Improving the Effectiveness of Penicillin Allergy De-labeling

Abstract

Approximately 10-20% of hospitalized patients are labeled as penicillin allergic, and this is associated with significant health and economic costs. We looked at the effectiveness of penicillin allergy de-labeling in clinical practice with the aim of deriving risk stratification models to guide testing strategies. Consecutive patients aged 15 years or more, referred to a Western Australian public hospital drug allergy service between 2008 and 2013 for beta-lactam allergy, were included. Follow-up surveys were conducted. Results of skin prick testing and intradermal testing (SPT/IDT) and oral challenge (OC), and follow-up of post testing antibiotic usage were the main outcomes. SPT/IDT was performed in 401 consecutive patients with immediate (IMM) (≤ 1 hour) (n = 151) and nonimmediate (NIM) (>1 hour) (n = 250) reactions. Of 341 patients, 42 (12.3%) were SPT/IDT+ to ≥ 1 penicillin reagents, including 35/114 (30.4%) in the IMM group and 7/227 (3.1%) in the NIM group (P < .0001). Of 355 SPT/IDT patients, 3 (0.8%), all in the IMM group, had nonserious positive OC reactions to single dose penicillin VK (SPT/IDT negative predictive value [NPV] 99.2%). Selective or unrestricted beta-lactam was recommended in almost 90% overall, including 238/250 (95.2%) in the NIM group and 126/151 (83.4%) in the IMM group (P = .0001). Of 182 patients, 137 (75.3%) were following the allergy label modifications (ALM) at the time of follow-up. Penicillin SPT/IDT/OC safely de-labels penicillin-allergic patients and identifies selective beta-lactam allergies; however, incomplete adherence to ALM recommendations impairs effectiveness. Infrequent SPT/IDT+ and absent OC reactions in patients with NIM reactions suggest OC alone to be a safe and cost-effective de-labeling strategy that could improve the coverage of penicillin allergy de-labeling in lower risk populations.