| Audience | Patients planning pregnancy, pregnant patients, partners, obstetric clinicians, primary care clinicians, and cannabis clinicians |
| Primary Topic | Preconception and early-pregnancy cannabis use in relation to preterm birth and birthweight-for-gestational-age outcomes |
| Source | Read the full source |
Early Pregnancy Cannabis Use and Preterm Birth: What a New Cohort Found
A new prospective cohort of 6,838 couples found that at-least-weekly cannabis use during early pregnancy was associated with more preterm birth than nonuse. The estimate was imprecise, and the observational study does not establish that cannabis caused the outcome.
| Study Type | Prospective internet-based preconception cohort |
| Population | 6,838 U.S. and Canadian couples who conceived between June 2013 and September 2024 |
| Female Eligibility | Ages 21 to 45 years and trying to conceive |
| Male Eligibility | Age 21 years or older |
| Exposure Windows | Female and male preconception use, plus female early-pregnancy use before 12 weeks |
| Outcomes | Preterm birth, small for gestational age, and large for gestational age |
| Frequent Early-Pregnancy Finding | At least weekly use versus nonuse: HR 1.87, 95% CI 1.06 to 3.29 |
| Less-Frequent Early-Pregnancy Finding | Less than weekly use versus nonuse: HR 0.85, 95% CI 0.36 to 2.02 |
| Preconception Finding | No appreciable association with preterm birth in either partner |
| Growth Finding | No reported association with small- or large-for-gestational-age birth |
| Journal | Paediatric and Perinatal Epidemiology |
| Publication | September 6, 2026 |
| PMID / DOI | 42703005 / 10.1111/ppe.70201 |
Wesselink and colleagues analyzed Pregnancy Study Online, an internet-based preconception cohort of couples trying to conceive in the United States and Canada. The analysis included 6,838 couples who conceived between June 2013 and September 2024.
Both partners completed baseline questionnaires. Female participants completed follow-up questionnaires before conception, during early pregnancy before 12 weeks, around 32 weeks, and six months after birth. Birthweight and gestational age came from birth certificates and questionnaires.
At-least-weekly cannabis use during early pregnancy was associated with a higher incidence of preterm birth than nonuse, with an adjusted hazard ratio of 1.87 and a 95% confidence interval from 1.06 to 3.29. Less-than-weekly use had a hazard ratio of 0.85, with a confidence interval from 0.36 to 2.02.
Female and male preconception cannabis use were not appreciably associated with preterm birth in the reported models. Cannabis use by either partner was not associated with small- or large-for-gestational-age birth.
The study separates cannabis use before conception from use during early pregnancy. It also treats preterm delivery, small-for-gestational-age birth, and large-for-gestational-age birth as different outcomes rather than combining them into a general label of adverse birth outcomes.
That separation is clinically useful because one estimate should not be stretched across every exposure window or every pregnancy outcome. The reported early-pregnancy preterm-birth association does not imply an association with fetal growth in this dataset.
The cohort is observational. The authors adjusted for potential confounders, but participants were not assigned to cannabis exposure. Self-reported use may be affected by recall, stigma, product variability, and incomplete measurement of dose, potency, route, or co-exposures.
The at-least-weekly estimate is compatible with a clinically important association, but its confidence interval is wide. The result supports caution and further study, not a precise prediction for an individual pregnancy or proof of a biological mechanism.
Clinicians can ask about timing, frequency, product type, route, reason for use, tobacco or alcohol co-use, and symptoms such as nausea, pain, anxiety, sleep difficulty, or withdrawal. A nonjudgmental approach is more likely to preserve honest disclosure and continued prenatal care.
Patients who are pregnant or planning pregnancy should discuss cannabis use with their obstetric and prescribing clinicians. Counseling should include evidence-based alternatives for symptoms, support for reducing or stopping use, and urgent assessment when severe vomiting, dehydration, bleeding, contractions, or other concerning symptoms occur.
Professional guidance generally recommends avoiding cannabis during pregnancy because safety has not been established and observational evidence raises concern. Counseling should still distinguish uncertainty from reassurance and avoid overstating a single estimate.
Pregnancy-related cannabis use may involve symptom treatment, habitual use, or cannabis use disorder. Effective care addresses the reason for use, offers realistic alternatives, and keeps the patient engaged rather than relying on stigma.
This study is useful because it separates preconception exposure from early-pregnancy exposure and reports several birth outcomes individually. That makes the clinical conversation more precise.
The result does not provide a product-specific risk calculator. It does reinforce a cautious recommendation to avoid cannabis during pregnancy and to make counseling practical, supportive, and tailored to the patient’s reason for use.
How to Read This Pregnancy Safety Signal
The study followed couples prospectively from before conception and reported exposure timing separately.
Four checks keep the result within the evidence.
A Four-Step Reading Frame
Separate the exposure windows
Preconception use and early-pregnancy use were measured and analyzed separately.
Separate the outcomes
Preterm birth and birthweight-for-gestational-age outcomes are not interchangeable.
Read the interval
The frequent-use estimate was elevated, but the confidence interval was wide.
Keep association distinct from cause
A prospective cohort improves timing but does not randomize exposure or eliminate residual confounding.
