Reevaluating Corticosteroid Classification Models in Patient Patch Testing

Abstract

Individuals with allergic contact dermatitis to one topical corticosteroid may also react to other corticosteroids. Corticosteroid classification models have been proposed to predict such copositivity, recommend representative screening corticosteroids, and guide allergen avoidance. To use patient data to determine copositivity patterns between corticosteroids and evaluate against previous corticosteroid classification models. This qualitative study included a retrospective analysis of the Mayo Clinic Contact Dermatitis Group corticosteroid patch test data from 2010 to 2019. Among patients undergoing patch testing with the Mayo Clinic's standard or steroid series who consented to research participation, 5637 patients were included in the analysis. Copositivity rates were determined between corticosteroids and analyzed by hierarchical clustering for comparison to previous classification models. The frequency of patch test positivity to each of the analyzed corticosteroids was noted and compared with previously published patch test positivity rates. Copositivity rates between each pair of corticosteroids were determined, and overall copositivity patterns were analyzed and evaluated against known steroid classes. A total of 49 472 individual patches were applied to 5637 patients, testing 18 corticosteroids. Patch test positivity rates ranged between 0.3% and 4.7%. The fluocinonide positivity rate corresponded to the highest copositivity rate with other corticosteroids (mean [SD], 50.7% [26.1%]). Tixocortol-21-pivalate, 0.1%, and tixocortol-21-pivalate, 1%, positivity rates corresponded to the lowest copositivity rates (mean [SD], 4.1% [1.7%] and 3.6% [1.4%], respectively). Hierarchical clustering elucidated patterns that did not support previous corticosteroid classification models. In this qualitative study, copositivity rates were variable between corticosteroids, and overall patch test positivity for allergy to topical corticosteroids was rare. Previously published corticosteroid classifications are not supported by real patient-derived data and may not be accurate in predicting corticosteroid copositivity.