Early Antimicrobial De-escalation Is Associated with Decreased Acute Kidney Injury in Patients Receiving Empiric Combination Therapy

BACKGROUND: Combination antimicrobial therapy is often initiated for empiric treatment of infections. Recently, vancomycin (VAN) and piperacillin-tazobactam (TZP) therapy has been linked to increases in acute kidney injury (AKI). The objective of this study was to determine whether antimicrobial de-escalation had significant impact on AKI rates. METHODS: Patient information from 7/1/2007 through 10/1/15 was collected from the University of Kentucky Center for Clinical and Translational Sciences Enterprise Data Trust. Patients were included if they received VAN and TZP in combination for ≥48 hours. Patients with history of CKD, who presented with AKI outside the treatment window or at baseline, or who were pregnant were excluded from this analysis. AKI was assessed with the Risk, Injury, Failure, Loss, End stage (RIFLE) criteria. De-escalation was defined as discontinuation of one or both agents. Patients were categorized based on time to de-escalation from treatment initiation with ≤3 days being early, 4 to 7 days being mid, and >7 days being late de-escalation. Basic descriptive statistics were performed in addition to multivariable logistic regression. RESULTS: A total of 7644 patients met all inclusion criteria, with 3664 having early, 2394 with mid, and 1586 with late de-escalation. Patients in the late de-escalation group were slightly older, had higher baseline comorbidity, and were more likely to be exposed to concomitant nephrotoxins than patients in the early and mid de-escalation groups. AKI was more common in the late group (42.7% vs. 23.0% and 13.9% for mid and early groups, respectively; P <0.00001). After controlling confounders, late de-escalation had significantly higher odds of AKI occurring (adjusted odds ratio [aOR]=3.3 [95% CI 2.84–3.83]) compared with patients who had early antimicrobial de-escalation. Additionally, patients in the mid de-escalation group were more likely to experience AKI compared with those in the early group (aOR 1.71 [95% CI 1.49–1.96]). CONCLUSION: Early antimicrobial de-escalation was associated with significant decreases in AKI incidence in patients receiving empiric VAN and TZP therapy. Controlled studies of early antimicrobial de-escalation and AKI are warranted. DISCLOSURES: All authors: No reported disclosures.