Metronidazole

Pregnancy

Muanda et al. (2017) reported 53 metronidazole-exposed miscarriage cases in a nested case-control study, not 53 total exposed pregnancies; the adjusted odds ratio was 1.70 (95% CI 1.27-2.26). That is an association, not proof of cause. A Norway registry cohort of 704,082 pregnancies reported an association with miscarriage after metronidazole exposure (HR 2.00, 95% CI 1.82-2.21); its bias analysis indicated that underlying infections may explain associations for common antimicrobials. The article appeared online March 28, 2025, although its journal issue is June 2026. Ajiji et al. (2021) reviewed 24 studies and found no clear increase in major malformations (pooled OR 1.15, 95% CI 0.98-1.34) or preterm birth (RR 1.10, 95% CI 0.78-1.55), but an association with spontaneous abortion (OR 1.72, 95% CI 1.40-2.12). A separate third-trimester subgroup estimate for preterm birth was RR 1.59 (95% CI 1.05-2.41), based on one study. Their meta-analysis also reported a hydrocephaly signal from two case-control studies (OR 4.06, 95% CI 1.75-9.42); both used the same Hungarian registry and were subject to recall bias, with possible overlap. Koss et al. (2012; 2,829 mother-infant pairs, 922 treated) found no association with preterm birth after first- or later-trimester exposure, low birth weight, or congenital anomalies; it did not study miscarriage. Taken together, the birth-defect and preterm findings are generally reassuring, while miscarriage associations remain unresolved and may be confounded by infection.

Breastfeeding

MotherToBaby reports that milk levels depend on dose and route, and that oral or intravenous use can produce milk amounts similar to infant treatment doses; it reports that most exposed babies have not had side effects. LactMed describes measurable infant exposure and case reports of diarrhea and Candida, while noting that no definitive human study has assessed the relevance of animal carcinogenicity findings. Passmore et al. (1988) measured milk levels in 12 mothers, all receiving other broad-spectrum antibiotics, and monitored 35 infants; the study found no significant increase in infant adverse effects attributable to maternal metronidazole therapy. LactMed says expert opinions vary and that some sources recommend a 12-to-24-hour pause after a single 2-gram dose, but not after lower doses. NHS Specialist Pharmacy Service says interruption is not required even after 2-gram oral doses and recommends infant monitoring for usual courses. The FLAGYL 375 capsule label gives pumping and discarding during treatment and for 48 hours after the last dose as an alternative; that is product-label wording, not a universal recommendation. These differences and small infant-outcome studies leave uncertainty about recommendations across doses, routes, and infants.

What was tested

Pregnancy evidence includes observational cohorts and case-control studies, meta-analyses, and randomized trials that often tested treatment of bacterial vaginosis or trichomoniasis rather than medication safety as their primary outcome. The 2026 multicentre randomized trial screened 914 high-risk pregnant women and analyzed 82 metronidazole and 84 clindamycin recipients; the preterm-delivery or late-miscarriage rates were 30% versus 24% with no statistically significant difference. The trial's primary endpoint was infection eradication, so it was not designed to establish medication safety. The Norway registry cohort included 704,082 pregnancies; its metronidazole association was subject to possible confounding by infection. Breastfeeding studies measured drug in milk and infant blood and reported infant effects, but samples were small. Passmore et al. followed 35 infants, while milk-level measurements were made for 12 mothers receiving other antibiotics as well.