Cannabis use during pregnancy is increasing. Data from the National Survey on Drug Use and Health shows that self-reported cannabis use among pregnant women rose from 3.4% in 2002 to 7.1% in 2022, with rates among first-trimester women exceeding 10% in some state-level surveys. Among those who use, the most commonly cited reasons are nausea relief during the first trimester, anxiety management, and sleep aid — all conditions where cannabis has demonstrated some efficacy in the general population.
The question is not whether pregnant women are using cannabis. They are. The question is what the best available evidence says about the consequences of that use for fetal development and birth outcomes. The answer is more complex than either side of the debate typically acknowledges.
The Endocannabinoid System in Fetal Development
To understand why cannabis use during pregnancy is a legitimate scientific concern, you need to understand the role of endocannabinoids in fetal development. This is not a case of a recreational substance happening to be harmful — it is a case of an exogenous compound directly interfacing with a signaling system that the developing fetus relies on for critical developmental processes.
The endocannabinoid system (ECS) is present and functionally active from remarkably early in human development. CB1 receptors appear in the fetal brain by approximately 14 weeks of gestation. The endocannabinoids anandamide and 2-AG are detectable in embryonic tissue even earlier, where they participate in processes that are fundamental to normal development.
Implantation and early placentation. Anandamide levels in the uterine lining play a direct role in the timing of embryo implantation. The uterus maintains a precise gradient of anandamide concentrations — lower at the implantation site, higher elsewhere — that guides the blastocyst to the correct location. Disruption of this gradient in animal models produces implantation failure and ectopic pregnancy.
Neural tube formation. Endocannabinoid signaling participates in neural tube closure, the embryological process that forms the spinal cord and brain during weeks 3 to 4 of gestation. CB1 receptor activation influences the proliferation and migration of neural progenitor cells during this critical window.
Axonal pathfinding. As the fetal brain develops, neurons must extend axons across considerable distances to connect with their target structures. Endocannabinoid signaling provides directional guidance for this process. CB1 receptor activation on growth cones — the navigational tips of extending axons — influences whether the axon turns toward or away from specific chemical signals. This pathfinding process establishes the fundamental wiring architecture of the brain.
Cortical development. In the second and third trimesters, endocannabinoid signaling participates in the organization of the cerebral cortex, including the migration of interneurons from their origin in deeper brain structures to their final positions in the cortical layers.
The critical implication is that THC, as a CB1 receptor agonist, has the potential to interfere with any of these processes by flooding the system with exogenous receptor activation at concentrations and timing patterns that the developmental program did not evolve to accommodate.
What Animal Studies Show
Animal research — primarily in rodents and, more recently, in non-human primates — has provided the clearest evidence of specific developmental disruptions from prenatal THC exposure.
Rat studies using doses roughly equivalent to moderate human cannabis use have demonstrated altered cortical layering, changes in dopaminergic neuron density in the ventral tegmental area (a brain region central to reward processing), and persistent changes in stress-response behavior in offspring. A 2023 study in Neuropharmacology found that prenatal THC exposure in rats produced measurable reductions in prefrontal cortex thickness that persisted into adulthood, accompanied by deficits in working memory tasks.
Non-human primate studies are fewer but arguably more relevant to humans. A series of experiments at the University of Maryland exposed pregnant macaques to THC at doses calibrated to produce blood levels comparable to a single joint session in humans. Offspring showed altered visual attention patterns, increased startle reactivity, and differences in social behavior that persisted through the first two years of life — the primate equivalent of early childhood.
The universal caveat with animal data applies: rodent brains develop differently from human brains, dose equivalency across species is imperfect, and the route of administration in controlled experiments does not perfectly replicate human cannabis use patterns. These studies demonstrate biological plausibility for harm, not definitive human risk quantification.
Human Epidemiological Evidence
Human studies on cannabis use during pregnancy are observational by necessity — no ethics board will approve randomizing pregnant women to THC exposure. This means all human evidence carries the inherent limitations of observational research: confounding variables, recall bias, difficulty separating cannabis effects from co-occurring factors (tobacco use, socioeconomic status, prenatal care access, nutrition).
With those limitations acknowledged, here is what the human data shows.
Birth weight and preterm birth. A 2020 meta-analysis in BMJ Open pooling 24 studies found that maternal cannabis use was associated with a 77% increased relative risk of low birth weight (below 2,500 grams) and a modest increase in preterm delivery risk. The absolute risk increase was approximately 4 to 5 percentage points above baseline rates. However, a substantial portion of the included studies did not adequately control for concurrent tobacco use, which is strongly associated with both low birth weight and cannabis use and makes isolating the cannabis-specific effect difficult.
Studies that attempted to isolate cannabis-only use (excluding tobacco co-use) generally found smaller effect sizes, with some showing no statistically significant birth weight difference after full adjustment. A 2024 Canadian cohort study of over 5,000 pregnancies with linked toxicology data found that cannabis-only exposure (confirmed negative for tobacco and other substances) was associated with a 100-gram reduction in average birth weight — statistically significant but clinically modest.
Neonatal outcomes. Some studies have reported increased rates of neonatal intensive care unit (NICU) admission among cannabis-exposed newborns. However, this finding is complicated by the fact that a positive maternal drug screen can itself trigger automatic observation protocols at many hospitals, inflating NICU admission rates regardless of actual clinical need.
Neurodevelopmental outcomes. This is the area of greatest scientific concern and greatest uncertainty. Two major longitudinal cohort studies have followed cannabis-exposed children over extended periods.
