GLP-1 during pregnancy: safety + registry data
GLP-1 drugs like semaglutide (Ozempic, Wegovy) and tirzepatide (Mounjaro, Zepbound) are now used by millions of women of childbearing age, and a lot of them get pregnant while taking one. The honest answer about safety is uncomfortable: there is no human trial proving these drugs are safe in pregnancy, every label says to stop them, and the data we do have comes from women who got pregnant by accident. This guide walks through what that real-world evidence actually shows, where it is reassuring, where it is thin, and what the registries are starting to tell us.
GLP-1 drugs like semaglutide (Ozempic, Wegovy) and tirzepatide (Mounjaro, Zepbound) are now used by millions of women of childbearing age, and a lot of them get pregnant while taking one. The honest answer about safety is uncomfortable: there is no human trial proving these drugs are safe in pregnancy, every label says to stop them, and the data we do have comes from women who got pregnant by accident. This guide walks through what that real-world evidence actually shows, where it is reassuring, where it is thin, and what the registries are starting to tell us.
Why This Question Matters Now
Prescriptions for GLP-1 receptor agonists among reproductive-age women have climbed sharply. These medications suppress appetite, slow stomach emptying, and improve insulin sensitivity. They can also restore ovulation in women with obesity or polycystic ovary syndrome, which is the source of the so-called "Ozempic baby" stories. A woman who has not had a regular cycle in years may suddenly become fertile a few months into treatment.
That creates a real problem. Many of these women are on contraception that the drug itself may make less reliable, or they assumed they could not conceive. So unplanned first-trimester exposure happens, and it happens often enough that researchers now have thousands of pregnancies to study after the fact.
None of these drugs are approved for use in pregnancy. Not one. The question is not "should you take a GLP-1 to lose weight while pregnant" — the answer to that is a flat no from every regulator and medical society. The real question is narrower: if a woman was on a GLP-1 and got pregnant before she knew, how worried should she be? That is where the new data comes in.
How GLP-1 Drugs Work and Why Pregnancy Is a Concern
GLP-1 (glucagon-like peptide-1) is a gut hormone. The drugs mimic it. They bind GLP-1 receptors in the pancreas, brain, and gut to boost insulin release when blood sugar is high, blunt hunger signals, and slow how fast food leaves the stomach. Tirzepatide adds a second target, the GIP receptor. For the full mechanism and the broader risk picture in non-pregnant adults, see our GLP-1 side effects complete guide.
There are three separate reasons pregnancy raises a red flag:
Animal studies showed harm. In rats and rabbits, semaglutide and tirzepatide caused birth defects, pregnancy loss, and reduced fetal growth at doses tied to drug exposure during organ formation. Regulators take animal teratology seriously even when human data is missing. It is unclear how much of the animal harm came from the drug directly versus the severe maternal weight loss the drug caused, but the signal is real and it is on every label.
Weight loss during pregnancy is not the goal. Pregnancy is a time to gain weight in a controlled way, not lose it. A drug that crushes appetite and causes nausea can push a mother toward not eating enough, which starves the fetus of nutrients. Even if the molecule were harmless, the effect on intake could harm the baby.
The drugs linger. Semaglutide has a half-life of about a week. It takes roughly five weeks to clear most of it. Tirzepatide is similar — MotherToBaby estimates up to 30 days on average for most of the drug to leave the body. That long tail is why labels tell women to stop well before trying to conceive, not just when they get a positive test.
What the FDA Labels Actually Say
The prescribing information is blunt and consistent across the class.
The Wegovy (semaglutide) FDA label instructs patients to discontinue the drug at least 2 months before a planned pregnancy because of the long half-life of semaglutide. It states that based on animal reproduction studies there may be potential risks to the fetus, and that weight loss offers no benefit to a pregnant patient and may cause fetal harm. The Ozempic label carries the same 2-month washout instruction.
Tirzepatide labels (Mounjaro, Zepbound) carry parallel warnings. The MotherToBaby tirzepatide fact sheet confirms that no human studies have been done to see whether tirzepatide increases birth defects, that animal studies found increased risk for some defects, and that it can take up to 30 days on average for most of the drug to clear the body.
