Piperacillin-tazobactam vs. carbapenem for treating blood stream infections due to extended spectrum β-lactamase producing bacteria: Systematic review and meta-analysis

BACKGROUND: Infections due to extended-spectrum β-lactamases-producing Enterobacteriaceae (ESBL-PE) pose a major public health threat due to poor outcomes and high mortality rates. Given the lack of randomized trials comparing PTZ to carbapenem in treating infections due to ESBLPE, we aimed to conduct a systematic review and meta-analysis to investigate the impact of PTZ on mortality of patients with ESBLPE bloodstream infections (BSI) compared with carbapenem. METHODS: MEDLINE, EMBASE, Scopus, and the Cochrane library were searched electronically for studies between 1950 and January 15, 2017 that have provided data for mortality and addressed the terms “extended spectrum β-lactamases or ESBL” and “PTZ or β-lactam/ β-lactamase inhibitor” and “carbapenem”. We also searched the reference sections of included studies looking for possible missed pertinent studies. Data extraction regarding study design, characteristics of the population, intervention, comparator, and outcomes was performed. The random-effects meta-analysis was performed with the use of StatsDirect statistical software (Version 3.0.190). RESULTS: Twenty-nine cohort or case–control studies were included and analyzed; 12 evaluated definitive treatment and 17 studied empiric therapy. PTZ was associated with a non-statistically significant higher 30-day mortality than carbapenem [odds ratio (OR) 1.28, 95% CI 0.88–1.86] for ESBLPE BSI treatment (Figure). No statistically significant differences in mortality were found between PTZ and carbapenem administered as definitive (OR 2.46, 95% 0.93–6.54) or empirical (RR 1.12, 95% CI 0.76–1.66) treatment. A subgroup analysis that included 3 studies that reported mortality based on PTZ MIC revealed that PTZ MIC >1/4 but ≤ 4/4 is associated with a non-significantly higher mortality compared with carbapenem with OR 1.33, 95% CI 0.29-6.03. All 17 patients who had a PTZ MIC≤ 0.5/4 survived after they were treated with PTZ, but the difference with carbapenem could not be estimated. CONCLUSION: PTZ was not significantly associated with higher overall 30-day mortality compared with carbapenem in treating EBLPE BSI. It may be considered as alternative treatment, especially if PTZ MIC is ≤ 0.5/4. There is a need for randomized controlled trials to better guide clinical practice and limit the use of carbapenem. DISCLOSURES: All authors: No reported disclosures.