Breastfeeding while on a GLP-1 drug is one of the questions these medicines' makers studied least, and the record is mostly absence. Three small studies have measured the drugs in human milk and found little or none. Nobody has published what happens to the babies or to milk supply. The labels, read side by side, do not agree with each other — and the dividing line turns out to be a tablet ingredient, not the drug.
This page sets out what has been measured, what the eight current labels say, and what nobody knows yet. All labels, registry records and papers were read on 4 October 2026. It reports the record; it does not tell anyone whether to breastfeed or take a medicine, which is a decision for a mother and her clinician.
What has actually been measured in human milk
| Drug | Study | Who | What was found |
|---|---|---|---|
| Semaglutide, injected | Diab et al., Nutrients 2024 (PMID 39275201) | 8 women; milk at 0, 12 and 24 h after a dose | Not detected in any sample (detection limit 1.7 ng/mL); authors' worst-case relative infant dose 1.26% |
| Semaglutide, tablet | Manufacturer study reported in the tablet labels; ClinicalTrials.gov lists one such Novo Nordisk study, NCT04817644 (completed, 14 enrolled) | Healthy breastfeeding women on the tablet | Semaglutide below the lower limit of quantification; SNAC and/or its metabolites present |
| Tirzepatide, injected | Lilly study reported in the Mounjaro and Zepbound labels | 11 healthy lactating women; one 5 mg dose; 171 samples over 28 days | Undetectable in 164 of 171; the other 7 together held <0.02% of the dose; last measurable on day 5 |
| Liraglutide, dulaglutide, orforglipron | — | — | No human milk data in their labels or in LactMed |
Three limits apply to all of it. The studies are tiny — 8, 14 and 11 women. The tirzepatide study used a single 5 mg dose, below the usual maintenance range, so it says nothing directly about months of weekly use at higher strengths. And none of them followed the babies: the Nutrients authors themselves list long-term infant outcomes, the mother's nutrient intake and the milk's nutrient content as open questions.
The 1.26% figure is a calculation, not a measurement: the authors assumed the milk held semaglutide at the lowest level their method could quantify, even though none was detected, and computed a ceiling from that. Their paper compares it with a 10% threshold commonly used in lactation pharmacology.
What the eight labels say
| Label (date read) | Human milk data in the label | Says breastfeeding is not recommended? |
|---|---|---|
| Ozempic injection (2026-07-30) | "There are no data on the presence of semaglutide in human milk" | No — weigh benefits against need |
| Wegovy injection (rev. 06/2026) | No data for the injection; "does not contain the SNAC metabolites" | No — weigh benefits against need |
| Wegovy tablets (rev. 06/2026) | Semaglutide below quantification; SNAC present | Yes |
| Ozempic tablets and Rybelsus (2026-01-30) | Semaglutide below quantification; SNAC present | Yes |
| Zepbound (2026-08-28) | The 11-woman single-dose study | No — weigh benefits against need |
| Mounjaro (2026-07-29) | The same study | No — weigh benefits against need |
| Saxenda, liraglutide (2026-02-25) | None; in rats ~50% of maternal plasma | No — weigh benefits against need |
| Foundayo, orforglipron (2026-07-29) | None; in rats 3-fold higher than plasma | Yes — "not recommended for nursing women" |
Section 8.2 of each label, read through openFDA and DailyMed. Trulicity's label (2023-08-05 in openFDA) also reports no human data and states that milk transfer was not measured in animals.
The tablet ingredient is what splits them
The semaglutide molecule is the same in the pen and the pill. What differs is SNAC (salcaprozate sodium), the absorption enhancer that lets a peptide survive the stomach; how it works is on oral GLP-1s. The tablet labels give their reasoning in full: SNAC or its metabolites are present in human milk, the enzymes that clear it "may be lower in infants compared to adults", and so "because there are alternative formulations of semaglutide that do not contain SNAC that can be used during lactation", breastfeeding is not recommended on the tablets. The Wegovy label, which now covers both forms, states plainly that the injection does not contain SNAC.
The National Library of Medicine's LactMed database reaches the same split and phrases it more directly: only injectable forms of semaglutide should be used during breastfeeding (entry updated 15 August 2026). Foundayo is a different case — it contains no SNAC, and its label's caution rests on its own rat data, where the drug reached milk at three times the plasma level, the opposite direction to semaglutide.
