Reflection paper on the use of third and fourth generation cephalosporins in food producing animals in the European Union: development of resistance and impact on human and animal health


  •   Scientific Advisory Group on Antimicrobials of the Committee for Medicinal Products for Veterinary Use

    Search for more papers by this author
    • *

      Members of the Scientific Advisory Group: C. Greko (Rapporteur), J. I. Badiola, B. Catry, E. van Duijkeren, M. A. Moreno, M. C. Matias Ferreira Pomba, S. Pyörälä, M. Ružauskas, P. Sanders, E. J. Threlfall, F. Ungemach, K.Törneke (Chairman) and J. Torren-Edo (EMEA secretariat). Acknowledgment to former SAGAM members; A. Caprioli, D. Mevius, J. Wallman

  • The information contained in this publication does not necessarily reflect the opinion or the position of the European Medicines Agency.

Christina Greko, Statens Veterinärmedicinska Anstalt, 751 89 Uppsala, Sweden. E-mail:


Resistance to third and fourth generation cephalosporins is rapidly increasing in bacteria infecting humans. Although many of these problems are linked to human to human transmission and to use of antimicrobials in human medicine, the potential role of community reservoirs such as food producing animals needs to be scrutinized. Resistance to third and fourth generation cephalosporins is emerging in enteric bacteria of food producing animals and also in food of animal origin. The genes encoding resistance to these cephalosporins are transferrable and often linked to other resistance genes. Systemic use of third and fourth cephalosporins selects for resistance, but co-selection by other antimicrobials is also likely to influence prevalence of resistance. Although there are many uncertainties, the potential consequences of a further increase of resistance to this critically important class of antimicrobials in bacteria colonising animals are serious. Measures to counter a further increase and spread of resistance among animals should therefore be considered.


Cephalosporins of the third and fourth generation represent subclasses of antimicrobials very important in the treatment of severe and invasive infections in humans; such antimicrobials are therefore of special interest from a public health perspective. The first nosocomial outbreaks of bacteria resisting these cephalosporins by production of beta-lactamases were described in the 1980s (Gniadkowski, 2001). Subsequently, the occurrence of infections with bacteria resistant to third and fourth generation cephalosporins, e.g. Klebsiella pneumoniae, Escherichia coli, Salmonella spp. and Pseudomonas aeruginosa has increased worldwide. Increased resistance implies either delayed adequate treatment or initial use of second- and third-line alternatives. Some of the latter carry a higher risk of adverse reactions (e.g. aminoglycosides) or are clearly more toxic (e.g. colistin). Because of delayed adequate therapy, the burden of infections with bacteria resistant to third and fourth generation cephalosporins can be substantial with severe outcomes, including both higher overall and higher infection-related mortality, increased length of hospital stay and higher costs (Schwaber & Carmeli, 2007).

In Europe, infections with bacteria resistant to third and fourth generation cephalosporins were previously mainly caused by K. pneumoniae, and were mostly diagnosed in specialist units (Livermore et al., 2007). In the last decade, this pattern has changed and resistance is rapidly emerging, not only in hospitals but also in community-acquired infections. Pathogens carrying genes encoding these resistance traits now include E. coli and Salmonella (Anonymous, 2006a; Canton & Coque, 2006; Livermore et al., 2007). Linkage to other resistance genes and co-selection by unrelated antimicrobials are important in the epidemiology of these resistance genes (Canton & Coque, 2006).

This change in epidemiology, with typically nosocomial organisms spreading to and from the community, indicates an exchange of organisms or genes with other, perhaps nonhuman bacterial reservoirs. Resistance has emerged in some countries in both Salmonella and E. coli from food producing animals. This suggests that food of animal origin may be one of the potential vectors for dissemination of transferable beta-lactamase-encoding genes in the community (Livermore et al., 2007; Carattoli, 2008). Third and fourth generation cephalosporins are used for food-producing animals and could potentially influence the prevalence of resistance. In addition, co-selection by other antimicrobials used for medication of large groups of animals (mass medication) may contribute to the occurrence and dissemination of resistance determinants in animals.

The objective of this document was to critically review recent information on the use of cephalosporins of the third and fourth generation in food-producing animals in the EU, its effect on development of resistance to this category of antimicrobial agents in bacterial species that are of importance for human and animal health and the potential impact on human and animal health1.


Mechanism of action, classification and spectrum of activity

The mechanism of antibacterial activity of the cephalosporins and cephamycins is essentially the same as for benzylpenicillin and other beta-lactam antimicrobials; they interfere with the formation of the cell wall by binding to enzymes that are active in the synthesis of peptidoglycans [transpeptidases, also called penicillin-binding proteins (PBPs)]. All true cephalosporins contain a 7-aminocephalosporanic acid molecule, composed of a beta-lactam ring essential for activity and a six-membered dihydrothiazine ring. A wide variety of cephalosporins has been generated by substitutions of various groups at different positions of the nucleus. The cephamycins differ from the true cephalosporins by the presence of a methoxy-group in the position 7 of the cephalosporin nucleus and are stable to many beta-lactamases.

The cephalosporins and cephamycins are grouped together and are classified on basis of their in vitro spectrum of activity, structural similarities and to some extent, the year of introduction. In this document, the term cephalosporins will be used to cover both true cephalosporins and cephamycins, unless specifically indicated. The traditional classification of these molecules into ‘generations’ will be followed and they are grouped according to the Anatomic Therapeutic Chemical (ATC) index of January 2005 (Anonymous, 2005) and ATCvet:

  • • First generation cephalosporins (e.g. cephalexin, cefadroxil, cephalotin) have the narrowest spectrum of activity. They have an excellent activity against Gram-positive cocci, including penicillinase-producing staphylococci but the activity against Gram-negative bacteria is limited.
  • • Second generation cephalosporins (e.g. cephaclor, cefoxitin2, cefuroxime) have an expanded spectrum of activity compared with first generation substances and are generally more active against Gram-negative bacteria.
  • • Third generation cephalosporins (e.g. cefotaxime, ceftiofur, cefoperazone, latamoxef2) generally have a broad spectrum of activity, with increased stability to many of the beta-lactamases that inactivate the earlier generation’s cephalosporins and other beta-lactam antimicrobials.
  • • Fourth generation cephalosporins (e.g. cefepime, cefpirome, cefquinome) have an even more extended activity against Gram-negative bacteria, as they have a further increased stability compared with the third generation compounds.

Use of cephalosporins in human medicine

Cephalosporins are widely used in human medicine, both in hospitals and in the community. In hospitals, third and fourth generation cephalosporins are used to treat, e.g. septicaemia, meningitis, hospital acquired pneumonia, intra-abdominal infections and complicated urinary tract infections (Paterson & Bonomo, 2005). The total hospital consumption of antimicrobials in 15 European countries in 2002 ranged from 1.28 to 3.89 defined daily doses (DDD)/1000 inhabitants per day (Vander Stichele et al., 2006). The proportion of cephalosporin use ranged from 8 to 31% of the total use, and within the group the proportion of third and fourth generation cephalosporins ranged between 10 and 50%. An increase in the hospital use of third and fourth generation cephalosporins between the years 1997 and 2002 was noted for all 15 countries.

