The rise of CTX-M extended-spectrum beta-lactamases (ESBLs) in Escherichia coli has been the major resistance shift of the past decade, affecting many countries including the UK. We have learnt much on (i) the enzymes themselves - predominantly CTX-M-2 in South America, CTX-M-9 in East Asia and CTX-M-15 in the UK, Europe, India and the Middle East - (ii) the clonality of producers and (iii) their other resistance mechanisms, pathogenicity factors and plasmids. Much, though, remain elusive.
Where did producers originate? CTX-M-15, now the commonest ESBL in much of the world, was first recorded from India in 1999 (Karim et al., FEMS Microbiol Lett. 2001; 201:237-41), and the earliest UK producer isolates now known date from the same year. CTX-M-15 is a ceftazidime-hydrolysing derivative of CTX-M-3, a genetic 'escape' from Kluyvera ascorbata. A plasmid encoding CTX-M-3 enzyme had disseminated in Poland before CTX-M-15 did so in the UK and Western Europe, but, any inference that the resistance came with extensive East European migration seems ill-founded since the genetic organisation around blaCTX-M-15 on most UK plasmids differs from that around blaCTX-M-3 in Polish plasmids.
CTX-M-15-encoding plasmids are often carried by epidemic E. coli strains, including multiple PFGE variants of ST131, a serotype O25 clone that that has become prevalent as far afield as Europe, India, Canada, the Middle East and South Korea. Not all producers belong to this clone, though, and there has also been 'epidemic' spread of plasmids encoding CTX-M-15 enzyme among strains of E. coli and K. pneumoniae. In contrast to the association of CTX-M-15 and ST131 the Far Eastern and Spanish E. coli strains with CTX-M-9 and -14 enzymes are essentially non-clonal.
There are big outstanding questions, too, about the ecology of E. coli strains with CTX-M enzymes. Many, but not all, infected patients have recent hospital contact or antibiotics and most are elderly, with underlying healthcare problems. It is widely agreed that gut colonization generally precedes E. coli infection, which most often involves the urinary tract, but it is unclear whether most gut colonization occurs in healthcare settings or in the community. Moreover, it is unclear if most transmission oro-faecal, from human to human, or whether the food chain is the direct vector. Finally, it is not clear whether the present ecology is stable, or just one step on a road that ends to CTX-M ESBLs becoming widespread in community E. coli from simple cystitis?
We can't answer these questions fully, but we can shed some light. In Belfast (Northern Ireland) faecal sampling showed that 40% of residents at 18 nursing homes had gut colonization with ESBL-producing E. coli, 90% of them carrying ST131 and half of these "strain" A, an ST131 variant commonly seen in infections in the city (Loughrey et al., Poster 1348, ECCMID 2007). These carriage rates use much higher than in the general community, suggesting transmission within the institutions, presumably via the oro-faecal route, allowing that food supplies cannot be radically different from those reaching the rest of the city. In addition, work by Rod Warren and colleagues (J Antimicrob Chemother 2008; 61, 504-8) examined E. coli isolates from raw chicken breasts bought at supermarkets in the UK's West Midlands: they found that only 1/69 samples with a UK origin had quinolone-resistant, ESBL-positive E. coli, compared with 10/16 of known overseas samples and 6/40 samples of untraceable origin. No isolate had CTX-M-15+ or ST131 isolates. Such data indicate that food has the potential to be a source of E. coli with CTX-M ESBLs, but they reveal no direct link to the current clinical problem. These results contrast somewhat with those from Japan, Spain, and Hong Kong, where the prevalent ESBL types in food and human infections are identical but where strain tracing is difficult because, even from human infections, the organisms are largely non clonal (Kojima et al., Antimicrob Agents Chemother 2005; 49:3533-7; Blanc et al., Vet Microbiol 2006; 118:299-304; Duan et al., Microbial Drug Resistance 2006; 12:145-8).
In short, human-to -human oro-faecal transmission seems the most likely vector of ESBL-producing E. coli; nevertheless wider food sampling and population gut carriage studies are needed. At a more molecular level analysis of the regions flanking the blaCTX-M genes will shed light on the origins and evolution of the plasmids that now carry these enzymes.