Association between prenatal cannabis and tobacco co‐exposure and cognition in childhood: Insights from the US adolescent brain cognitive development study

BACKGROUND AND AIMS: While evidence on the individual effects of prenatal cannabis and tobacco exposure on offspring cognition is accumulating, their combined impact remains understudied. Therefore, we examined the interaction between prenatal cannabis and tobacco exposure on fluid (e.g. reasoning abilities) and crystallized (e.g. acquired knowledge) cognition in middle childhood. We also compared cognition in children with prenatal co-exposure to those with cannabis-only exposure, tobacco-only exposure and no exposure. DESIGN: Cross-sectional exploratory analysis of baseline data from an ongoing longitudinal cohort study. SETTING: Data collected between 2016 and 2018 from the United States-based Adolescent Brain Cognitive Development study were extracted. PARTICIPANTS: Children (n = 8706), aged 9-11 years old, were parsed into four prenatal exposure groups: co-exposure (n = 236); cannabis only (n = 195); tobacco only (n = 803); and no exposure (n = 7472). MEASUREMENTS: Prenatal co-exposure was measured as a binary variable (yes/no) based on maternal/caregiver retrospective report. Fluid and crystallized scores were derived from the National Institutes of Health (NIH) Toolbox cognitive battery; scores ranged from 44 to 131 and 51115 for fluid and crystallized cognition, respectively. We employed linear mixed-effects models to examine interactions between cannabis and tobacco exposure and group differences in fluid and crystallized scores, adjusting for covariates selected a priori using a Directed Acyclic Graph. RESULTS: For fluid cognition, a statistically significant interaction emerged between prenatal cannabis and tobacco exposure [beta coefficient (β) = -2.08, 95% confidence interval (CI) = -4.13 to -0.03]. Cannabis only (β = 2.13, 95% CI = 0.21-4.05) and unexposed children (β = 1.39, 95% CI = 0.03-2.76) had statistically significantly higher fluid scores than co-exposed children, while the difference with tobacco-only exposed children was not statistically significant (β = 1.34, 95% CI = -0.14 to 2.82). No statistically significant interaction was observed between prenatal co-exposure and crystallized cognition (β = -0.62, 95% CI = -1.87 to 0.64), and crystallized scores did not statistically significantly differ between the co-exposed group and the cannabis-only (β = 0.66, 95% CI = -0.51 to 1.84), tobacco-only (β = 0.19, 95% CI = -0.71 to 1.1) or unexposed groups (β = 0.24, 95% CI = -0.6 to 1.08). SUMMARY: Prenatal co-exposure to cannabis and tobacco appears to be associated with poorer fluid cognition among children, but not poorer crystallized cognition, compared with cannabis-only and unexposed groups. These findings should be interpreted with caution given the potential for residual confounding by shared genetic liability, substance exposure misclassification and selection bias. The cross-sectional design of this study precludes causal inference.