What do medical sources say about Cannabis (Marijuana) during pregnancy?

1st Trimester: MotherToBaby says it is unknown whether cannabis raises miscarriage risk. A 2018 study found higher odds for a combined stillbirth-or-miscarriage outcome, but it had only six losses and a very wide estimate. A 2025 review found moderate-certainty links to preterm birth, low birth weight, and small babies for gestational age after adjustment for other factors. A 2025 cohort found signals for two rare abdominal-wall defects, but none for 36 others. ACOG, CDC, and FDA advise against cannabis use during pregnancy. Ask your provider about your situation. 2nd Trimester: CDC recommends that pregnant women do not use cannabis. A 2025 review found moderate-certainty links to preterm birth, low birth weight, and small-for-gestational-age birth after adjustment for other factors. A 2026 cohort of 6,838 couples linked weekly early-pregnancy use to more births before 37 weeks, but the estimate was imprecise; it found no link to small or large birth size for gestational age. ACOG advises stopping cannabis throughout pregnancy, and sources did not identify a trimester-specific lower-risk window. Ask your provider about your situation. 3rd Trimester: MotherToBaby reports newborn jitteriness after prenatal exposure, but also reports exposed newborns without these signs. A 1987 study of 250 infants found more tremors and startles and poorer ability to adjust to repeated visual cues; it did not diagnose withdrawal. ACOG reports changes in alertness and regulation during the first month after prenatal exposure. A 2026 study of 460 pairs found a link between THC at delivery and higher attention and behavior scores at age five; it was observational. ACOG advises cannabis cessation during pregnancy. Ask your provider about your situation.

What do medical sources say about Cannabis (Marijuana) while breastfeeding?

LactMed reports THC in milk from six days to over six weeks. In a 2018 study, THC appeared in 34 of 54 milk samples up to six days after last reported use; a 2021 study projected clearance beyond six weeks for seven women who abstained. An observational 1990 study reported lower motor scores at age one in 68 milk-exposed infants. ACOG advises stopping cannabis during lactation, but says continued use is not a reason to stop breastfeeding or discourage it. Ask your provider about your situation.

Where do sources disagree on Cannabis (Marijuana) in pregnancy or breastfeeding?

Miscarriage and pregnancy loss MotherToBaby says it is not known whether cannabis raises miscarriage risk and notes that one study found a higher chance while other studies did not. Coleman-Cowger et al. reported higher odds for a combined stillbirth-or-miscarriage outcome in a small cannabis-only group; Mumford et al. found no association between preconception use and pregnancy loss in women with a history of prior loss. The 2018 study combined stillbirth and miscarriage, used a small birth-outcome subset, and had only six loss events, producing an extremely wide confidence interval. The 2021 study evaluated preconception use and had too few early-pregnancy users to estimate that risk. The evidence does not establish a maternal cannabis effect on miscarriage. The 2018 combined-outcome signal is not a precise miscarriage estimate, while the later prospective study does not resolve risk from use during early pregnancy. The public summary keeps the uncertainty and does not assign the 2018 result to miscarriage alone.

Back to search

Cannabis (Marijuana)

Everyday exposureAlcohol, tobacco & drugs

Also called · 9 names▾

marijuana, weed, pot, THC, edibles, smoking weed, vaping marijuana, pot brownies, cannabis use

Recreational substance

General information · effects depend on amount, timing, and how it is used. About these sources

