Prolonged Infusion Antibiotics for Suspected Gram-Negative Infections in the ICU: A Before-After Study

BACKGROUND: β-Lactam antibiotics demonstrate time-dependent killing. Prolonged infusion of these agents is commonly performed to optimize the time the unbound concentration of an antibiotic remains greater than the minimum inhibitory concentration and decrease costs, despite limited evidence suggesting improved clinical results. OBJECTIVE: To determine whether prolonged infusion of β-lactam antibiotics improves outcomes in critically ill patients with suspected gram-negative infection. METHODS: We conducted a single-center, before-after, comparative effectiveness trial between January 2010 and January 2011 in the intensive care units at Barnes-Jewish Hospital, an urban teaching hospital affiliated with the Washington University School of Medicine in St. Louis, MO. Outcomes were compared between patients who received standardized dosing of meropenem, piperacillin-tazobactam, or cefepime as an intermittent infusion over 30 minutes (January 1, 2010, to June 30, 2010) and patients who received prolonged infusion over 3 hours (August 1, 2010, to January 31, 2011). RESULTS: A total of 503 patients (intermittent infusion, n = 242; prolonged infusion, n = 261) treated for gram-negative infection were included in the clinically evaluable population. Approximately 50% of patients in each group received cefepime and 20% received piperacillin-tazobactam. More patients in the intermittent infusion group received meropenem (35.5% vs 24.5%; p = 0.007). Baseline characteristics were similar between groups, with the exception of a greater occurrence of chronic obstructive pulmonary disease (COPD) in the intermittent infusion group. Treatment success rates in the clinically evaluable group were 56.6% for intermittent infusion and 51.0% for prolonged infusion (p = 0.204), and in the microbiologically evaluable population, 55.2% for intermittent infusion and 49.5% for prolonged infusion (p = 0.486). Fourteen-day, 30-day, and inhospital mortality rates in the clinically evaluable population for the intermittent and prolonged infusion groups were 13.2% versus 18.0% (p = 0.141), 23.6% versus 25.7% (p = 0.582), and 19.4% versus 23.0% (p = 0.329). CONCLUSIONS: Routine use of prolonged infusion of time-dependent antibiotics for the empiric treatment of gram-negative bacterial infections offers no advantage over intermittent infusion antibiotic therapy with regard to treatment success, mortality, or hospital length of stay. These results were confirmed after controlling for potentia