GLP-1 Drugs and Pregnancy: The Safety Data
Peptides in Pregnancy and Lactation
2 months minimum washout
FDA labeling recommends discontinuing semaglutide at least 2 months before planned pregnancy due to its 7-day half-life and animal reproductive toxicity data.
FDA Prescribing Information, Wegovy/Ozempic
FDA Prescribing Information, Wegovy/Ozempic
View as imageMillions of women of reproductive age now take GLP-1 receptor agonists for diabetes or weight management. The drugs come with a clear labeling instruction: stop before pregnancy. FDA prescribing information for semaglutide recommends discontinuation at least 2 months before a planned pregnancy, based on the drug's approximately 7-day half-life and animal reproductive toxicity data showing fetal harm at high doses. But the actual human evidence is more nuanced than the label suggests. Recent observational data from over 1,000 pregnancies exposed to GLP-1 agonists in the first trimester has not found an increased rate of major birth defects. At the same time, reports of unexpected pregnancies in women on semaglutide, the so-called "Ozempic babies" phenomenon, have raised questions about whether these drugs inadvertently increase fertility through weight-loss-mediated restoration of ovulation. This article is part of the peptides in pregnancy and lactation cluster examining how peptide hormones and peptide-based drugs interact with reproductive biology.
Key Takeaways
- FDA labeling recommends stopping semaglutide at least 2 months before planned pregnancy; liraglutide requires at least 3 days; tirzepatide requires 25-35 days
- Animal studies show fetal skeletal anomalies and growth restriction at maternally toxic doses, but these have not been confirmed as direct teratogenic effects in humans
- A multicenter observational study of GLP-1 RA exposure in early pregnancy (6 Teratology Information Services) found no increased risk of major birth defects compared to diabetes or obesity reference groups
- In pooled FDA clinical trial data, congenital abnormalities occurred in 2.7% of GLP-1 RA-exposed pregnancies versus 5.3% in the placebo group
- The "Ozempic babies" phenomenon likely reflects weight loss restoring ovulation in previously anovulatory women with obesity or PCOS, not a direct fertility-enhancing drug effect
- GLP-1 agonists have high molecular weight (~4,114 Da for semaglutide), and placental transfer in the first trimester is generally not anticipated based on molecular size
What the animal data actually shows
The basis for pregnancy contraindications comes from preclinical reproductive toxicology studies required by regulators before drug approval. These studies test drugs at multiple dose levels in pregnant animals, typically rats and rabbits.
For semaglutide, animal studies showed fetal skeletal anomalies and growth restriction. For liraglutide, exenatide, and dulaglutide, similar findings appeared. The critical context: these effects occurred at maternally toxic doses, meaning doses high enough to cause toxicity in the mother animal. At therapeutic-equivalent doses, the reproductive toxicity signals were absent or minimal.
The distinction between teratogenic (directly causing birth defects through fetal exposure) and maternally toxic (causing fetal harm indirectly through maternal toxicity) matters. A drug that harms a fetus because the mother is severely ill from drug toxicity tells you something different from a drug that crosses the placenta and directly disrupts fetal development. Based on available animal data, GLP-1 receptor agonists have not been classified as direct teratogens. The current contraindication is based on the precautionary principle: given insufficient human safety data, the regulatory default is to recommend avoidance.
The human evidence: more reassuring than expected
Teratology Information Services multicenter study
A multicenter, observational, prospective cohort study based on six Teratology Information Services databases examined outcomes of pregnancies exposed to GLP-1 receptor agonists in early pregnancy. The results: exposure was not associated with increased risk of major birth defects when compared to pregnancies in women with diabetes or in overweight/obese women without GLP-1 exposure. Cox proportional hazard models found no increased risk of pregnancy loss in the GLP-1 group.