In 2003, the total outpatient use of antimicrobials (i.e. for nonhospitalized patients) in 34 European countries ranged from 9.78 to 31.40 DDD/1000 inhabitants per day (Ferech et al., 2006). The cephalosporin use in 25 of these countries ranged from 0.02 to 6.18 DDD/1000 inhabitants per day, which equates a factor of 270 between the highest and lowest using country (Coenen et al., 2006). In many countries, most or almost all of this use was first and second generation cephalosporins. Cephalosporins of higher generations have mainly been available for injection or infusion, and their use has probably mostly been limited to patients in e.g. elderly care with severe community acquired pneumonia or complicated urinary tract infections. Some products for oral use have been introduced on the market. In three countries, use of third and fourth generation cephalosporins was more than 40% of the total outpatient use of cephalosporins in 2003, compared with others with almost no such use at all. This extreme variation is probably explained by inappropriate use of third generation cephalosporins for uncomplicated urinary and respiratory tract infections in some countries (Coenen et al., 2006).

The total hospital consumption of antimicrobials has been estimated to be about 5–10% of the total consumption (Vander Stichele et al., 2006). Hospital exposure is more concentrated in terms of number of patients in the population exposed and the intensity of treatment. This provides a selective pressure that if combined with inadequate infection control and with an inherent accumulation of vulnerable patients, creates conditions for emergence and spread of infections with resistant bacteria within the hospital and eventually to the community.

Cephalosporins for food-producing animals

Cephalosporins authorized for animals in the EU

Cephalosporins have been authorized for use in food-producing animals via national procedure, mutual recognition or centralized procedure. One product with cephalosporins (ceftiofur) is authorized centrally. The number of products containing the various active ingredients for some EU countries plus Iceland, authorized for food producing animals by national procedure or by mutual recognition procedure, is illustrated in Figs 1 (systemic use) & 2 (intramammary use). In addition, cephalexin is authorized for use in water or milk-replacers in at least two Member States and cefapirin and cefquinome in several Member States for intrauterine use.

Figure 1.

 Number of products formulated for injection per antimicrobial substance and Member State (data from 2006).

Figure 2.

 Number of products formulated for intramammary use per antimicrobial substance and Member State (data from 2006).

Maximum residue limits (MRLs) have been established for cattle for all substances shown in Figs 1 & 2, and in addition for sheep and goat for cefazolin, for pig for ceftiofur [the current entry for this substance in the Annexes of Council Regulation (EEC) No 2377/90 includes all mammalian food-producing species] and for pigs and horses for cefquinome. Presently, no MRLs have been established for poultry.

Regarding potential effects of use of cephalosporins on resistance in bacteria, systemic use particularly for groups or flocks of animals, is likely to have the major impact. This is because microbiota in multiple body sites is exposed. This document will therefore focus on potential effects of systemic use. Of the substances authorized for systemic use, cephalexin is a first, ceftiofur a third and cefquinome a fourth generation cephalosporin.

Intramammary use (i.e. local use) leads to a lesser exposure of the normal microbiota of the target animal. It is recognized that cephalosporins are widely used as lactating or dry cow therapy in many Member States (Fig. 2). Such use may have an impact on selection of resistance in target bacteria. Furthermore, if milk from recently treated cows is fed to calves before the withdrawal time has elapsed, the intestinal microbiota of the latter may be exposed. In some countries, these practises might be widespread (Sampimon et al., 2008).

Ceftiofur (free acid) is centrally authorized for subcutaneous administration in pigs with an extended dosage interval (‘long acting’) for treatment of respiratory tract infections, septicaemia and polyarthritis and polyserositis caused by defined pathogens. Ceftiofur hydrochloride (not ‘long acting’) is authorized in most countries for intramuscular administration in cattle and pigs with indications for treatment of respiratory disease, and in cattle also for interdigital necrobacillosis and puerperal metritis. In some Member States, ceftiofur was previously authorized for injection of day-old chickens for prevention of septicaemia (Bertrand et al., 2006). There are currently no cephalosporin-containing products authorized for poultry species in the EU.

Cefquinome (fourth generation) is available in some Member States for systemic use in cattle, pigs and horses. The indications for use are mainly respiratory infections, interdigital necrobacillosis in cattle, septicaemia caused by E. coli in calves and foals, and streptococcal infections in horses. In some countries, indications such as bovine mastitis caused by E. coli, postpartum dysgalactia syndrome in sows [previously ‘MMA’ syndrome (mastitis, metritis and agalactia)] and meningitis and arthritis in piglets are also included. For ceftiofur, formulations of cefquinome for subcutaneous administration with extended dosage intervals are authorized in some Member States.

Use of cephalosporins for animals in the EU

Information on the consumption of antimicrobial agents for food-producing animals is not readily available for most Member States, although the availability of such data is slowly improving. Reported data are mostly compiled for all animal species, including dogs and cats. Further, data are often reported as ‘beta-lactam antimicrobials’, i.e. including also benzylpenicillin and phenoximethylpenicillin, ampicillin, etc. Finally, in the few reports where information on penicillins and cephalosporins are given separately, data are not further divided into generations. It is therefore not possible to compile comparable and relevant data on the use of cephalosporins of different generations in the Member States.

As an example of amounts used, available data on use of all cephalosporins in some Member States are presented in Table 1. Although these figures cannot be regarded as representative for all Member States, some observations can be made. A substantial part (55–98%) of all use of cephalosporins in these countries is for pets. Until recently, only first generation-cephalosporins have been authorized for pets. By contrast, the systemic use for food-producing animals is likely to be dominated by third and fourth generation cephalosporins. In the reports on use of antimicrobials from Denmark and France, data are given per animal species. The amounts for systemic use for pigs were 98 and 1310 kg active substance, respectively, representing 51% and 89% of the total use for food-producing animals.

Table 1.   Sales of cephalosporins (all generations) for systemic use in food-producing animals and pets, and for intramammary use in Denmark, Finland, France and Sweden, expressed as kg active substance (years 2005 or 2006)*
CountrySystemic useIntramammary useTotal
Food-producing animalsPets
  1. *Sources: AFSSA 2005, DANMAP 2006, 2007, FINRES-vet, 2005–2006, SVARM, 2007.


Ceftiofur and cefquinome are both mainly administered parenterally. It has been argued that this would limit their use to special situations. Other factors such as the very broad spectrum, short or zero withdrawal times for milk and the availability of ‘long acting’ formulations for certain indications are factors which could also make these drugs a convenient choice in many situations.

