Maternal carriage of *Prevotella* during pregnancy associates with protection against food allergy in the offspring

Thirty years after the hygiene hypothesis was proposed linking larger household size to reduced allergic disease, accumulating evidence implicates diminished early-life microbial exposure and microbiome alterations in allergy risk. Losses of specific ancestral microbiota and reduced dietary fiber may contribute to immune-related diseases. The maternal gut microbiome during pregnancy influences fetal immune development, and cord blood immune phenotypes at birth associate with later allergy, yet no human study had investigated maternal gut microbiota during pregnancy and subsequent offspring allergic disease using culture-independent techniques. Prevotella is less abundant in westernized populations; it ferments dietary fiber to produce acetate and succinate and generates endotoxins that could shape fetal immune tolerance via TLR4-dependent pathways. The study aimed to investigate early-life exposures—including maternal diet and household size—maternal and infant gut microbiota, and risk of infant allergic disease. The hypothesis was that higher maternal dietary fiber intake would reduce offspring allergy, mediated by SCFA-producing organisms, particularly Prevotella. The authors show in an Australian prebirth cohort that larger household size predicts maternal P. copri carriage during pregnancy, which associates strongly with protection against food allergy in the offspring.

Background literature highlights: (1) Westernized populations show reduced abundance of Prevotella compared to traditional communities. (2) Prevotella are gram-negative anaerobes that ferment dietary fiber to produce acetate, a short-chain fatty acid with anti-inflammatory effects and promotion of IL-10–producing regulatory T cells. In pregnant mice, microbiota-derived acetate crosses the placenta and attenuates postnatal allergic responses. (3) Prevotella also produce succinate, which stimulates innate immune cell development, migration, and function, and endotoxins that can influence fetal immune development via TLR4 pathways. (4) Prior human work has linked under-representation of Prevotella during infancy and childhood with IgE-related allergic disease and atopic dermatitis. However, no prior human study had directly assessed maternal pregnancy gut microbiota via culture-independent methods in relation to offspring allergy.

Study design and population: The Barwon Infant Study (BIS) is an Australian birth cohort (n=1064 mothers/1074 infants) recruited 2010–2013. Mothers provided third-trimester fecal samples and completed a food frequency questionnaire. Infants were reviewed at 1, 3, 6, 9, and 12 months. A case-cohort design compared infants with clinically proven IgE-mediated food allergy at 1 year (n=60 mothers/61 infants) to a random subcohort (n=321 mothers/324 infants). Within the random subcohort, atopic wheeze and atopic eczema were also evaluated. Ethics approval was obtained and informed consent provided. Clinical outcomes: Sensitization assessed by skin prick testing (SPT) at 12 months to common food and aeroallergens. Food allergy defined by positive SPT plus clinical history and/or in-hospital open food challenge using predetermined stopping criteria. Atopic wheeze: parent-reported wheeze during first year plus sensitization. Atopic eczema: UK Working Party infant criteria plus sensitization. Diet and exposures: Maternal diet in third trimester via DQES2 to estimate macro- and micronutrients. Maternal antibiotic intake and microbial exposures recorded at enrolment (28–32 weeks). Fecal collection and processing: Maternal samples at 36 weeks; infant samples at 1, 6, 12 months; stored at −80°C. DNA extracted (Qiagen PowerSoil). 16S rRNA gene V4 region amplified (292 bp) and sequenced on Illumina MiSeq. USEARCH used to merge, filter, cluster reads at 97% identity into OTUs, identify representative sequences, remove chimeras. mothur used for taxonomic assignment against SILVA v123 Nr99. Samples with <2500 read pairs excluded. Technical replicates used for QC; one sample per individual analyzed. Targeted qPCR using P. copri–specific primers/probe for V4 region to confirm 16S findings and define substantial carriage (>1% expression estimate). Short-chain fatty acids (SCFAs) quantified by capillary gas chromatography at CSIRO. Statistical analysis: Analyses conducted in R using phyloseq and limma. Raw OTU counts normalized and weighted (voom) to account for sequencing depth variability; normalized abundance compared between groups via moderated t-tests. Multiple testing corrected by Benjamini–Hochberg (q-values). Adjusted for processing variables (sequencing batch; storage duration at −80°C; prior home freezer storage). Confounders determined via DAGs and included household size, maternal PUFA intake, paternal allergy, pet ownership, Caucasian ancestry; additional perinatal/postnatal variables (mode of delivery, perinatal antibiotics, breastfeeding, age at solids) also considered. qPCR expression analyzed as log-transformed continuous and as binary (substantial carriage >1%) using binomial regression with log link and inverse probability weighting (IPW) for case-subcohort design. Odds ratios for contingency tables via median unbiased estimation with midpoint-exact CIs. Association between household size and maternal P. copri carriage assessed by logarithmic regression (R package survey). Mediation analysis (R package medflex) using neWeight to incorporate IPW, evaluating infant 6-month carriage as mediator of maternal carriage effect on food allergy. Two-sided tests throughout.

