Edibles During Pregnancy and Breastfeeding: What Was Measured, and Why the Medical Bodies Are Right

This is the one question on EdibleRank where following the evidence and following the medical bodies land in the same place. ACOG tells clinicians to say there are no medical indications for cannabis in pregnancy or postpartum, the FDA strongly advises against THC and CBD in any form, and the label on prescription THC capsules says to avoid use in pregnant women.

The harm evidence is observational and confounded, and we would lean on those weaknesses if anything sat on the other side of the scale. Nothing does. Nausea is why many people end up on this question, and no trial has tested a cannabinoid for it. The breast-milk numbers are measured and uncomfortable. And the edible, the form budtenders most often call safe, has no route-specific pregnancy data behind that reputation.

Nobody reading this needs a lecture. A lot of people reading this are vomiting, scared of a drug test, or both, so the rest of the page sticks to what was measured.

Is it safe to use edibles while pregnant?

No dose has been shown to be safe, and every medical body that has looked advises against it. ACOG's 2025 consensus says cannabis has no medical indication in pregnancy, the FDA strongly advises against it in any form, and adjusted cohort data tie use to lower birth weight and earlier delivery.

ACOG's Committee Opinion 722, the 2017 document you will still see quoted, now redirects to the college's withdrawn-document page. The live text is Clinical Consensus No. 10, from October 2025: advise stopping, ask about the symptom without judgment, look for another way to treat it. The AAP's 2018 clinical report is among the guidelines it cites for abstaining or cutting down.

The closest thing to a gummy with a federal label is dronabinol: swallowed THC, in a capsule, at a dose somebody checked. The Marinol label says to avoid it in pregnant women. It reports no birth defects in mice and rats, fetal deaths at doses toxic to the mother, and THC in the cord blood of some exposed infants.

What did the big pregnancy studies actually find?

Lower birth weight, earlier delivery and more NICU admissions among exposed babies, in effect sizes that often shrink after adjustment but mostly survive it. A 2025 meta-analysis of 51 studies and about 21 million participants rates that evidence moderate certainty. A 2016 analysis that adjusted for tobacco found nothing significant, and it is now the minority result.

The table sets the main studies side by side: what each measured, how big it was, what it found, and what it cannot settle.

Conner is the best result for the skeptical side: cannabis and cigarettes travel together, and once it adjusted for that the signal faded. Metz answered it better than a questionnaire can, by testing urine for nicotine, and the association held. Self-report also undercounts: 54.9 percent of pregnant Kaiser patients who screened positive were positive on toxicology alone, an error Corsi's authors say pulls associations toward zero.

The confounder nobody has removed is the nausea itself. Lo's review could not adjust for severe vomiting, which independently raises the risk of low birth weight and preterm birth and is also what sends people to cannabis. Part of the cannabis signal might be the nausea. That is the best argument the other side has, and the dose response still runs the wrong way: in Lo's heavy-use subgroups the odds climbed to 2.36 for low birth weight and 1.95 for preterm birth. Nobody will ever randomize pregnant people to gummies, so this is the evidence there is going to be.

Later childhood is thinner and calmer. Sorkhou's 2024 review of 57 studies found little evidence of harm in early childhood apart from attention and externalizing problems, and ACOG calls long-term risk hard to quantify.

Is half a gummy every other day going to hurt the baby?

Nobody can answer that, because no study we found in PubMed has measured it. The cohorts record cannabis as a yes, a frequency or a urine level, never as milligrams in a gummy. ACOG says risk rises with dose, which is an argument for less and says nothing about where a safe floor sits, if there is one.

A poster on What to Expect's Ganja Mamas board (48.5K members on November 7) asked nearly that on October 31, 2026: half of a 10mg gummy every other day, helping with anxiety and with keeping food down. The one reply came from someone who used gummies through most of two pregnancies and reports healthy kids: true for that person, and no use to anyone else as data.

The nearest published number is in Metz, where risk climbed once cumulative urine metabolite passed 1024 ng/mL, a level prior studies tie to use at least every other day. With no route or milligrams recorded, it cannot be converted into gummies.

