Using Serum IgE Antibodies to Predict Esophageal Eosinophilia in Children.

ABSTRACT

OBJECTIVES: Symptoms of eosinophilic esophagitis are variable and can be nonspecific. Food-specific serum immunoglobulin E (IgE) antibodies are frequently found in patients with eosinophilic esophagitis and are obtained using a widely available blood test. Our objective was to evaluate the ability of food-specific IgE antibodies to predict the presence of esophageal eosinophilia. METHODS: We reviewed 144 medical records for pediatric patients having esophageal biopsy and serum analysis for IgE antibodies to food (exploratory group). We performed logistic regression using sex and number of positive food-specific IgE tests to develop a model that predicts ≥15 eosinophils/high-power field (hpf) in the esophagus. We tested the model using 142 additional patients (validation group). RESULTS: The probability of having ≥15 eosinophils/hpf in the esophagus was higher in boys and increased with the number of positive food-specific IgE tests from 12% (95% confidence interval 4.8-26) in girls with 0 foods positive to 86% (95% confidence interval 71-94) for boys with 4 or 5 foods positive. The statistical model using sex and number of positive IgE tests to predict patients having ≥15 eosinophils/hpf showed acceptable discriminative ability (area under the receiver operating characteristic curve 0.80). The performance metrics for the model to predict ≥15 eosinophils/hpf in the validation group were similar (area under the receiver operating characteristic curve 0.75). CONCLUSIONS: Requiring only a blood test and a simple algorithm, analysis for IgE antibodies to food may expedite an esophagogastroduodenoscopy and decrease delays in the diagnosis and treatment of patients with nonspecific gastrointestinal symptoms who have increased eosinophils in the esophagus.