Pooled FDA clinical trial data
Parker et al. analyzed pregnancy outcomes from regulatory clinical trial databases for GLP-1 agonists. Among GLP-1 RA-exposed pregnancies, 43% resulted in healthy children compared to 34% in the placebo group. Congenital abnormalities occurred in 2.7% of GLP-1 RA-exposed pregnancies versus 5.3% in placebo pregnancies. GLP-1 RA exposure was linked to a lower rate of congenital heart defects specifically.
These numbers must be interpreted carefully. Clinical trial pregnancies are typically discovered incidentally (not planned), and exposure is usually limited to early pregnancy before the drug is stopped. The comparison groups are small and the data is observational within trial populations, not from dedicated pregnancy safety trials. Ethical considerations make it unlikely that a randomized trial of GLP-1 drugs in pregnant women will ever be conducted.
Semaglutide-specific outcomes
A 2025 study examining pregnancy outcomes after semaglutide exposure found no increased risk of major malformations compared to insulin-exposed or unexposed pregnancies. Semaglutide-exposed pregnancies did show higher rates of preterm birth, large-for-gestational-age infants, neonatal hypoglycemia, and jaundice. However, these rates were similar to those in insulin-exposed pregnancies, suggesting they reflect the underlying diabetic condition rather than semaglutide-specific toxicity.
Why first-trimester transfer is unlikely
GLP-1 receptor agonists are large peptide molecules. Semaglutide has a molecular weight of approximately 4,114 Da. Placental transfer of molecules this large is generally not anticipated during the first trimester, when the placental barrier is relatively impermeable to large molecules. By the time the placenta develops sufficient transport capacity for larger molecules (second and third trimester), most women have already discontinued the drug after discovering pregnancy.
The washout math: different drugs, different timelines
Each GLP-1 agonist has a different half-life, which determines how long the drug remains in circulation after the last dose. The standard pharmacological rule: it takes approximately 5 half-lives to clear 97% of a drug from the body.
| Drug | Half-life | Time to clear (5 half-lives) | Recommended stop before conception |
|---|---|---|---|
| Semaglutide (Ozempic/Wegovy) | ~7 days | ~35 days | 2 months (FDA label) |
| Tirzepatide (Mounjaro/Zepbound) | ~5 days | ~25 days | 25-35 days (guidance) |
| Liraglutide (Victoza/Saxenda) | ~13 hours | ~3 days | At least 3 days |
| Exenatide (Byetta) | ~2.4 hours | ~12 hours | Minimal washout needed |
| Dulaglutide (Trulicity) | ~5 days | ~25 days | Similar to tirzepatide |
The FDA's 2-month recommendation for semaglutide is conservative, providing roughly 8 half-lives of clearance rather than the minimum 5. This reflects the precautionary approach: in the absence of definitive human safety data, regulators add margin.
"Ozempic babies": weight loss restoring fertility
Reports of unexpected pregnancies in women taking semaglutide became widespread in 2023-2024. The mechanism is almost certainly indirect: weight loss improving reproductive function rather than semaglutide directly enhancing fertility.
Obesity disrupts ovulation through multiple pathways. Excess adipose tissue produces estrogen, which suppresses gonadotropin-releasing hormone (GnRH) pulsatility. Insulin resistance, common in obesity, increases ovarian androgen production and disrupts follicular development. In women with polycystic ovary syndrome (PCOS), even modest weight loss of 5-10% can restore regular ovulation.
Semaglutide produces mean weight loss of 15%. In a woman with obesity-related anovulation or PCOS, this magnitude of weight loss can rapidly restore menstrual cyclicity and fertility. Women who believed they were infertile or who relied on irregular cycles as de facto contraception suddenly became pregnant, sometimes within the first few months of GLP-1 therapy.
This phenomenon has been widely reported in media, patient forums, and clinical practice. The fertility restoration is particularly striking in women with PCOS, where anovulation may have persisted for years before GLP-1 therapy.
This has clinical implications for contraception counseling. Women of reproductive age starting GLP-1 agonists who do not wish to become pregnant need reliable contraception from the first dose, not after weight loss occurs. The fertility restoration can occur faster than the 2-month washout period required before a safe pregnancy, creating a window of risk where a woman is newly fertile while still on a drug that should be stopped before conception.
