The food challenge risk index: Predicting positive open food challenges to milk, egg and peanuts in children

Abstract

Rationale Suspected food allergy is commonly encountered. Proper diagnosis may require skin testing, RAST testing and ultimately a food challenge. In previous studies, food specific IgE threshold levels have predicted greater than 95% of the patients with a positive food challenge. However, in patients with food specific IgE antibodies less than the threshold, a food challenge is still required to exclude food allergy. Oral food challenges are time consuming and may subject patients to allergic reactions. We developed a food challenge risk index (FCRI) to predict positive food challenges. Methods Fifty-five open food challenges were retrospectively reviewed. Skin prick tests, food specific IgE levels, and time since last reaction were utilized to develop a formula and SPSS program was used to generate ROC curves, and correlations. The FCRI=[(skin prick test wheal (mm) x3) + (RASTx10) - (time since last reaction (months)]. The FCRI value was calculated and compared to the food challenge results. Threshold levels determined were 15, 20, and 20 for milk, egg and peanut respectively. Results Comparison of the FCRI to the actual food challenges across all groups revealed a sensitivity of 0.842 and a specificity of 0.861. An overall negative predictive value indicated that the proposed risk index is accurate 91.2% of the time. Conclusions We have derived a simple and accurate formula which uses readily available clinical data, to predict positive food challenges. The food challenge risk index may be used in a prospective fashion to reduce frequency of food challenges necessary for diagnosis of milk, egg, and peanut allergy.