Last verified: April 2026

Pregnancy — Universal Medical Opposition

There is no area of cannabis science where the medical consensus is more unified. Every major medical and public health body — the American College of Obstetricians and Gynecologists (ACOG), the American Academy of Pediatrics (AAP), the U.S. Food and Drug Administration (FDA), the Centers for Disease Control and Prevention (CDC), and the World Health Organization (WHO) — recommends against cannabis use during pregnancy and breastfeeding.

This consensus is based on both biological plausibility and accumulating (though imperfect) human evidence. The endocannabinoid system (ECS) plays critical roles in fetal development, including implantation, placentation, neural proliferation, axonal migration, and synaptogenesis. CB1 receptors are expressed in the fetal brain from early gestation, and endocannabinoid signaling guides fundamental processes of neural circuit formation. Exogenous THC — which activates CB1 with greater intensity and different kinetics than endogenous anandamide — could theoretically disrupt these precisely timed developmental events.

Human observational data, while limited by confounding (cannabis-using pregnant women are more likely to also use tobacco, alcohol, and other substances, and differ in socioeconomic and nutritional status), consistently suggest associations between prenatal cannabis exposure and lower birth weight, preterm birth, and subtle neurodevelopmental differences in exposed offspring. The Ottawa Prenatal Prospective Study, the Maternal Health Practices and Child Development Study, and more recently the ABCD Study have all reported associations, though effect sizes vary and residual confounding cannot be fully excluded.

THC in Breast Milk — The Pharmacokinetic Problem

THC’s extreme lipophilicity creates a specific pharmacokinetic problem for breastfeeding. In 2018, Teresa Baker and colleagues published findings demonstrating that THC concentrations in breast milk can reach approximately 8 times the concurrent plasma concentration — breast milk preferentially concentrates THC because it is a lipid-rich medium and THC partitions aggressively into fat.

The half-life of THC in breast milk is prolonged. Studies have detected THC in breast milk up to six days after the last use in moderate users, meaning that "pump and dump" strategies (expressing and discarding milk after cannabis use) are ineffective for acute exposures — unlike alcohol, which equilibrates rapidly with plasma and clears within hours, THC remains trapped in the lipid compartment of breast milk long after psychoactive effects have resolved.

The clinical significance of infant THC exposure through breast milk is not fully characterized. Neonatal CB1 receptor expression, immature hepatic metabolism (limited CYP2C9 and CYP3A4 activity in infants), and the developing blood-brain barrier all suggest that infants may be more sensitive to THC effects than adults. However, large-scale outcome studies of breastfed infants exposed to cannabis are lacking, and the ethical impossibility of a randomized trial means the evidence will remain observational.

Driving — The ~2x Crash Risk

Meta-analyses of epidemiological studies consistently estimate that cannabis use approximately doubles the risk of a motor vehicle crash (pooled odds ratio ~2.0). This represents a meaningful increase in absolute risk, but it is important to contextualize: alcohol at the legal limit of 0.08% blood alcohol concentration (BAC) increases crash risk by approximately 7–13-fold. Cannabis-impaired driving is genuinely dangerous; it is also substantially less dangerous than alcohol-impaired driving by every available metric.

The primary impairments from acute THC intoxication affect tracking ability (maintaining lane position), divided attention (managing multiple simultaneous tasks), and reaction time. Importantly, laboratory driving simulator studies and closed-course studies consistently show that cannabis-impaired drivers tend to compensate for perceived impairment — driving more slowly, increasing following distance, and avoiding risky maneuvers — in marked contrast to alcohol-impaired drivers, who tend to underestimate their impairment and increase risk-taking behavior. This behavioral compensation partially (but not fully) offsets the pharmacological impairment.

The combined use of cannabis and alcohol is substantially more dangerous than either alone. The impairments are additive to supra-additive, and the compensatory behaviors seen with cannabis alone are abolished by concurrent alcohol intoxication. Public health messaging should emphasize this interaction particularly.

