Prenatal Cannabis and Tobacco Together Tied to Lower Cognition Scores

Children exposed before birth to both cannabis and tobacco scored modestly lower on a composite measure of reasoning, attention, memory and processing speed than children exposed to cannabis alone or to neither substance, according to a study published online November 2 in the journal Addiction.

The analysis drew on baseline data from the Adolescent Brain Cognitive Development study, a long-running U.S. cohort, and included 8,706 children aged 9 to 11 whose cognitive scores were measured with the National Institutes of Health Toolbox Cognition Battery. Neither substance on its own produced a statistically significant association with fluid cognition in the main analysis; the finding that held was the interaction between the two.

What the researchers compared

The study sorted children into four groups based on retrospective reports from their biological mothers or caregivers: 236 exposed to both cannabis and tobacco, 195 to cannabis only, 803 to tobacco only, and 7,472 to neither. Exposure counted if a mother reported use either before or after learning she was pregnant.

Fluid cognition — the capacity to solve novel problems, hold and manipulate information, and shift attention — was scored from five subtests covering cognitive flexibility, attention, memory and processing speed. Crystallized cognition, which reflects accumulated knowledge such as vocabulary and reading skill, came from two subtests.

After adjustment for factors including alcohol use during pregnancy, family income, maternal age and education, pregnancy planning and parental mental health symptoms, children exposed to cannabis only scored about 2.1 points higher on fluid cognition than co-exposed children, and children exposed to neither substance scored about 1.4 points higher. The tobacco-only group scored about 1.3 points higher, a difference the authors described as not statistically significant. Crystallized cognition showed no meaningful differences between groups.

An interaction, not two separate harms

The pattern is worth reading carefully, because it complicates the way prenatal substance exposure is usually discussed. In the main-effects models, neither cannabis nor tobacco exposure was independently associated with lower fluid cognition. The signal appeared only when the two were combined, and the interaction term was the study's central question.

That structure matters for polysubstance exposure, which is the rule rather than the exception among pregnant people who use drugs. A clinical message built on one substance at a time — stop vaping, or stop smoking cigarettes — does not describe what the data show. The authors also reported that the group with combined exposure had the youngest mothers, the fewest years of education and the highest scores for parental mental health symptoms, a reminder that co-exposure travels alongside other stressors rather than in isolation.

How much weight the finding carries

The effect sizes were small. The authors characterized the differences in fluid cognition as modest, and the interaction estimate itself was uncertain, narrowly clearing the threshold for statistical significance. With 236 children in the co-exposure group, the analysis had limited precision, which is also why the tobacco-only comparison fell short of significance despite a similar point estimate.

The study is a cross-sectional exploratory analysis of one wave of data, so it cannot establish that prenatal exposure caused the difference in scores, only that the two travel together after adjustment. Exposure was measured by maternal report years after pregnancy, which allows for recall error in both directions. And the cohort's baseline data were collected between 2016 and 2018, a period before the sharp rise in high-potency cannabis concentrates and vape products that now dominate the legal market — a limitation the authors do not need to state for readers to notice.

What it adds to prenatal care

The practical argument for tracking both substances is that prenatal visits already ask about tobacco, and increasingly ask about cannabis, but rarely frame them as a pair with a shared developmental question. Recommendations to quit smoking in pregnancy are standard; guidance on cannabis has been vaguer, in part because evidence on its own effects has been mixed.

Children in this cohort were 9 to 11 years old when their cognition was assessed, which places the finding in middle childhood rather than infancy — the stage where differences in executive function start to show up in school performance. For clinicians working with pregnant patients who use both substances, the study supports treating the combination as the exposure of interest and referring to substance use disorder treatment that addresses nicotine and cannabis together rather than sequentially.

The authors conclude that co-exposure is associated with modestly lower fluid cognition and call for longitudinal work that can test whether the association persists as the cohort ages. The ABCD study is designed to follow these children into early adulthood, which makes that follow-up a matter of time rather than a new study.