Abstract
Background: Extended-spectrum β-lactamase-producing Escherichia coli isolates (ESBL-E coli) cause more than 5000 cases of bacteraemias annually in the UK. The contribution of the food chain to these infections is debated. We aimed to identify the most important reservoirs of ESBL-E coli that colonise and infect humans to identify strategic intervention points. Methods: Sampling for ESBL-E coli was done between Aug 1, 2013, and Dec 15, 2014. We used selective media to seek ESBL-E coli in routinely submitted samples from human faeces, and prospectively collected samples from sewage, farm slurry, and retail foodstuffs in London, East Anglia, northwest England, Scotland, and Wales. We sequenced recovered isolates and compared these isolates with 293 bloodstream and 83 veterinary surveillance ESBL-E coli isolates from the same regions. Findings: 2157 (11%) of 20 243 human faeces samples contained ESBL-E coli, including 678 (17%) of 3995 in London. ESBL-E coli also were frequent in sewage and retail chicken (104 [65%] of 159 meat samples), but were rare in other meats and absent from plant-based foods (0 of 400 fruit and vegetable samples). Sequence type (ST) 131 dominated among ESBL-E coli from human blood (188 [64%] of 293 isolates), faeces (128 [36%] of 360), and sewage (14 [22%] of 65) with STs 38 and 648 also widespread; CTX-M-15 was the predominant ESBL in these lineages (319 [77%] of 416). By contrast, STs 602, 23, and 117—mostly with CTX-M-1 ESBL—dominated among food and veterinary isolates (68 [31%] of 218), with only two ST131 organisms recovered. ST10 occurred in both animals and humans, being frequent in surveillance bovines (11 [22%] of 51 cattle) and representing 15 (4%) of 360 human faecal isolates (but only three [1%] of 293 from bacteraemias); however, both human and animal ST10 isolates were diverse in serotype. Interpretation: Most human bacteraemias with ESBL-E coli in the UK involve internationally prevalent human-associated STs, particularly ST131; non-human reservoirs made little contribution to invasive human disease. Any interventions that seek to target food or livestock can affect the numbers of human infections caused by ESBL-E coli; prevention of the spread of resistant lineages among humans is more vital. Funding: NIHR Policy Research.
| Original language | English |
|---|---|
| Pages (from-to) | 1325-1335 |
| Number of pages | 11 |
| Journal | The Lancet Infectious Diseases |
| Volume | 19 |
| Issue number | 12 |
| DOIs | |
| Publication status | Published - Dec 2019 |
Bibliographical note
Funding Information:We compared ESBL- E coli from human bacteraemias with those from human faeces, sewage, food, slurry, and animals across five regions in the UK. Bloodstream isolates followed expected patterns; they were mostly found in older patients with community-associated infection of genitourinary or gastrointestinal origin. 2 Faecal ESBL- E coli were often linked to foreign travel (particularly to south or southeast Asia) or previous use of antibiotics, which is consistent with the literature. 20,21 Greater contamination of chicken than other meats concurs with previous findings. 22 Typing and ESBL results showed commonality between human bloodstream ESBL- ). ST117 was widely found in isolates from both bovines and chickens. Little contamination was seen for foodstuffs other than chicken. E coli and those from faeces and sewage, with STs 131 (especially), 38, and 648 prominent in all these sources, largely with CTX-M-15 enzyme. There was also commonality between the lineages from surveillance chickens and chicken meat, with STs 23 and 602 dominating, often with CTX-M-1 ESBL, and between cattle and their slurry, where ST10 (with CTX-M-14 or CTX-M-15) dominated. There was little crossover between types from humans, chickens, and bovines, with only serotype diverse ST10 among the top ten most common types from humans, animals, and meat ( figure Our findings do not support the assertion that invasive ESBL- E coli are disseminating via the food chain. Rather, they suggest that host-adapted ESBL- E coli lineages are circulating, with infrequent interspecies transmission. This conclusion agrees with most studies included in a 2015 meta-analysis. 10 ST131, which dominated among human-related isolates, is well known and often multidrug resistant. 6,23 Although ST131 occasionally occurs in food animals (as was seen in two instances in our analysis), the animal ST131 clades are generally different. 