Background: There is uncertainty about the prevalence of food allergy in communities. Objective: To assess the prevalence of food allergy by performing a meta-analysis according to the method of assessment used. Methods: The foods assessed were cow's milk, hen's egg, peanut, fish, shellfish, and an overall estimate of food allergy. We summarized the information in 5 categories: self-reported symptoms, specific IgE positive, specific skin prick test positive, symptoms combined with sensitization, and food challenge studies. We systematically searched MEDLINE and EMBASE for publications since 1990. The meta-analysis included only original studies. They were stratified by age groups: infant/preschool, school children, and adults. Results: A total of 934 articles were identified, but only 51 were considered appropriate for inclusion. The prevalence of self-reported food allergy was very high compared with objective measures. There was marked heterogeneity between studies regardless of type of assessment or food item considered, and in most analyses this persisted after age stratification. Self-reported prevalence of food allergy varied from 1.2% to 17% for milk, 0.2% to 7% for egg, 0% to 2% for peanuts and fish, 0% to 10% for shellfish, and 3% to 35% for any food. Conclusion: There is a marked heterogeneity in the prevalence of food allergy that could be a result of differences in study design or methodology, or differences between populations. Clinical implications: We recommend that measurements be made by using standardized methods, if possible food challenge. We need to be cautious in estimates of prevalence based only on self-reported food allergy.
Journal of Allergy and Clinical Immunology, 2007, Vol 120, Issue 3, p. 638-646