24 studies, 68 effects) verified a significant and positive association between exposure to risk factors in social networks and various disorders in adolescents and young adults (aggregate correlation r = 0.2173, 95% CI [0.1826, 0.2520], p ≤ 0.0001), although with high heterogeneity (I² = 99.66%). Prevention strategies were indicated by revealing data from contexts in which 40% of adolescents who died by suicide had developed online identities focused on suicidal thoughts.

Background: Problematic digital use in adolescents and young adults has emerged as a relevant exposure for mental health and neurobiological vulnerability. However, its relationship with peripheral DNA methylation differences in neurobiologically relevant genes remains poorly synthesized. Objective: This systematic review and exploratory multilevel meta-analysis aimed to synthesize available scientific evidence on associations between phenotypes of problematic internet or smartphone use and DNA methylation differences in neurobiologically relevant genes, and to quantitatively explore the average magnitude of these associations from available data. Methods: Primary human studies were identified through a search strategy with no initial publication-year restriction; the eligible studies retrieved and included in the review were published between 2018 and 2024. The review followed PRISMA criteria and a broad, predefined eligibility strategy to capture both direct and indirect evidence. Seven studies were included in the systematic review; four provided quantifiable data for statistical synthesis, from which 41 effect sizes were harmonized. Analyses were performed in R using multilevel meta-analysis and meta-regression models with REML estimation, supplemented by sensitivity analyses with robust variance estimation. Results: The main meta-analysis revealed a significant positive association between problematic digital exposure and DNA methylation differences (Fisher’s z = 0.2957; r = 0.287; 95% CI [0.1146, 0.4769]; p = 0.0020), with high heterogeneity (total I² = 93.15%). Meta-regression indicated that the type of effect and the biological direction of the finding explained a substantial proportion of the observed heterogeneity (total R² ≈ 97.4%). Significance: These findings should be interpreted as exploratory evidence of a biologically plausible association between problematic digital exposure and peripheral DNA methylation differences in neurobiologically relevant genes, rather than as causal or clinically validated biomarker evidence. The available evidence remains methodologically heterogeneous, preliminary, and dependent on a small number of independent studies, underscoring the need for larger, longitudinal, and more methodologically rigorous research in this field.