202 • CID 2009:49 (15 July) • EDITORIAL COMMENTARY

E D I T O R I A L C O M M E N T A R Y

Resistant Escherichia coli—We Are What We Eat

Peter CollignonInfectious Diseases Unit and Microbiology Department, The Canberra Hospital, and School of Clinical Medicine, Australian National University, Canberra, Australia

(See the article by Johnson et al. on pages 195–201)

Received 30 March 2009; accepted 30 March 2009;electronically published 9 June 2009.

Reprints or correspondence: Prof. Peter Collignon, In-fectious Diseases Unit and Microbiology Department, TheCanberra Hospital, PO Box 11, Woden, ACT 2606, Australia(peter.collignon@act.gov.au).

Clinical Infectious Diseases 2009; 49:202–4� 2009 by the Infectious Diseases Society of America. Allrights reserved.1058-4838/2009/4902-0006$15.00DOI: 10.1086/599831

Escherichia coli is a frequent cause of life-

threatening bloodstream infections [1]

and other common infections, such as uri-

nary tract infections. Antibiotic resistance

rates in E. coli are rapidly rising, especial-

ly with regard to fluoroquinolones and

third- and fourth-generation cephalospo-

rins. Surprisingly, most of these multi-

drug-resistant strains are acquired in the

community rather than in healthcare set-

tings [2, 3].

Drug-resistant E. coli are readily ac-

quired via the diet (food and water), and

there is a major turnover of drug-resistant

E. coli each day [4]. When people eat ster-

ile food, there is a rapid and substantial

fall in the numbers of drug-resistant E. coli

these people carry [4].

What remains unclear is where the

drug-resistant E. coli in our food are com-

ing from. Are they mainly human strains

that contaminate our food (and water), or

are these strains mainly derived from food

animals?

The study by Johnson et al [5] in this

issue helps to answer this important ques-

tion. They analyzed 287 E. coli isolates re-

covered from meats for a large number of

virulence factors and resistance markers.

They showed that drug-resistant isolates

had characteristics very similar to those of

susceptible isolates recovered from the

same types of meat but quite different

from those of isolates found in other types

of meat. Put another way, the drug-resis-

tant isolates found on retail poultry prod-

ucts are very similar to susceptible isolates

found on retail poultry products, com-

pared with drug-resistant or susceptible

isolates of E. coli recovered from pork or

beef. Thus, it is very likely that most drug-

resistant isolates found on poultry meat,

for example, are the result of antibiotic use

in poultry (rather than the result of the

introduction of strains from people or

other animals into poultry flocks or the

cross-contamination of poultry meat after

slaughter). Although this fact may seem

obvious, it has not been readily accepted

by many people associated with the agri-

culture and pharmaceutical sectors.

There are limitations in this study. The

authors only looked at resistance to a few

antibiotics and relatively small numbers of

isolates (especially for some food). Gro-

cery stores in only 4 United States geo-

graphical sites were sampled. However, it

is hard to see how an expansion of their

study to many more sites and an increased

sample size would have altered their high-

ly convincing and statistically significant

findings. The applicability of these find-

ings to the E. coli found on food from

regions or countries where human fecal

material may frequently contaminate food

or water will, however, need further study.

Some people argue that antibiotic-re-

sistant strains that develop in food animals

are largely irrelevant to human health be-

cause E. coli strains are relatively species-

specific and so will not cause disease in

people. This current study [5] shows that

argument is flawed. A large proportion of

isolates in these meats possessed virulence

factors associated with causing disease in

people. Even if some E. coli animal strains

are relatively species-specific, the antibi-

otic resistance genes carried in animal

strains are likely to be much less host-

specific and thus transferrable to bacteria

carried more frequently by people. We

carry large numbers of E. coli in the bowel

(more than 1 � 106 organisms per gram

of feces) [4]. Johnson et al [6] have shown

elsewhere that the majority of drug-resis-

tant E. coli carried by people are likely

derived from food animals (especially

poultry). Indeed, susceptible strains car-

ried by people are very different in char-

acteristics to the drug-resistant E. coli

strains they carry. Indeed, the latter are

much more similar to susceptible poultry

strains than to susceptible human strains.

