The burden of antibiotic allergies in adults in an Australian intensive care unit: the BASIS study

Objective: To determine the prevalence and impact of patient-reported antibiotic allergies in the intensive care unit (ICU), which are currently poorly defined. Antibiotic allergy labels (AALs) are associated with inappropriate antibiotic prescribing and with inferior patient, microbiological and hospital outcomes. Design: Prospective, single-centre case-control study. Setting: Mixed ICU, Austin Hospital, Melbourne. Participants: All adults (>= 18 years old) admitted to the ICU who received at least two doses of systemic antibiotics between 12 February and 20 April 2018. Main outcome measures: Demographic data, infection and allergy history, antibiotic prescriptions and ICU interventions and outcomes. Results: Of the 247 patients (79.9%) who received systemic antibiotics, 43 patients (17.4%) had an AAL and 204 (82.6%) did not. A higher proportion of patients with AAL were female (P = 0.032) and received vancomycin (37.2% AAL v 18.6% no antibiotic allergies [NAAL]; P = 0.014), and a lower proportion of patients received narrow spectrum beta-lactams (39.5% AAL v 58.8% NAAL; P = 0.028). On multivariable logistic regression, the AAL cohort had twice higher odds of receiving vancomycin (odds ratio [OR], 2.04; 95% CI, 1.07-3.86; P = 0.029) and half the odds of receiving a narrow spectrum beta-lactam (OR, 0.52; 95% CI, 0.29-0.94; P = 0.03). AAL distribution on the electronic medical record included 17% type A (predictable), 13% type B-I (immediate), 2% type B-IV (delayed), 35% type B (unspecified), and 32% unknown. An interview clarifying allergy phenotype found that 59.5% of AALs matched their documented description. Conclusion: Patients with AALs had twice the odds of receiving intravenous vancomycin and half the odds of receiving narrow spectrum beta-lactams, which highlights the continued need for antimicrobial stewardship initiatives.