1957. Impact of β-Lactam Antibiotic Allergy on Antimicrobial Use, Clinical Outcomes, and Costs for Hospitalized Children

Abstract

BackgroundMost β-lactam antibiotic allergies (BLA) are incorrectly diagnosed and could be de-labeled. Adult patients with BLA are more likely to receive broader-spectrum antimicrobials and experience worse health outcomes than nonallergic patients. Similar studies on the impact of BLA on antimicrobial use and clinical outcomes are limited in pediatrics. Our objective was to compare antimicrobial use, and clinical and economic outcomes between hospitalized children with and without BLA.MethodsThis was a retrospective cohort of pediatric patients hospitalized at an Intermountain Healthcare (IH) hospital from 2007 to 2017. IH has 22 hospitals including one children’s hospital. Patients aged 30 days-17 years who received ≥1 dose of an antimicrobial during hospitalization were included. The exposure variable was the presence of BLA (penicillins or cephalosporins) in the allergy field of the medical record. Patients with BLA were matched to nonallergic controls on age, sex, race, clinical service line, admission date, children’s hospital or other hospital, and co-morbid conditions. We used multivariable log-transformed-linear and logistic regression models to compare patients with BLA to controls in terms of antibiotic selection and total antimicrobial days, antimicrobial cost, length-of-stay (LOS) and 30-day readmission. For antibiotic selection we examined the odds of receiving the following broader-spectrum agents individually and in composite: vancomycin, fluoroquinolones, clindamycin, carbapenems, and macrolides.Results39,785 patients were identified including 2897 (7%) with BLA. The prevalence of BLA increased with age (Figure 1). 2459 (85%) patients with BLA were matched to a control. Patients with BLA had higher odds of receiving broader-spectrum antibiotics (OR 2.35, 95% CI: 2.07–2.67) and had greater antimicrobial costs (1.21-fold increase, 95% CI: 1.08–1.35) than nonallergic patients (Figure 2). There were no differences in LOS, total antimicrobial days, or 30-day readmission (Figure 2).ConclusionPediatric patients with BLA are more likely to receive broader-spectrum antibiotics and incur higher antimicrobial costs than matched controls. De-labeling interventions could reduce unnecessary exposure to these agents and lower costs. Disclosures All Authors: No reported Disclosures.