Acute Kidney Injury in Patients on Concomitant Vancomycin and Piperacillin/Tazobactam: Are Extended Infusions to Blame?

Background. Empirical combination therapy for hospital-acquired infections often consists of vancomycin (V) and piperacillin/tazobactam (PT). Acute kidney injury (AKI) is a treatment-limiting adverse event with V, and recent evidence has suggested that combination therapy with PT significantly increases this risk. At the Detroit Medical Center (DMC) we noted a rise in perceived V-associated AKI which coincided with two changes: a switch of preferred empiric Gram-negative coverage from cefepime to PT; and a switch from standard infusions (SI) to extended infusions (EI) of β-lactams. We hypothesized that EI PT, as opposed to SI PT, when given in combination with V was the driver of AKI. We therefore developed a retrospective cohort study to investigate the impact of EI PT in combination with V on AKI. Methods. A retrospective cohort analysis was performed at the DMC from December 2011 to December 2013. Patients were eligible for inclusion if they received a minimum of 48 hrs of combination treatment with V and PT started within 24 hours of one other. Patients were excluded if they had creatinine >1.2 mg/dL at the start of therapy or if they were receiving renal replacement therapy. Patients were divided into cases (those receiving EI PT with V) and controls (those receiving SI PT with V). EI cases were matched with SI controls by severity of illness, ICU status, duration of combination therapy, daily dose of V and number of concomitant nephrotoxic agents. The primary outcome was incidence of nephrotoxicity (NTX) according to RIFLE criteria. A propensity score was developed measuring the likelihood of receiving EI as opposed to SI PT. A propensity-adjustment analysis was then performed to determine the impact of EI PT on NTX. Results. Three hundred twenty patients were included in the analysis (160 EI, 160 SI). Mean age for the cohort was 54.8 years, 150 (46.9%) were women, and 204 (63.8%) were African-American. RIFLE defined AKI occurred in 52 (32.5%) of patients receiving EI and 53 (33.1%) of patients receiving SI (p = 0.69). Characteristics of EI and SI patients were very similar. In propensity-adjusted analysis, there was no association between EI PT and AKI (odds ratio = 0.90; 95% CI 0.53–1.53). Conclusion. Among patients receiving concomitant PT and V, PT delivered via EI was associated with a similar risk of NTX as PT delivered by SI. Disclosures. All authors: No reported disclosures.