I took edibles before I knew I was pregnant. What now?

Tell your clinician, and know that the best timing data offer some reassurance, with limits. In Metz's urine-tested cohort, cannabis use confined to the first trimester was not significantly associated with the composite of small babies, indicated preterm birth, stillbirth and hypertensive disorders. Use continuing past that point was.

The limits matter: 197 people, a raw composite rate of 24.1 percent against 17.4 in the unexposed, and no measure of brain development. For regular users who stop, ACOG notes withdrawal can begin 24 to 72 hours after the last use and run up to two weeks.

Are edibles safer than smoking during pregnancy?

They remove the smoke, which is a real difference, and nobody has shown they reduce what reaches the fetus. A PubMed search on November 7, 2026 found no cohort reporting birth outcomes by route and no human study measuring fetal THC after an edible. Route is the biggest blank in this literature.

The counter has filled that blank anyway. When mystery shoppers posing as pregnant reached 505 California budtenders, edibles were the form most often endorsed as safe, and 20.6 percent called prenatal cannabis safe outright. In Arizona, 71.2 percent of 104 budtenders recommended a product for morning sickness, and 34.6 percent suggested a doctor unprompted. Colorado's figure in 2018 was 69 percent, mostly on personal opinion. Three states, plenty of retail opinion, and no route data under any of it.

Two things give reason for unease. One is time: our duration guide, which owns these numbers, puts an edible's main effects at 4 to 8 hours, up to 12 at higher doses, against 2 to 3 hours for smoked cannabis, the longest effect window of the common routes. The other is the metabolite. Swallowed THC goes through the liver first and comes out as 11-hydroxy-THC in quantity. In Kumar's 2023 perfused term placenta experiments, THC was pushed back toward the maternal side while 11-hydroxy-THC appeared to cross by passive diffusion. So an edible might deliver more active metabolite than smoke does, an inference from a lab placenta that nobody has tested in a pregnant person.

None of this is a niche concern: 32.6 percent of pregnant Kaiser patients reporting cannabis use in 2021 and 2022 used edibles. The closest thing to a controlled edible exposure we found involved rhesus macaques, five on a daily THC edible and five controls, with subtle differences in fetal brain growth.

Does cannabis help morning sickness?

Nobody has tested it. A PubMed search on November 7, 2026 for cannabinoids, pregnancy and nausea, filtered to trials, returned six records, and none gives a cannabinoid to anyone for pregnancy nausea. The relief reports are real and come from surveys, and the cohort data point at least partly the other way.

The antiemetic reputation was earned elsewhere, in chemotherapy patients. The best pregnancy data in favor is a 2022 survey of 550 people recruited through the Hyperemesis Education and Research Foundation's social media. Among the 14 percent who used cannabis, 82 percent reported relief against 60 percent of prescription antiemetic users, and 71 percent had turned to it because prescribed drugs failed; 59 percent smoked it and 19 percent swallowed it. The authors call the sample unrepresentative. Those respondents are not imagining the relief, and ACOG's own nausea bulletin concedes the condition gets minimized and undertreated. A gummy adds a practical problem: onset runs 30 minutes to 2 hours, per the duration guide, and it has to stay down to work.

Now the awkward direction. Across 220,510 Kaiser pregnancies, prenatal use ran 11.3 percent with severe nausea against 4.5 percent with none (adjusted odds ratio 3.80), and the authors offer three readings: self-medication, cannabis causing nausea, or clinicians diagnosing it more often in users. Metz's urine data agree on the pattern, with moderate-to-severe nausea in 20.7 percent when metabolite was detected against 15.5 percent when it was not. A 2025 analysis of 356,343 pregnancies then found that daily use before conception, before morning sickness had a chance to prompt it, carried an adjusted odds ratio of 2.61 for severe nausea.

That is the shape you would expect if some of this vomiting is cannabinoid hyperemesis syndrome, the cyclic vomiting of long-term heavy use. A 2025 case series argues it is under-reported in pregnancy, and ACOG says more than 95 percent of people with it resolve on stopping. In a 2024 analysis of Ganja Mamas posts, one poster described an ER doctor suggesting, rudely, that the weed was behind the nausea.

