For semaglutide, at least 2 months. The Wegovy label says to stop that long before a planned pregnancy, because the drug lasts [1]. For tirzepatide, the Zepbound label sets no washout [2]. It says to stop once a pregnancy is recognized. Its 5 to 6 day half-life points to about a month, but that is arithmetic, not a label rule.
A pregnancy can also arrive unplanned. Oral contraception has its own caution on these drugs, set out in whether a GLP-1 affects birth control. This page covers the planned case: when to take the last dose.
What the Wegovy label says
Section 8.3 is short [1]. Stop Wegovy at least 2 months before a planned pregnancy. The stated reason is the long half-life of semaglutide.
The label puts that half-life at about 1 week. It says semaglutide stays in the circulation for about 5 to 7 weeks after the last dose. Two months is about 8 to 9 weeks. So the label’s rule sits past that window, with a margin.
If a pregnancy is recognized on Wegovy taken for weight, the label says stop. It calls the human data insufficient to establish a drug-associated risk.
What the Zepbound label says
No fixed washout [2]. The pregnancy section says to discontinue Zepbound when a pregnancy is recognized. It calls the human data insufficient to evaluate risk. Animal studies suggest a possible risk to a fetus.
Section 8.3 of this label covers contraception only. People on the pill are told to switch to a non-oral method, or add a barrier method. That lasts 4 weeks after starting, and 4 weeks after each dose increase.
The half-life arithmetic
One half-life halves the drug in the body. After five, about 3% of the last dose is left. Five half-lives is a common rule of thumb for “cleared”.
Semaglutide at about 7 days gives 35 days. Tirzepatide at 5 to 6 days gives 25 to 30 days. Clearance varies between people, so these are central estimates.
A 2025 narrative review of 9 studies ran the same sums [7]. It proposed at least 35 days for semaglutide and 25 to 35 days for tirzepatide. For semaglutide, the Wegovy label asks for longer. The label is the rule a prescriber works to.
After stopping: the JAMA cohort
One academic health system reviewed 149,790 pregnancies [3]. Exposure meant a GLP-1 order between 3 years before and 90 days after conception. Each of 448 exposed pregnancies was matched with three unexposed ones, 1,344 in all. In the exposed group, 84% had obesity.
The exposed group gained more: 13.7 kg against 10.5 kg, a 3.3 kg difference. Excess weight gain was 65% against 49%. Preterm delivery was 17% against 13%. Gestational diabetes was 20% against 15%. Hypertensive disorders were 46% against 36%.
Cesarean delivery and large or small babies did not differ. The authors tie the results to discontinuation before or early in pregnancy. It was one health system, and it did not compare stop dates.
Stopping earlier did not help: a semaglutide cohort
A national records study split semaglutide users into two groups [5]. 429 were exposed into pregnancy, for a median 44 days. 801 had stopped before conceiving. 2,203 nonusers served as the comparison.
Against nonusers, the women who stopped first still had higher odds. Excessive weight gain was 1.98, gestational diabetes 1.43, and cesarean delivery 3.92. The pregnancy-exposed group was similar. The two drug groups did not differ significantly on any outcome.
The authors read that as rebound after stopping, not harm from exposure. The rebound was not measured directly. How fast weight returns off the drug is in how much weight comes back after stopping.
The cohort that points the other way
A US records network matched 4,267 people to 4,267 [4]. One group had a GLP-1 prescription in the 24 months before pregnancy. The other had never been prescribed one.
The prescribed group did better. Gestational diabetes was 15.2% against 18.2%, odds ratio 0.81. Hypertensive disorders were 19.9% against 22.8%. Preterm delivery was 3.0% against 4.4%. Cesarean delivery was 17.6% against 19.7%.
That contrast is ever against never. It does not say when anyone stopped. People prescribed a GLP-1 may differ from people never prescribed one in ways records miss.
Why the studies disagree
They asked different questions. The JAMA cohort looked at stopping close to conception. The records study looked at any prescription in a two-year window. Neither randomized anyone, and neither tested a 2-month stop against a later one.
One reading fits both, and it is unproven. Weight loss before pregnancy may help. A fast rebound during pregnancy may undo some of it. How to taper is a separate question, covered in whether a lower dose can hold the weight.
If exposure happens early by accident
A multinational cohort looked at birth defects in women with type 2 diabetes [6]. Exposure ran from 90 days before pregnancy to the end of the first trimester. Among 938 GLP-1 exposed infants, major malformations were 8.3%.
Against insulin, the adjusted relative risk was 0.95 (95% CI 0.72 to 1.26). The authors found no large increased risk and asked for confirmation. That is reassurance after an accident. It is not a reason to skip the label’s washout.
Who decides
The prescriber and the obstetrician. The label gives the floor for semaglutide. The cohorts give reasons to plan the months around it, not a date.
Bring the drug name, the dose and the date of the last injection. Say whether it was compounded. The JAMA cohort counted orders up to 3 years before conception. Mention any use in that window.
What it means for a prepaid plan
A multi-month plan can outlast a pregnancy plan. On semaglutide, the last dose falls 2 months before trying. Doses prepaid past that date go unused unless the plan refunds them. The trade-offs are in whether to prepay for a GLP-1.
A restart after delivery or breastfeeding is a new prescription decision. The steps are in how to restart a GLP-1.