Levothyroxine

Pregnancy

FDA labeling reports that human clinical and postmarketing experience with oral levothyroxine used to maintain normal thyroid function has not reported increased rates of major birth defects, miscarriage, or other maternal or fetal outcomes. The label also says untreated hypothyroidism is associated with pregnancy complications and says treatment should not be discontinued during pregnancy. Human observational studies are mixed. In a Swedish register study of 9,866 mothers using thyroid hormones and 10,055 infants, Leung et al. reported a small congenital-malformation association (OR 1.14, 95% CI 1.05-1.26); the study also reported higher preeclampsia and diabetes rates and a marginal preterm-birth association. In an expanded Swedish analysis of 23,259 infants exposed to maternal thyroxin, Källén et al. reported an adjusted overall OR of 1.06 (95% CI 0.98-1.14), not statistically significant. The same abstract reported signals for a few rare malformation categories, including choanal atresia (RR 3.14, 95% CI 1.26-6.47); it reported anal atresia as RR 1.85 with a 95% CI of 1.00-1.85, as printed in the abstract. Andersen et al. found the prevalence of congenital anomalies was the same in exposed and unexposed pregnancies in a Danish registry cohort; its total cohort included 8,318 pregnancies exposed to antithyroid drugs or levothyroxine and 969,303 unexposed pregnancies. These registry studies cannot fully separate medication exposure from the underlying thyroid condition and related factors. The 2004 prospective dose study included 20 pregnancies in 19 women and found that 17 pregnancies required a dose increase; mean levothyroxine requirements increased 47% during the first half of pregnancy. The authors proposed approximately a 30% increase after a positive pregnancy test. The 2017 ATA summary recommends two additional tablets weekly after pregnancy is confirmed. The 2026 ATA guideline says to consider pre-pregnancy TSH and dose when deciding on an early increase because standardized increases may overtreat some patients. The May 2026 FDA-approved LEVO-T label says to measure TSH and free T4 at pregnancy confirmation and at least each trimester; for pre-existing primary hypothyroidism with TSH above the trimester range, increase by 12.5-25 mcg/day and monitor TSH every four weeks until stable, then return to the pre-pregnancy dose after delivery and check TSH four to eight weeks postpartum.

Breastfeeding

LactMed, last revised 15 September 2026, describes levothyroxine as a normal component of human milk and says limited information on maternal replacement doses has reported no adverse infant effects. It notes that infant-level information was not found and that a mild rash was reported in one exclusively breastfed 2.5-month-old infant while the mother took both sertraline and levothyroxine. The underlying 2025 consultation study included 200 breastfeeding or pregnant people seeking medication consultations; nine potential infant adverse events were recorded. For the rash case, its table lists sertraline and levothyroxine, a Naranjo score of 2, and an LCAT classification of possible. The article's narrative says the Naranjo algorithm did not support a causal link between any infant event and maternal medication, so causality remains uncertain and the report does not isolate levothyroxine. The FDA label says levothyroxine is present in human milk, no adverse effects on breastfed infants have been reported, and information on milk production is unavailable; it says adequate treatment may normalize low milk supply when hypothyroidism is the cause. A 2026 retrospective matched case-control study compared 103 term infants of mothers with gestational hypothyroidism treated with levothyroxine with 103 controls; exclusive breastfeeding at one month was 58.3% versus 81.6%. Because the study groups were defined by maternal disease and treatment, the feeding difference cannot be attributed to levothyroxine alone. A 2026 prospective cohort compared milk from 19 mothers with hypothyroidism taking levothyroxine throughout pregnancy and postpartum with 30 mothers without hypothyroidism. Overall milk macronutrients and major fatty-acid groups did not differ, while some individual fatty acids differed; the study did not measure an infant medication effect. A publisher correction changed the corresponding-author designation, not the study findings.

What was tested

Pregnancy evidence includes postmarketing labeling summaries, national registry cohorts, and a small prospective study measuring dose requirements. The registry studies evaluate maternal thyroid-hormone prescriptions or recorded use and pregnancy or infant outcomes; they do not randomize treatment and cannot fully separate medicine exposure from hypothyroidism. Breastfeeding evidence includes LactMed's review of small human studies and case reports, a 2025 pharmacist-consultation cohort with one infant rash report involving two maternal medicines, a 2026 matched case-control study of gestational hypothyroidism treated with levothyroxine and breastfeeding outcomes, and a 2026 milk-composition cohort. The recent breastfeeding studies do not establish that levothyroxine caused the observed feeding or milk-composition findings.