Two labels for one molecule disagree on a fact
The Ozempic injection label, revised in July 2026, still says there are no data on semaglutide in human milk. The semaglutide tablet labels from the same manufacturer report a human lactation study in which semaglutide was below the limit of quantification, and an independent study of the injection was published in 2024. The injection label is not wrong about its own product — the manufacturer's study used the tablet — but a reader comparing documents will find one molecule described with data in one place and without it in another. The pregnancy sections of these labels diverge in a similar way, by brand rather than formulation; that record is on GLP-1 drugs and pregnancy.
Who has published a position
- LactMed (National Library of Medicine), 15 August 2026. Semaglutide: not detectable in the milk of mothers injecting it, and "their breastfed infants experienced no adverse effects"; only injectable forms during breastfeeding. Tirzepatide: "usually undetectable" in milk at subcutaneous doses up to 5 mg; "not a reason to discontinue breastfeeding", but to be used "with caution … especially while nursing a newborn or preterm infant". Liraglutide: no information; the amount in milk is "likely to be very low". For all three it notes that some experts recommend monitoring infant growth and development and the mother's nutrition.
- A 2026 systematic review of type 2 diabetes drugs in breastfeeding (Richardson et al., Clinical Pharmacology & Therapeutics, PMID 41670332) found clinical milk-transfer data for only 5 of 20 drugs, semaglutide among them; most published guideline recommendations (78%) advised against antiglycaemic drugs while breastfeeding, and only 4.4% of recommendations rested on clinical evidence.
- A 2026 systematic review of GLP-1 drugs in pregnancy and lactation (Ozbek et al., Diabetes, Obesity and Metabolism, PMID 41885132) screened the literature to September 2025 and described the lactation data as "sparse": one pharmacokinetic study, no detectable semaglutide transfer.
What nobody has measured
- The breastfed baby. Every label says there are no data on effects on the infant. No controlled study has followed infants of mothers taking these drugs.
- Milk supply and composition. No label reports data on milk production. These drugs reduce appetite and food intake, and what that does to the volume or fat content of milk has not been published. The first registered attempt to measure it, NCT07720414 at the University of Colorado, plans 18 mothers who have been prescribed an injected GLP-1 drug and measures hourly milk output and composition before and after they start; it began recruiting in July 2026.
- Weekly use at maintenance doses. The only tirzepatide data are from one 5 mg dose. A search of ClinicalTrials.gov for tirzepatide and milk returns only the Colorado study; the labels do not give a registry number for the 11-woman study.
- The newer and unapproved molecules. Orforglipron has rat data only. Compounded and research-market products have no lactation data at all, and their contents are not verified; the difference is on compounded vs branded semaglutide.
Before deciding anything
The decision about any medicine while breastfeeding belongs with the prescriber, and ideally a lactation specialist, who can weigh the baby's age (LactMed singles out newborns and preterm infants), the mother's reason for treatment and the formulation. The labels' own framing is that breastfeeding's benefits are weighed against the mother's clinical need. For those planning another pregnancy, the separate two-month washout guidance in the semaglutide labels is covered on GLP-1 drugs and pregnancy, and how long the drugs stay in the body after the last dose is on stopping GLP-1.
Sources
Section 8.2 of the current labels for Ozempic injection (2026-07-30), Ozempic tablets and Rybelsus (2026-01-30), Zepbound (2026-08-28), Mounjaro (2026-07-29), Saxenda (2026-02-25), Foundayo (2026-07-29) and Trulicity (2023-08-05), via openFDA; Wegovy injection and tablets (revised 06/2026), via DailyMed · Diab H et al., Nutrients 2024;16:2886 (PMID 39275201) · LactMed entries for semaglutide, tirzepatide and liraglutide (PMIDs 30000039, 35759552, 30000036), updated 15 August 2026 · Richardson K et al., Clin Pharmacol Ther 2026 (PMID 41670332) · Ozbek L et al., Diabetes Obes Metab 2026 (PMID 41885132) · ClinicalTrials.gov NCT04817644 and NCT07720414. All read 4 October 2026.