GLP-1 Pregnancy Washout: 2 Months Before Conception — Why?

Every FDA-approved GLP-1 receptor agonist used for chronic weight management or type 2 diabetes carries the same labeled instruction for women planning pregnancy: discontinue at least two months before a planned conception. The number is not arbitrary. It is the elimination time required to clear roughly eight half-lives of semaglutide and an even higher multiple of tirzepatide half-lives, anchored to rodent embryofetal toxicity signals seen at maternally-toxic doses (Drummond 2025^[5]). This article walks through the verbatim label text, the half-life arithmetic, what the human pregnancy outcome data actually look like, the oral-contraceptive interaction that affects tirzepatide titration, and the practical preconception timeline that fits all of it together.

The honest summary

• The labels are explicit and identical in spirit. Wegovy^[1], Zepbound^[2], Ozempic^[3], and Mounjaro^[4] all instruct patients to discontinue at least two months before a planned pregnancy. The two-month window is the time needed for near- complete drug elimination given each compound’s half-life.
• The math is straightforward. Semaglutide has a terminal half-life of roughly 7 days; eight weeks is about eight half-lives, leaving less than 0.4% of the original steady-state concentration. Tirzepatide’s ~5-day half-life gives an even larger multiple at eight weeks (Pharmacokinetics Review, 2025^[8]).
• Human pregnancy data so far are reassuring but limited. The 2026 systematic review and meta-analysis of periconceptional exposure (Diop 2026, Am J Obstet Gynecol^[6]) and the 2026 safety review (Diabetes Obes Metab^[7]) do not show a clear signal of major malformations above background, but neither is powered to rule out small effects. None of these data change the labeled instruction.
• Inadvertent pregnancy on a GLP-1 is not an indication for termination. Standard practice is to stop the drug immediately, refer to maternal-fetal medicine (MFM), and continue routine prenatal monitoring (Drummond 2025^[5]).

What the FDA labels actually say

The Wegovy DailyMed label (SetID ee06186f)^[1], Section 8.1, states the instruction in two sentences: Wegovy may cause fetal harm and should be discontinued when pregnancy is recognized; because of the long washout period for semaglutide, the label adds the now-familiar verbatim guidance to discontinue at least two months before a planned pregnancy. The Ozempic label^[3] carries an identical preconception instruction in Section 8.1, anchored to the same molecule and the same half-life. The Zepbound^[2] and Mounjaro^[4] labels carry the same two-month rule for tirzepatide. The Saxenda (liraglutide) label uses a shorter washout window in practice because liraglutide’s half-life is roughly 13 hours rather than days, but the spirit of the rule — discontinue at pregnancy recognition and ideally before conception — is the same.

The half-life arithmetic

Five elimination half-lives reduce a steady-state plasma concentration to roughly 3% of starting value; eight half-lives reduce it to about 0.4%. The FDA convention for “clinically eliminated” is generally five half-lives. The two-month rule is purposely generous — well past five half-lives for either semaglutide or tirzepatide — to keep the residual exposure during organogenesis (gestational weeks 3–8 post-conception) functionally negligible.

• Semaglutide terminal half-life ~7 days (Pharmacokinetics Review^[8]). Eight weeks = ~8 half-lives. Residual < 0.4%.
• Tirzepatide terminal half-life ~5 days^[8]. Eight weeks = ~11 half-lives. Residual < 0.05%.
• Liraglutide half-life ~13 hours^[8]. Three to four days of washout reaches the same elimination depth.
• Dulaglutide half-life ~5 days^[8]; same elimination math as tirzepatide.

Magnitude: % of steady-state semaglutide remaining by week

What the rodent embryofetal data showed

The label-driving preclinical signal for GLP-1 receptor agonists comes from rodent reproductive toxicology studies in which fetal malformations and reductions in fetal weight were seen at maternally-toxic doses. The Bjerre Knudsen 2010 rodent program^[9] is best known for the thyroid C-cell signal that drove the boxed warning on medullary thyroid carcinoma, but the broader GLP-1 reproductive-tox package across semaglutide and tirzepatide development showed dose-related effects on fetal growth and skeletal/cardiovascular malformations at maternally-toxic exposures. Whether the rodent signal translates to humans at therapeutic exposures is unknown; the regulatory response was to require the preconception discontinuation language across the class (Drummond 2025^[5]).

What the human pregnancy data actually look like

The 2026 Diop systematic review and meta-analysis^[6] pooled observational data on periconceptional GLP-1 RA exposure and reported no statistically significant elevation in major congenital malformations versus matched controls. The 2026 safety systematic review in Diabetes Obesity & Metabolism^[7] reached the same directional conclusion across preconception, pregnancy, and lactation windows. The 2025 Drummond and Seif narrative review in the American Journal of Obstetrics and Gynecology^[5] explicitly summarizes the labeling rationale and the post-marketing experience that informs current MFM practice.

Three caveats on those conclusions: (1) sample sizes are modest, so small absolute risk differences cannot be excluded; (2) most exposures captured are early first-trimester and inadvertent, not deliberate; and (3) confounding by maternal obesity and type 2 diabetes — both independent risk factors for adverse obstetric outcomes — complicates causal inference. The honest reading: human data are reassuring in the aggregate, do not justify deviating from the labeled 8-week rule, and do not constitute grounds for termination if exposure happens to occur.

