Prospective Assessment of Diagnostic Tests for Pediatric Penicillin Allergy: From Clinical History to Challenge Tests

Melinda M. Rathkopf

doi:10.1542/peds.2019-2461Y

MD Ibáñez, P Rodríguez Del Río, EM Lasa; Penicillin Allergy in Children (APENIN) Task Force; Pediatric Allergy Committee, Spanish Society of Allergy and Clinical Immunology (SEAIC). Ann Allergy Asthma Immunol. 2018;121(2):235–244.e3

Purpose of the Study:

To determine the efficacy of available diagnostic tools in penicillin allergy relative to the results of drug provocation tests (DPTs).

Study Population:

The population included 732 children ≤14 years of age presenting to participating allergy departments who reported an adverse drug reaction with benzylpenicillin (penicillin G), phenoxymethyl penicillin (penicillin V), amoxicillin-clavulanic acid, or cloxacillin. Patients with a history of severe skin conditions during the index reaction (IR), presence of skin disease not allowing skin testing, and/or any condition for which epinephrine use was contraindicated were excluded.

Methods:

This was a multicenter study (38 allergists from 23 centers) across Spanish allergy departments (from the Spanish Society of Allergy and Clinical Immunology). After a detailed history, IRs were classified as immediate (symptoms within 1 hour of the first dose) or nonimmediate. Researchers then classified reactions as a “clear-positive” case, “doubtful” case, or a “clear-negative” case to assess accuracy of clinical history alone. Then, every patient underwent a uniform protocol of skin tests (skin prick test and intradermal test), serum-specific immunoglobulin E (IgE) tests, and then, regardless of the results of these tests, DPTs.

Results:

Of the 732 participants who completed the study, 35 (4.8%) were diagnosed with a penicillin allergy at the end of the tests: 6 (17% of the patients with a penicillin allergy) with immediate reactions and 29 (83%) with nonimmediate reactions. A family history of drug allergy was associated with a higher risk of penicillin allergy (odds ratio: 2.56). Amoxicillin alone or with clavulanic acid was the most frequent trigger. The most frequently reported symptom was an exanthema and was more commonly associated with nonimmediate IR (58.4% vs 30% in immediate IR). Urticaria was next most common, associated more with an immediate IR (55.5% vs 37.6% of nonimmediate IRs). Thirty-one cases (4.2%) were classified as clear positive on the basis of history alone, 518 were classified (70.8%) as clear negative, and 183 (25%) were classified as doubtful. One of the clear-positive cases turned out to be a true penicillin allergy on the basis of DPTs. Twenty-three of the cases initially classified as clear negative and 12 of the doubtful cases turned out to be a true allergy. Only 1 positive skin prick test result was recorded for benzylpenicilloyl polylysine, and none were recorded for the other tests. Few had positive intradermal test results. Overall skin testing had low sensitivity (0% to 20%) and high specificity (97% to 100%) for skin testing. For serum-specific IgE, sensitivity was low (2.9% to 17%), but specificity was high (94.4% to 99.9%).

Conclusions:

Currently available diagnostic methods (clinical history, skin testing, and serum-specific IgE) performed poorly compared with DPT. DPTs performed in specialized centers should be used as the first diagnostic tool to evaluate children with mild to moderate penicillin IRs when anaphylaxis and severe skin syndromes have been excluded.

Reviewer Comments:

With the knowledge that most adults labeled as penicillin allergic are not and a push to confirm the diagnosis, the allergist is performing more penicillin allergy evaluations. With regards to a pediatric population, this study shows the need for DPTs to best evaluate children labeled as penicillin allergic, despite the logistic difficulties in performing this diagnostic test.