Ozempic and Fertility: What to Do Before You Start Trying

Talk to whoever prescribes it before you start trying, and do not stop it on your own. Current guidance is that these medicines come off before you conceive, not when a test turns positive — and how far in advance depends on which drug you are on. It is a planning problem, not an emergency.

The order to do things in:

1. Check which molecule you are on. Ozempic, Rybelsus and Wegovy are all semaglutide. Mounjaro and Zepbound are tirzepatide — a different drug with different labelled advice.
2. Ask what interval that drug's label gives before a planned pregnancy, and what you take instead in the meantime if it is treating diabetes.
3. Ask whether your contraception needs to change while you are still on it. For one of these drugs, it might.
4. Then set a date for trying, counting backwards from it.

It depends on the drug, which is why a single number would mislead you. Semaglutide's US label is specific: stop at least two months before a planned pregnancy, because the drug takes a long time to clear. Tirzepatide's US label gives no pre-conception interval at all. The ADA Standards of Care 2026 flags that gap and notes the Canadian tirzepatide label advises stopping at least one month before pregnancy.

So ask which interval applies to your prescription rather than assuming two months covers everything. If the medicine is treating diabetes, the ADA adds that several months is the realistic figure, because you also need time to move onto a treatment approved in pregnancy and get glucose levels where they should be first.

Yes, by two separate routes that get run together. The fix is different for each.

The first is that losing weight can restart ovulation. The 2025 review in the American Journal of Obstetrics & Gynecology describes exactly this: women diagnosed with oligomenorrhoea who had not been able to conceive became pregnant unexpectedly after losing weight on these medicines. No contraception failed. The body became fertile again and nobody had planned around it. That review recommends everyone taking a GLP-1 use contraception.

The second route is the drug interfering with the pill, and here the two molecules genuinely differ. Tirzepatide's US label says it may reduce the efficacy of oral hormonal contraceptives, because it slows the stomach emptying, and advises switching to a non-oral method or adding a barrier method for four weeks after starting and four weeks after every dose increase — the effect being largest after the first dose. Hormonal methods that are not swallowed, such as an implant, an injection or a hormonal IUD, are not expected to be affected. Semaglutide is different: tested alongside an ethinylestradiol and levonorgestrel pill, it showed no clinically relevant effect on absorption.

In India this is not a hypothetical distinction. Both molecules are on the market here, and the contraceptive warning sits on the tirzepatide label.

On the published evidence you cannot tell, because the trials did not measure whether anyone had a baby. A 2026 systematic review pooling 18 randomised trials in PCOS found GLP-1 agonists lowered BMI and improved insulin resistance at moderate certainty, with no significant effect on total testosterone. Those outcomes are body measurements and blood markers. A separate network meta-analysis of 16 trials said so about itself in plain terms: reproductive outcomes were not evaluated, so no claim of overall superiority in PCOS can be made.

That is the gap. A medicine can reliably shift your weight and your insulin numbers without anyone having shown it shifts your chance of a pregnancy — and the second thing is what you came for. Mechanism is not outcome.

It also matters that GLP-1s are not part of the recognised fertility route for PCOS. The 2023 international PCOS guideline keeps lifestyle first and letrozole first-line for ovulation induction, with metformin for metabolic management rather than for getting you ovulating. Our page on what actually changes the outcome in PCOS goes through that ladder, and how ovulation induction works covers the step GLP-1s get mistaken for.

Not automatically, and one trial should give anyone pause before assuming otherwise. LIFEstyle, published in the New England Journal of Medicine in 2016, randomised infertile women with a BMI of 29 or above either to six months of a lifestyle programme followed by eighteen months of fertility treatment, or to prompt treatment for the full twenty-four. The women who did the weight-loss programme first lost more weight — and ended up with fewer healthy term births, not more.

That trial tested diet and exercise, not a GLP-1, and part of what it tested was the delay itself: the prompt-treatment group had more months of treatment in the same window. It does not make weight loss useless. It shows the time spent losing weight has a cost that can outweigh the benefit — a trade-off worth saying out loud if age or ovarian reserve is already a question. The 2016 paper carries a published erratum which we could not retrieve; the figures here are as originally published.

Our page on weight and female fertility covers what weight does and does not change, and the evidence on keto hits the same drug-versus-weight-loss problem.

Tell the prescribing clinician promptly, and do not stop any other prescribed medicine on your own while you wait to be seen. Exposure before a positive test is common and it is not a failure on your part — it is what happens with a drug that works before anyone knows there is a pregnancy.

A 2026 systematic review of 36 studies found exposure around conception or in early pregnancy was not consistently linked to more major congenital malformations, fetal growth restriction, stillbirth or newborn deaths, compared with women treated with insulin or with similar conditions. That is reassuring as far as it goes, and not the same as being established as safe. The AJOG review makes the limit concrete: the largest cohort, of 938 pregnancies, held no information on how well the mothers' glucose was controlled — and poor glucose control is itself what causes much of the harm being measured.

The caution in the labels comes from animal studies: in pregnant rats, rabbits and monkeys given semaglutide, researchers saw reduced growth, skeletal and internal abnormalities and early pregnancy loss, alongside marked maternal weight loss. Animal findings do not transfer directly to humans, but with human data this thin they carry more weight than they otherwise would. Expect your clinician to acknowledge the uncertainty rather than resolve it, and ask what monitoring is recommended.

This topic attracts confident claims in both directions, and the confident ones are the least supported. What is genuinely unknown:

• Whether GLP-1 agonists raise anyone's chance of a pregnancy or a live birth. No trial in the meta-analyses above measured either. They are not approved as fertility treatments and are not prescribed as ones.
• Whether they improve egg or sperm quality. Nothing retrieved here shows that.
• Whether they help PCOS androgens. The pooled effect on total testosterone was not statistically significant, at very low certainty.
• Whether restored ovulation lasts once the drug stops, or whether regained weight takes it away again.
• Whether they are safe in pregnancy. Observational data have not thrown up a consistent signal, which is weaker than safety, and the labels still advise against use in pregnancy.
• What continued use through a whole pregnancy does, as opposed to accidental early exposure. The 2026 review says that data is still missing.
• Use while breastfeeding. The lactation evidence is sparse enough that the labels do not rely on it (see the FAQs below).

Being told to stop a medicine that is finally working, to try for something that might take a while and might not work, is a genuinely hard trade — and saying so plainly to your doctor is reasonable, not a complaint.

This area moves quickly. Labels and guidance get revised as human pregnancy data accumulate, so confirm current advice with the clinician prescribing it rather than relying on any article, including this one.