24 At the upper edge of the reported prevalence range, Johnson and colleagues 25 in the USA found five of 25 ESBL- E coli from chickens or chicken meat belonged to ST131. By contrast, we—and a previous investigation covering the UK, Germany, and the Netherlands 22 —found only occasional ST131 isolates from food and animals. This rarity is supported by a major review, 6 cataloguing many individual detections of ST131 from food or food animals, but no dissemination. Other common types from bacteraemia—ST38 and ST648, each accounting for about 5% of cases versus 64% for ST131—were absent from food or animals. ST38 (with CMY-2, rather than ESBLs) has been found in poultry, humans, and wildlife; 26 ST648 is also largely reported from humans, although carriage was seen in horses and dogs. 27 Among the major meat and animal types, ST23 was reported from an outbreak in a French hospital, 28 with various further one-off reports but, as we report here, is mostly found in poultry, 29 as is ST117, 23 which has spread in Nordic broiler production. 30 ST602, although common in our study, has not been widely reported in previous studies. 22 ST10, as the sole lineage to appear in the top ten of both human bloodstream and meat-associated groups has been repeatedly noted by other studies 22 in both animals and humans. Nonetheless, the serotype diversity seen in our analysis argues against simple direct flows of ST10 along the food chain. Our results are consistent with those of a comparison of ESBL- E coli from human bacteraemias and livestock in the east of England, one of the regions we surveyed, which also found that these isolate groups and their resistance determinants are largely distinct. 31 Rather than the food chain, the human to human oral-faecal route is likely to be the most frequent route of transmission for human-adapted ESBL- E coli . This route would account not only for the strain and enzyme distributions we have summarised, but also the regional variation in gut carriage of ESBL- E coli with higher rates in London than elsewhere, where sampling was solely from the Royal London Hospital, which predominantly serves poor, crowded areas and populations with frequent travel to and from south Asia. A study in the West Midlands, UK, similarly showed that human gut carriage of ESBL- E coli was more prevalent in inner city conurbations (ie, around Birmingham) than in rural Shropshire. 32 We cannot exclude the possibility that some small minority of human infections might have a direct origin from food, nor that local clusters can occur. In Canada, 33,34 near-identical ST131 and ST117 E coli (ESBL-producing or not) have been found in both retail chicken meat and human infections; nevertheless these putative crossovers accounted for only a small minority of all the human and animal E coli collected. Further, we cannot exclude the possibility that some future multidrug resistant E coli lineage from one or more food animal species will also prove adept at colonising and infecting humans. One further caveat remains: we do not know when, where, or how often bla CTX-M genes escaped from Kluyvera spp (where they are endogenous and chromosomal) to mobile DNA, nor the chain of transmission to human-adapted E coli lineages. However, it seems logical that the hazard of such gene escape will multiply with the range of animal species and intestinal microbiotas exposed to selective antibiotics. 35 Our findings suggest that efforts to stop the rise of ESBL- E coli in invasive infections should concentrate upon disrupting oral-faecal transmission by good post-toilet hygiene (eg, in care homes), on prevention of urinary tract infections by good hydration and catheter care, and on prompt effective treatment of preceding urinary tract infections. Vaccines could provide a solution in the future, with promising early results for cystitis in younger women. 36 Efforts to counter the spread of ESBL- E coli in food production seem unlikely to affect greatly the tally of invasive human infections but remain important in ensuring that veterinary infections remain tractable. Contributors MJD, KLH, MT, LR, CT, and CW designed the study. MJD also led the central laboratory processing and sequencing of isolates from all sources. KLH was also overall project manager. DWW led the design, analysis, and co-ordination of the faecal screening programme and managed local aspects of the project in London. MT also managed the project in Wales and led analysis of the sewage data. NE managed all non-meat food sampling and sewage analyses. LR and CT also managed