So the regulatory bottom line is settled: stop the drug, ideally two months before trying to conceive. The labels do not say the drug definitely causes harm in humans — they say we do not know, the animal data is worrying, and the precautionary move is to stop.
| Drug (brand) | Approved in pregnancy? | Washout before trying to conceive | Animal teratogenicity signal | Human pregnancy data |
|---|---|---|---|---|
| Semaglutide (Ozempic, Wegovy) | No | At least 2 months | Yes (rats, rabbits) | Observational only, growing |
| Tirzepatide (Mounjaro, Zepbound) | No | ~30 days to clear | Yes (animal studies) | Very limited |
| Liraglutide (Victoza, Saxenda) | No | Per label, discontinue | Yes | Some cohort data |
| Dulaglutide (Trulicity) | No | Per label | Yes | Sparse |
The Human Evidence: What We Actually Know
Here is the key thing to understand. Because no one can ethically run a randomized trial giving pregnant women a drug labeled as potentially harmful, every piece of human evidence is observational. It comes from insurance claims, birth registries, and accidental exposures in trials. These studies cannot prove safety. They can only look for warning signs — and so far, the bigger studies have not found a clear one. That is genuinely reassuring, but it is not the same as proof of safety.
The JAMA Internal Medicine multinational cohort (2024)
This is the largest and most cited study. Cesta and colleagues pooled data from the Nordic countries, Israel, and the United States, covering more than 50,000 pregnancies in women with type 2 diabetes. They identified 938 pregnancies with periconceptional GLP-1 exposure and compared them against women using insulin.
The result: no increased risk of major birth defects. The risk ratio for any major congenital malformation was 0.95 (95% CI, 0.72–1.26), and for cardiac malformations 0.68 (95% CI, 0.42–1.12). Roughly 8.2% of GLP-1-exposed infants had a major malformation versus 7.8% in the insulin group — essentially the same, and both reflect the elevated baseline risk that diabetes itself carries. You can read the full study via PubMed (Cesta et al., JAMA Internal Medicine, 2024).
This is the strongest reassurance we have. The caveat: it used insulin as the comparison, so it tells us GLP-1 is not clearly worse than insulin in diabetic mothers. It does not tell us much about healthy-weight women without diabetes, and confidence intervals still leave room for a modest effect.
The Annals of Internal Medicine target trial emulation (2026)
The most recent major analysis used a "target trial emulation" design on US insurance claims from 2011 to 2024. Researchers compared women who continued a GLP-1 into the first trimester (filled at least one more prescription after conception) against women who stopped. The cohort was 3,572 pregnancies.
The finding: continuing the drug into early pregnancy did not substantially raise the risk of nonlive birth, abnormal fetal growth, or major congenital malformation. Nonlive birth risk was 29.7% with continuation versus 27.1% without — a small, statistically unconvincing difference. The authors were careful to flag that estimates for rare outcomes like major malformations were imprecise, so more research is needed. Details are on PubMed (Annals of Internal Medicine, 2026).
The Danish nationwide cohort (2026)
A Danish study tracked 756,636 singleton pregnancies, of which 529 involved periconceptional GLP-1 exposure (mostly liraglutide and semaglutide). After propensity-score matching, it found a raised preterm-birth risk — but only in women using the drug for diabetes (semaglutide adjusted odds ratio 1.84, 95% CI 1.24–2.71), not in women using it for weight management (adjusted odds ratio 0.71, 95% CI 0.30–1.70). The authors concluded the diabetes itself, not the drug, likely drove the preterm signal. See PubMed (Human Reproduction Open, 2026).
This is an important nuance. When a study lumps diabetic and non-diabetic women together, the underlying disease can masquerade as a drug effect. Separating them out made the apparent risk disappear in the weight-management group.
The systematic review of clinical-trial exposures (2025)
Parker and colleagues pulled FDA data on women who got pregnant by accident during GLP-1 trials. Numbers were small — a handful of pregnancies per drug program — with a mix of healthy births, spontaneous abortions, ectopic pregnancies, and elective terminations, and no clear pattern of birth defects. The authors stressed that the class remains contraindicated until better data arrives. Full text on PubMed (Diabetes, Obesity and Metabolism, 2025).
The meta-analysis of registry cohorts (2024)
Kuitunen reviewed register-based cohorts covering roughly 938 pregnancies and likewise found no association between early-pregnancy GLP-1 exposure and congenital malformations — while flagging that sample sizes were small and confidence intervals wide. See PubMed (Acta Paediatrica, 2024).
