RESUMO
PURPOSE: Febrile neutropenia (FN) is a life-threatening complication of cancer therapy, and initial ineffective therapy is associated with poor outcomes. Piperacillin/tazobactam (PTZ) is a commonly used empiric antibiotic for the treatment of FN, but resistance among Gram-negative pathogens is well described. We conducted a retrospective case-control study to identify risk factors for PTZ-resistant (PTZ-R) Gram-negative isolates. METHODS: Hematology/oncology patients with FN from November 2007 to November 2013 with a positive culture for Gram-negative bacilli were divided into two groups: PTZ-sensitive (PTZ-S) and PTZ-R. A multivariable model using logistic regression was constructed to identify risk factors for PTZ-R. RESULTS: A total of 171 patients were included (25 PTZ-R, 146 PTZ-S), yielding a 14.6 % resistance rate. Thirty-day all-cause mortality was significantly higher in the PTZ-R group (29 vs 11 %, P = 0.024). Multivariable analysis yielded intensive care unit (ICU) status (odds ratio (OR) 20.18; 95 % confidence interval (CI) 1.03-397.35; P = 0.048), antibiotics for > 14 days in the previous 90 days (OR 6.02; CI 1.17-30.93; P = 0.032), and respiratory source (OR 13.65; CI 1.14-163.57; P = 0.039) as significant risk factors for PTZ-R, and the receiver operating characteristic area under the curve of the model was 0.894. Among PTZ-R isolates, 88 % were sensitive to meropenem and 100 % were sensitive to amikacin. CONCLUSIONS: Given the high mortality rates in the PTZ-R group, a risk-factor-guided approach driven by this multivariable model may help identify patients that could benefit from amikacin combination therapy to help optimize empiric therapy in this setting.