Prenatal Cannabis Exposure Was Not Linked to Neurodevelopmental Differences in a Long-Term Cohort Study
A long-term study of children born in Western Australia found no statistically significant association between prenatal cannabis exposure and neuropsychological performance in late childhood or autistic traits in early adulthood. The findings do not establish that cannabis is safe during pregnancy, but they add to an inconsistent body of observational research on the subject.
Published in 2023 in Paediatric and Perinatal Epidemiology, the study analyzed participants from the Raine Study, a birth cohort of children born between 1989 and 1992. Researchers included 2,804 children whose mothers had provided information about cannabis use during pregnancy; 285 children, or 10.2 percent, were classified as prenatally exposed.
The primary outcome was performance at age 10 on the Clinical Evaluation of Language Fundamentals, a standardized assessment of language abilities. The researchers also examined vocabulary, behavior, motor development, nonverbal reasoning and processing speed at age 10, as well as Autism Spectrum Quotient scores collected at ages 19 to 20.
To reduce the influence of differences between families who reported prenatal cannabis exposure and those who did not, the researchers used propensity-score matching. They also used multiple imputation to address missing covariate information and inverse probability-of-censoring weighting to account for missing outcome data.
After these adjustments, the exposed and unexposed groups had similar scores on the language assessment. Prenatal cannabis exposure was also not significantly associated with the study’s secondary neuropsychological outcomes or with an increased risk of reaching a defined clinical deficit threshold.
The study was observational, however, so it cannot prove that cannabis exposure has no effect on development. Cannabis use was assessed through maternal reports, and exposure often occurs alongside other factors—including tobacco or alcohol use, socioeconomic conditions, maternal health and differences in the home environment—that can be difficult to measure completely. The cohort also reflects cannabis use and product potency from the late 1980s and early 1990s, which may limit how directly the results apply to current products and patterns of use.
The authors’ conclusion was therefore narrower than the original headline might suggest: after adjustment for measured sociodemographic and clinical characteristics, prenatal marijuana exposure was not associated with worse test scores at age 10 or more autistic traits at ages 19 to 20 in this cohort. “No association detected” is not the same as evidence that exposure is harmless.
That distinction is important because health authorities continue to recommend avoiding cannabis during pregnancy while research develops. The Centers for Disease Control and Prevention’s guidance on cannabis and pregnancy says available studies raise concerns about fetal and child development and advises pregnant people who use cannabis to discuss it with a healthcare professional. The American College of Obstetricians and Gynecologists’ 2025 clinical consensus likewise recommends cessation during pregnancy and emphasizes supportive, evidence-based counseling rather than relying solely on punitive responses.
The Raine Study provides useful long-term evidence, particularly because it assessed several domains of development and followed participants into early adulthood. Still, decisions about cannabis use during pregnancy should not rest on one observational study. Larger contemporary cohorts, better measurement of dose and timing, and careful control of co-occurring exposures are needed to clarify the potential effects of prenatal cannabis exposure.
