Methods: This single healthcare system, retrospective, observational, matched cohort study evaluated adults hospitalized with GN BSI from January 1, 2010 to June 30, 2015. All patients with PA were enrolled and matched with control patients with NPA in a 1:1 fashion based on: age ±10 years, sex, site of acquisition (community-acquired [CA], vs. healthcare-associated), and pre- vs. post- ASST interventions (beginning January 1, 2014). Multivariable logistic regression was used to evaluate the impact of PA and ASST interventions on empiric carbapenem use.
Results: A total of 280 patients were enrolled. Overall, the mean age was 71 years, 208 (74%) were women, 142 (51%) had CA BSI, and 64 (23%) were in the post-ASST intervention period. Empiric carbapenem use was greater in the PA vs. NPA groups (27% vs. 12%). After adjustment for covariates, including baseline differences in race and indwelling urinary catheters, PA remained independently associated with empiric carbapenem use (aOR 3.98, 95% CI 1.98-8.45, p<0.001). Other predictors of empiric carbapenems included male sex (aOR 2.18, 95% CI 1.08-4.36), Bloodstream Infection Mortality Risk Score (aOR 1.19 per point, 95% CI 1.08-1.32), and prior beta-lactam use within 90 days (aOR 2.72, 95% CI 1.07-6.64). ASST interventions were associated with reduction in empiric carbapenem use (aOR 0.41, 95% CI 0.16-0.40). Time to carbapenem de-escalation in the PA group was also significantly shorter in the post- vs. pre-ASST intervention period (2.0 vs. 4.2 days, p<0.001).
Conclusion: Penicillin allergy is an independent risk factor for empiric carbapenem use for community-onset GN BSI. ASST interventions successfully reduced carbapenem utilization.
J. Kohn, None
P. B. Bookstaver, Allergan Pharmaceuticals: Scientific Advisor , Research grant
C. Troficanto, None
J. A. Justo, None