Commensal Escherichia coli Isolate Resistant to Eight Classes of Antimicrobial Agents in the United States

Introduction

E scherichia coli is a commensal bacterium commonly found in the intestinal tract of healthy humans and animals. In the United States, E. coli is a common cause of urinary tract infections and sepsis in humans, and accounts for an estimated 36,000 deaths from sepsis (Russo and Johnson, 2003; Johnson et al., 2005; Manges et al., 2007). Antimicrobial resistance among commensal bacteria is an indicator of antimicrobial use in a population and can predict resistance in pathogens (van den Bogaard and Stobberingh, 2000).

Infection with E. coli resistant to antimicrobial agents, including ceftriaxone, may complicate treatment (Russo and Johnson, 2003). Resistance determinants are often encoded on plasmids or other mobile elements such as integrons. Several resistance genes can be encoded on a single multidrug-resistant mobile element, which can allow for the dissemination of these resistance genes to enteric bacteria, including E. coli.

Risk factors for resistant E. coli infections include farm or animal exposure, meat consumption, prior antimicrobial use, and international travel (Sotto et al., 2001; Akwar et al., 2007; Hillier et al., 2007; Manges et al., 2007; Price et al., 2007; Sannes et al., 2008). To increase understanding of community-acquired resistance, the National Antimicrobial Resistance Monitoring System began antimicrobial susceptibility of commensal E. coli in 2004. We report an extensively resistant commensal E. coli that was resistant to all eight classes of antimicrobial agents routinely tested.

Materials and Methods

From 2004 to 2008, collaborators in Maryland and Michigan each cultured 10 stool specimens a month for ceftriaxone-resistant E. coli. Specimens were collected from adults recruited from outpatients who had submitted a stool specimen to a clinical laboratory for bacterial analysis; persons who had been hospitalized in the 2 weeks before stool collection were ineligible. Stools were plated on Eosin Methylene Blue containing 2 μg/mL cefotaxime (Becton Dickinson and Co.). If presumptive E. coli was present, one colony was selected, species-confirmed with API 20E (bioMérieux), and sent to Centers for Disease Control and Prevention (CDC) for susceptibility testing to 15 agents representing 8 Clinical and Laboratory Standards Institute (CLSI) antimicrobial classes (CLSI, 2008) except for ceftiofur defined as minimum inhibitory concentration ≥8 μg/mL (CLSI, 2010). Using National Antimicrobial Resistance Monitoring System quality control standards, minimum inhibitory concentrations were determined by broth microdilution (Sensititre^®; Trek Diagnostics) according to manufacturer's instructions and interpreted using CLSI criteria when available (Folster et al., 2009). Highly resistant isolates were additionally tested for susceptibility to imipenem and cefepime by broth microdilution (Sensititre^®; Trek Diagnostics).

Plasmid DNA was isolated using a modified protocol and the QiaFilter Midi Kit (Qiagen) (Folster et al., 2009). Plasmid DNA was transferred by electroporation to a laboratory E. coli strain (DH10B; Invitrogen). Antimicrobial susceptibility testing was conducted on the transformed E. coli strain, and presence of resistance genes was confirmed by polymerase chain reaction (Ng et al., 1999; Madsen et al., 2000; Winokur et al., 2001; Chen et al., 2004; Levings et al., 2005). Plasmid Inc/rep typing was performed as previously described (Carattoli et al., 2005).

The surveillance study protocol was approved by institutional review boards of CDC and participating sites. Verbal consent was obtained. Interviews were conducted within 60 days after stool collection using a standard questionnaire.

Results

Stool samples were collected from 477 patients during 2004 to 2008; 68 (12.3%) specimens yielded E. coli on 2 μg/mL cefotaxime selective media. At CDC, the isolate from 7 (1.5%) patients was confirmed as ceftriaxone-resistant E. coli. All ceftriaxone-resistant isolates were additionally resistant to ampicillin and ceftiofur; 6 (85.7%) to sulfisoxazole, trimethoprim-sulfamethoxazole, and tetracycline; 4 (57.1%) to streptomycin and chloramphenicol; 3 (42.8%) to nalidixic acid and ciprofloxacin; 2 (28.6%) to amoxicillin-clavulanic acid, cefoxitin, kanamycin, and gentamicin; and 1 (14.3%) to amikacin. One isolate (1.5% [n = 68]) was resistant to all 15 agents from the 8 antimicrobial classes routinely tested (Table 1). Upon additional testing, the highly resistant isolate was susceptible to imipenem and cefepime (Sjolund-Karlsson et al., 2010).

The extensively resistant isolate was from a specimen collected in 2004 from a 50-year-old Michigan woman who denied any antimicrobial use, foreign travel, hospital or nursing home exposure, and contact with farm animals in the 6 months before stool collection. She was not a health-care provider. The reason for her visit to the physician and submission of a stool specimen was to follow-up on a suspicious colonoscopy result. She reported eating meat (beef, chicken, and pork) and eggs within 1 month before stool collection. She had a diagnosis of non-Hodgkin's lymphoma in 1993; more information about the illness was not recorded.

Resistance to seven of the eight CLSI classes of antimicrobial agents was mediated by an incompatibility type A/C plasmid of ∼170 kb in size. The following resistance genes were identified on the plasmid: strA and strB (aminoglycoside [streptomycin]), bla _CMY (penicillins, β-lactam/β-lactamase inhibitor combinations, and cephems), sul1 and sul2 (folate pathway inhibitors [sulfonamides]), floR (phenicols), and tetA (tetracyclines). We were unable to identify genes that could account for resistance to the folate pathway inhibitor, trimethoprim, or the aminoglycosides (amikacin, gentamicin, and kanamycin). Resistance to the eighth class, quinolones, was not transferred, suggesting that it is was not located on this MDR plasmid. The quinolone resistance mechanism has yet to be further characterized.

Discussion

Using selective media, we isolated ceftriaxone-resistant E. coli from 1.5% of participants. Further, one ceftriaxone-resistant E. coli isolate, from an apparently healthy 50-year-old person, was resistant to eight CLSI classes. Most of the resistance genes in the extensively resistant isolate were contained on an incompatibility type A/C plasmid. Incompatibility type A/C plasmids are common among Enterobacteriaceae and can carry multiple resistance determinants, resulting in multidrug resistance (Lindsey et al., 2009). The potential spread of this plasmid or the individual resistance determinants it carries to other bacteria is concerning.

The source of the extensively resistant E. coli is not known, but the strain may have been acquired through food. Studies have demonstrated that the food supply is a common source of E. coli (Danish Integrated Antimicrobial Resistance Monitoring and Research Program, 2008; Dutil et al., 2010). Introduced E. coli strains may persist for months in the intestine and may at a later time appear as a cause of symptomatic infection (Manges et al., 2007). Additional studies are warranted, such as among E. coli isolated from retail meat and humans, to further understand the role of the food supply for resistant E. coli in humans.