The Ottawa Prenatal Prospective Study, which began in 1978, followed children of cannabis-using mothers through age 18 to 22. It found subtle but persistent differences in executive function tasks (planning, impulse control, sustained attention) among heavily exposed children, emerging around age 4 and persisting through adolescence. These differences were small in magnitude — roughly 0.2 to 0.3 standard deviations below non-exposed peers on specific cognitive measures — but were present after controlling for multiple confounders.
The Generation R study in the Netherlands, using a larger modern cohort with better covariate data, found associations between prenatal cannabis exposure and slightly increased externalizing behavior (aggression, hyperactivity) at ages 3 to 5, but these associations attenuated substantially after adjusting for postnatal parental cannabis use and home environment factors.
The honest scientific summary is this: there is consistent evidence of small-magnitude associations between prenatal cannabis exposure and subtle neurodevelopmental differences. Whether these associations are causal effects of THC on fetal brain development or are partially or wholly attributable to residual confounding remains genuinely unresolved.
The First-Trimester Nausea Problem
The most common context for cannabis use during pregnancy is first-trimester nausea. Hyperemesis gravidarum — the severe, persistent form of pregnancy nausea — affects 0.3% to 2% of pregnancies and can result in hospitalization, dehydration, and significant weight loss. Standard antiemetics (ondansetron, promethazine, metoclopramide) carry their own risk profiles during pregnancy and are not always effective.
In this context, some pregnant women turn to cannabis as a nausea remedy. Survey data from Colorado and California suggest that 15% to 20% of women who use cannabis during pregnancy do so specifically for nausea, with the highest rates among those experiencing severe or medication-refractory symptoms.
The clinical dilemma is real. Severe persistent nausea itself carries risks: dehydration, malnutrition, electrolyte imbalances, weight loss, and in extreme cases, Wernicke encephalopathy. For a woman who is vomiting twenty times a day and cannot keep prescription antiemetics down, the risk calculus is not as straightforward as “just don’t use cannabis.”
This does not mean cannabis is recommended for pregnancy nausea. Every major obstetric organization — ACOG, SOGC, RCOG — advises against cannabis use during pregnancy. But the clinical advice should include honest acknowledgment of why some pregnant women are making this choice and engagement with their actual risk-benefit calculations rather than blanket prohibition without alternatives.
What About CBD?
CBD has a different pharmacological profile than THC and does not activate CB1 receptors in the same way. This has led some to assume that CBD-only products are safer during pregnancy. This assumption is premature.
CBD does interact with the endocannabinoid system — it modulates FAAH (the enzyme that breaks down anandamide), which means it can alter anandamide levels in tissues where precise anandamide concentrations guide developmental processes. A 2023 study in Reproductive Toxicology found that CBD at pharmacologically relevant concentrations disrupted trophoblast migration in human placental cell cultures — an in vitro finding that suggests potential effects on placentation.
The human safety data for CBD during pregnancy is essentially nonexistent. No clinical trial has tested CBD in pregnant populations, and observational data rarely distinguishes between THC and CBD exposure. The absence of evidence of harm is not evidence of absence of harm.
The responsible position is that CBD products should be treated with the same caution as THC products during pregnancy until data demonstrating safety exists. The burden of proof should be on demonstrating safety, not on proving harm.
Breastfeeding Considerations
THC is lipophilic (fat-soluble) and accumulates in breast milk. Studies measuring THC in breast milk of cannabis-using mothers have detected concentrations sufficient to produce meaningful infant exposure. A 2018 study in Pediatrics found detectable THC in 63% of breast milk samples from cannabis-using mothers, with concentrations that would deliver an estimated 2.5% of the maternal dose to the infant per feeding session.
The infant’s capacity to metabolize THC is substantially lower than an adult’s, which means that even relatively small absolute doses may produce proportionally larger effects. Additionally, the first months of postnatal life represent another period of intense brain development where endocannabinoid signaling plays regulatory roles.
Current recommendations from the American Academy of Pediatrics advise against cannabis use during breastfeeding. However, they also note that the benefits of breastfeeding generally outweigh the risks of cannabis exposure in breast milk, and that cannabis use alone should not be the basis for recommending formula over breastfeeding.
The Bottom Line
The current state of evidence on cannabis and pregnancy does not support the conclusion that moderate prenatal cannabis exposure causes severe developmental damage. It also does not support the conclusion that it is safe.
What the evidence does support: THC crosses the placenta, reaches the fetal brain, and interacts with a signaling system that is actively guiding critical developmental processes. Animal data demonstrates plausible mechanisms of harm. Human epidemiological data shows small but consistent associations with reduced birth weight and subtle neurodevelopmental differences. The magnitude of these associations is modest, and the degree to which they reflect causal effects of cannabis versus confounding factors remains uncertain.
The most scientifically defensible advice is to avoid cannabis during pregnancy if possible, to use conventional antiemetics as first-line treatment for pregnancy nausea, and to have honest conversations with healthcare providers about risk-benefit tradeoffs if cannabis use is being considered for refractory symptoms. What is not scientifically defensible is treating cannabis exposure during pregnancy as equivalent to alcohol or opioid exposure, or using it as a basis for punitive action against pregnant women — neither of which is supported by the current evidence base.
The research continues to evolve, and larger, better-controlled longitudinal studies are actively underway. For now, precaution is warranted, but so is honesty about what we do and do not actually know.