Currently, there is no harmonized approach on prudent use of cephalosporins in animals in different Member States. In some marketing authorizations in the EU, special precautions for use have been added to the Summary of Product Characteristics of cephalosporin products. The guideline of the Federation of Veterinarians of Europe on prudent use is at a general level and states: ‘where an appropriate narrow spectrum agent is available, it should be selected in preference to a broad spectrum agent’ (FVE, 1999). Guidance on prudent use of antimicrobials for animals have been published in many countries (Passantino, 2007) but most are on a general level and cephalosporins are not specifically mentioned.

Some national guidelines give specific recommendations for the use of cephalosporins. For example, in the Dutch guidelines for therapy (Formularia) issued by the Royal Veterinary Association (KNMvD), drugs of eminent importance to public health are considered third choice drugs for treatment of infections in food animals (KNMvD, 2007). Third choice means: only to be used if no alternative therapy is possible, based on susceptibility test of the target pathogens. For individual animals with severe invasive infections, third choice drugs may be used for empiric first choice therapy. According to the German guidelines for prudent use of antimicrobials in veterinary medicine issued by the Federal Veterinary Association and the Working Group of Chief Veterinary Officers (2000), it is mandatory that reserve antimicrobials with last resort character in human medicine are used restrictively in individual animals on a short-term basis and only in cases where they are strictly indicated. The Finnish guidelines are even more specific; recommendations are given for specific indications in different animal species. Third generation cephalosporins are advised, with specific cautions, only for treatment of foal septicaemia (Anonymous, 2003). National legislation in Finland prohibits the use of third and fourth generation cephalosporins for animals unless a veterinary medicinal product containing these substances has a marketing authorization or a special licence. Off-label use of these products is prohibited.

Even when specific guidelines exist their implementation is generally not monitored. In specific and in general guidelines, off-label use is restricted to situations where no other suitable product is available and should be carefully justified (Passantino, 2007). Despite these recommendations, off-label use for nonauthorized indications can be common. For example, in 10 pig farms in Denmark, off-label use of ceftiofur was common (Jorgensen et al., 2007). The authorized indications for the use of ceftiofur in pigs is treatment of respiratory tract infections, septicaemia, polyserositis and polyarthritis, but eight farms used the drug for systemic prophylaxis in newborn piglets and one for treatment of diarrhoea. Information on indications authorized in countries outside EU, such as injection in ovo or of day-old poultry, or claimed advantages such as routine prophylactic use of ceftiofur for weaning pigs is easily available on the internet and this might sometimes influence the veterinarians’ choice of therapy.

Resistance mechanisms and genetics

Resistance to cephalosporins in staphylococci

In staphylococci, resistance to penicillinase-sensitive penicillins (benzylpenicillin, phenoximethylpenicillin and aminopenicillins) is caused by narrow spectrum beta-lactamases (Li et al., 2007). Resistance to all beta-lactams, including the cephalosporins, is caused by the alteration of the PBPs. The altered PBP has a low affinity for beta-lactam antimicrobials. This mechanism is generally referred to as methicillin3 resistance (Li et al., 2007). The gene encoding this mechanism, mecA, is chromosomally located as part of the Staphylococcal Cassette Chromosome (SCCmec). SCCmec is horizontally transferable between staphylococci and is commonly present in some species of coagulase negative Staphylococcus spp. present in humans and in animals.

Resistance to cephalosporins in Enterobacteriaceae

Resistance to cephalosporins in Enterobacteriaceae (e.g. Salmonella, E. coli) is primarily caused by production of beta-lactamases with broad or extended spectrum, e.g. with substrate specificity not only for the penicillins but also for cephalosporins. To date, several hundred variants of beta-lactamases have been distinguished. The enzymes are classified according to different schemes. The most commonly used are those by Bush et al. (1995) based on functional properties, the Ambler’s (1980) system based on structural similarities or a combination of both. For the purpose of this document, a simplified overview of beta-lactamases is provided based on bacterial host and functional differences (Table 2). Unfortunately, some of the enzymes have been given more than one name. A list of common homonyms and a comprehensive list of the origin of the names has recently been published (Jacoby, 2006) (for a complete list see

Table 2.   Main types of beta-lactamases among staphylococci, Enterobacteriaceae, Pseudomonas and Acinetobacter (adapted from Jacoby & Munoz-Price, 2005)
Bacterial speciesBeta-lactamaseExamples Substrate specificity/activity patternSusceptible to clavulanic acid*
  1. *The number of plus signs denotes relative susceptibility of the families to inhibitors. Note that within the generally susceptible ESBL families, inhibitor-resistant variants occur.

Staphylococcus aureusPenicillinasePC-1Benzylpenicillin, aminopenicillins++++
Enterobacteriaceae, Pseudomonas and AcinetobacterBroad-spectrumTEM-1, TEM-2, SHV-1Benzylpenicillin, amino-penicillin, ureido-penicillins, carboxy-penicillins, 1st generation cephalosporins (e.g. cephalothin).+++
OXA family (some)As for the broad spectrum group plus penicillinase stable penicillins such as oxacillin+
Extended SpectrumTEM family and SHV-familySubstrates of the broad-spectrum group plus oxiamino-cephalosporins (cefotaxime, ceftazidime, ceftiofur) and monobactams++++
Others (BES, GES/IBC family, PER, VEB, TLA, SFO)As for TEM-family and SHV-family++++
CTX-M familySubstrates of the expanded-spectrum group plus, for some enzymes, cefepime++++
OXA-family (some)Same as for CTX-M family+
AmpCACC, ACT, CMY, DHA, FOX-family, LAT-family, MIR, MOXSubstrates of the extended spectrum group except 4th generation cephalosporins, but in addition cephamycins (e.g. cefotetan, cefoxitin)0
CarbapenemaseIMP-family, VIM-family, GIM and SPMSubstrates of the extended-spectrum group plus cephamycins and carbapenems0
KPCSame as for IMP-family, VIM-family, GIM and SPM+++
OXA-23, -24, -25, -26, -27, -40, -48Same as for IMP-family, VIM-family, GIM and SPM+

Extended spectrum beta-lactamases

The extended spectrum beta-lactamases (ESBLs) often evolve from enzymes with a narrower spectrum such as the wide-spread TEM-1 and SHV-1. Amino acid substitutions or insertions by mutations in the genes encoding these enzymes lead to extended substrate specificity or increased hydrolytic rate (Gniadkowski, 2001; Jacoby & Munoz-Price, 2005). The number of known variants of the TEM and SHV families is constantly growing and most has emerged (and are emerging) by stepwise mutations. The selection of a particular variant in a given hospital has often been related to a specific profile of use (Gniadkowski, 2001). In the last decade, plasmid mediated CTX-M enzymes (hydrolysing cefotaxime) have emerged and spread rapidly in Enterobacteriaceae in many parts of the world, including Europe (Canton & Coque, 2006). The CTX-M family can be sub-divided in several clusters and, as for the TEM- and SHV-families, mutational events lead to emergence of variants within each cluster (Livermore et al., 2007).

