Semaglutide in Breast Milk: What the Eight-Mother Study Measured

The 2024 injected-semaglutide milk study offers an early exposure measurement, not a complete answer about breastfeeding safety. Its most useful result is specific: no semaglutide was detected in the collected milk samples.¹

What the milk test measured

Eight mothers supplied samples at 0, 12 and 24 hours post-injection while using 0.25–1 mg weekly. Sampling followed at least three administrations at the same dose, sometimes before steady state; the study covered early escalation, not every treatment stage. One participant reported a compounded product. The detection limit was 1.7 ng/mL; the quantification limit was 5.7 ng/mL. Nondetection does not prove zero exposure. Calibration assumed added drug behaved like drug arriving in milk; a mismatch could affect accuracy.¹

Who was represented

Table 2 lists infants aged 4–6, 7–11 and 12–23 months, although the Discussion says 7–23 months. Their diets included other foods or fluids; some received formula. Mothers reported generally unremarkable growth and milestones. One described her own gastrointestinal symptoms alongside her infant’s transient diarrhea and decreased appetite, which resolved; causation was not established. Reported exposure included three-plus weeks and an interrupted nine weeks. These questionnaires were not long-term clinical follow-up.¹

What remains unanswered

Milk production, composition and infant drug levels were not measured. The small, later-postpartum sample cannot establish newborn or exclusive-breastfeeding outcomes. The authors’ modeled exposure estimates used assumed milk concentrations and pharmacokinetics, not measured infant absorption. Our reading is that this is preliminary transfer evidence, with nutritional and longer-term outcomes still unresolved.¹

FDA’s May 2019 study-design draft separately considers milk transfer, infant safety information and milk-production effects. For chronic dosing, it recommends steady-state sampling; for intervals longer than 24 hours, it considers the full interval or 24-hour sampling near the expected plasma peak. It is dated, nonbinding draft guidance for research, not a personal feeding schedule or evidence that FDA endorsed this study.²

For product and route distinctions, see our oral-versus-injectable guide. For the separate question of clinical follow-up, see our monitoring guide.

This is an adapted summary of Diab and colleagues’ 2024 study (© 2024 by the authors), licensed under CC BY 4.0. No author endorsement is implied.¹