the meat and slurry work, and sourced the veterinary surveillance isolates. PC designed and undertook all statistical analyses and managed the project in northwest England. CW also managed all aspects of the study in Scotland. MD and MJE did the bioinformatic analyses of whole genome sequencing data. NW wrote the original funding application and led the overall project design and co-ordination. DML co-ordinated the project in East Anglia and led the writing and revising of this paper. All authors commented on the draft manuscript and contributed to the final version. Declaration of interests DML reports advisory board or ad-hoc consultancy fees from Accelerate, Allecra, Antabio, BioVersys, Centauri, Entasis, Integra-Holdings, Meiji, Melinta, Menarini, Mutabilis, Nordic, ParaPharm, Pfizer, QPEX, Roche, Shionogi, Taxis, T.A.Z., Tetraphase, VenatoRx, Wockhardt, and Zambon; lecture fees from Accelerate, Astellas, bioMerieux, Beckman Coulter, Cepheid, Correvio, Merck, Menarini, Pfizer, and Nordic; and shares in Dechra, GSK, Merck, Perkin Elmer, Pfizer, and T.A.Z, amounting to less than 10% of portfolio value. NW reports grants from Momentum Bioscience, Tetraphase Pharmaceuticals, Bio-Rad, bioMerieux, Meiji Seika Pharma, Accelerate Diagnostics, Wockhardt, Check-Points, Helperby Therapeutics, Merck Sharpe & Dohme, Roche, VenatoRx, AstraZeneca, Paratek, Shionogi, Neem Bioteck, GlaxoSmithKline, Innovate UK, Kalidex Pharmaceuticals, Melinta, Mobidiag, Rokitan, Trius Therapeutics, and Rabiotics Rx. KLH, on behalf of PubIic Health England's AMRHAI Reference Unit, has received financial support for conference attendance, lectures, research projects, or contracted evaluations from Accelerate Diagnostics, Achaogen, Allecra, Amplex, AstraZeneca UK, AusDiagnostics, Basilea Pharmaceutica, Becton Dickinson Diagnostics, bioMérieux, Bio-Rad Laboratories, The BSAC, Cepheid, Check-Points, Cubist Pharmaceuticals, Enigma Diagnostics, European Centre for Disease Prevention and Control, GlaxoSmithKline, Helperby Therapeutics, Henry Stewart Talks, IHMA Ltd, Innovate UK, Kalidex Pharmaceuticals, Melinta Therapeutics, Merck Sharpe & Dohme, Meiji Seika Pharma, Mobidiag, Momentum Biosciences, Neem Biotech, Nordic Pharma, Norgine Pharmaceuticals, Rempex Pharmaceuticals, Roche, Rokitan, Smith & Nephew, Shionogi, VenatoRx, Wockhardt, and WHO. MJE undertook work as a member of Public Health England's AMRHAI Reference Unit, which has received financial support for conference attendance, lectures, research projects, or contracted evaluations from numerous sources, including Accelerate Diagnostics, Achaogen, Allecra Therapeutics, Amplex, AstraZeneca UK, AusDiagnostics, Basilea Pharmaceutica, Becton Dickinson Diagnostics, bioMérieux, Bio-Rad Laboratories, BSAC, Cepheid, Check-Points BV, Cubist Pharmaceuticals, Department of Health, Enigma Diagnostics, ECDC, Food Standards Agency, GlaxoSmithKline Services Ltd, Helperby Therapeutics, Henry Stewart Talks, IHMA, Innovate UK, Kalidex Pharmaceuticals, Melinta Therapeutics, Merck Sharpe & Dohme, Meiji Seika Pharma, Mobidiag, Momentum Biosciences, Neem Biotech, NIHR, Nordic Pharma, Norgine Pharmaceuticals, Rempex Pharmaceuticals, Roche, Rokitan, Smith & Nephew UK, Shionogi & Co, Trius Therapeutics, VenatoRx Pharmaceuticals, Wockhardt, and WHO. PC and MAT reports grants from the UK Department of Health. All other authors declare no competing interests. Acknowledgments This work is based on independent research commissioned and funded by the National Institute for Health Research (NIHR) Policy Research Programme. The views expressed in the publication are those of the authors and not necessarily those of the UK National Health Service, the NIHR, the Department of Health and Social Care, arms-length bodies, or other government departments. The research team would like to thank staff at the following hospitals, agencies, and companies for assisting in collection of isolates and samples: Barts Health NHS Trust; Addenbrookes Hospital, Cambridge; Norfolk & Norwich University Hospital; Princess Alexandra Hospital, Harlow; Ipswich Hospital; Southend Hospital; Manchester Royal Infirmary; Lancashire Teaching Hospitals NHS Trust; Royal Gwent Hospital; NHS Greater Glasgow and Clyde; Royal Infirmary Edinburgh; PubIic Health England East of England Field Epidemiology Services; PubIic Health England Food, Water, and Environment Laboratory; Public Health Wales; Department of Epidemiological Studies, APHA; United Utilities, Warrington; Thames Water, Reading; and Welsh Water.
