2259. Predictors of Empiric Carbapenem Therapy in Complicated Intra-Abdominal Infections in the United States, 2013–2017: A Retrospective Cohort Study

BACKGROUND: Complicated intra-abdominal infections (cIAI) remain an important cause for hospitalization. Evidence-based guidelines recommend reserving broad-spectrum antibiotic coverage for high-risk cases in order to reduce overuse of certain antibiotic classes, particularly in the face of emerging carbapanem resistance. We examined the factors associated with use of empiric carbapenem treatment (ECT) among hospitalized patients with cIAI. METHODS: We performed a multicenter retrospective cohort study in the Premier database of approximately 180 hospitals, 2013–2017. Using an ICD-9/10 based algorithm including a requirement for a laparotomy/laparoscopy, we identified all adult patients hospitalized with cIAI and included those with a positive blood or abdominal culture. We derived and tested a multivariable logistic regression model to examine predictors of ECT. RESULTS: Among 321,317 hospitalized patients with cIAI, 4,453 (1.4%) were culture-positive, 1,185 (26.6%) of whom received ECT. Among those given ECT, >50% (682) had no risk factors for resistance, and in only 120 (10.1%) was an organism resistant to a third-generation cephalosporin (C3R extended spectrum β-lactamase [ESBL] phenotype) isolated. The top 5 variables associated with ECT use were: pre-cIAI anti-fungal therapy (OR 2.57, 95% CI 1.91, 3.45) urgent (vs. emergent) admission (OR 1.56, 95% CI 1.21, 2.01), corticosteroids (OR 1.50, 95% CI 1.13, 1.99), ICU admission (OR 1.46, 95% CI 1.17, 1.82), and presence of sepsis/septic shock (OR 1.43, 95% CI 1.18, 1.74). The model had a moderately good fit (c-statistic = 0.683; 95% CI (0.665, 0.700), Hosmer-Lemeshow P value = 0.411). CONCLUSION: Among patients hospitalized with a cIAI, 26.6% received ECT despite >50% lacking risk factors for resistance, and an only 10% prevalence of C3R in this cohort. This suggests that there remains an opportunity for carbapanem-sparing strategies. Further stratification of the risk for resistance is needed among patients with markers of high illness severity, such as those identified in our model. DISCLOSURES: All authors: No reported disclosures.