Pregnancy

A 2025 living systematic review and meta-analysis synthesized 51 adjusted cohort or case-control studies (N=21,146,938). It rated the evidence moderate for associations with low birth weight, preterm birth, and small-for-gestational-age birth after accounting for tobacco and other factors, and low for perinatal mortality. Perinatal mortality is not a specific miscarriage estimate. The 2025 birth-defects cohort included 363,952 infants, with 22,494 exposed. It found a model-dependent gastroschisis association, an omphalocele association in its adjusted models, and no association with the other 36 structural defects. Because the cases were rare, the broad live-entry wording about heart, brain, and other defects is not supported by this recent cohort and is narrowed in the summary. For miscarriage, MotherToBaby says the relationship is unknown. A 2018 cross-sectional study of 500 pregnant patients reported higher odds for a combined stillbirth-or-miscarriage outcome in its cannabis-only group, but the birth-outcome analysis covered 360 pregnancies and had only six loss events; its adjusted estimate was very imprecise (OR 12.1, 95% CI 1.03-141.8). A 2021 prospective cohort of 1,228 women with a history of pregnancy loss found no association between preconception cannabis use and pregnancy loss (RR 0.81, 95% CI 0.46-1.42); only 11 of 789 participants used cannabis in the first eight weeks, so early-pregnancy risk could not be estimated. A 2024 cohort of 119,976 children found no association between early prenatal use and speech, global, or motor developmental delays through age 5.5. A 2026 cohort of 6,838 couples found an association between weekly early-pregnancy use and preterm birth (HR 1.87, 95% CI 1.06-3.29), with no association for small- or large-for-gestational-age birth. A separate 2026 cohort of 460 mother-child pairs found timing-dependent associations between THC measured at delivery and higher attention, ADHD-symptom, and pervasive-development problem scores at age five; it was observational and does not establish cause. A 2024 analysis of U.S. Drug Enforcement Administration seizure samples collected from 2013 to 2022 found high-THC samples were generally above 10% THC, with some above 20%. Seizure samples are not a measure of every retail product or of pregnancy exposure, but they show why older low-potency studies may not represent current products. ACOG advises cannabis cessation during pregnancy; CDC and FDA also advise against use.

Breastfeeding

THC transfers into breast milk. LactMed's August 2026 review reports detection durations ranging from six days to more than six weeks. Bertrand et al. measured 54 samples from 50 breastfeeding women; THC was detected in 34 samples (63%), up to about six days after last reported use. Wymore et al. enrolled 25 postpartum participants intending to breastfeed; plasma testing verified abstinence in 12. All participants had detectable milk THC during the study, and among seven women abstinent for more than five weeks the estimated mean half-life was 17 days, with projected elimination beyond six weeks. These results do not establish a universal clearance interval. The 1990 Astley and Little study compared 68 infants exposed through milk with 68 matched controls and reported lower motor development at one year. It is an old observational signal; its abstract does not report adjustment for prenatal cannabis exposure. MotherToBaby and LactMed describe possible changes to prolactin and milk supply. MotherToBaby also says cannabis might change milk fat and protein; LactMed summarizes mixed findings on milk nutrients and says their clinical relevance is questionable. The 2024 AAP nursery survey measured hospital practices and clinician perceptions, not infant outcomes. ACOG advises cessation during lactation but says continued use is not a contraindication to breastfeeding and breastfeeding should not be discouraged.

Miscarriage and pregnancy loss

Source A says: MotherToBaby says it is not known whether cannabis raises miscarriage risk and notes that one study found a higher chance while other studies did not. (MotherToBaby: Marijuana (Cannabis))
Source B says: Coleman-Cowger et al. reported higher odds for a combined stillbirth-or-miscarriage outcome in a small cannabis-only group; Mumford et al. found no association between preconception use and pregnancy loss in women with a history of prior loss. (Coleman-Cowger et al., Neurotoxicology and Teratology 2018, Mumford et al., Human Reproduction 2021 (prospective cohort))
Why they differ: The 2018 study combined stillbirth and miscarriage, used a small birth-outcome subset, and had only six loss events, producing an extremely wide confidence interval. The 2021 study evaluated preconception use and had too few early-pregnancy users to estimate that risk.
What the evidence shows overall: The evidence does not establish a maternal cannabis effect on miscarriage. The 2018 combined-outcome signal is not a precise miscarriage estimate, while the later prospective study does not resolve risk from use during early pregnancy. The public summary keeps the uncertainty and does not assign the 2018 result to miscarriage alone.