Evidence Grading: An Honest Scorecard
This is where it pays to be straight. The data is reassuring but not conclusive, and the quality varies a lot by question. Here is how the evidence stacks up.
| Outcome | What the data shows | Evidence strength | Honest caveat |
|---|---|---|---|
| Major birth defects | No clear increase vs. insulin or vs. stopping | Moderate | Largest study compares only to insulin; rare-outcome estimates imprecise |
| Cardiac defects | No clear increase (RR 0.68) | Low-moderate | Wide confidence intervals; cannot rule out a small effect |
| Miscarriage / nonlive birth | No substantial increase | Low-moderate | Background rate already high; hard to separate from obesity/diabetes |
| Preterm birth | Raised only in diabetic users; not in weight-loss users | Low | Likely confounded by underlying diabetes |
| Fetal growth restriction | No clear signal; theoretical concern from appetite loss | Very low | Few events; biologically plausible mechanism remains |
| Long-term child outcomes | Essentially no data | None | No follow-up studies of exposed children into childhood |
Two things drive the "low" and "very low" grades. First, almost no study can fully separate the drug from the conditions it treats — obesity and diabetes both independently raise the risk of miscarriage and birth defects, so a "safe" result might just mean the drug offsets some of that baseline risk. Second, the studies are observational and underpowered for rare outcomes. A drug can look clean across a few thousand pregnancies and still carry a small risk that only shows up across tens of thousands.
Be skeptical of any source — including manufacturer materials or weight-loss clinic blogs — that frames this evidence as "GLP-1s are safe in pregnancy." That overstates it. The accurate framing is: "early exposure has not shown a clear danger signal in the studies done so far, but the drugs are not approved and should be stopped."
What To Do If You Are Trying To Conceive
If you are on a GLP-1 and want to get pregnant, the plan is straightforward and every major society agrees on it.
Stop the drug before you try. Because of the long half-life, the labels call for stopping at least 2 months before conception for semaglutide and roughly a month of clearance for tirzepatide. Talk to your prescriber about timing so your weight and blood sugar are managed by other means before you conceive.
Switch diabetes care to insulin. Both ACOG and the American Diabetes Association recommend insulin as first-line therapy for pregestational diabetes when trying to conceive and during pregnancy. Insulin has decades of safety data in pregnancy that GLP-1 drugs simply do not.
Get contraception counseling first. GLP-1 drugs can restore fertility faster than expected and may blunt the effectiveness of oral birth control, partly through delayed stomach emptying. If you are not ready to conceive, clinicians often suggest a long-acting reversible method. This matters because the surprise pregnancies driving the safety question are usually the result of fertility returning before the woman planned for it. Whether a GLP-1 fits your situation at all is a separate question — our GLP-1 eligibility guide covers who these drugs are and are not for.
If you have PCOS, plan extra carefully. Women with polycystic ovary syndrome are a high-risk group here because the drug can trigger ovulation in someone who assumed she could not conceive. We cover the off-label use in detail in our piece on semaglutide for PCOS.
If You Already Got Pregnant On a GLP-1
This is the scenario most women actually face. You took the drug not knowing you were pregnant, and now you are worried.
The honest reassurance: the data above should lower your alarm. The largest studies have not found a clear rise in birth defects from first-trimester exposure. Most organ formation that matters happens in weeks 3 to 8, so an early stop limits how long the fetus is exposed.
The action steps are simple. Stop the drug as soon as you know. Tell your obstetrician right away so they can adjust your diabetes or weight management and schedule appropriate fetal screening, such as a detailed anatomy ultrasound around 18 to 20 weeks. And enroll in the pregnancy registry.
That registry matters. Novo Nordisk runs a Wegovy pregnancy registry that tracks outcomes in women exposed during pregnancy. Every enrolled pregnancy adds to the evidence base that future women will rely on. The reason we have any reassuring data at all is that women before you reported their outcomes. Reporting yours is how the gaps in this guide get filled.
