Exposure to Systemic Antimicrobials During Pregnancy and Risk of Miscarriage: A Population-Based Registry Study
Common urinary infection antibiotics were not linked to higher miscarriage risk, while several other antimicrobial drugs showed higher risk that may partly reflect the infection being treated.
*Population-based cohort study; Level 2b (OCEBM).
*Population-based cohort study; Level 2b (OCEBM).
Citation
Boissiere-O'Neill T, van Gelder MMHJ, Engjom HM, Nordeng HME. Exposure to systemic antimicrobials during pregnancy and risk of miscarriage: a population-based registry study. BJOG: An International Journal of Obstetrics & Gynaecology. 2026;133:1394–1404. doi:10.1111/1471-0528.181551.
Background
Antimicrobial medicines are frequently used in early pregnancy, but past studies have been hard to interpret because infection itself may raise miscarriage risk and because of timing-related biases.
Patients
704,082 pregnancies in Norway (2009–2018). Excluded ectopic or molar pregnancies and pregnancies exposed to known birth-defect–causing medicines near conception.
Intervention
Filled prescriptions for whole-body (not topical) antibacterial, antifungal, or antiprotozoal medicines in early pregnancy, treated as changing over time; prescriptions within 14 days before a pregnancy outcome were not counted.
Control
No filled prescription for these medicines during the same time window.
Outcome
Miscarriage before 20 weeks of pregnancy, identified from linked national registries.
Follow-up Period
From last menstrual period to 20 weeks.
Results
| Exposure (weighted analysis) | Association with miscarriage | Hazard ratio (95% confidence interval) |
|---|---|---|
| Nitrofurantoin | Lower risk | 0.75 (0.66–0.84) |
| Pivmecillinam | Lower risk | 0.91 (0.87–0.95) |
| Metronidazole | Higher risk | 2.00 (1.82–2.21) |
| Ciprofloxacin | Higher risk | 1.89 (1.62–2.20) |
| Cephalexin | Higher risk | 1.87 (1.57–2.22) |
| Fluconazole (oral) | Higher risk | 1.61 (1.45–1.78) |
| Trimethoprim with sulfa medicine | Higher risk | 1.49 (1.36–1.63) |
Elective terminations were treated as censoring events. Bias analyses suggested several drug–miscarriage links could be explained by the underlying infection.
Limitations
Observational data cannot fully separate medicine effects from infection severity. Very early miscarriages not seen in healthcare records were missed. Some antifungal use was over-the-counter and not captured. A filled prescription may not mean the medicine was taken.
Funding
University of Oslo; Norwegian Research Council visiting grant.
Clinical Application
When treating early-pregnancy infections, prefer nitrofurantoin or pivmecillinam for urinary infection when appropriate; be cautious with oral fluconazole, metronidazole, quinolones, and cephalexin.
Journal Club
Discussion questions and an EBM quiz for this paper. How to run a journal club →
Discussion question (1)
In this population-based registry cohort, metronidazole exposure in early pregnancy was associated with higher miscarriage risk (adjusted HR 2.00, 95% CI 1.82–2.21); how much should this influence prescribing given confounding by indication and the authors’ bias analyses suggesting infection may drive associations?
EBM quiz (1 question)
1. The authors treated antimicrobial use as a time-varying exposure in their cohort analysis. What bias is this primarily intended to reduce?
Discussion
Sign in to join the discussion.
No comments yet. Be the first to share your thoughts.