Sensitivity and specificity of double-blinded penicillin skin testing in relation to oral provocation with amoxicillin in children

Abstract

Current recommendations for the management of penicillin allergy are to perform penicillin skin testing (PST) with penicilloyl-polylysine (PPL) and benzylpenicillin (BP) prior to drug challenge with amoxicillin. However, the role of PST is increasingly questioned in the pediatric setting. To resolve the question of PST's diagnostic accuracy, consecutive children with a history of non-life-threatening penicillin allergy referred to a tertiary-care allergy center were recruited to undergo double-blinded PST with PPL and BP prior to drug provocation to amoxicillin. Five of 158 participants (3.2%) presented with an immediate or accelerated reaction upon amoxicillin challenge, none of which were severe. Only one of these had positive PST (20%), compared to 15 of 153 amoxicillin tolerant participants (9.8%). The sensitivity and specificity of PST with PPL and BP for reacting upon amoxicillin challenge were 20% (95% CI: 0.5-71.6%) and 90% (95% CI: 84.4-94.4%), respectively. These results argue against the routine use of PST as a preliminary step to drug provocation with amoxicillin in this population, as it is unlikely to significantly alter pre-test probability of reacting to challenge.