Early-Pregnancy Cannabis Exposure Through Eight Clinical Lenses
What the cohort can inform for patients, clinicians, partners, researchers, and careful readers
A Signal Is Not a Verdict
If you used cannabis before realizing you were pregnant, this study cannot determine what will happen in your pregnancy. Its elevated estimate applies to a group reporting at-least-weekly early-pregnancy use, not to a specific person, product, or episode.
Tell your obstetric clinician what you used, when, how often, and why. The goal is supportive care, symptom alternatives, and appropriate prenatal follow-up. Avoiding shame helps preserve honest communication and makes a practical reduction or stopping plan more achievable.
Ask Beyond Yes or No
A binary cannabis question misses clinically useful detail. Ask about timing relative to conception, current frequency, route, THC and CBD content when known, tobacco or alcohol co-use, the symptom target, and any difficulty reducing use.
The prospective signal supports counseling toward avoidance during pregnancy, but counseling should not overstate causal certainty. Offer symptom-specific alternatives, screen for cannabis use disorder when indicated, document shared decisions, and coordinate with obstetric, primary care, psychiatric, or addiction care.
Partner Exposure Is a Separate Question
The study included male preconception cannabis use and did not report an appreciable association with preterm birth in that analysis. That finding should not be converted into proof that paternal or partner use has no reproductive, household, or caregiving implications.
Partners can help by supporting a lower-exposure home, avoiding smoke around the pregnant person, securing products, and sharing responsibility for symptom care and appointments. The most useful response is practical support, not surveillance, accusation, or false reassurance.
Prospective Still Means Observational
Following participants from before conception improves temporal clarity and reduces some forms of recall error. It does not randomize cannabis use. People who use frequently during early pregnancy may differ from nonusers in measured and unmeasured ways that also influence preterm birth.
The authors adjusted for potential confounders, according to the abstract, but full methods were unavailable for this review. Residual confounding, exposure misclassification, missing information, and selection into an internet-based cohort remain important questions rather than reasons to dismiss the result.
The Confidence Interval Matters
The hazard ratio for at-least-weekly early-pregnancy use was 1.87, with a 95% confidence interval from 1.06 to 3.29. The interval excludes 1.0 but spans effects of very different clinical magnitude, which signals limited precision.
Less-than-weekly use produced a different, highly imprecise estimate whose interval included both lower and higher incidence. These categories do not establish a clean dose-response relationship. Readers should resist turning one significant category into an exact risk estimate or safe threshold.
Safety Counseling Should Not Become Punishment
Pregnancy cannabis disclosure can trigger fear of stigma, legal consequences, or loss of trust. Those pressures can reduce accurate reporting and delay care. An observational safety signal should be used to improve access to support, not justify biased assumptions or automatic punitive responses.
Consistent counseling standards matter across race, income, geography, and insurance status. Clinicians should explain what is known, what remains uncertain, and what alternatives are available. Patients should retain dignity, informed participation, and access to prenatal and substance-use care.
Measure Product and Exposure More Precisely
Future studies should capture THC and CBD content, route, dose, frequency, timing, co-use, indication, and changes across pregnancy. Biological measures can complement self-report, but they also require careful consent, interpretation, and safeguards against punitive misuse.
Larger samples could improve precision for frequent exposure and examine spontaneous versus medically indicated preterm birth. Transparent missing-data analyses, sensitivity analyses for residual confounding, and replication across diverse populations would clarify how robust and generalizable the association is.
Address the Reason for Use
A recommendation to stop is more actionable when it includes a plan for the problem cannabis was addressing. Nausea, sleep difficulty, pain, anxiety, withdrawal, and habitual use require different assessments and evidence-based alternatives.
Clinicians should ask what has helped, what barriers exist, and whether the patient can reduce or stop safely with outpatient support. Severe vomiting, dehydration, psychiatric symptoms, or inability to control use may require more urgent or specialized care. Follow-up turns advice into care.
Join the Conversation
Have a question about how this applies to your situation? Ask Dr. Caplan
Want to discuss this topic with other patients and caregivers? Join the forum discussion
Frequently Asked Questions
Did the study prove that cannabis causes preterm birth?
No. It was a prospective observational cohort, so it can identify associations but cannot establish that cannabis caused the outcome.
How many couples were included?
The analysis included 6,838 U.S. and Canadian couples who conceived between June 2013 and September 2024.
What exposure was associated with preterm birth?
At-least-weekly cannabis use during early pregnancy was associated with higher preterm-birth incidence than nonuse in the adjusted analysis.
How precise was the frequent-use estimate?
The hazard ratio was 1.87, with a 95% confidence interval from 1.06 to 3.29, indicating substantial uncertainty about the effect size.
Was preconception cannabis use associated with preterm birth?
The abstract reports no appreciable association for female or male preconception use in the analyzed categories.
Was cannabis associated with small- or large-for-gestational-age birth?
The abstract reports no association with either small- or large-for-gestational-age birth.
Does the study identify a safe amount of cannabis during pregnancy?
No. The exposure categories do not establish a safe dose, product, route, frequency, or timing.
What should someone do after cannabis exposure early in pregnancy?
Discuss the timing, frequency, product, and reason for use with an obstetric clinician, and make a supported plan without self-blame or delayed care.
Why does self-reported exposure matter?
Recall, stigma, product uncertainty, and incomplete dose information can introduce exposure misclassification.
Was the full paper reviewed for this report?
No. The full article was not openly accessible, so this report confines study-specific claims to the peer-reviewed indexed abstract.
