What Tobacco Does During Pregnancy to the Teen Brain: A Breakdown of the Large ABCD Study

Can the effects of smoking during pregnancy be seen ten years later in the child's brain scans? In a large longitudinal study of 5,417 children (the ABCD program), scientists tracked how the thickness and area of the cortex changed from ages 9-10 to 11-12, and compared the trajectories of those who were exposed to tobacco and/or alcohol in utero. The conclusion is stark and fairly clear: prenatal exposure to tobacco is associated with faster thinning of the cortex, primarily in the frontal areas, while no significant links were found for alcohol in this age range. Thinning itself is a normal part of brain maturation in adolescence, but in those "exposed to tobacco" it occurs faster and/or earlier, and this correlates with external behavioral problems and sleep disorders.

Background

Prenatal exposure to tobacco (PTE) and alcohol (PAE) are among the most common and preventable risk factors for the developing brain. Nicotine and other components of tobacco smoke readily cross the placenta, affecting vascular tone, fetal oxygen supply, and the formation of neural networks. Ethanol is a recognized teratogen; at high doses, it causes a spectrum of fetal alcohol disorders (FASD), and at lower doses, more subtle neurocognitive and behavioral effects. Clinical observations have long associated PTE/PAE with a higher risk of behavioral problems, attention deficits, sleep disturbances, and learning difficulties in children and adolescents.

The cerebral cortex normally undergoes a “restructuring” during puberty: the cortex gradually becomes thinner (synaptic “pruning” and myelination occur), and the cortex area changes unevenly across regions. Therefore, longitudinal data are especially important for assessing the influence of prenatal factors - not just “how much thinner the cortex is at a certain age,” but how its thickness and area change over time. Previously, many studies were cross-sectional, with small samples and mixed exposures (tobacco and alcohol together), which made it difficult to separate the contribution of each factor and understand whether the “normal” adolescent trajectory is accelerating or its start is shifting in time.

An additional methodological challenge is assessing the exposure itself: post hoc surveys of mothers are more common, rarely confirming the data with biomarkers (e.g., cotinine). Associated factors also play a role: socioeconomic status, parental mental health, use of other substances, passive smoking. All this requires large, representative cohorts with multiple MRI examinations, standardized image processing, and rigorous statistics adjusted for multiple comparisons.

This is the niche that the data from the ABCD project fills - the world's largest longitudinal study of brain development, where thousands of children undergo repeated MRI, cognitive and behavioral tests. On such an array, it is possible to separate the effects of PTE and PAE, look at region-specific trajectories of cortex thickness/area in the key age window of 9-12 years and link them with external manifestations - impulsivity, behavioral symptoms, sleep quality. The practical motivation is obvious: if traces of intrauterine tobacco manifest themselves as an acceleration of cortical changes specifically in the frontal areas, this is an argument in favor of strict smoking cessation programs when planning and during pregnancy, as well as a reason to purposefully monitor the sleep and behavior of children with confirmed PTE. Theoretically, such findings are consistent with the idea of "accelerated biological/epigenetic aging" with tobacco exposure and with the effects of tobacco toxicants on microglia and synaptic pruning - hypotheses that require further testing in long-term follow-up.

Who and how was studied

• Cohort: 5417 ABCD participants (21 US centers). Average age at baseline was 9.9 years; mean age at follow-up was 11.9 years. Approximately 2 years between visits.
• Exposures: Prenatal alcohol exposure (PAE) and tobacco exposure (PTE) as assessed by caregiver surveys - before and after recognition of pregnancy.
• Outcomes: cortical thickness and area in 68 partial zones, behavioral scales (CBCL, BIS/BAS, UPPS), sleep disturbance scale. Analysis - with correction for multiple comparisons (FDR).

First, the developmental norm. On average, all children have thinner cortexes with age, and the area in different zones can grow or shrink - these are natural trajectories of adolescent maturation. Against this background, the researchers looked at whether these trajectories changed in children with PAE/PTE relative to their "unaffected" peers.

Key Results

• Alcohol: at 9-12 years, no significant associations were found with either cortical thickness/area or their change over time.
• Tobacco:
  • already “on the cut” - the cortex is thinner in the paramedian areas (bilateral parahippocampal cortex, left lateral orbitofrontal cortex; partial r≈0.04, P <0.001, FDR correction);
  • over time - more rapid thinning in 11 frontal and 2 temporal regions (including bilateral rostral middle frontal, superior frontal, medial orbitofrontal, rostral anterior cingulate; right pars orbitalis and pars triangularis, etc.; |r|≈0.04, P <0.001).
• Behavior: the faster the thinning, the higher the scores for externalizing behavior problems, impulsivity (negative urgency), fun-seeking, and sleep disturbances - the associations are weak but reproducible (typically |r|≈0.03-0.05), and mainly in children with PTE.

The authors interpret these connections cautiously: perhaps it is not just a matter of “faster” thinning, but an earlier start of the same processes – conventionally, “the curve is shifted to the left.” This is supported by the literature on accelerated epigenetic aging in people with prenatal/current tobacco exposure and the relationship between DNA methylation and cortical thickness and synaptic “pruning.” This is still a hypothesis, but it explains why tobacco gives a broader and more persistent signal than alcohol in early adolescence.

What is important for practice and policy

• There is no safe level of tobacco exposure in pregnancy. The results support aggressive smoking cessation programs when planning and in early pregnancy - the effect of PTE is broader and more stable than PAE at this age.
• Developmental monitoring: In children with confirmed PTE, behavior and sleep should be monitored more closely - this is where associations with faster cortical thinning are seen.
• Communication with parents. It is important to explain that the "thin cortex" is not a diagnosis, but a biomarker of the trajectory, and it is necessary to work not with the MRI number, but with specific problems (sleep, impulsivity, external symptoms).

A few methodological details - why this data can be trusted

• The longitudinal design (two MRI points with an interval of ~2 years) instead of one “slice” reduces the risk of confusing age differences with true dynamics.
• Large sample and strict statistics: 5417 children, analysis of 68 regions for each hemisphere, FDR control.
• Post hoc checks: when dividing into "continued to smoke after learning about pregnancy"/"no" the effects are weaker - probably some mothers underestimate their use, or differences in exposure time are important. This does not cancel out the main conclusion.

Restrictions

• Self-reported exposure. No biomarkers (e.g. cotinine) mean there may be errors in PTE/PAE assessment. The authors explicitly acknowledge this and propose to correct it in future ABCD waves.
• Observation window: The study covers early adolescence; alcohol effects may manifest earlier/later, and subcortical effects were not analyzed here.
• Association ≠ causation. These are carefully controlled associations, not experiments; confirmation of mechanisms (including epigenetic ones) is required.

Where will science go next?

• Add exposure biomarkers (cotinine) and extend the observation window into late adolescence.
• Link anatomy to function: cognitive tests, behavioral tasks, nocturnal polysomnography - to understand which functions are more “sensitive” to accelerated thinning.
• To test the epigenetic bridge (DNA methylation ↔ rate of cortical changes) at the individual data level.

Conclusion

Traces of intrauterine tobacco in the brain are visible years later - as an acceleration/early start of the normal adolescent thinning of the cortex, especially in the frontal areas; it is this trajectory that is associated with more frequent behavioral and sleep problems. For alcohol at 9-12 years, such a signal is not visible.

Source: Marshall AT et al. Prenatal Tobacco and Alcohol Exposure and Cortical Change Among Youths. JAMA Network Open, 2025;8(6):e2516729. doi:10.1001/jamanetworkopen.2025.16729