Can prenatal antibiotics leave a mark on a child's skin?
We often treat antibiotics in pregnancy as a short trade: infection cleared, problem solved. A meta-analysis of 30 studies¹ suggests the story may reach the child's skin, with a modest but consistent rise in atopic dermatitis after prenatal exposure.
Lay public section
Find here your dedicated section
Sources
This article is based on scientific information
About this article
Author
Antibiotics are among the most common drugs in pregnancy: an earlier meta-analysis of about 34 million pregnancies put exposure at 23.6%. Their effect on the mother is well documented: antibiotics induce a gut microbiome disruption. Their effect on the child's early immune development is far less clear. A new systematic review in Scientific Reports, from researchers at the University of Szeged, tackles that gap for atopic dermatitis, a condition your patients' children may already know well.
Peripartum prophylactic antibiotic therapy decreases bifidobacterium levels in breast milk
Thirty studies, one modest signal
Searching four databases up to August 2025 under a registered PROSPERO protocol, the team pooled 30 observational studies covering more than 4.1 million mothers and 4.3 million children, 26 studies assessing maternal antibiotic use at any point in pregnancy. The available data did not allow reliable comparisons between exposure in the first trimester, second trimester and third trimester. Prenatal exposure was associated with higher odds of atopic dermatitis in childhood: an adjusted odds ratio (aOR) of 1.32 (95% CI 1.12 to 1.56) across 14 studies, and a pooled adjusted hazard ratio of 1.10 (1.05 to 1.14) across four. The age of children at diagnosis also varied between studies. Evidence of a potential publication bias was detected, so the authors applied a trim-and-fill correction. The aOR fell to 1.22 (1.03 to 1.44) but stayed significant: the raw figure may have overstated the effect without erasing it. The association remained statistically significant in leave-one-out analyses, even when individual studies at serious risk of bias were excluded. The authors themselves call the effect modest. These relative effect estimates do not quantify the absolute increase in risk or imply that every exposed child will develop atopic dermatitis.
Prenatal exposure means antibiotics taken at any point in pregnancy. Intrapartum exposure means prophylaxis given during labour and delivery, as in the group B Streptococcus prophylaxis studies included here. In this meta-analysis, the first showed a modest association with atopic dermatitis and the second did not.
Atopic dermatitis is a chronic inflammatory skin disease with red, dry, itchy skin, driven by a defective skin barrier. Eczema is the broader, non-specific term, but the two are often used as synonyms, which is why the authors counted both as outcomes.
An odds ratio recalculated to account for confounders such as infant sex, maternal age, maternal allergy or atopy, and smoking in pregnancy. An aOR of 1.32 means the odds of atopic dermatitis were 32% higher in exposed children, after those adjustments.
Where the signal holds, and where it fades
Risk rose with the number of antibiotics courses. Pooled adjusted hazard ratios were 1.07 for one course, 1.10 for two and 1.16 for more than two, although only three studies contributed. Geography mattered: European studies showed a significant association (aOR 1.40, 95% CI 1.08 to 1.83), while Asian and US studies did not reach significance. Of course, an absence of significance does not imply an absolute absence of association between prenatal antibiotic use and eczema in those regions. Results were also stronger with parent-reported eczema diagnoses than with medical records, and for the largest studies (over 10,000 children), the pooled adjusted association was not statistically significant. Overall, heterogeneity was very high (I² up to 99.7%) and GRADE certainty was low. Intrapartum exposure behaved differently: a pooled analysis of four intrapartum studies found no significant association (OR 1.64, 95% CI 0.84 to 3.17). However, individual findings were inconsistent, heterogeneity was substantial, and the certainty of evidence was very low. Findings also differed by birth mode, with some studies reporting different results for vaginal and caesarean deliveries. The window of intrapartum exposure is much narrower, which may mean only a transient effect on the infant's microbiota, though one study did find a signal when exposure exceeded 24 hours.
The period from conception to about age 2, when the microbiome and immune system are taking shape. It is the window in which the exposures discussed here are thought to matter most.
A statistical method that estimates which studies may be missing because of publication bias and recalculates the pooled effect. Here it lowered the aOR from 1.32 to 1.22, which stayed significant.
A framework that rates how much confidence we can place in a body of evidence. The prenatal findings were rated low and the intrapartum findings very low, because of bias, inconsistency and imprecision.
A plausible mechanism, not a proven one
The meta-analysis tested exposure and outcome, not the microbiota itself. It runs like this: antibiotics alter the mother's gut and vaginal communities, those communities act as the "starter culture" for the newborn's gut at birth, and a different start could shape early immune programming during the first 1,000 days. Birth mode and feeding practices may also influence how that microbiota develops. Reduced intestinal diversity has already been tied to atopic dermatitis in some other studies. The caveats are real. Postnatal antibiotic use, a plausible driver in its own right, was not consistently adjusted for, and observational data cannot establish causality. The authors urge antibiotics only when clearly indicated, for the shortest effective course, favouring narrow-spectrum antibiotics when appropriate. The takeaway for us is not to withhold antibiotics from pregnant patients with an infection, which is still essential to treat, but to treat each prescription as a risk-benefit decision, with a possible effect on the child's skin now part of the equation.
1. Petrás S, Vargané Szabó B, Kiss T, Bahar MA, Csupor D, Tóth B. Prenatal maternal antibiotic use increases the risk of childhood eczema: a systematic review and meta-analysis. Sci Rep. 2026;16:21156. doi:10.1038/s41598-026-51441-x.
See also