How GLP-1 Pregnancy Risk Compares To Alternatives
It helps to put the risk in context against the other paths a woman in this situation might take.
| Option | Pregnancy safety | Trade-off |
|---|---|---|
| Stop GLP-1, switch to insulin (if diabetic) | Best-established safety record | Requires injections, glucose monitoring; weight may rise |
| Stop GLP-1, manage weight with diet/lifestyle | No fetal drug exposure | Slower, harder; some regain |
| Continue GLP-1 into pregnancy | Not recommended; no clear harm signal but unproven | Against all labels and guidelines |
| Metformin (for diabetes/PCOS) | Long history of use in pregnancy, generally considered acceptable | Less weight effect than GLP-1; GI side effects |
The pattern is clear. For a woman who is or wants to be pregnant, the safest move is to come off the GLP-1 and manage her condition with agents that have real pregnancy track records. The GLP-1 is a tool for before and after pregnancy, not during it. For how subgroups like prediabetes and cardiovascular patients weigh these drugs outside of pregnancy, see our GLP-1 by subgroup analysis, and for safe dosing when you do restart, our semaglutide dosing schedule guide.
Breastfeeding: A Separate Question
Pregnancy and breastfeeding are different risk pictures, and the breastfeeding one is a bit less worrying on paper. GLP-1 drugs are large peptide molecules. Large molecules generally do not pass into breast milk in meaningful amounts, and any that did would likely be broken down in the infant's digestive tract rather than absorbed. The MotherToBaby fact sheet notes there is no human data confirming this, but the biology points toward low transfer.
Still, no GLP-1 is approved for use while breastfeeding, and the same intake concern applies — appetite suppression in the mother could affect her nutrition and milk supply. The standard advice is to discuss it with your provider rather than assume it is fine.
The Bottom Line
GLP-1 drugs are not approved in pregnancy, every label tells women to stop them well before conceiving, and animal studies showed real harm. That is the cautious frame, and it is the correct one. At the same time, the human observational data that has piled up through 2026 — from JAMA Internal Medicine, Annals of Internal Medicine, and national registries — has not turned up a clear danger signal from accidental first-trimester exposure. Both of those statements are true at once.
If you are planning a pregnancy, stop the drug ahead of time and switch to a pregnancy-proven option. If you got pregnant on one, stop now, tell your doctor, and enroll in the registry. And treat anyone who tells you these drugs are simply "safe" in pregnancy with suspicion — the evidence does not support that claim, and it may not for years.
Frequently Asked Questions
Is it safe to take Ozempic or Wegovy while pregnant?
No GLP-1 drug is approved or recommended during pregnancy. The FDA labels tell women to stop semaglutide at least 2 months before trying to conceive. Observational studies of accidental early exposure have not found a clear rise in birth defects, but that is not proof of safety, and weight loss during pregnancy can itself harm the fetus.
What should I do if I got pregnant while taking a GLP-1?
Stop the drug as soon as you find out and contact your obstetrician right away. The reassuring news is that the largest studies have not shown a clear increase in major birth defects from first-trimester exposure. Your doctor will adjust your diabetes or weight management and arrange appropriate fetal screening. Enrolling in the manufacturer's pregnancy registry also helps future patients.
How long before pregnancy should I stop a GLP-1 medication?
For semaglutide (Ozempic, Wegovy), the label says stop at least 2 months before a planned pregnancy because of the drug's long half-life. For tirzepatide, MotherToBaby estimates it takes up to about 30 days on average for most of the drug to clear. Always confirm the timing with your prescriber.
Can GLP-1 drugs cause birth defects?
Animal studies showed birth defects, which is why the labels carry warnings. In humans, the larger cohort studies — including a JAMA Internal Medicine analysis of 938 exposed pregnancies — have not found a clear increase in major malformations compared with insulin. The data is reassuring but limited, and rare outcomes are still hard to measure precisely.
Can I take a GLP-1 while breastfeeding?
No GLP-1 is approved for breastfeeding. The biology suggests little of these large-molecule drugs passes into breast milk, and any that did would likely break down in the baby's gut, but there is no human data to confirm this. Discuss it with your provider before restarting.
This article is for general information only and is not medical advice. GLP-1 medications are not approved for use during pregnancy. Always consult your doctor or obstetrician about medication decisions before, during, and after pregnancy.
On Google
Get our answers in your Google results.
Add The GLP-1 Daily as a preferred source and Google will surface our reporting more often — in Top Stories and AI answers, marked with a preferred badge. One tap, free, undo anytime.
Add us as a preferred sourceOpens Google's source preferences for theglp1daily.com. No sign-up with us — it's a Google setting.