Prospective assessment of diagnostic tests for pediatric penicillin allergy From clinical history to challenge tests

Background: Diagnostic guidelines for penicillin allergy in children recommend cumbersome protocols based partially on data from adults, which may be suboptimal for pediatric use. Objective: To assess the accuracy of tools for diagnosis of penicillin allergy in children. Methods: A prospective, multicenter study was conducted in children with reported adverse events related to penicillin, excluding severe reactions. All patients underwent a uniform diagnostic protocol that consisted of clinical history, skin tests, serum specific IgE (sIgE), and, regardless of these results, drug provocation tests (DPTs). Results: A total of 732 children (mean age, 5.5 years; 51.2% males) completed the allergy workup, including DPTs. Amoxicillin triggered 96.9% of all reactions. None of the patients with an immediate index reaction (IR) developed a reaction on DPT. Penicillin allergy was confirmed in 35 children (4.8%): 6 immediate reactions (17%) and 29 nonimmediate reactions (83%) on the DPT. No severe reactions were recorded. The allergist diagnosis based on the clinical history was not associated with the DPT final outcome. In 30 of 33 allergic patients (91%), the results of all skin tests and sIgE tests were negative. A logistic regression model identified the following to be associated with penicillin allergy: a family history of drug allergy (odds ratio [OR], 3.03; 95% confidence interval [CI], 1.33-6.89; P = .008), an IR lasting more than 3 days vs 24 hours or less (OR, 8.96; 95% CI, 2.01-39.86; P = .004), and an IR treated with corticosteroids (OR, 2.68; 95% CI, 1.30-5.54; P = .007). Conclusion: Conventional predictors of allergy to penicillin performed weakly. The authors propose straightforward penicillin provocation testing in controlled, experienced centers for the diagnosis of nonsevere penicillin allergy in children. (C) 2018 American College of Allergy, Asthma & Immunology. Published by Elsevier Inc. All rights reserved.