Ethnicity associations with food sensitisation are mediated by gut microbiota development in the first year of life

What do we already know about this subject?

The number of children with food allergies is increasing rapidly, currently representing 28% of children aged 1-5 years in the United States. The development of the gut microbiota (GM) in the first months of life may be involved in this sensitisation to food allergens [2]. Many factors influence the establishment of GM, such as the mode of delivery (caesarean versus vaginal), type of breastfeeding (breast or formula) and use of antibiotics [3, 4]. A recent study showed that GM structure also varied widely between different ethnic groups [5].

Transferring GM from healthy children to mice was shown to protect them from cow’s milk allergy. Low GM richness in young infants and a high ratio of Enterobacteriaceae/Bacteroidaceae (E/B) in early and late infancy are predictors of food allergen sensitisation [6].

What are the main insights from this study?

The study included 1,422 children from the CHILD (Canadian Healthy Infant Longitudinal Development) cohort. Prick tests were performed (inhalant and food allergens) at the ages of 1 and 3 years. Stool samples were collected in early (3.5 ± 0.9 months) and late (12.2 ± 0.3 months) infancy.

Atopy prevalence was 12% at 1 year and 12.8% at 3 years, with 9.5% and 5.8% of food sensitisation and 3.3% and 10.1% of sensitisation to inhalant allergens at ages 1 and 3 years, respectively.

Late infancy GM had lower beta-diversity and intra-individual variability compared to early infancy GM (p < 0.001). Late infancy gut microbiota were enriched with Bacteroides, Faecalibacterium, Lachnospira, Prevotella, unclassified Lachnospiraceae and unclassified Clostridiales, but depleted with Clostridium, Veillonella, Bifidobacterium and unclassified Enterobacteriaceae. The principal component analysis identified two clusters (C1 and C2, Figure 1). C1 was composed of 75.5% of early infancy samples and C2 of 63.7% of late infancy samples. These early and late infancy samples representing vaginal births without intrapartum antibiotic prophylaxis were of type C2, dominated by the genus Bacteroides (Figure 2).

The authors identified four trajectories according to the type of early and late infancy cluster: C1-C1, C1-C2, C2-C1 and C2-C2. The C1-C1 trajectory was more common among Asian than Caucasian infants (p < 0.05), as well as in atopic-risk children compared to the C2-C2 (OR 1.9; 95% CI 1.15-3.14) or C1-C2 (OR 2.38; 95% CI 1.43-3.96) trajectory. Infants in the C1-C1 trajectory were twice as likely to have food sensitisation at the age of 3 years compared to those in the C2-C2 trajectory (OR 2.34; 95% CI 1.20- 4.56) and C1-C2 trajectory (OR 2.60; 95% CI 1.33-5.09), especially to peanuts (vs C2-C2 = OR 2.82; 95% CI 1.13-6.01 and vs C1-C2 = OR 2.01; 95% CI 0.85-4.78) (Figure 3). Children who had not acquired peanut sensitisation at the age of 3 years had persistently higher levels of Bacteroides
(p = 0.044), lower levels of unclassified Enterobacteriaceae (p = 0.001) and a lower E/B ratio (p = 0.013) throughout childhood.

The C1-C1 trajectory mediated the risk of food and peanut sensitisation in Asian children. The association was even high for peanuts (OR 7.87; 95% CI 2.75-22.55). Infants in the C1-C1 trajectory were more often colonised with C. difficile. These same children, with both the C1-C1 characteristic and colonised with C. difficile, had an extra risk of food (OR 5.69; 95% CI 1.62-19.99) and peanut (OR 5.89; 95% CI 1.16-29.87) sensitisation.

Finally, microbiota of the C1-C1 trajectory had a deficiency in sphingolipid metabolism and glycosphingolipid biosynthesis-related functions.

What are the consequences in practice?

This study allows us to envisage GM-targeted therapies for food allergies in infants, either as a preventative or therapeutic option.