This study describes the characteristics of ST131 isolates among carbapenemase-producing E. coli strains collected globally by 2 research groups during 2008–2013. The Merck Study for Monitoring Antimicrobial Resistance Trends (SMART) (http://www.merck.com/about/featured-stories/infectious_disease.html) started in 2002 and AstraZeneca's global surveillance study of antimicrobial resistance (unpublished data) began in 2012, to monitor global antimicrobial resistance trends among gram-negative bacteria (Technical Appendix). Antimicrobial susceptibilities of different antimicrobial agents (Table 1) were determined by using frozen broth microdilution panels according to 2013 Clinical and Laboratory Standards Institute and European Committee on Antimicrobial Susceptibility Testing guidelines (6). Established PCR and sequencing methods were used to identify β-lactamase genes (7,8) and define O25b:H4, O16:H5 ST131, fimH30 lineage, H30-Rx sublineage (9–11), and virotypes (12).

Overall, 47,843 E. coli isolates were collected and tested for susceptibility; 407 were found to be nonsusceptible to ertapenem or imipenem and were molecularly characterized for β-lactamase genes. A total of 116 of the 407 isolates were positive for NDM, KPC, OXA-48-like, VIM, and IMP types of carbapenemases. Various gene types were identified: 44 (38%) were positive for bla_NDM, 38 (33%) for bla_KPC, 30 (26%) for bla_OXA-48-like, 2 (2%) for bla_VIM-1 and 2 (2%) were positive for bla_IMP (Table 1).

The countries from which the E. coli isolates were obtained are shown in Table 2. The isolates were cultured from intraabdominal specimens (37%), peritoneal fluid (16%), biliary fluid (10%), urine (30%), and from miscellaneous sources such as sputum and tissue (9%).

PCR testing for O25b:H4, O16:H5, and MLST showed that 41/116 (35%) belonged to the sequence type ST131. Antimicrobial susceptibilities, types of β-lactamases, the presence of the fimH30 lineage, and virotypes are shown in Table 1. ST131strains were more likely than non-ST131 strains to be nonsusceptible to ciprofloxacin and to be positive for bla_KPC, the H30 lineage, and virotype C; non-ST131 isolates were more likely to be positive for bla_NDM.

The majority, i.e., 24 (58%), of ST131strains were positive for bla_KPC, 13 (32%) for bla_OXA-48-like, 3 (7%) for bla_NDM-1, and 1 (2%) for bla_IMP-14. ST131 was present in 16 countries spanning 5 continents (Table 2). The distribution of ST131 during 2008–2013 is shown in Table 3.

Various fimH alleles were identified among ST131 isolates: 24 H30 (58%), 3 H41 (7%), 3 H54 (7%), 2 H22 (5%), 2 H27 (5%), and 2 H191 (5%); and 1 each (2%) belonging to H24, H32, H65, and the new fimH alleles H434 and H435. Of the 24 H30 ST131 strains, 19 (79%) belonged to the H30-R sublineage and 5(21%) to the H30-Rx sublineage.

NDM variants were the most common carbapenemase identified and were especially prevalent in E. coli strains from India and Vietnam (Table 2). KPCs, which were the second most common carbapenemase identified, were distributed globally, i.e., in South America, Central America, North America, Europe, the Middle East, and Asia (Table 2). This was unexpected because KPCs have been relatively rarely reported among E. coli (5).

Because of the unprecedented global success of ST131, the presence of carbapenemases had been carefully monitored by molecular epidemiologists but has been limited to case reports from several countries (1). The largest collections of ST131 with carbapenemases were reported from Hangzhou, Zhejiang Province, China (13) and Pittsburgh, Pennsylvania, USA (14). Of note, 24/38 (63%) of E. coli strains with bla_KPC belonged to ST131, as opposed to 3/44 (7%) for NDMs and 13/30 (43%) for OXA-48-like strains. Our results suggest that ST131 is most likely responsible for the global distribution of E. coli with bla_KPC.

The expansion of the H30 lineage and its H30-R and H30-Rx sublineages have contributed substantially to the spread of ST131 E. coli (11,15). In our study, H30-R, which belongs to virotype C, was the most common lineage among ST131 strains (i.e., 58%); it was associated with bla_KPC and was especially prominent during 2012–2013. The increase of the ST131 H30 lineage with bla_KPC during 2012–13 is cause for concern.

E. coli ST131 has received comparatively less attention than other antimicrobial-resistant pathogens. Retrospective molecular surveillance studies have shown that ST131 with bla_CTX-M-15 was rare during the early 2000s, but that an explosive global increase followed during the mid-to-late 2000s (1). The results of this study show a similar scenario with E. coli ST131 and bla_KPC; a low prevalence combined with a global distribution. This study is of special concern because we documented resistance to “last resort” antimicrobial drugs (i.e., carbapenems) in most regions of the world, in a common community and hospital pathogen (i.e., E. coli) among a very successful sequence type (i.e., ST131). We urgently need well-designed epidemiologic and molecular studies to clarify the dynamics of transmission, risk factors, and reservoirs for ST131.

The medical community can ill afford to ignore E. coli ST131strains with carbapenemases. This sequence type poses a major threat to public health because of its worldwide distribution and association with the dominant H30 lineage. This sequence type among E. coli has the potential to cause widespread resistance to carbapenems.