The ESBL-encoding genes are often associated with genetic structures that are highly mobile. These may be large self-transmissible plasmids and/or transposons and integrons. Mobility and expression is further promoted by the association of many of the CTX-M genes with insertion sequences (Canton & Coque, 2006; Jacoby, 2006). The insertion sequences are probably responsible for their mobilization from progenitors and can contribute to further dissemination. Integrons are genetic elements that are able to capture individual antimicrobial resistance gene cassettes. The integrons carrying genes encoding CTX-M-type enzymes are mostly of class 1 type, which are in turn associated with insertion sequences and often with transposons and plasmids (Canton & Coque, 2006).

Chromosomal and plasmid-mediated AmpC type beta-lactamases

The AmpC type beta-lactamases form a large group of originally species-specific enzymes encoded chromosomally in various Gram-negative bacteria. The amount of enzyme that is inherently produced varies between species depending on the mechanism of regulation. For example, in E. coli the chromosomally encoded production of AmpC is normally repressed and the levels of the enzyme are then insufficient to confer ampicillin resistance. However, mutations in the promoter region can lead to derepression of the AmpC gene resulting in hyper-production of the enzyme, with clinical resistance to ampicillin and cephalosporins as a consequence (Batchelor et al., 2005c; Gootz, 2004; Li et al., 2007; Pfaller & Segreti, 2006).

Although enzymes of the AmpC type do not normally affect fourth generation cephalosporins, reports from human medicine indicate that on rare occasions, mutations in the genes encoding these enzymes can lead to an expanded spectrum AmpC which includes resistance to fourth generation cephalosporins (Ahmed & Shimamoto, 2008; Le Turnier et al., 2009;Mammeri et al., 2007; Mammeri et al., 2008; Wachino et al., 2006).

In Enterobacteriaceae genes encoding AmpC-type beta-lactamases are increasingly associated with plasmids (Gootz, 2004). Apparently these genes have been mobilized from the chromosome of certain bacterial species in which they are inherent, evolved further and are now spread horizontally between different species of Enterobacteriaceae (Alvarez et al., 2004; Biedenbach et al., 2006; Gootz, 2004). For example, Salmonella does not inherently carry this type of enzymes but over the last decade, genes encoding variants of the enzyme CMY (cefamycinase) have been identified on plasmids in a large number of different Salmonella serovars (Arlet et al., 2006). Plasmid encoded CMY production has also been identified in, e.g. E. coli of animal origin (Blanc et al., 2006; Brinas et al., 2005, 2003a,b; Donaldson et al., 2006; Kojima et al., 2005).


The genes encoding ESBLs are often physically linked in integrons, transposons and/or plasmids with genes encoding resistance to other, structurally unrelated resistance genes (Canton & Coque, 2006). Co-resistance to, e.g. aminoglycosides, tetracyclines and sulphonamides is frequent, not only in isolates from nosocomial outbreaks but also in isolates of Salmonella (Batchelor et al., 2005b; Bertrand et al., 2006; Hasman et al., 2005; Li et al., 2007; Michael et al., 2006; Politi et al., 2005; Weill et al., 2004) and other Enterobacteriaceae from animals (Blanc et al., 2006; Brinas et al., 2003b; Kojima et al., 2005). Multiresistant CTX-M-producing strains from humans have been shown to carry transferable quinolone resistance genes [qnr and/or aac(6′)-Ib-cr] (Canton & Coque, 2006; Robicsek et al., 2006a). The latter of these genes encodes the enzyme AAC(6′)-Ib-cr, a variant of an aminoglycoside acetyltransferase that also modifies some fluoroquinolones via N-acetylation at the amino nitrogen on its piperazinyl substituent (Robicsek et al., 2006b).

For CTX-M, genes encoding CMY and other plasmid-mediated AmpC type resistance are frequently associated with other genes encoding resistance to structurally unrelated antimicrobials (Batchelor et al., 2005c; Jacoby & Munoz-Price, 2005). Co-resistance with several other antimicrobials, e.g. aminoglycosides, chloramphenicol and florfenicol, sulphonamides, tetracycline and/or trimethoprim is common and has been documented in Salmonella and E. coli from animals and food (Alcaine et al., 2005; Allen & Poppe, 2002; Berge et al., 2004; Lopes et al., 2006; White et al., 2001; Zhao et al., 2003) and E. coli from animals (Brinas et al., 2003b; Donaldson et al., 2006).

Laboratory detection of ESBL and AmpC-type beta-lactamases

Reliable laboratory detection of resistance mediated by ESBLs depends on screening for decreased susceptibility with several different cephalosporins. Use of both cefotaxime and ceftazidime or cefpodoxime and use of low break-points has been recommended for the testing of Enterobacteriaceae (Livermore & Brown, 2001). For surveillance purposes, epidemiological cut-off values such as those set by European Committee for Antimicrobial Susceptibility Testing (EUCAST) are more sensitive in detecting organisms that harbour ESBLs or AmpCs, than the mostly higher clinical break-points (Kahlmeter, 2008). Recently EFSA advised that testing cefotaxime and use of epidemiological cut-off values should be sufficient to detect principally all ESBLs and AmpC-type beta-lactamases (Anonymous, 2006b).

Resistance in bacteria from food-producing animals

Methicillin-resistant Staphylococcus aureus

Infections with methicillin-resistant Staphylococcus aureus (MRSA) in hospitals but also increasingly in the community are a major public health problem worldwide (Boyce et al., 2005). Colonization and infection with MRSA has been increasingly reported among pets and horses, and more recently in food-producing animals (for a review see Leonard & Markey, 2008). In the Netherlands, high prevalence of a particular clone of MRSA, ST398, has been reported in pigs, but also in other animals (de Neeling et al., 2007). The same clone has also been reported in animals in other European countries (Guardabassi et al., 2007; Witte et al., 2007), and from infections in humans in several countries. The ‘pig clone’ was previously absent in human infections, which suggests that it has emerged in animals.

Resistance in Enterobacteriaceae

In Tables 3 & 4, phenotypic data on resistance to third generation cephalosporins in E. coli from healthy animals and in Salmonella have been compiled from the national zoonoses reports submitted to EFSA ( Data from Enter-net (an EU-funded international surveillance network for the enteric infections Salmonella and VTEC O157; indicate that in year 2005, the overall average figure of resistance to cefotaxime in S. Typhimurium isolated from humans in EU was 0.6% (Anonymous, 2006a). Comparability is hampered by differences in inclusion criteria, testing methodology and choice of interpretation criteria. In many cases (both in veterinary and human medicine), use of clinical break-points has probably lead to an underestimation of the occurrence of ESBLs (Kahlmeter, 2008).

Table 3.   Reported resistance to cefotaxime or ceftiofur in Escherichia coli isolated in healthy animals in 2007 (no. investigated isolates and per cent reported as resistant; based on data in national zoonoses reports submitted to EFSA in accordance with Directive 2003/99/EC,; only entries with results from more than 10 isolates were included)
Country reportingMethodCut-off*
CattleFowl (Gallus gallus)Pigs
  1. Dil, (micro)dilution; DD, disk diffusion.

