627 Prevalence of and Risk Factors for Carriage of Carbapenem-Resistant Enterobacteriaceae (CRE) in the Endemic Setting

Sunday, April 3, 2011: 2:15 PM
Coronado A (Hilton Anatole)
Mahesh Swaminathan, MD , Mount Sinai School of Medicine, New York, NY
Saarika Sharma, MD , New York University Medical Center, New York, NY
Stephanie Poliansky Blash, MPH , Mount Sinai School of Medicine, New York, NY
Gopi Patel, MD , Mount Sinai School of Medicine, New York, NY
Michael Phillips, MD , New York University Langone Medical Center, New York, NY
Vincent Labombardi, PhD , Mount Sinai School of Medicine, New York, NY
Karen F. Anderson, BS , Centers for Disease Control and Prevention, Atlanta, GA
Brandon Kitchel, MS , Centers for Disease Control and Prevention, Atlanta, GA
Arjun Srinivasan, MD , Centers for Disease Control and Prevention, Atlanta, GA
David P. Calfee, MD, MS , Weill Cornell Medical College, New York, NY
Background: CRE are rapidly emerging as an important cause of health care-associated infections. Retrospective studies of patients with clinical CRE isolates and patients involved in CRE outbreaks have identified several risk factors for CRE carriage and/or infection.  These findings may not, however, represent the epidemiology of CRE carriage in the endemic setting.

Objective: To describe the epidemiology of CRE carriage in high-risk hospital patients in an endemic area.

Methods: Active surveillance for CRE was performed on patients admitted to acute rehabilitation, intensive care and general medical units at two New York City hospitals from 2/1/09 to 1/31/10.   Peri-rectal swabs were collected at admission and then weekly.  A case-control study was performed to identify factors associated with CRE carriage.  Cases were those in whom CRE carriage was newly identified by surveillance or clinical culture.  Controls had no microbiologic evidence of CRE and at least one negative surveillance test.   Cases and controls were matched by sex and restricted based on hospital unit.  Clinical data were abstracted from the medical record. 

Results: New CRE carriage was identified in 175 (3.1%) of 5,676 patients of which 106 (61%) were male.   143 (81.7%) were identified by surveillance and 32 (18.3%) by clinical culture.  Carbapenemase production was confirmed by modified Hodge test in 92% of 143 tested isolates.  The most common CRE species was K. pneumoniae (148 cases, 84.6%).  PFGE revealed that most K. pneumoniae isolates were genetically similar to one another and to isolates from other parts of the US. In univariable analysis, factors associated with CRE carriage included: length of hospital and ICU stay (prior to positive culture for cases); dialysis; central lines, urinary catheters, and mechanical ventilation; and hospitalization, surgical procedures, and exposure to antibiotics within 6 months.  The total number of days of antibiotics and several specific classes of antibiotics were associated with CRE carriage.  In multivariable analysis, pulmonary disease (OR = 2.8, 95% CI 1.2-6.8, p = .018), mechanical ventilation (OR = 4.8, 95% CI 1.6-13.9, p = .004) and days of antibiotic exposure (OR = 1.025/day, 95% CI 1.01-1.04, p< .001) were independently associated with CRE carriage.  More cases than controls died but this did not reach statistical significance (21.1% vs. 13.7%, p = 0.067).

Conclusions: Factors associated with carriage of CRE in the endemic setting include chronic pulmonary disease, invasive mechanical ventilation (which is likely a marker of severity of illness), and the duration of exposure to antibiotics. Of these three factors, antibiotic exposure is the most amenable to modification.  Thus, aggressive antimicrobial stewardship activities in addition to routine infection control measures may provide the most effective approach to preventing acquisition of these important pathogens.