489Accuracy of Clinicians' Empiric Treatment Choices for Resistant Gram-Negative Uropathogens

Background. Emergence of multi-drug resistant uropathogens (MDRUTI) is making treatment of UTI more challenging. We sought to evaluate the accuracy of empiric therapy for gram-negative MDR UTI and the utility of prior culture data in improving accuracy. Methods. We performed a retrospective review of electronic health record for treatment of MDR UTI from 3 VA facilities from 2010-2013. MDR UTI was defined as a clinician diagnosed infection resulting in therapy for a gram-negative uropathogen resistant to 3 or more classes of antibiotics. Previous culture data and antimicrobial use were captured from inpatient and outpatient settings. Results. Among 111 episodes of MDR UTI treated by clinicians, 58 (52.2%) empiric therapy choices were active against the pathogen. 81 had prior microbiologic data available for a gram negative organism within the last 2 years, with 69% within the last 6 months. 46/81 (56%) patients received empiric therapy concordant with prior available culture data whereas 35/81 (44%) received discordant therapy. When empiric therapy was concordant with prior culture data (even if current MDR UTI was not the same organism), the antibiotic covered the uropathogen in 36/46 (78%) events. If therapy was discordant from previous microbiology, 12/35 (34%) empiric therapy choices were active (OR 6.9; 95% CI 2.6;18.6, p < .001). Empiric therapy was equally as likely to be active when concordant with recent (within 6 months) or remote (2 years) culture data (75.8% vs 84.6%, p = .70 Fishers). Genitourinary (GU)-directed agents (nitrofurantoin or sulfa; OR 18.0, 95% CI 5.4;59.1, p <.001) or broad-spectrum agents (carbapenems, 4gen cephalosporins; OR 33.1, 95% CI 8.6;127.1, p <.001) were more likely to be active than fluoroquinolones and earlier generation cephalosporins. Conclusion: Only half of empiric therapy choices were active against MDR UTI. Choosing an agent concordant with previous susceptibility data significantly increased the chance of activity for the current MDRUTI, even if the previous uropathogen was a different species and the data was remote. Either GU-directed or broad therapy choices were more likely to be active than other regimens. Accuracy of empiric therapy could be improved using these simple rules. Disclosures. K. Gupta, Paratek: Consultant, Consulting fee