  2. *Cut-off used to define resistance to cefotaxime unless otherwise indicated; no. isolates tested; per cent of tested isolates reported as resistant; §cut-off for ceftiofur; figure in table is for bovines unspecified. Also reports veal n = 87, 1% resistance and dairy cattle n = 18, 5.6% resistance.

EstoniaDil0.25214.0  190
FinlandDil0.25    1350
ItalyDil0.25  372.71490.7
The NetherlandsDil0.251520439.01691.2
NorwayDil0.25    1981
SloveniaDDNA220  342.9
SwedenDil0.25  2961.0  
SwitzerlandDil4§  2840.4980
United KingdomDDNA16526.7    
Table 4.   Reported resistance to cefotaxime or ceftiofur in Salmonella isolated from animals or food products in 2007 (no. investigated isolates and per cent reported as resistant; based on data in national zoonoses reports submitted to EFSA in accordance with Directive 2003/99/EC,; only entries with results from more than 10 isolates were included)
Country reportingMethodCut-off*
Salmonella EnteritidisSalmonella TyphimuriumSalmonella spp.
Fowl (Gallus gallus)CattleFowl (Gallus gallus)PigsCattleGallus gallus (fowl)PigsMeat from poultryMeat from pigs
  1. Dil, (micro)dilution; DD, disk diffusion.

  2. *Cut-off used to define resistance to cefotaxime unless otherwise indicated; no. isolates tested; per cent of tested isolates reported as resistant; §cut-off for ceftiofur; isolates from broiler meat.

AustriaDDNA780  130140          
Czech republicDil0.51950    210          
DenmarkDil2§  130  5750        710
FinlandDil0.5  140              
IrelandDil0.5      410          
ItalyDil0.5590  140310120    258.01290
The NetherlandsDil0.5190  110840  13416.4430    
SlovakiaDil0.5790      1701370280120  
SpainDil0.51851.1    190          
SwedenDil0.5    120150    140    
United KindomDDNA1230860  8890    500    

As can be seen in Tables 3 & 4, some countries report a comparatively high prevalence of resistance to third generation cephalosporins for E. coli from poultry and, for one country also cattle. Individual countries also report increasing resistance in Salmonella from different animal sources. As an example of a strikingly rapid emergence, data on the occurrence of resistance to cefotaxime, defined by the epidemiological cut-off value of EUCAST (>0.25 mg/L), in E. coli from healthy broilers, and in Salmonella Paratyphi B var. Java from broilers in the Netherlands are shown in Fig. 3 (data from MARAN 2005; Dik Mevius, personal communication, 2007).

Figure 3.

 Occurrence of resistance to cefotaxime in E. coli and Salmonella Paratyphi B var. Java from broilers in the Netherlands.

The resistance data in Tables 3 & 4 consist of percentages of resistance among bacterial isolates investigated, i.e. the epidemiological unit of concern is the bacterial species colonizing a particular animal. By use of selective screening techniques, data on prevalence on animal-level can be obtained. In studies where such techniques have been used, the prevalence of pigs and poultry carrying E. coli with decreased susceptibility to third generation cephalosporins varies widely, from 10 to 93% (Girlich et al., 2007; Jorgensen et al., 2007; Moreno et al., 2007). Even in cases when the prevalence of animals with E. coli showing decreased susceptibility was high, only a low percentage of E. coli in each sample displayed this trait.

As laboratory detection of these types of resistance can be problematic and under-reporting may have occurred, the information above indicates that the resistance to third and fourth generation cephalosporins in E. coli and Salmonella from animals is rapidly emerging in some European countries. In others, information is still incomplete and trends are difficult to assess.

Emergence of ESBLs

Nontyphoidal Salmonella spp. with ESBL-type resistance appeared in the late 1980s and early 1990s. Subsequently, the number serovars, enzymes and countries involved have increased steadily (Arlet et al., 2006; Miriagou et al., 2004). In 2004, various ESBLs of the SHV, TEM or CTX-M families had been described in more than 30 serovars isolated from humans or animals in more than 30 countries, whereof 13 European (Arlet et al., 2006). Judging by the number of reports, in Europe TEM-52, SHV-2, -5 and -12 and a wide array CTX-M-enzymes are the most commonly encountered (Arlet et al., 2006). This is also true if only reports including isolates from food-producing animals or food are included.

In a comprehensive study from the Netherlands including nonduplicate isolates of Salmonella from poultry and poultry meat from 2001–2002, Hasman et al. (2005) showed a great diversity of serovars and ESBLs. TEM-52 was the most common ESBL, occurring in various serovars including S. Enteritidis and S. Typhimurium. In 2002, a multiresistant Salmonella Virchow producing CTX-M-9 was isolated from a girl in France and similar isolates were recovered in 2003 from six chicken farms and one hatchery supplying these farms, as well as from poultry meat at retail (Weill et al., 2004). Production of CTX-M-9 was also reported from Spain in S. Virchow and S. Enteritidis isolated in 2003 and 2004, a study that also reported SHV-12 in a S. Rissen isolated from a pig (Riano et al., 2006). From Greece, multiresistant S. Virchow producing CTX-M-32 were isolated in 2001 from two batches of poultry products from the same company (Politi et al., 2005). The clonal emergence during 2000–2003 of a multiresistant S. Virchow-producing CTX-M-2 was described from Belgium and France (Bertrand et al., 2006). An isolate of S. Virchow with a similar antibiogram was also described from poultry (isolated 2002) in the study from the Netherlands (Hasman et al., 2005).

A series of publications based on materials from the Spanish Veterinary-Antimicrobial-Resistance-Surveillance (VAV) Network describe an increase in the percentage and variety of ESBLs (and of AmpC-type resistance, see below) over time. In a study on E. coli isolated from healthy animals and foods in Spain during 1997–1999, genes encoding TEM, SHV and OXA were demonstrated. Some of the enzymes were inhibitor resistant, but none were of ESBL-type (Brinas et al., 2002). In a study on isolates from 2000–2001 from healthy chickens, one isolate carrying a gene encoding CTX-M-14 was detected (Brinas et al., 2003b). During 2003, several CTX-M-variants (CTX-M-1, CTX-M-9, CTX-M-14 and CTX-M-32) and SHV-12 were demonstrated in E. coli from materials from sick food-producing animals and from healthy chickens (Brinas et al., 2005). The strains carrying genes encoding CTX-M-type enzymes appeared to be clonally unrelated. In a separate study on E. coli isolated from Catalonian poultry, pig and rabbit farms various CTX-M-type enzymes were demonstrated as well as SHV-2 and TEM-52 (from poultry) (Blanc et al., 2006). Different patterns were observed depending on animal species, and as in the previous study, the strains showed a low clonal relationship. The lack of clonality indicates horizontal spread of plasmids or other transferable genetic elements. In human medicine, dissemination of genes encoding CTX-M-9 is associated with the large conjugative plasmids carrying often also conferring resistance to aminoglycosides and trimethoprim. Similarly, epidemic plasmids carrying genes encoding CTX-M-14 or CTX-M-32 have been described (Canton & Coque, 2006; Livermore et al., 2007).

