GLP-1 medicines, birth control and pregnancy: what labels and studies say

What the four labels say about birth control pills

This page reports what labels, a regulator and published studies say. It does not tell anyone what to use, and a prescriber or gynecologic clinician decides. The labels the Atlas read are Mounjaro (8/2026), Zepbound (8/2026), Ozempic (5/2026) and Wegovy (6/2026).

The wording is paraphrased from the DailyMed labels. Labels change, and the current text may differ. The tirzepatide page and the semaglutide page have more on each medicine.

What the tirzepatide labels measured

The pill advice rests on one study described in the Mounjaro and Zepbound labels. People took a single 5 mg dose of tirzepatide together with a combined oral contraceptive containing 0.035 mg of ethinyl estradiol and 0.25 mg of norgestimate. The labels report that tirzepatide lowered the peak and total blood levels of the pill's hormones and delayed the peak by 2.5 to 4.5 hours.

Source: the Mounjaro and Zepbound labels, which use identical wording. The labels say the delay in stomach emptying is largest after the first dose and diminishes over time, and they give 4-week windows after starting and after each dose increase. They report blood levels, not pregnancy rates.

What UK regulators say

The UK Medicines and Healthcare products Regulatory Agency (MHRA) posted guidance on GLP-1 medicines in June 2025, and the version the Atlas read carries a revision date of February 5, 2026. In that version:

• It says GLP-1 medicines should not be taken during pregnancy or just before trying to get pregnant, because there is not enough safety data to know whether they can harm a baby. It adds that some animal studies found harm to the unborn offspring.
• It says anyone who could become pregnant should take steps to avoid pregnancy while using these medicines, including during the washout period after stopping.
• For people on Mounjaro who use the pill, it says to use a barrier method such as condoms for four weeks after starting and after any dose increase, and describes switching to a non-oral method such as the coil or implant as an alternative. It gives the reason that GLP-1 medicines may reduce the effectiveness of oral contraceptives in those who are overweight or obese.
• It says GLP-1 medicines should not be taken by people who are breastfeeding, because of the lack of safety data.

What sources say about stopping before a planned pregnancy

Semaglutide stays in the body for a long time, and the sources that give a waiting period do not agree for every medicine.

The two sources match for semaglutide and differ for tirzepatide: the MHRA gives 1 month and the U.S. labels the Atlas read give no figure. The Atlas does not advise on timing.

What the labels say about pregnancy and breastfeeding

• Mounjaro: human data are not enough to judge the risk of major birth defects, miscarriage or other outcomes. In animal studies, rats had slower fetal growth and fetal abnormalities at clinical exposure, and rabbits had slower fetal growth. The label says to use it in pregnancy only if the potential benefit justifies the potential risk.
• Zepbound: the label says "Weight loss offers no benefit to a pregnant patient and may cause fetal harm," and says to discontinue Zepbound when a pregnancy is recognized. Its animal findings match Mounjaro's.
• Ozempic: the label says data in pregnant women are limited. Rats had embryo and fetal death, structural abnormalities and growth changes at clinical exposure, and rabbits and monkeys had early pregnancy loss or structural abnormalities. The label says the rat and rabbit findings came with marked weight loss in the mothers. The label gives background rates in the U.S. general population of 2% to 4% for major birth defects and 15% to 20% for miscarriage.
• Wegovy: the label says the pregnancy data from reports and trials are not enough to establish risk, repeats that weight loss offers no benefit to a pregnant patient, and says to discontinue Wegovy in pregnant patients using it for weight reduction. A pregnancy exposure registry exists.
• Breastfeeding: in the studies described in the Mounjaro and Zepbound labels, 11 healthy breastfeeding women took a single 5 mg tirzepatide dose, and the drug was undetectable in 164 of 171 milk samples. The other 7 samples held less than 0.02% of the dose. The effect on a breastfed infant was not studied. Ozempic's label reports no data on semaglutide in human milk. Wegovy's label says breastfeeding is not recommended during treatment with the tablets and reports no human-milk data for the injection.

What do studies of early-pregnancy exposure show?

Researchers cannot assign pregnant people to a weight-loss medicine, so nearly everything here comes from people who were exposed around conception or in early pregnancy, often by accident, and were then followed. The table covers the newest reviews and three cohort studies.

The Atlas read the abstracts of D'Antonio, Hakim, Collier and Lau, and the full texts of Hattler, Uysal, Dao and Cesta. Odds ratios and risk ratios near 1 mean similar rates in the exposed and comparison groups; the ranges in parentheses are 95% confidence intervals.

Why these studies cannot settle the question

• Nearly all are observational. Hattler's pooled studies include one long-term follow-up of a randomized trial, but in the rest people were not randomly assigned, so differences in diabetes severity, weight and other conditions can distort the comparison.
• Exposure usually means a filled prescription around conception, and studies often cannot say when use ended. In the Dao cohort the median stop was at 5 weeks of pregnancy, and the Hattler authors note that only three of their studies could confirm exact exposure end points.
• Most of the data are on liraglutide and semaglutide. The Hattler authors say tirzepatide and dual agonists are underrepresented.
• Several studies include only people with type 2 diabetes, whose pregnancies carry different baseline risks than those of people using a medicine for weight.
• Rare outcomes need large numbers. Wide confidence intervals, such as Dao's 0.16 to 5.82, leave room for anything from a large drop to a several-fold rise in risk.

