Old but not forgotten: Antibiotic allergies in General Medicine (the AGM Study)

Objectives: To determine the nature, prevalence and description accuracy of recorded antibiotic allergy labels (AALs) in a cohort of general medical inpatients, and to assess the feasibility of an oral antibiotic re-challenge study. Design: Multicentre cross-sectional study. Setting and participants: All patients admitted to the general medical units of Austin Health and Alfred Health, 18 May-5 June 2015. Main outcome measures: Baseline demographics, medical and allergy history, infection diagnoses and antibiotic prescribing data for general medical inpatients were collected. A questionnaire was administered to clarify AAL history, followed by correlation of responses with electronic and admissions record descriptions. A hypothetical oral re-challenge in a supervised setting was offered to patients with low risk allergy phenotypes (non-immediate reaction, non-severe cutaneous adverse reaction, or unknown reaction more than 10 years ago). Results: Of the 453 inpatients, 107 (24%) had an AAL (median age, 82 years; interquartile range, 74-87 years); 160 individual AALs were recorded, and there was a mismatch in AAL description between recording platforms in 25% of cases. Most patients with an AAL were women (64%; P< 0.001), and more presented with concurrent immunosuppression than those without an AAL (23% v 8%; P< 0.001). beta-Lactam penicillins were employed less frequently in patients with an AAL (16% v 35%; P = 0.02), while ceftriaxone (32% v 20%; P = 0.02) and fluoroquinolones (6% v 2%; P = 0.04) were used more often. Fifty-four per cent of patients with AALs were willing to undergo oral re-challenge, of whom 48% had a low risk allergy phenotype. Conclusions: AAL prevalence in general medical inpatients was 24%, and was associated with excessive use of broad spectrum antibiotics. Allergies in a large proportion of patients with AALs were incorrectly documented, and were non-immune-mediated and potentially amenable to oral re-challenge. A direct oral re-challenge study in carefully selected patients with low risk allergy phenotypes appears feasible.