The Eli Lilly tirzepatide pregnancy registry

Patients with inadvertent tirzepatide exposure during pregnancy can enroll in the Lilly-sponsored pregnancy surveillance registry; Novo Nordisk maintains analogous post-marketing pharmacovigilance for semaglutide. The Drummond 2025 review^[5] describes both registries as the current primary mechanisms for collecting prospective pregnancy outcome data on these molecules. Patients should ask their prescribing clinician about enrollment if they become pregnant on a GLP-1.

The oral-contraceptive interaction (tirzepatide only)

Tirzepatide’s mechanism of slowed gastric emptying meaningfully reduces oral-contraceptive exposure. The Lilly clinical-pharmacology data summarized in the Zepbound^[2] and Mounjaro^[4] labels (Section 7.2) and reviewed in the 2025 pharmacokinetics paper^[8] show that a single 5 mg tirzepatide dose reduces oral combined oral-contraceptive (COC) Cmax by roughly 55–66% and AUC by ~20–23%. The labeled mitigation is either to switch to a non-oral contraceptive (IUD, implant, injection) or to add a barrier method for four weeks after starting tirzepatide and for four weeks after each dose escalation. Semaglutide does not produce a clinically meaningful oral-contraceptive interaction in the same magnitude^[8].

For women trying to conceive in the planned 8-week window leading up to conception, the contraceptive question becomes moot once the GLP-1 is stopped — but the four weeks before discontinuation still carry the interaction. The practical implication is to plan the washout calendar against the contraceptive method, not just the conception attempt.

Weight regain after stopping: STEP-4 and the trajectory

STEP-4 (Rubino 2021 JAMA^[9]) randomized patients who had reached steady-state weight loss on semaglutide 2.4 mg at week 20 to either continue or switch to placebo for another 48 weeks. The continue arm lost a further ~7.9% of body weight; the placebo arm regained ~6.9%. The trajectory of regain begins almost immediately after discontinuation and is roughly two-thirds recovered by 12 months off drug. The clinical implication for preconception planning is that the post-stop weight is not the patient’s “true” weight; it is a moving target that will likely drift upward through the eight-week washout and into early pregnancy. STEP-1 (Wilding 2021 NEJM^[10]) and SURMOUNT-1 (Jastreboff 2022 NEJM, magnitude −20.9%) give the on-drug benchmarks against which the post-stop regain is measured.

For women with polycystic ovary syndrome, Carmina 2023^[10] (PMID 37762862) reported restoration of ovulatory cycles in obese PCOS patients treated with semaglutide. The clinically important corollary is that fertility may improve substantially on a GLP-1 — raising the risk of unplanned conception in patients who assumed anovulation was their default. PCOS patients on a GLP-1 who do not want pregnancy need a reliable contraceptive plan that accounts for the restored ovulatory capacity.

The practical preconception timeline

1. Month −4 (~16 weeks before TTC): Pre-conception planning visit with OB-GYN or MFM. Confirm target conception window. Start folic acid 400 mcg daily (per ACOG — folate stores need 4 weeks to saturate before neural tube closure at gestational week 4 post- conception).
2. Month −3: Final assessment of weight and metabolic status while still on the GLP-1. Do not expect the post-stop weight to remain at the on-drug nadir.
3. Month −2 (8 weeks before TTC): Discontinue the GLP-1. Continue folic acid. If on tirzepatide and using oral contraception, confirm non-oral method or barrier coverage during any final dose interval.
4. Month −1: Approximately 4 half-lives for semaglutide; residual ~6%. Continue prenatal vitamin. Address any weight-regain related concerns with the clinician.
5. Month 0 (start trying to conceive): ~8 half-lives of semaglutide cleared. Residual < 0.5%. The label window has been met.
6. If pregnancy occurs sooner than planned: Stop the GLP-1 immediately, contact the prescriber, enroll in the manufacturer pregnancy registry if offered, and continue routine prenatal monitoring. The labels and the 2025 Drummond review^[5] are explicit that inadvertent exposure is not an indication for termination.

Lactation and the postpartum question

The Wegovy^[1] and Ozempic^[3] labels state that there are no data on the presence of semaglutide in human milk, the effects on the breastfed infant, or the effects on milk production; semaglutide was present in the milk of lactating rats. The Zepbound^[2] and Mounjaro^[4] labels report the same posture for tirzepatide. The 2026 Diabetes Obes Metab safety review^[7] confirms the absence of robust human lactation data across the class. The practical implication is that GLP-1 therapy is typically resumed after weaning rather than during breastfeeding.

Related research and tools

• GLP-1, pregnancy, PCOS and fertility — broader companion to this washout-timing piece
• BHRT + GLP-1 in perimenopause — hormone replacement stacking in midlife patients
• GLP-1 muscle loss prevention protocol — the protein and resistance training protocol for any off-ramp (preconception, surgery, pause)
• GLP-1 protein calculator — carry a sustainable target into the off-drug window

Important disclaimer. This article is educational and does not constitute medical advice. Preconception planning, contraception choices, and decisions about continuing or discontinuing prescription medication must be made in consultation with a qualified clinician. Patients who become pregnant while on a GLP-1 should contact their prescriber promptly and consider enrollment in the manufacturer pregnancy registry. PMIDs were verified live against the PubMed E-utilities API on 2026-05-28.

Last verified: 2026-05-28. Next review: every 12 months, or sooner if a new prospective registry analysis, FDA labeling change, or systematic review materially updates the preconception-washout evidence base.