Background: There is a paucity of data examining the natural history of and risk factors for egg allergy persistence, the most common IgE-mediated food allergy in infants. Objective: We aimed to assess the natural history of egg allergy and identify clinical predictors for persistent egg allergy in a population-based cohort. Methods: The HealthNuts study is a prospective, population-based cohort study of 5276 infants who underwent skin prick tests to 4 allergens, including egg. Infants with a detectable wheal were offered hospital-based oral food challenges (OFCs) to egg, irrespective of skin prick test wheal sizes. Infants with challenge-confirmed raw egg allergy were offered baked egg OFCs at age 1 year and follow-up at age 2 years, with repeat OFCs to raw egg. Results: One hundred forty infants with challenge-confirmed egg allergy at age 1 year participated in the follow-up. Egg allergy resolved in 66 (47%) infants (95% CI, 37% to 56%) by 2 years of age; however, resolution was lower in children with baked egg allergy at age 1 year compared with baked egg tolerance (13% and 56%, respectively; adjusted odds ratio, 5.27; 95% CI, 1.36-20.50; P=.02). In the subgroup of infants who were tolerant to baked egg at age 1 year, frequent ingestion of baked egg (>= 5 times per month) compared with infrequent ingestion (0-4 times per month) increased the likelihood of tolerance (adjusted odds ratio, 3.52; 95% CI, 1.38-8.98; P=.009). Mutation in the filaggrin gene was not associated with the resolution of either egg allergy or egg sensitization at age 2 years. Conclusion: Phenotyping of egg allergy (baked egg tolerant vs allergic) should be considered in the management of this allergy because it has prognostic implications and eases dietary restrictions. Randomized controlled trials for egg oral immunotherapy should consider stratifying at baseline by the baked egg subphenotype to account for the differential rate of tolerance development.