A Retrospective Critical Analysis and Risk Stratification of Penicillin Allergy Delabeling in a UK Specialist Regional Allergy Service.

Background: A spurious label of penicillin allergy (Pen-A) negatively impacts on antibiotic stewardship and health care costs. Recent studies have proposed a guideline-steered direct penicillin challenge without undertaking allergy tests when "true allergy" is unlikely.

Objective: To critically analyze Pen-A clinical presentation, perform risk stratification, and determine clinical predictors for "true allergy."

Method: Data were extracted retrospectively from clinical and electronic patient records.

Results: A total of 231 patients (M = 82; F =149; mean age 51.22 [standard deviation ± 18.07] years) were analyzed. Based on clinical history, patients were categorized as likely type I hypersensitivity reaction (HSR) (n = 27), likely type IV HSR (n = 65), indeterminate (n = 111), and HSR unlikely (n = 28). Based on an index reaction and comorbidities, patients were classified into "low risk" (n = 143) and "high risk" (n = 78). Pen-A was excluded in 74% of patients assessed having likely type I HSR, 91% with likely type IV HSR, 93% of indeterminate, and 100% of HSR unlikely patients. The negative predictive value for successful delabeling in the "low risk" group was 94% (odds ratio [OR] = 2.9; P = .02). Predictors for "true Pen-A" were history of anaphylaxis (OR = 30.6; P < .001), hospitalization (OR = 7; P < .001), ≤5 years since the index reaction (OR = 3; P = .04).

Conclusions: Systematic clinical characterization and risk stratification has an important role in Pen-A delabeling. These data provide proof of concept for a guideline-based selection of patients labeled with Pen-A for a direct penicillin challenge. Patients in the "low risk" group seem suitable for this intervention, although a rigorous prospective evaluation is needed in a multicenter study.