Background: Increasing evidence has suggested that combination antibiotic therapy is more effective than monotherapy and leads to better outcomes.However, results from observational researches were contradictory and did not support the combination antibiotic therapy in this context. Here, a meta-analysis of published studies determined whether a combination of two or more antimicrobial agents reduces mortality in patients with Gram-negative bacteremia. Objective: The aim of the review was to comprehensively summarize the evidence and evaluate the impact of antibiotic combinations or monotherapy on infection caused by carbapenem-resistant Gram-negative bacteria. Method: A comprehensive computerized search was performed through PubMed, Cochrane library, web of science and Embase up to March 15, 2012, to identify published studies. We would include randomized controlled trials, non-randomised studies without consideration of publication status, time of publication, blinding status or language. Case reports, review articles were excluded. Studies were considered available if they compared monotherapy regimens with combination therapy regimens for treating infection due to carbapenem-resistant Gram-negative bacteria. Results: A total of 52 studies were included enrolling 6524 participants, which included 5 randomized controlled trials and 47 non-randomised studies of the effects of interventions. The results showed higher mortality (OR=1.38, 95%CI=1.17-1.64) in monotherapy groups, lower clinical success (OR=0.63, 95%CI=0.45-0.89) and microbiological eradication (OR=0.60, 95%CI=0.43-0.84) in monotherapy for treating carbapenem-resistant Gram-negative bacteremia. Similarly, significant higher mortality (OR=1.58, 95%CI=1.22-2.06), lower clinical success (OR=0.36, 95%CI=0.24-0.53) and microbiological eradication (OR=0.36, 95%CI=0.18-0.72) in monotherapy for treating carbapenem-resistant Enterobacteriaceae. There were only few clinical researches comparing the effect of combination therapy versus monotherapy, most of them being case-control and retrospective cohort studies limited by small study size. Well- argued randomised controlled trials are lacking. Conclusions: Infection patients caused by carbapenem-resistant Gram-negative bacteria may benefit from combination therapy. We suggest choosing combination therapy to treat severe infection due to carbapenem-resistant Gram-negative bacteria before larger and better-devised randomized controlled trials completed. Because of risk and the design of the available studies, well-powered randomized controlled trials are needed.