Per Se THC Limits — Scientifically Indefensible

Several U.S. states have enacted per se THC limits for driving — typically 5 ng/mL of blood THC, modeled on blood alcohol concentration limits. These laws are legally convenient but pharmacologically indefensible, a position formally endorsed by the AAA Foundation for Traffic Safety.

A quantitative threshold for per se laws for THC following marijuana use cannot be scientifically supported.

AAA Foundation for Traffic Safety, 2016

The fundamental problem is that blood THC levels do not correlate reliably with impairment. THC pharmacokinetics differ radically from alcohol. After inhalation, blood THC peaks within minutes and falls rapidly due to redistribution into fatty tissues, not metabolism — a user can be profoundly impaired with declining blood levels or minimally impaired with elevated residual levels. Chronic daily users maintain detectable blood THC levels for 30+ days after complete cessation due to slow release from adipose stores, meaning they can test positive at per se thresholds while completely sober.

The testing logistics compound the problem. Roadside blood draws are invasive, require trained phlebotomists, and typically occur 90–165 minutes after a traffic stop — by which time blood THC has declined substantially from the level present during the driving behavior in question. Oral fluid testing is faster but measures a different compartment with its own pharmacokinetic complications.

The scientifically honest position is that cannabis-impaired driving is a real public health problem that cannot be effectively addressed by borrowing alcohol’s regulatory framework. Impairment-based assessment (standardized field sobriety testing adapted for cannabis, Drug Recognition Expert evaluation) is more defensible than per se chemical testing, though it introduces its own reliability and subjectivity concerns.

Contaminant Risks — Myclobutanil, Heavy Metals, and Aspergillus

Beyond the pharmacological risks of THC itself, cannabis products may contain contaminants that pose additional dangers to vulnerable populations — pregnant women, immunocompromised patients, and those with respiratory disease.

Myclobutanil is a systemic fungicide widely used in cannabis cultivation (often illicitly, as it is not approved for use on inhaled products in any jurisdiction). When heated to combustion or vaporization temperatures, myclobutanil undergoes pyrolysis to produce hydrogen cyanide (HCN) — a potent respiratory and mitochondrial toxin. The concentrations produced from typical contaminated cannabis are unlikely to be acutely lethal, but chronic low-level HCN exposure from repeated inhalation of contaminated products has not been studied.

Heavy metals — including lead, cadmium, arsenic, and mercury — can accumulate in cannabis through contaminated soil, water, or growing media. Cannabis is a known bioaccumulator (it was planted at Chernobyl for phytoremediation), meaning it concentrates environmental metals in its tissues. Heavy metal exposure during pregnancy is associated with neurodevelopmental toxicity, and any additional source of exposure is clinically concerning.

Aspergillus contamination of cannabis flower is potentially fatal for immunocompromised patients. Invasive pulmonary aspergillosis in HIV/AIDS patients, organ transplant recipients, and chemotherapy patients who inhale Aspergillus-contaminated cannabis has been documented in case reports and series. These patients should use only rigorously lab-tested products or, preferably, non-inhalation routes entirely.

The Risk Calculus

Pregnancy and driving represent the two contexts where the risk-benefit calculation for cannabis use is most unfavorable. In pregnancy, there is no established therapeutic indication that justifies fetal THC exposure when alternatives exist for every condition cannabis might treat (nausea, pain, anxiety, insomnia). In driving, the impairment is real and the consequences potentially fatal, even if the magnitude of risk is smaller than alcohol.

The contaminant dimension adds urgency. Even if THC itself posed no risks, unregulated or poorly tested products can deliver myclobutanil-derived cyanide, bioaccumulated heavy metals, and Aspergillus spores — hazards that regulatory testing is intended to catch but, as the lab science evidence shows, frequently misses. Vulnerable populations deserve both honest pharmacological risk assessment and robust product safety standards.