Understanding Food Allergy Risk Factors: Current Knowledge and Recent Advances Using a Large Retrospective Cohort Analysis

Abstract

Introduction: Immunoglobulin E-mediated food allergies (FAs) is a growing concern in the Western world. Established risk factors include a personal or familial history of other atopic comorbidities, genetic predisposition, male sex, and hygienic environment. However, these factors fail to explain most FA cases. The study aimed to discover additional risk factors for FA by analyzing a large database.

Methods: This retrospective study utilized data from Clalit Health Services, Israel's largest healthcare provider. Employing explainable artificial intelligence methods, the investigation sought to identify variables in the first 6 months of life and parental factors correlated with FA diagnosis, among children born from January 1, 2006, to August 31, 2021.

Results: The analysis encompassed 370,298 children, with FA diagnosed in 6,911 infants (1.87%). Key findings revealed that high socioeconomic score (SES) (relative risk [RR] = 9.92), Jewish origin (RR = 3.83), and personal history of atopic dermatitis (RR = 5.3) were the most significant variables associated with FA development. Other variables with lesser correlation strength included parental atopic comorbidities, blood type, and antibiotics and anti-acid drugs use. Surprisingly, prematurity and birthweight <2,500 g were correlated with a lower risk of FA development.

Conclusions: The correlations found between those variables and FA do not explain most FA cases. Moreover, the impact of SESs and ethnicity might be explained by differences in cultural behaviors that influence the development of FA. This could not be determined as data such as the age of allergenic foods introduction were unavailable in the database. Further investigation is warranted to establish causal links and the clinical significance of these suggested risk factors.

Keywords: Atopic comorbidities; Explainable AI; Food allergy; Infants; Risk factors.