ABSTRACT
OBJECTIVES: Despite the increasing incidence of carbapenem-intermediate or -resistant Enterobacteriaceae (CIRE), risk factors associated with CIRE infections have not been well defined. This study characterizes factors associated with CIRE among two different source populations. METHODS: A case-control study was performed at a tertiary care medical centre between January 2005 and December 2009. Cases were adults with a culture-confirmed Enterobacteriaceae infection with reduced susceptibility to meropenem or ertapenem. The CIRE cases were matched 1:1 to patients from two different control series: (i) those with carbapenem-susceptible Enterobacteriaceae (CSE) infections; and (ii) inpatients residing on the same ward within 30 days of CIRE culture date. Logistic regression was used to identify variables independently associated with CIRE among each source population. Restricted multivariate analyses were performed to determine if covariates predictive of CIRE varied by infecting organism or presence of the bla(KPC) gene. RESULTS: There were 102 cases of CIRE during the study period. The only covariate independently associated with CIRE in all multivariate analyses was the cumulative number of prior antibiotic exposures. Compared with CSE controls, the odds ratios (95% confidence interval) were 1.43 (1.19-1.72), 2.05 (1.70-2.47) and 2.93 (2.43-3.53) for 1, 2 and ≥ 3 antibiotic exposures, respectively. The strength of this association was comparable for the hospitalized control group and analyses stratified by organism and presence of the bla(KPC) gene. CONCLUSIONS: A patient's cumulative antibiotic exposure history is likely to be more important than any one specific exposure when determining the likelihood of developing a CIRE infection.
