Experiencias adversas en la infancia y desenlaces relacionados con el uso de sustancias en adolescentes y adultos: revisión sistematizada con síntesis narrativa estructurada y metaanálisis exploratorio

Abstract

Introduction: Adverse childhood experiences (ACEs), including childhood trauma, have been associated with multiple substance-related outcomes. Heterogeneity in exposures, populations, metrics, and potentially dependent publications limits a single global estimate and supports a domain-based synthesis. Objective: To critically synthesize observational evidence on adverse childhood experiences and substance use disorder, initiation, use patterns, relapse, craving, severity, and psychiatric comorbidity in adolescents and adults. Methods: A systematized review of publications from January 1, 2016 through August 17, 2026 was conducted using PECO criteria, DOI-level auditing, control of dependence across publications, methodological appraisal, and domain-based narrative synthesis. The audit identified 200 unique DOIs; 187 publications were retained in the evidence map and 13 were excluded. The analytical corpus comprised 75 reports with documented and audited methodological appraisal. No literature-wide pooled effect was estimated. A secondary exploratory meta-analysis combined three independent samples from Kristjansson et al. (2016) using REML and modified Knapp-Hartung inference. Results: Risk of bias in the analytical corpus was moderate in 39 reports, low-moderate in 21, moderate-high in 12, and high in 3. Positive associations predominated for substance use disorder diagnosis, earlier initiation, use patterns, and severity, while null and attenuated findings after adjustment were also retained. The exploratory meta-analysis showed positive associations for alcohol, nicotine, cannabis, stimulants, and opiates. Conclusion: Adverse childhood experiences are a relevant marker of vulnerability to substance-related outcomes, but clinical and methodological heterogeneity argues against a single global effect and supports domain-specific interpretation.