• Maternal P. copri carriage during pregnancy strongly associated with protection against infant food allergy. Differential abundance was most evident for OTU41 and OTU697 mapping to P. copri. OTU41 sequence matched P. copri strain JCM 13464 16S rRNA gene (253 bp region) 100%.
• Within the random subgroup, both presence and abundance-when-present of P. copri (OTU41) were substantially greater in mothers of non-allergic infants than case mothers (p<0.001; q=0.003; limma moderated t-test). Only one case mother had P. copri >0.03%.
• Association remained after adjustment for processing and confounding variables, and after further adjustment for perinatal/postnatal factors (p<0.001).
• qPCR confirmation: Each doubling of maternal fecal P. copri expression was associated with 8% lower risk of food allergy (adjusted risk ratio [aRR] 0.921, 95% CI 0.870–0.974, p=0.005; n=276). Substantial carriage (>1% by qPCR) associated with 83% lower risk (aRR 0.165, 95% CI 0.039–0.693, p=0.02; n=276).
• SCFA concentrations in maternal feces did not differ by maternal P. copri carriage (no evidence of association).
• In the random subcohort, substantial maternal P. copri expression (qPCR) associated with complete absence of offspring atopic wheeze (OR 0, 95% CI 0–0.936, p=0.03; n=283) and a non-significant trend toward reduced atopic eczema (RR 0.280, 95% CI 0.038–2.074, p=0.2; n=291). No associations with wheeze or eczema when ignoring sensitization status.
• Infant P. copri carriage: At 6 months, infant normalized abundance associated with decreased food allergy (p<0.001; q=0.04; n=312), but this attenuated after adjusting for maternal carriage during pregnancy (p=0.85; q=0.96). Maternal carriage effect persisted after adjusting for infant carriage (p<0.001; q=0.19; n=277). Mediation analysis estimated 12.1% of the maternal effect mediated by infant carriage at 6 months.
• Maternal diet: No independent relationship between maternal fiber intake alone and P. copri carriage or offspring allergy. High intake of both fat and fiber (top two quintiles: fiber >22.4 g/day; fat >77.3 g/day) associated with lower infant food allergy risk, with the greatest protective effect among women with substantial P. copri carriage.
• Household size: Strong positive association with maternal P. copri carriage. Predicted probability increased by RR 1.36 per additional household member via 16S (95% CI 1.13–1.65, p=0.001) and RR 1.50 per additional member via qPCR (95% CI 1.33–1.70, p<0.001). Trend toward decreased offspring allergy mediated by maternal P. copri carriage.
• Antibiotics: None of 11 mothers in the random subcohort reporting third-trimester antibiotics were P. copri carriers (Fisher exact p=0.13, n=273).

This study is the first human, culture-independent investigation linking the maternal gut microbiota during pregnancy to clinically confirmed infant food allergy. The data provide strong, dose-responsive evidence that maternal carriage of P. copri is associated with a markedly reduced risk of IgE-mediated food allergy in infancy, independent of numerous potential confounders and perinatal factors. The protective association was most pronounced among mothers consuming diets high in both fat and fiber, consistent with Prevotella metabolism generating immunomodulatory metabolites such as succinate (and potentially SCFAs), although fecal SCFA levels did not reflect differences by carriage. Household size and the absence of recent maternal antibiotic exposure were associated with increased maternal P. copri carriage, aligning with the hygiene hypothesis and shared microbiota within households. The findings suggest that maternal microbial exposures—specifically P. copri—during pregnancy may shape fetal immune development and reduce allergic disease risk. Potential mechanisms include metabolite-mediated effects (e.g., succinate), transplacental transfer of microbial epitopes bound to maternal IgG, or endotoxin-mediated modulation of TLR4 signaling. While infant P. copri carriage at 6 months correlated with reduced allergy, mediation analysis indicates that only a small proportion of the effect is mediated via infant carriage, underscoring the primacy of maternal pregnancy carriage. The results have public health implications, supporting antibiotic stewardship during pregnancy and dietary approaches that support a healthy maternal gut microbiome, while recognizing that P. copri may have context-dependent pro-inflammatory effects in other conditions.

Maternal carriage of Prevotella copri during pregnancy is strongly associated with protection against clinically confirmed food allergy in infants, with evidence of dose-response and independence from multiple confounders. Larger household size predicts maternal P. copri carriage, and diets high in both fat and fiber show the greatest protective association when substantial P. copri carriage is present. The primary effect arises from maternal pregnancy carriage rather than infant carriage. These findings suggest P. copri as a potential biomarker and candidate probiotic target for allergy prevention, contingent on safety considerations. Future research should replicate findings across populations, elucidate mechanistic pathways (including metabolite profiles such as succinate), incorporate metagenomic and metabolomic analyses, assess strain-level heterogeneity and diet–microbe interactions, and evaluate inter-household transmission dynamics and intervention feasibility.

• Lack of metagenomic and metabolomic data limits inference on microbiome function and mechanistic pathways.
• 16S-based OTU matching to a 253 bp region of P. copri does not exclude strain-level genomic heterogeneity; P. copri properties may vary between strains.
• Dietary assessment via food frequency questionnaire may inadequately capture fiber intake.
• Maternal microbiota composition changes across pregnancy; samples were from late gestation, and temporal dynamics were not assessed longitudinally across trimesters.
• Fecal SCFA concentrations may poorly reflect fetal exposure due to rapid consumption in the colon.
• Low prevalence of P. copri carriage in modern communities may necessitate large samples to evaluate mediation of household size effects on allergy.
• Observational design cannot definitively establish causality; residual confounding is possible.