GLP-1s, Pregnancy and Trying to Conceive: What the Labels Say

If you are pregnant, or there is any chance you might be, tell a clinician this week and bring the name of the drug with you. That is the first sentence for a reason: none of what follows is a substitute for it. What this article can do is set out what the five approved labels actually say — about pregnancy, about the interval before a planned one, and about breastfeeding — so that the conversation starts from the same text your prescriber is working from. We read all five on August 15, 2026.

The interval before a planned pregnancy, drug by drug

This is the fact people come looking for, and it is not the same across the class. All three semaglutide labels name a period. Neither tirzepatide label names one.

Where the two months comes from

The labels tie the interval explicitly to how long the drug lingers. Semaglutide takes about a week to fall by half, and the Ozempic label says the drug is still circulating roughly five weeks after a last dose.^[2] Wegovy's clinical pharmacology section stretches that to five to seven weeks at its higher strengths.^[1] Two months is that tail plus a margin.

Tirzepatide clears faster. Its labels put the elimination half-life at approximately five days, and Zepbound's at five to six days in people with overweight or obesity.^[4][5] A shorter tail is a plausible reason for a shorter interval. It is not a reason to assume none, and neither Lilly label converts it into one. Our account of where the tirzepatide evidence stops covers the same habit in a different section of the label: a warning and a counted rate are separate objects, and so are a silence and a clearance.

What the animal studies found

Every one of the five labels rests its pregnancy caution on animal reproduction data, and every one is candid that the human data are insufficient.

• Semaglutide in rats: embryofetal death, structural abnormalities and altered growth, at maternal exposures no higher than the ceiling of the human dose range.^[2]
• Semaglutide in rabbits and cynomolgus monkeys: early pregnancy losses and structural abnormalities, at clinically relevant exposures in rabbits.^[1][2]
• Tirzepatide in rats: external, visceral and skeletal malformations and reduced fetal weights at the highest dose given during organogenesis.^[5]
• Tirzepatide in rabbits: a few animals died or aborted, at every dose level tested, and fetal weights fell at the top dose.^[5]
• SNAC, the absorption enhancer in the semaglutide tablets, crosses the placenta in rats; a rat study reported longer gestation, more stillbirths and lower pup viability.^[3]

Both manufacturers add the same qualification, and it matters. In the animals, these effects coincided with marked maternal weight loss and reduced food intake — the drugs doing to the mother what they are designed to do. That does not make the findings irrelevant. It does mean the effects may be downstream of reduced maternal intake rather than direct toxicity, and neither label claims to know which.^[1][5]

What the human record shows so far

Two 2026 analyses are the largest available, and neither settles the question.

The first emulated a trial inside U.S. insurance claims covering 2011 to 2024, comparing 3,572 pregnancies in women who had a GLP-1 prescription filled in the 90 days before their last menstrual period, split by whether they filled another one during the first trimester. Nonlive birth occurred in 29.7% of the continuation group against 27.1% of the noncontinuation group, an adjusted risk ratio of 1.09. Prevalence ratios were 1.29 for small-for-gestational-age, 1.08 for large-for-gestational-age and 1.21 for major congenital malformation. The authors' own conclusion is worth repeating exactly: risks were not definitively higher, but the malformation and growth estimates were imprecise and were compatible with both no increased risk and a clinically relevant one.^[6]

The second pooled 22 studies covering 49,395 pregnancies with exposure around conception. It found no statistically significant association with major congenital malformations, cardiac malformations, gestational age or weight, preterm delivery, live birth, pregnancy loss, hypertensive disorders or cesarean delivery. One signal did reach significance: renal malformations, at an odds ratio of 1.23, which the authors attribute largely to a single large cohort with substantial baseline imbalances.^[7]

Read together, they say a large and obvious signal has not appeared in what has been collected so far. They do not say these drugs are safe in pregnancy, and no label treats them as such. This is a genuinely thin evidence base being read carefully, which is a different thing from a reassuring one.