What has been tested is Diclegis, doxylamine plus B6, an FDA-approved drug for this: a round white tablet printed with a pink image of a pregnant woman, randomized against placebo in 261 women. The result was modest, a 0.7-point edge on the PUQE nausea score at day 15, which helps explain why people go looking.

Modest and measured still beats unmeasured.

How long does THC stay in breast milk?

Six days to more than six weeks, depending on how much was used before stopping. LactMed gives that range. Amounts are small in absolute terms, about 2.5 percent of the maternal dose in the one controlled-dose study, and concentrated relative to blood, with milk running roughly six to seven times plasma.

The table lists the milk studies, who was measured, how they used cannabis, and what turned up.

The half-life spread is a genuine disagreement. LactMed's modeling predicts 12 to 39 hours; Wymore measured 17 days in people who had used regularly through pregnancy. The likeliest reading is that both hold for different bodies, since THC dissolves in fat and a heavy user has more stored. Pumping and discarding does not get around this, as our postpartum recovery page lays out.

Almost none of it is about edibles. Bertrand is the one place non-inhaled use shows up, seven mothers of 50, and route did not predict concentration, with far too few people to settle anything. A PubMed search on November 7 found no study measuring milk after a known edible dose. Baker's dose works out to 23mg of THC on paper, and no conversion to gummies exists, since the Marinol label has only 10 to 20 percent of a swallowed dose reaching the blood.

Do I have to stop breastfeeding if I use cannabis?

ACOG does not say that. Its 2025 text says clinicians should advise stopping cannabis during lactation, then says continued use is not a contraindication to breastfeeding and that breastfeeding should not be discouraged. The advice is to drop the cannabis and keep the milk, a distinction that gets lost.

The infant evidence is thin in both directions. Astley's 1990 comparison of 136 breastfed babies tied first-month exposure to lower motor scores at one year, in an era of weaker flower. A 2024 case report describes a 5-week-old with recurrent apnea and a positive THC screen who improved on formula, causation unproven. A 2026 cohort of 1,520 infants whose mothers tested positive at delivery found development no different by breastfeeding status after adjustment (odds ratio 1.24 for delay among the non-breastfed, 95% CI 0.93 to 1.67). LactMed flags a practical problem as well: a breastfed infant's positive urine test can carry legal consequences.

Is CBD safer while pregnant or breastfeeding?

Nobody has shown that, and the FDA strongly advises against it. The agency says no comprehensive research exists on CBD in pregnancy or a breastfed baby. LactMed says it has not been studied in nursing women on Epidiolex and has been detected in the milk of some cannabis users, and product labels add their own uncertainty.

Moss measured a milk-to-plasma ratio of 2.6 for CBD, ACOG cites testing in which one in five products labeled CBD-only contained THC, and a 2025 mouse study of oral THC or CBD oil at 20 mg/kg, far above a gummy per kilogram, saw impaired fetal growth with both. CBD was also one of the two things Arizona budtenders recommended most, which is a lot of confidence in a molecule the FDA describes this way. Our interactions page covers its enzyme effects.

Will my OB report me if I tell them?

That varies by state and hospital, and ACOG says so plainly. Its 2025 consensus tells clinicians to screen everyone by interview rather than urine test, to get informed consent before any drug test, and to know their local child protective services rules, because those rules land unevenly and can carry real consequences.

The fear is rational, and it runs through the forums where people ask this. ACOG's own figures back the worry: Black patients screen positive for substance use at the same rate as White patients and are 4 to 10 times more likely to have a positive result reported to child protective services.

This page cannot tell you what your hospital does. What hiding the history costs is easier to state: a clinician who does not know about daily use cannot consider cannabinoid hyperemesis or work out what the cannabis was treating. Testing policy is a fair question before delivery day, and ACOG expects clinicians to know the answer. Detection windows have their own page, and ACOG notes clearance in pregnancy has not been studied extensively.