To the Editor: Several clones of extended-spectrum β-lactamase (ESBL)–producing extraintestinal pathogenic Escherichia coli (ExPEC) have globally expanded their distribution, including multilocus sequence types (MLSTs) ST38, ST131, ST405, and ST648 (1). ExPEC infections often originate from the patient’s own intestinal flora, although the degree of overlap between diarrheagenic E. coli and ExPEC pathotypes is unclear. Relatively little is known about antimicrobial drug resistance in the most common diarrheagenic E. coli groups, including enteroaggregative E. coli (EAEC), and bacterial gastroenteritis is generally managed without use of antimicrobial drugs.

The ability of diarrheagenic E. coli to cause extraintestinal infections has been shown in previous studies: a study among children in Nigeria linked EAEC to uropathogenic clonal group A (2), and a study in Brazil showed that EAEC markers were present in 7.1% of the E. coli isolates from urinary tract infections (3). Neither of these studies identified clonal lineages of EAEC specifically associated with extraintestinal infections.

We conducted this study to establish the presence and characteristics of ESBL-producing EAEC in a well-defined collection of ESBL-producing isolates (4). The isolates were from human and animal sources in Germany, the Netherlands, and the United Kingdom. The study was conducted at Public Health England during January–April 2013.

DNA from 359 ESBL isolates (4) was screened for the presence of the EAEC transport regulator gene (aggR), located on the EAEC plasmid, by using a real-time PCR assay and the following primers and probe: AggR_F 5′-CCATTTATCGCAATCAGATTAA-3′ AggR_R 5′-CAAGCATCTACTTTTGATATTCC-3′, AggR_P Cy5-CAGCGATACATTAAGACGCCTAAAGGA-BHQ. The amplification parameters were 50°C for 2 min, 95°C for 2 min, and 40 cycles at 95°C for 10 s and at 60°C for 20 s. Isolates positive for aggR were confirmed to be E. coli by using the Omnilog GenIII MicroPlate (Biolog, Hayward, CA, USA). Serotyping was done by using standard methods (5).

The phylogroup was determined for each isolate, and isolates were then assigned to 1 of the 4 major E. coli groups: A, B1, B2, and D (6). A microarray was used to detect ESBL genes, such as bla_CTX-M, at the group level, as previously described (4). The antimicrobial drug susceptibilities of EAEC isolates were determined by using the agar incorporation method, as described in the British Society for Antimicrobial Chemotherapy guidelines (7).

Virulence factors associated with intestinal and extraintestinal infection (8) and with EAEC were investigated as previously described (9). We assigned a virulence score (total number of virulence factor genes detected; maximum possible score 22) and a resistance score (total number of drug classes; maximum score 11) to each isolate.

We isolated 11 EAEC from humans. Eight of the EAEC were isolated from urine specimens, and 1 was isolated from a blood culture; 63% belonged to phylogroup D (Table). EAEC ST38, the most common (55%) ST, was significantly associated with extraintestinal sites in the subset of 140 human isolates (Fisher exact test, p<0.0001).

In this study, we identified multidrug-resistant EAEC isolates belonging to ST38; the isolates had various somatic antigens and bla_CTX-M genes (Table). The multiple somatic antigens, variety of antimicrobial drug–resistance scores, and variety of gene complements in this successful ST indicate multiple acquisitions of virulence markers, rather than clonal expansion from a single source (Table; Technical Appendix Figure).

In the MLST public database, which contained 5,143 E. coli entries in June 2013, ST38 is predominantly associated with urinary tract infections, but in-house MLST studies at the Gastrointestinal Bacteria Reference Unit, Public Health England, have shown that ST38 is a successful EAEC group. The presence of EAEC virulence factors, such as aggregative adherence fimbria AAFI and aggR, can mediate adherence of E. coli to bladder epithelial cells, but the virulence factors do not impart uropathogenic properties to all EAEC isolates (10). The ST38 strain described here probably originated from the gut and independently acquired the 2 phenotypes (uropathogenic E. coli [UPEC] and EAEC), which would suggest the emergence of a UPEC/EAEC hybrid strain. It seems likely that an ST38 E. coli strain adapted to EAEC plasmid carriage (a change that would help survival in the gut through increased adherence) has acquired UPEC virulence factors, facilitating the exploitation of an extraintestinal niche, the urinary tract.

Despite the characterization of numerous virulence factors, no single genetic feature currently defines EAEC or UPEC isolates. Because the EAEC ST38 strain had 4–7 ExPEC-associated virulence factors, we suggest that, on the basis of epidemiologic, microbiological, and molecular characteristics, the EAEC ST38 described in this study should be considered an ExPEC associated with uropathogenic infections. It is possible that the multidrug-resistant EAEC ExPEC group has expanded globally but is currently underreported. We therefore urge testing for the EAEC genotype in all clinical studies of E. coli pathotypes.

Our findings show the potential for EAEC, previously considered a gut pathogen, to cause extraintestinal infection. We suggest that the UPEC/EAEC pathotype may be an evolving clonal group. In particular, a single sequence type, ST38, was associated with multidrug resistance and with urinary tract infection in humans.