From the UK, the first isolates of E. coli producing a CTX-M-14-like enzyme were reported from diarrhoeic calves in 2005 (Teale et al., 2005). In France, CTX-M-1 and CTX-M-15 were detected in clinical isolates from cattle, swine and poultry (Meunier et al., 2006) and in Portugal from broilers (Pomba et al., 2008). Further, in a screening study using selective techniques, 12 of 112 healthy poultry sampled at slaughter carried CTX-M-1 producing E. coli (Girlich et al., 2007). In Belgium, CTX-M-1 was also the most common ESBL enzyme among healthy broilers at five commercial farms, but also TEM-52, CTX-M-2, CTX-M-14 and TEM-106 were also present (Smet et al., 2008). In a Portuguese study, TEM-52, SHV-2 and CTX-M-1 were detected in Enterobacteriaceae in samples from chicken meat and broilers and SHV-12 in samples from healthy swine (Machado et al., 2008). CTX-M-1 has also been described in clinical isolates of E. coli and Klebsiella pneumoniae from horses in the Netherlands (Vo et al., 2007). Occurrence of CTX-M-2 and CTX-M-18 in E. coli from healthy chickens and CTX-M-2 from cattle has been reported from Japan (Kojima et al., 2005; Shiraki et al., 2004).

Taken together, a wide array of genes encoding ESBLs has emerged and is now present in enteric bacteria from animals, although the different enzymes found in animals are markedly less diverse than those found in isolates from humans. In particular, different CTX-M type genes are increasingly reported both in E. coli and in Salmonella from food-producing animals in Europe in recent years. As noted previously, the genes encoding ESBLs are often encoded on plasmids and/or other transferable genetic elements and are often linked to multiple other resistance genes. Spread of resistance can be clonal or horizontal, or both.

Emergence of transferable AmpC-type beta-lactamases

The first occurrence of plasmid-mediated AmpC-type beta-lactamase (CMY-2) was described from humans in Algeria in an isolate of S. Senftenberg from 1994 (Koeck et al., 1997). Later, AmpC production was reported in a multidrug-resistant Salmonella isolate in the US (Horton et al., 1999). This was rapidly followed by a number of reports from the US and Canada on production of CMY-2 by different serovars of Salmonella spp. isolated from animals and food, in particular in S. Typhimurium and S. Newport (Alcaine et al., 2005; Allen & Poppe, 2002; Carattoli et al., 2002; Chen et al., 2004; Fey et al., 2000; Gray et al., 2004; Pitout et al., 2003; White et al., 2001; Winokur et al., 2000, 2001; Zhao et al., 2003). In most reports, the isolates were multiresistant (Biedenbach et al., 2006). From 1999 to 2003, resistance to ceftiofur in Salmonella increased from 4 to 19%, with S. Newport being the most common serovar (Frye & Fedorka-Cray, 2007). Resistance was predominantly associated with CMY-2 encoding plasmids. Available information indicates that the increase of MDR-AmpC S. Newport is explained by the spread of one clone among animals and humans (Berge et al., 2004; Zhao et al., 2003).

There are to date only two reports from Europe on production of CMY-2 in Salmonella from animals, and in both cases there was a link to imported animals (Aarestrup et al., 2004; Liebana et al., 2004). Likewise, in an outbreak of MDR-AmpC S. Newport in France, imported horsemeat was implicated (Espie et al., 2004). Another AmpC-type enzyme, ACC-1, was described in Salmonella spp in a study from the Netherlands (Hasman et al., 2005). The gene encoding ACC-1 was present in S. Bareilly, S. Braenderup and S. Infantis and it was carried on indistinguishable plasmids.

In the US, plasmid-mediated genes encoding CMY-2 appear to be widely disseminated in E. coli from diseased and healthy food-producing animals and food (Bradford et al., 1999; Donaldson et al., 2006; Winokur et al., 2001; Zhao et al., 2001). In one of these studies, 15% of 377 isolates of E. coli from clinical submissions from cattle and pigs were carrying CMY-2 (Winokur et al., 2001). More than 90% of these isolates were also resistant to tetracycline, sulfonamides and streptomycin, almost 70% to gentamicin and 15% to ciprofloxacin (resistance defined by clinical break-points).

In a Belgian study on occurrence of ceftiofur-resistance among E. coli isolated from broilers at five farms, CMY-2 was as common as ESBL-type enzymes. Eight per cent of the ceftiofur-resistant isolates had both CMY-2 and CTX-M-1 (Smet et al., 2008). Plasmid-mediated AmpC-type beta-lactamases have been reported from Spain and Portugal at low frequency in E. coli isolated from diagnostic submissions from cattle, pigs and broiler (Brinas et al., 2003a; Pomba et al., 2008) and from Spain from healthy chickens and rabbits (Blanc et al., 2006; Brinas et al., 2005). Batchelor et al. (2005a) reported the isolation of CMY-2 positive E. coli from one of 140 samples from healthy cattle in the UK. Two different types of E. coli harbouring indistinguishable large CMY-2 carrying plasmids were isolated from the same animal. In the Netherlands, E coli with plasmid-mediated CMY-2 has recently been reported in a clinical isolate of E. coli from a horse (Vo et al., 2007). Plasmid-mediated CMY-2 have also been reported from healthy food-producing animals in Japan (Kojima et al., 2005).

In summary, plasmid-mediated CMY-2 resistance has become widespread among Salmonella and E. coli in animals in North America. In Europe, most reports indicate a limited occurrence, but high frequencies in a recent study on broilers (Smet et al., 2008) indicate that also this type of enzymes may be increasing. As for the ESBLs, the plasmid borne genes encoding CMY-2 are mostly linked to multiple other resistance genes. Spread of resistance can be clonal or horizontal, or both.

Influence of use of antimicrobials on the emergence and spread of resistance

Following systemic administration, ceftiofur and cefquinome are mainly excreted in urine with a limited portion, excreted in faeces (EMEA/MRL/005/95, Summary of Product Characteristics, Annex 1; Naxcel 100 mg/mL). Information on what that means in terms of active concentrations over time in intestinal contents is not available in sources in the public domain. Such information is essential to evaluate the exposure of the gastro-intestinal microbiota of the target animal to the parent drug or active metabolites (CVMP/VICH GL 27; Guidance on the preapproval information for registration of new veterinary medicinal products for food producing animals with respect to antimicrobial resistance, CVMP/VICH/644/01-final).