Fertility and "Ozempic babies"

"Ozempic babies" is a popular phrase for unplanned pregnancies in people using semaglutide. The Atlas found no published study that counted unplanned pregnancies after people started a GLP-1 medicine, so how common they are is not known. A 2026 review in BJOG says GLP-1 medicines may improve fertility outcomes in selected groups, including women with obesity and polycystic metabolic ovarian syndrome (PMOS, the term the review uses for polycystic ovary syndrome).

The trial evidence is small and limited to one condition. In an open-label trial in China, 100 women with polycystic ovary syndrome and overweight or obesity took metformin with or without semaglutide for 16 weeks. Between weeks 16 and 40, the natural pregnancy rate was 35% in the semaglutide group and 15% in the metformin-only group. A 2026 meta-analysis of three trials of GLP-1-based treatment, alone or with metformin, against metformin, with 252 women, reported an odds ratio of 3.10 (95% CI 1.74 to 5.49) for natural pregnancy and called the signal promising but preliminary. Its pooled odds ratios for overall pregnancy (1.41, 0.81 to 2.45) and for pregnancy after IVF (1.01, 0.36 to 2.85) were not statistically significant.

A retrospective study of 1,936 women with overweight or obesity who were already in fertility care points the other way. Semaglutide use in the year before care was associated with lower and delayed pregnancy and live birth, and the authors call for confirmation. The findings differ by group, and none of these studies measures accidental pregnancy.

Quick answers

Do GLP-1 medicines make birth control pills less effective?

For tirzepatide, the Mounjaro and Zepbound labels say it may reduce the effectiveness of oral hormonal contraceptives, based on lower blood levels of the pill's hormones in a single-dose study. The labels tell patients on the pill to switch to a non-oral method or add a barrier method for 4 weeks after starting and after each dose increase. The Ozempic and Wegovy labels carry no such advice. A prescriber or gynecologic clinician decides.

Does semaglutide affect the pill?

Ozempic's label says its interaction studies found no clinically relevant interaction, and they included a pill with ethinylestradiol and levonorgestrel. Wegovy's label has no contraceptive advice. It says delayed stomach emptying may affect how oral medicines are absorbed, and it also reports that the semaglutide injection did not affect the absorption of orally administered medications and that its tablet interaction studies found no clinically significant differences for ethinyl estradiol and levonorgestrel.

How long before a planned pregnancy do the sources say to stop?

The Ozempic and Wegovy labels say at least 2 months, and so does the MHRA for semaglutide. The MHRA says at least 1 month for tirzepatide, and the U.S. tirzepatide labels the Atlas read give no figure. The MHRA says liraglutide has no waiting period. The Atlas does not advise on timing.

Are GLP-1 medicines safe in pregnancy?

The Atlas cannot say. The labels say human data are limited or insufficient, and the Zepbound and Wegovy labels say weight loss offers no benefit in pregnancy. Reviews of people exposed around conception have not shown a clear rise in birth defects, and their authors say the data cannot exclude risk.

What do studies show about birth defects after early exposure?

In Cesta's cohort of women with type 2 diabetes, the adjusted risk ratio for major malformations was 0.95 (95% CI 0.72 to 1.26) against insulin. In Dao's cohort, major birth defects were 2.6% with GLP-1 exposure and 2.3% and 3.9% in the comparison groups. Meta-analyses found no clearly detectable increase, with a few signals the authors attribute to confounding.

What do the labels say about breastfeeding?

The Mounjaro and Zepbound labels report tirzepatide undetectable in 164 of 171 milk samples from 11 women, with the effect on infants not studied. Ozempic's label reports no human-milk data. Wegovy's label says breastfeeding is not recommended during treatment with the tablets. The MHRA says GLP-1 medicines should not be taken during breastfeeding.

What are "Ozempic babies"?

A popular phrase for unplanned pregnancies in people using semaglutide. The Atlas found no study that counted them. A small trial in China of women with polycystic ovary syndrome reported more natural pregnancies with semaglutide plus metformin than with metformin alone. A 2026 meta-analysis of three trials of GLP-1-based treatment pointed the same way for natural pregnancy but not for overall pregnancy.

What do sources say about becoming pregnant while on one?

The Zepbound label says to discontinue Zepbound when a pregnancy is recognized, and the Wegovy label says to discontinue Wegovy in pregnant patients using it for weight reduction. The MHRA says to speak to a healthcare professional straight away. A clinician who knows your history decides.

Questions to ask

• Which medicine am I on, and does its label say anything about birth control pills?
• If I use the pill, which method would the label text point to for me, and for how long?
• If I might want to become pregnant, what do you suggest about timing and the medicine?
• What would we do if I found out I was pregnant while taking it?
• Does anything change if I am breastfeeding or plan to?

What is still unknown

• The Atlas found no published study that counted unplanned pregnancies after people started a GLP-1 medicine.
• The tirzepatide pill finding comes from a single-dose study of blood levels. The labels do not report pregnancy rates, and the Atlas found no study of pill failure in people on tirzepatide.
• Studies of early-pregnancy exposure mostly cover liraglutide and semaglutide, often without knowing when use stopped, and tirzepatide is underrepresented.
• The waiting periods before a planned pregnancy differ between sources for tirzepatide, and the Atlas does not advise on them.
• The Atlas read only the abstracts of several of these studies, not the full papers.