What does this mean? It is very likely

that a large proportion of the drug-resis-

tant strains of E. coli carried by people are

acquired via food and especially food an-

imals. We know that highly drug-resistant

strains of E. coli (resistant to fluoroquin-

olones, to extended-spectrum b-lactams,

etc) are spreading into the community

at NA

TIO

NA

L SU

N Y

AT

-SEN

UN

IVE

RSIT

Y on A

ugust 24, 2014http://cid.oxfordjournals.org/

Dow

nloaded from

EDITORIAL COMMENTARY • CID 2009:49 (15 July) • 203

[2,3]. The food chain appears to be the

most likely explanation for this spread.

There is widespread use of antibiotics

in food animals (often for inappropriate

practices such as growth promotion or

ongoing mass prophylactic medication).

In many countries, there is widespread

use of third- and fourth-generation ceph-

alosporins (ceftiofur and cefquinome)

and fluoroquinolones (enrofloxacin) in

food animals. Although many countries,

especially some developing countries, use

more of these critically important anti-

biotics than the United States does, we

know that large numbers of meat chick-

ens produced in the United States and

Canada are exposed to third-generation

cephalosporins [7,8]. Recently, when the

US Food and Drug Administration tried

to limit the use of ceftiofur [7], especially

for mass medication and off-label use in

poultry, lobbying by agriculture and

pharmaceutical interests resulted in this

decision being deferred (although, one

hopes, not reversed).

There are arguments that sections of ag-

riculture can’t survive without the use of

these types of broad-spectrum antibiotics.

However, this argument appears falla-

cious. Fluoroquinolones have been with-

drawn from poultry use in the United

States without major production prob-

lems. In Australia, fluoroquinolone use in

food animals was never approved, yet

there remain viable export meat industries

and expanding meat production. There

are almost no fluoroquinolone-resistant E.

coli isolates in food animals or locally pro-

duced retail meats in Australia [9, 10].

More importantly, very little fluoroquin-

olone resistance (!5%) is found in Aus-

tralia in E. coli isolates recovered from peo-

ple, even from people who are hospitalized

[1, 11], despite fluoroquinolone use in

people and pets for 120 years. Thus, it is

very likely that the low level of fluoro-

quinolone resistance in E. coli isolates re-

covered from people in Australia is related

in major part to the lack of use of fluo-

roquinolones in food animals. This rela-

tionship has implications on what to do

to decrease the antibiotic resistance bur-

den elsewhere.

Although we worry about the rapidly

increasing number of extended-spectrum

b-lactamase–producing and fluoroquino-

lone-resistant E. coli in developed coun-

tries, the problem is much worse in de-

veloping countries [1,2], where resources,

controls, and surveillance are often lack-

ing. Huge numbers of people who live in

developing countries are infected with

multidrug-resistant E. coli that may be ef-

fectively untreatable. This is also an issue

for travelers. Travel is a major risk factor

for acquisition of extended-spectrum b-

lactamase–producing E. coli and other an-

tibiotic-resistant strains of E. coli [2]. This

is also an issue for food imported from

many countries, because these can fre-

quently have extended-spectrum b-lacta-

mase–producing and fluoroquinolone-re-

sistant strains of E. coli present [12].

The drug-resistant strains that people

in the community carry will have an effect

on the types and numbers of drug-resis-

tant bacteria seen in healthcare facilities.

The drug resistance seen in bacteria re-

covered from patients in healthcare set-

tings is normally attributed to the often

poor infection control and antibiotic prac-

tices of the hospital. However, when drug-

resistant strains are carried widely in the

community and these people are exposed

to antibiotics (eg, to treat pneumonia or

to provide surgical prophylaxis), then any

drug-resistant strains they carry have the

opportunity to proliferate and spread in

healthcare settings.

We need to stop using fluoroquinolones

and third- or fourth-generation cephalo-

sporins in food animals. This seems to be

most important in poultry [6], where

there is often mass medication by adding

antibiotics to feed or water and by using

automated injections into the eggs of meat

chickens just before they hatch. At the very

least, as an immediate measure worldwide,

we need significant decreases in the usage

of these types of antibiotics, and we can

achieve these decreases by banning all off-

label use in food animals (eg, as proposed

by the US Food and Drug Administration

for ceftiofur).

