Nomogram for predicting food allergy in infants with feeding problems and malnutrition

Abstract BACKGROUND: It is difficult to diagnose food allergy in infants in General Outpatient Department of Pediatrics as oral food challenge (OFC) cannot be performed routinely.OBJECTIVE: To construct a nomogram to screen for a high probability of food allergy in infants with feeding problems and malnutrition. METHODS: From August 2018 to December 2021, 289 infants with feeding problems and malnutrition from 7 hospitals in Shanghai, China were recruited. Food allergy was defined as positive response of skin prick test or OFC with gastrointestinal, dermatologic, or respiratory symptoms. Demographic characteristics, the cow's milk-related symptom scores (CoMiSS) and blood eosinophils (EOS) were obtained for evaluating their correlation with food allergy. Variables identified by multivariate logistic regression were included to develop the nomogram model with the bootstrapped-concordance index as an assess index.RESULTS: Altogether, 249 of 289 infants with feeding problems and malnutrition had food allergy (86.2%). After analyzing, feeding pattern (Odds ratio (OR) = 5.284; 95% confidence interval (CI): 2.133-13.086), family history of allergy (OR=1.793; 95%CI: 0.713-4.508), CoMiSS score (OR=1.450; 95%CI: 1.187-1.771) and EOS percentage (OR=1.332; 95%CI: 1.111-1.598) were included to develop the model, which had a good performance with an area under the curve of 0.868 (95% CI: 0.792-0.944) and a bootstrapped-concordance index of 0.868. Among them, EOS percentage was the strongest factor producing a highly probability in the model. CONCLUSION: Nomogram model can be a useful tool to screen infants with high probability of food allergy among infants with feeding problems and malnutrition for further food allergy diagnosis.