Fruitful or unfruitful: strawberry and tomato specific immunoglobulin E testing at a tertiary pediatric center

Suspected strawberry and tomato (S/T) food allergy (FA) can be evaluated using specific immunoglobulin E (sIgE) testing despite its low specificity and positive predictive value.

This study aims to understand ordering patterns for S/T sIgE testing and identify relevant factors to clinical decision-making.

We retrospectively reviewed 814 patients with sIgE testing available for strawberries (651), tomatoes (276), or both (113) from January 2012 to May 2022 at a tertiary pediatric hospital. Patient demographics, provider specialty, and reasons for testing were collected. Student's t-test and multiple regression analyses were performed to test the association between the S/T sIgE level and clinically relevant outcome (CRO) status. Fisher's exact test and general linear models were used to evaluate and compare potential predictive factors for CRO status.

Allergy and immunology, gastroenterology, and general pediatrics ordered most S/T sIgE testing. Testing was ordered most frequently for non-IgE-mediated gastrointestinal symptoms, mild possible IgE-mediated reactions, and eczema. Testing was most often ordered for infants and school-age children. Mean sIgE levels were higher for S/T tests resulting in a CRO when controlling for other predictor variables (p = 0.015; p = 0.002 for S/T, respectively). Only 2.2% and 5.4% of tests resulted in a CRO for S/T, and severe allergy was rare. Testing for non-IgE-mediated GI symptoms or eczema, or in non-atopic patients, yielded no CROs. Exposure and reaction history of present illness (ERH) was associated with CROs (p < 0.001; p = 0.04) with a high negative predictive value (99.5%; 100%) and low positive predictive value (11.5%; 15.0%). ERH (p < 0.001, η² = 0.073; p = 0.009, η² = 0.123) was a more significant predictor than the sIgE level (p = 0.002, η² = 0.037; p = 0.212, η² = 0.030) for CRO status.

The diagnosis of S/T food allergy is made primarily based on clinical history. S/T sIgE testing for children and adolescents should be avoided for patients without an ERH and in the workup of non-IgE-mediated GI symptoms. Testing for eczema and non-atopic patients is likely low-yield.