A Phase 1 Prognostic Trial for Predicting Paediatric Allergy Using the Placenta at Birth.

Allergy is the most common and earliest onset non-communicable disease in children. The early identification of children who are at risk of allergy would allow early intervention to prevent the onset of the disease or reduce its impact. The origins of allergy are hypothesized to be derived from early developmental events which include prenatal events and early childhood exposures. We have previously identified that many placental glucocorticoid-regulated genes were also associated with child allergy. We have questioned if it is possible to predict which children are at risk of allergy based on their placental glucocorticoid-regulated mRNA profile. Placentae from two different populations located in South Australia (n = 105) and Victoria, Australia (n = 261) were included in the study. Glucocorticoid regulated genes were measured by qPCR. Statistical modelling was executed in R (version 4.4.1). Evaluation of the importance of each gene in the model identified AFF1, ARID5B, IER3, ATF4 and SLC19A2 as the top five ranking genes. The best-performing random forest model achieved an AUC of 0.664, indicating moderate ability to distinguish between positive and negative allergic susceptibility. While our models demonstrate both measurable specificity and sensitivity, the glucocorticoid genes particularly excel at identifying children who will not develop an allergy. This study demonstrates that placental glucocorticoid-regulated genes possess significant predictive power, especially in identifying individuals unlikely to develop allergies. It offers further evidence that the placentae of nonallergic children and allergic children are transcriptionally distinct.