Antibiotic resistance in E. coli strains

carried and acquired in the community is

rapidly rising, especially resistance to crit-

ically important antibiotics. We need to

do all we can to stop these strains from

developing and spreading further. We

need to stop using antibiotics such as

third-generation cephalosporins and flu-

oroquinolones in food animals.

Acknowledgments

Potential conflicts of interest. P.C.: noconflicts.

References

1. Kennedy KJ, Roberts JL, Collignon PJ. Esch-erichia coli bacteraemia in Canberra: incidenceand clinical features. Med J Aust 2008; 188:209–13.

2. Laupland KB, Church DL, Vidakovich J, Mu-censki M, Pitout JD. Community-onset ex-tended-spectrum b-lactamase (ESBL)–pro-ducing Escherichia coli: importance of in-ternational travel. J Infect 2008; 57:441–8.

3. Mesa RJ, Blanc V, Blanch AR, et al. Extended-spectrum b-lactamase–producing Enterobac-teriaceae in different environments (humans,food, animal farms, and sewage). J AntimicrobChemother 2006; 58:211–5.

4. Corpet DE. Antibiotic resistance from food.N Engl J Med 1988; 318:1206–7.

5. Johnson JR, McCabe JS, White DG, JohnstonB, Kuskowski MA, McDermott P. Molecularanalysis of Escherichia coli from retail meats(2002–2004) from the United States NationalAntimicrobial Resistance Monitoring System.Clin Infect Dis 2009; 49:195–201 (in thisissue).

6. Johnson JR, Sannes MR, Croy C, et al. An-timicrobial drug-resistant Escherichia colifrom humans and poultry products, Minne-sota and Wisconsin, 2002–2004. Emerg In-fect Dis 2007; 13:838–46.

7. US Department of Health and Human Ser-vices, Food and Drug Administration, Centerfor Veterinary Medicine. New animal drugs;cephalosporin drugs; extralabel animal druguse; order of prohibition. 21 CFR part 530.24 June 2008. Effective 1 October 2008. Avail-able at:http://edocket.access.gpo.gov/2008/E8-15052.htm. Accessed 24 March 2009.

8. Canadian Integrated Program for Antimicro-bial Resistance Surveillance. Salmonella Hei-delberg ceftiofur-related resistance in humanand retail chicken isolates. Modified 16 Au-gust 2007. Available at: http://www.phac-aspc.gc.ca/cipars-picra/heidelberg/heidelberg-eng.php. Accessed 29 March 2009.

9. Australian Government Department of Ag-riculture, Fisheries and Forestry (DAFF).

at NA

TIO

NA

L SU

N Y

AT

-SEN

UN

IVE

RSIT

Y on A

ugust 24, 2014http://cid.oxfordjournals.org/

Dow

nloaded from

204 • CID 2009:49 (15 July) • EDITORIAL COMMENTARY

Antimicrobial resistance—DAFF. Reviewed15 January 2009. Available at: http://www.daff.gov.au/agriculture-food/food/regulation-safety/antimicrobial-resistance Accessed 29March 2009.

10. Barlow R, Gobius K. Pilot survey for an-timicrobial-resistant (AMR) bacteria inAustralian food, prepared for the Austra-lian Government Department of Health

and Ageing. Modified 19 January 2009.Available at: http://www.health.gov.au/in-ternet/main/publishing.nsf/Content/foodsecretariat-standing-priority-list. Accessed24 March 2009.

11. Pearson J, Turnidge J, Franklin C, Bell J; Aus-tralian Group on Antimicrobial Resistance.Prevalence of antimicrobial resistances incommon pathogenic Enterobacteriaceae in

Australia, 2004: report from the AustralianGroup on Antimicrobial Resistance. Com-mun Dis Intell 2007; 31:106–12.

12. Warren RE, Ensor VM, O’Neill P, et al. Im-ported chicken meat as a potential source ofquinolone-resistant Escherichia coli producingextended-spectrum b-lactamases in the UK. JAntimicrob Chemother 2008; 61:504–8.

at NA

TIO

NA

L SU

N Y

AT

-SEN

UN

IVE

RSIT

Y on A

ugust 24, 2014http://cid.oxfordjournals.org/

Dow

nloaded from