Influence of cephalosporin use on occurrence of MRSA

As MRSA are resistant to all beta-lactams, use of any substance in that group may provide selective pressure. In human medicine, use of cephalosporins, other beta-lactams, fluoroquinolones and glycopeptides have been shown by metanalysis to be associated with an increased risk of acquisition of MRSA (Tacconelli et al., 2008). In view of the increasing occurrence of MRSA in animals, the risk associated with use of substances with a potential to select for MRSA-colonization of animals should be further examined. The potential influence of the use of products formulated as ‘long acting’, with long excretion times deserve special attention, as the time when concentrations are close to the MIC of intestinal and skin microbiota can be long. This document is focused on resistance with particular relevance for the third and fourth generation cephalosporins rather than all cephalosporins and other penicillinase stable beta-lactams. Emphasis will, therefore be on resistance in Gram-negative enteric bacteria and MRSA will not be further discussed in detail.

Influence of cephalosporin use on the evolution of genes encoding beta-lactamases

The beta-lactamases TEM-1 and SHV-1 are common in bacteria from various animals. These enzymes do not confer resistance but mutations in the genes encoding these enzymes lead to structural changes that can extend or alter the substrate specificity (Gniadkowski, 2008). Similarly, mutations in the genes encoding AmpC-type enzymes can give rise to extended spectrum AmpC with activity also against 4th generation cephalosporins (Ahmed & Shimamoto, 2008; Le Turnier et al., 2009; Mammeri et al., 2007, 2008; Wachino et al., 2006).

The evolution of ESBLs has been attributed to the selective pressure exerted by use of higher generation cephalosporins (Medeiros, 1997). There are a number of studies in human clinical settings in support of that (Gniadkowski, 2008). Blásquez et al. (2000) have suggested a broader view: that in vivo evolution of ESBLs is driven by the constant fluctuating pressure of various beta-lactams, including also penicillins and first generation cephalosporins. This may explain why many of the enzymes generated in vitro never occur naturally – only ESBLs with a truly broad-based resistance would survive and be selected for in an environment where different beta-lactams are used.

Current knowledge on use of cephalosporins as a driver of the evolution of ESBLs and AmpCs is based on laboratory studies and studies in human clinical settings. It is probable that the general principle applies also to animal production thus the use of third and fourth generation cephalosporins in animal populations, and possibly also of other beta-lactams, may favour the evolution of beta-lactamases in exposed bacterial populations.

Influence of cephalosporin use on selection and amplification of genes encoding beta-lactamases

Use of third generation cephalosporins is a recognized risk factor for ESBL colonization of patients in the human hospitals (Asensio et al., 1996;Quale et al., 2002; Saurina et al., 2000; Urbanek et al., 2007). Several authors have suggested that the use of ceftiofur in cattle and turkeys may have contributed to the spread in Salmonella in North America, of plasmid mediated AmpC-type beta-lactamases (Allen & Poppe, 2002; Dunne et al., 2000; Fey et al., 2000; White et al., 2001; Winokur et al., 2000). Until recently, there have been no specific studies on the influence of the use of third generation cephalosporins on resistance in Enterobacteriae in food producing animals.

In an experimental study, administration of a single dose of ceftiofur to turkey poults without detectable ceftiofur-resistant strains did not result in the emergence of resistant strains (Poppe et al., 2005). In a parallel experiment, the poults were dosed both with susceptible S. Newport and with E. coli carrying a large plasmid encoding AmpC-type beta-lactamases. The plasmid was readily transferred in the intestine to the Salmonella strain and also to a serotype of E. coli different from the donor in absence of any selective pressure. The experiment did not include a group receiving both antimicrobials and bacteria carrying resistance; hence the influence of ceftiofur on transfer and shedding of bacteria carrying resistance genes was not evaluated.

In a small experimental study, pigs were injected with ceftiofur, cefquinome or amoxicillin once daily for 3 days (Cavaco et al., 2008). Untreated animals served as controls. Almost all animals were positive for E. coli with CTX-M-1 before the start of the experiment. Animals were also inoculated with a CTX-M-1 producing strain of E. coli before the start of treatments. Significantly higher counts of cefotaxime-resistant coliforms were observed in all treated groups compared with controls for up to 22 days after the end of treatment. The cephalosporins had a more pronounced effect than amoxicillin.

Tragesser et al. (2006) studied the occurrence of ceftriaxone-resistant E. coli in dairy herds in Ohio, USA and linked the results to information on use of ceftiofur in the studied herds. Most of the isolates that showed reduced susceptibility to ceftriaxone carried a plasmid coding for CMY-2. Such isolates were recovered from at least one of the sampled cows in 10 of the 12 herds reporting use of ceftiofur, and in two of the seven herds reporting non-use (odds ratio 25, P = 0.01). The mean within-herd prevalence was 40% for herds reporting use of ceftiofur, compared with 9% for those reporting non-use. There was neither association at individual cow-level nor a linear relation between within-herd prevalence and treatment frequency. There was no attempt to analyse the influence of use of other antimicrobials on the farm. All CMY-2-producing isolates of E. coli were co-resistant to streptomycin, sulphonamides and tetracycline, and in addition, commonly also to gentamicin, kanamycin and trimethoprim–sulphonamides. Co-selection by other antimicrobials as well as management factors could account for the lack of linear relation between within-herd prevalence and use of ceftiofur.

In a study from Denmark, pigs in farms using and not using ceftiofur were sampled (Jorgensen et al., 2007). Escherichia coli with reduced susceptibility to third generation cephalosporins was demonstrated in 69 of 200 sampled pigs (five of 10 farms) but only in three of 200 animals in control farms (one of 10 farms). The difference was statistically significant (P = 0.02). Production of ESBL (CTX-M-1) was demonstrated in 19 isolates from two of the ceftiofur-using farms (not statistically different from farms not using ceftiofur). The study did not examine other drug-use practices in the farms.

Lowrance et al. (2007) studied the influence of administration of ceftiofur crystalline free acid (‘long acting’) to steers. Ceftiofur was administered subcutaneously to different cohorts at 6.6, 4.4 (single doses) and at 6.6 mg/kg thrice with 6 days interval). Untreated steers served as controls. Ceftiofur-resistant faecal E.coli were present at the start of the study, and administration of ceftiofur was associated with an increase in the proportion of E. coli resistant to ceftiofur during treatment in all treated groups. No changes in proportion of resistant isolates recovered using nonselective techniques were observed in co-mingled control animals. Almost all resistant isolates were co-resistant to at least chloramphenicol, streptomycin, sulphonamides and tetracyclines, a pattern associated with a multiresistance plasmid described in AmpC-producing Salmonella and E. coli (Winokur et al., 2001).

The influence of general use of antimicrobials on antimicrobial resistance in bacteria from calves in the US was studied in a field trial (Berge et al., 2006). Individual treatments transiently increased the shedding of multiply resistant E. coli compared with nontreated calves. The isolates were resistant to ceftiofur, which was the antimicrobial used for most of the individual treatments.