Which structural birth defects are associated with prenatal cannabis use

Source A says: MotherToBaby reports that some studies suggest associations with several defect types while other studies do not, and it says exposure measurement and co-use limit interpretation. (MotherToBaby: Marijuana (Cannabis))
Source B says: A 2025 cohort found an association with omphalocele in adjusted models and a gastroschisis association that depended on the model; it found no association with the other 36 structural defects studied. (Avalos et al., Birth Defects Research 2025 (structural birth defects cohort))
Why they differ: The MotherToBaby summary discusses studies with different exposure definitions and defect groups. The 2025 cohort used universal screening in one health system and evaluated 38 named defects; the two abdominal-wall findings were rare, and the gastroschisis association was not present in every adjusted model.
What the evidence shows overall: Do not retain the live summary's broad reference to heart, brain, and other defects. The corrected summary names only the two rare abdominal-wall signals from the 2025 cohort and notes the absence of associations for other defects in that study.

THC detection time in breast milk

Source A says: FDA's consumer page, whose content is current as of October 16, 2019, says breast milk can contain THC for up to six days after use. (FDA consumer update on cannabis use during pregnancy and breastfeeding, Bertrand et al., Pediatrics 2018 (milk cannabinoid study))
Source B says: LactMed's August 2026 review reports detection from six days to more than six weeks; Wymore et al. measured milk THC in a small verified-abstinence cohort and projected elimination beyond six weeks for some participants. (LactMed: Cannabis, Wymore et al., JAMA Pediatrics 2021 (milk pharmacokinetic cohort))
Why they differ: The FDA page predates the 2021 pharmacokinetic study. The studies measured different populations and exposure patterns, and the longer-duration estimate came from seven abstinent participants.
What the evidence shows overall: Detection time varies, and the evidence does not establish a single clearance interval. The updated summary uses LactMed's range and does not present six days as a universal endpoint.

How to support breastfeeding if cannabis use continues

Source A says: MotherToBaby, LactMed, CDC, FDA, and AAP 2018 advise avoiding cannabis use during lactation because THC reaches milk and infant-outcome evidence is limited. (MotherToBaby: Marijuana (Cannabis), LactMed: Cannabis, CDC: Cannabis and Pregnancy, FDA consumer update on cannabis use during pregnancy and breastfeeding, AAP: Marijuana Use During Pregnancy and Breastfeeding (2018 report summary))
Source B says: ACOG 2025 advises cessation of cannabis during lactation but says continued use is not a contraindication to breastfeeding and breastfeeding should not be discouraged. The AAP 2024 article likewise describes AAP policy as not making cannabis use a breastfeeding contraindication. (ACOG Clinical Consensus No. 10: Cannabis Use During Pregnancy and Lactation, Chang et al., Pediatrics 2024 (AAP nursery breastfeeding survey))
Why they differ: The guidance shares a goal of reducing cannabis exposure but varies in how explicitly it separates cessation counseling from the decision to continue breastfeeding when use does not stop.
What the evidence shows overall: The entry rates cannabis use during lactation as avoid, but that rating must not be read as a recommendation to stop breastfeeding. The public summary states ACOG's distinction directly.

Later child development after prenatal exposure

Source A says: A 2024 cohort of 119,976 children found no association between early prenatal cannabis use and speech, global, or motor developmental delays through age 5.5. (Avalos et al., JAMA Network Open 2024 (child developmental-delay cohort))
Source B says: A 2026 cohort of 460 mother-child pairs found timing-dependent associations between THC measured at delivery and higher attention and behavior problem scores at age five. (Moore et al., Journal of Pediatrics 2026 (child neurodevelopment cohort))
Why they differ: The studies measured different outcomes and exposures. One assessed developmental-delay diagnoses after early-pregnancy screening; the other used THC biomarkers at several gestational windows and behavior scores at age five. Both are observational, and the smaller study's findings do not establish causation.
What the evidence shows overall: Evidence on later child outcomes is mixed. The corrected summaries attribute findings to their studies and avoid describing a specific long-term outcome as established.

Last reviewed 2026-10-16

Independent check 2026-10-16

Not yet clinician-reviewed

Ask your provider about your situation.

Cannabis (Marijuana)

Cannabis (Marijuana) — Breastfeeding

What the research shows

What's in it / what was tested

Sources disagree

Options sources describe

Beyond pregnancy

All sources

Related