Losing weight during a pregnancy is not the goal

The obesity labels are unusually blunt. Wegovy's risk summary states that there is no benefit to a pregnant patient in losing weight and that it may harm the fetus, and it directs prescribers to stop the drug for anyone taking it to reduce weight once a pregnancy has been identified.^[1] Zepbound opens its own risk summary with the same sentence.^[4]

Both then point at the reason. Guidance still calls for weight to be gained during pregnancy, in an amount set by the weight someone started at, and that holds for people who already have overweight or obesity, because maternal tissue itself adds weight.^[1][4] Whatever case there was for treatment before conception, the target changes.

Wegovy is the one exception the label carves out, and it is narrow: for the liver indication, where the untreated condition carries its own risks to mother and fetus, the label permits treatment during pregnancy only where the likely benefit outweighs the possible risk, and states that whether it reduces those risks is unknown.^[1]

Breastfeeding: the labels split by formulation

This is where the answer stops tracking the molecule and starts tracking what else is in the product.

• The semaglutide tablets — Rybelsus, Ozempic tablets and Wegovy tablets — advise that breastfeeding is not recommended during treatment. Semaglutide itself was below the limit of quantification in human milk; the concern is SNAC and its metabolites, which are present in milk, and which infants may clear more slowly than adults.^[1][3]
• Injectable semaglutide carries no such advice. Wegovy injection and Ozempic have no human milk data at all, and both fall back on the standard formula: set the health and developmental value of breastfeeding against how much the mother needs the drug.^[1][2]
• Tirzepatide has an actual lactation study behind it. After a single 5 mg study dose given to 11 lactating women, tirzepatide was undetectable in 164 of 171 milk samples; across the 7 that were measurable, the cumulative amount over 28 days came to under 0.02% of the dose the mother received. There are still no data on effects on a breastfed infant or on milk production, and both labels use the same weighing exercise.^[4][5]

Fertility can come back before the plan does

This is the practical point most likely to matter, and it is the one least likely to be raised in an appointment. Losing a meaningful amount of weight can restore ovulation in people who were not ovulating reliably, particularly with polycystic ovary syndrome, and menstrual cycles that had been irregular can become regular. Reviews of the reproductive evidence describe exactly that: ovulatory function improving and conception becoming likelier, arriving alongside a human safety record in early pregnancy that is still small and heterogeneous.^[8][9]

So the sequence that catches people out is not carelessness. Someone starts treatment while subfertile, contraception is either absent or was never a live concern, fertility improves faster than the plan around it, and a pregnancy begins during treatment. If you have been told conceiving would be difficult, treat that assessment as out of date once the weight starts moving, and settle the contraception question before it becomes retrospective. The labeled contraceptive interaction is a separate issue and it points the same way.

The pregnancy registries

Two of the five labels ask for something back. Novo Nordisk runs a pregnancy exposure registry for Wegovy, reachable on 1-877-390-2760 or at wegovypregnancyregistry.com.^[1] The Zepbound label states that a registry will monitor outcomes in women exposed during pregnancy and directs patients and clinicians to Eli Lilly on 1-800-LillyRx.^[4] Ozempic, the semaglutide tablets and Mounjaro carry no registry text.

Enrolling is voluntary and it does nothing for your own pregnancy. It is how the evidence base above stops being thin, and right now it is thin because the exposures that have happened were mostly never recorded.

Questions worth taking to the appointment

1. Which drug am I on, and which label applies to me? The pre-pregnancy interval exists for semaglutide and not for tirzepatide, so the name decides the conversation.
2. Why am I taking it? Diabetes, weight, cardiovascular risk and liver disease are different indications, and the labels weigh pregnancy against each differently.
3. If I want to conceive, what is the plan and what replaces the treatment? Stopping is not a standalone decision, particularly with diabetes, where the labels are explicit that poorly controlled disease carries its own risks in pregnancy.
4. If I am already pregnant, what happens now? Ask about the registry at the same time.
5. If I am breastfeeding, does the formulation change the answer? For the tablets, the labels say it does.
6. Is my contraception adequate for the drug I am on? Ask it even if you are not planning a pregnancy.