Abstract: 　Objective　To explore the value of allergy risk score obtained by combining multiple risk factors in predicting high risk of food allergy in children. Methods　Questionnaire survey was performed to obtain information of allergic risk factors in infants under 3 years old. Multivariate logistic stepwise regression analysis was used to analyze the risk factors of food allergy in infants. The positive predictive value of each risk factor for food allergy was calculated. Allergy risk score was calculated based on regression equation. The value of allergy risk score in screening children having high risk of food allergy was evaluated by ROC curve. Results The information was retrospectively obtained from 78 infants diagnosed with food allergy and 156 infants under 3 years old with nonallergic diseases. Family history of allergic diseases, cesarean section, use of antibiotics or antipyretic analgesics during pregnancy, smoking or exposure to smoking environment during pregnancy, exposure to insecticides or mosquito repellents during pregnancy, and pets at home are all risk factors for food allergy in infants and young children (all P<0.05). The positive predictive value of family history of food allergy was 63.2%, and the area under ROC curve (AUC) was 0.696 (95% CI: 0.620~0.771). When the allergy risk score obtained by combining multiple factors was more than 2.85, the AUC for predicting food allergy was 0.804 (95% CI: 0.746~0.863), with a sensitivity of 0.526, a specificity of 0.910 and a Youden index of 0.436. Conclusion　The predictive value of multi-risk factors combination is higher for screening high-risk children of food allergy than that of family history of allergy.

Key words: 　food allergy;　high-risk children;　positive predictive value;　allergy risk score