Oral contraceptive interaction
Semaglutide delays gastric emptying, which can alter the absorption of oral medications including combined oral contraceptive pills. The clinical significance of this interaction for contraceptive efficacy is not definitively established, but theoretical concern exists that delayed absorption could reduce peak contraceptive hormone levels. Non-oral contraceptive methods (IUDs, implants, injectables) are not affected by gastric emptying changes.
This interaction becomes practically relevant because the same women who need reliable contraception while taking GLP-1 agonists (reproductive-age women who may experience restored fertility) may have reduced oral contraceptive efficacy if they rely on the pill. The interaction is pharmacokinetic (absorption-related), not pharmacodynamic, so it applies to all orally administered estrogen/progestin combinations.
Breastfeeding: even less data
Data on GLP-1 agonists during breastfeeding is essentially absent. Animal studies show semaglutide is excreted in milk in rats, but human lactation data does not exist. Given the large molecular size of GLP-1 agonists, significant transfer into breast milk is considered unlikely, but "unlikely" is not the same as "studied." Current labeling recommends against use during breastfeeding. Oxytocin, the endogenous peptide that drives lactation, operates through entirely different receptors and pathways; GLP-1 agonists do not appear to affect milk production or let-down.
What happens if you get pregnant while on semaglutide
Many pregnancies are unplanned. In the United States, approximately 45% of pregnancies are unintended. Given that millions of reproductive-age women now take GLP-1 agonists, pregnancies occurring during active treatment are inevitable regardless of labeling instructions.
The clinical approach when pregnancy is discovered during GLP-1 therapy: stop the drug immediately. Most pregnancies are detected at 4-6 weeks gestation. If semaglutide was last dosed within the previous week, the drug will take approximately 5 weeks to clear fully. During this clearance period, fetal exposure occurs primarily through maternal circulation, but the large molecular size limits placental transfer in early pregnancy.
The observational data reviewed above comes almost entirely from this scenario: pregnancies where GLP-1 exposure occurred before the woman knew she was pregnant. The absence of excess birth defects in this real-world exposure population provides the most practically relevant safety data. It does not prove safety at the level a randomized trial would, but it addresses the exact clinical situation that patients and clinicians face.
Registry-based surveillance is ongoing. The manufacturer of semaglutide maintains a pregnancy exposure registry, and multiple countries are tracking outcomes through national health databases. As these datasets grow, the evidence will become more definitive. For now, the data is reassuring for incidental early exposure, even as the precautionary labeling against planned use during pregnancy remains appropriate.
The safety review context
The comprehensive safety review by Smits and Van Raalte noted that semaglutide's safety profile across SUSTAIN and PIONEER trials was dominated by transient GI effects, with no unexpected safety signals[1]. Peter and Bain's safety review of injectable semaglutide similarly found a profile typical of the GLP-1 class[2]. Neither review identified pregnancy-specific concerns beyond those already in the labeling, but both acknowledged that pregnant women were excluded from all clinical trials by design.
The broader GLP-1 safety landscape from the STEP trials[3] shows a drug class with a well-characterized adverse event profile in non-pregnant adults. Translating this safety profile to pregnancy requires the kind of large-scale pregnancy registry data that is only now being accumulated.
The Bottom Line
GLP-1 agonists are contraindicated in pregnancy based on animal reproductive toxicity data at maternally toxic doses and the precautionary principle. Human observational data from over 1,000 exposed pregnancies has not shown increased birth defect rates. The "Ozempic babies" phenomenon reflects weight-loss-mediated fertility restoration, not direct drug fertility enhancement. Semaglutide requires a 2-month washout before conception; liraglutide needs only 3 days. The evidence gap is closing as pregnancy registries accumulate data, but prospective safety trials in pregnant women are unlikely to be conducted for ethical reasons.