The prevalence of isolation of imipenem-resistant Enterobacter (IRE) strains is rising, with potential serious consequences in terms of patients' outcomes and general care. The study objective was to define the various epidemiological aspects of the isolation of these strains in comparison to cases of isolation of imipenem-susceptible Enterobacter (ISE) strains. Molecular analysis of IRE strains included genotyping and defining the presence of carbapenemases. We conducted a matched retrospective case-control study of patients hospitalized from April 2003 to December 2006. Each IRE case was matched with an ISE case by age and source of isolation. A multivariate analysis using conditional logistic regression was performed to compare the two patient groups. There were 33 cases of IRE isolations during the study period. Twenty isolates were analyzed and found to belong to three distinct pulsotypes. Cell extracts of all of these isolates hydrolyzed imipenem. PCR and sequencing revealed that these isolates harbored a KPC-2 gene. In multivariate analysis, a high invasive-device score (P = 0.02) remained a predictor of IRE isolation. The mortality in the IRE group was 33%, compared to 9% among controls. Being an IRE case was significantly associated with increased mortality after controlling for confounders in a multivariate model (odds ratio, 8.3 +/- 8.6; 95% confidence interval, 1.07 to 64; P = 0.043). Resistance to imipenem due to bla(KPC-2) among Enterobacter isolates has occurred in several clones in Tel Aviv, affecting particularly patients with multiple invasive devices compared to ISE controls. IRE infections are associated with increased mortality. Enhanced measures to control the hospital spread of IRE are warranted.