A longitudinal study over 5 months on healthy young calves on a dairy farm showed a persistent high prevalence (65–100%) of calves shedding ceftiofur-resistant, CMY-2 producing, E. coli (Donaldson et al., 2006). The isolates were all multi resistant and belonged to 59 clonal types. The farm reported use of various antimicrobials including ceftiofur but kept no individual records; thus no attempt to correlate use with resistance could be made.

The persistence of ESBLs of CTX-M type on a dairy farm in the absence of use of cephalosporins and other beta-lactams has been documented from the UK (Liebana et al., 2006). All use of beta-lactam antimicrobials, apart from intramammary use of cefquinome, was stopped during the study period in an attempt to remove the selective pressure. The prevalence of animals shedding CTX-M positive E. coli remained high over 6 months. As in studies on CMY-2, there was a diversity of clones but an almost complete predominance of one plasmid type carrying a gene encoding CTX-M and in addition, streptomycin resistance. Occurrence of E. coli producing ESBL and CMY in the apparent absence of use of cephalosporins has also been reported in broiler (Smet et al., 2008; MARAN 2005).

It has been argued that active concentrations of ceftiofur in the intestines of treated animals are very low, and that the substance is rapidly metabolized by the intestinal microbiota (Hornish & Kotarski, 2002). The studies quoted above show that the concentrations are sufficient to select for E. coli with resistance to third generation cephalosporins. The lack of clonality of resistant isolates reported in several studies clearly indicates horizontal dissemination of resistance genes.

Co-selection of resistance in Enterobacteriaceae by noncephalosporin antimicrobials

As discussed above, ESBL- or AmpC-producing bacteria are often also resistant to multiple other antimicrobials. In most cases, the genes encoding these unrelated resistance traits are linked on the same plasmid or transferable genetic element as the ESBLs. Many of the antimicrobials in question are used in veterinary medicine, e.g. neomycin, streptomycin, tetracycline, trimethoprim, sulphonamides and fluoroquinolones. A few of these substances are also used as growth promoters in some parts of the world. Of particular concern is the described association between CTX-M or AmpC encoding genes and plasmid-mediated quinolone resistance (Robicsek et al., 2006a). In human hospitals, use of fluoroquinolones has been identified as a risk factor for spread of CTX-M (Ben-Ami et al., 2006).

The frequent linkage of resistance genes implies that once the ESBL- or CMY-encoding genes have entered a bacterial population in a production unit, a broad range of antimicrobials, including beta-lactams such as amoxicillin, but also of structurally unrelated antimicrobials can favour their selection and spread between animals and between bacterial strains (co-selection). In the Netherlands, Salmonella and E. coli-producing ESBL have emerged and increased in prevalence in poultry without prior use of cephalosporins (MARAN, 2005). Similarly, a high prevalence of ESBL- and CMY-producing E. coli has been reported in Belgian broiler flocks (Smet et al., 2008). It is probable that use of either beta-lactams such as amoxicillin or non-beta-lactam antimicrobials, have contributed to the observed increase.

Exposure of humans to resistant bacteria and resistance genes from animal sources

Exposure to resistant bacteria from animals

As noted previously occurrence of ESBL- or AmpC-producing Salmonella of different serovars from animals and from food of animal origin has been demonstrated in a number of studies (for references see previous sections).

Most salmonella infections in humans are attributed to food-borne transmission. Person-to-person transmission is uncommon, except in outbreaks with nosocomial spread. Therefore, exposure of humans to Salmonella resistant to cephalosporins via food, direct contact with infected animals or indirectly via the environment will have a significant influence on the occurrence of ESBL- or AmpC-producing Salmonella in humans. This is supported by the observation in the US of a temporal association between emergence of AmpC-type beta-lactamases in Salmonella from various animals and an increased prevalence of such infections in humans (Frye & Fedorka-Cray, 2007; Gupta et al., 2003; Lopes et al., 2006). Several outbreaks of cephalosporin-resistant Salmonella (AmpC or ESBL producing) have been linked to consumption of animal products (Bertrand et al., 2006; Espie et al., 2004; Weill et al., 2004). An outbreak of human infections with multiresistant, CMY-2-producing S. Newport implicating handling of pet treats containing dried beef as the source has been described in Canada (Pitout et al., 2003).

Direct spread of a multiresistant S. Newport, also resistant to third generation cephalosporins, from cattle to a person was documented by Fey et al. (2000). In a retrospective case–control study, patients infected with multidrug-resistant (including AmpC-type resistance) S. Newport were more likely to have had direct contact with cattle than either patients infected with susceptible S. Newport or matched healthy controls (Gupta et al., 2003).

Taken together, it is clear humans are exposed to cephalosporin-resistant salmonella via food or via direct contact with infected animals, and that this may result in clinical infections. This is also discussed in a recently published opinion by EFSA (Anonymous, 2008).

Exposure to resistance genes from bacteria associated with food of animal origin

In the study by Corpet (1988), volunteers eating sterilized food had significantly less coliforms resistant to tetracyclines in their faeces than when eating nonsterilized food. This type of study has not been repeated since, but data on occurrence of resistant E. coli on food, including vegetables, indicate that humans ingest resistant bacteria on a daily basis. During the passage through the intestine, such bacteria may transfer their resistance genes to host-adapted bacteria or to zoonotic pathogens. Exchange of resistance genes between bacteria from different sources has also been demonstrated in water, soil, on kitchen towels, on cutting boards and on the surface of food (Kruse & Sørum, 1994). Evidence for horizontal transfer of plasmids or resistance genes other than cephalosporin resistance between bacteria colonizing animals and those colonizing humans has been documented in several studies (Chaslus-Dancla et al., 1991; Hunter et al., 1994; Lester et al., 2006; Levy et al., 1976; Nikolich et al., 1994; Tschäpe, 1994).

Genes encoding ESBL- or AmpC-type resistance have been demonstrated, not only in Salmonella isolated from food (see above) but also in E. coli (Brinas et al., 2002; Zhao et al., 2001). As discussed, these genes are mostly carried on mobile genetic elements. The number of studies is still limited, as is information on prevalence of resistance to cephalosporins in E. coli isolated in meat in Europe (see Table 3). However, available information suffices to conclude that humans can be exposed to genes encoding ESBL or AmpC-type resistance via food.

Indistinguishable plasmids or other genetic elements coding for ESBLs or AmpC-type resistance have been described from different bacterial species and different animal and human hosts (Batchelor et al., 2005a; Hasman et al., 2005; Poppe et al., 2005; Winokur et al., 2001). Thus, there is evidence that the plasmids carrying genes encoding ESBLs and AmpC-type beta-lactamases are transferred horizontally between different bacterial species of different hosts. Certain plasmids carrying genes encoding CMY-2 are disseminated among both Salmonella and E. coli from both animals and humans, and the pattern indicates that certain plasmids are epidemic (Hopkins et al., 2006; Mulvey et al., 2008). Further, there are some reports indicating acquisition of resistance plasmids by E. coli and Salmonella in the human gut (Su et al., 2003; Yan et al., 2005). A plasmid-encoding ESBL was identified in E. coli and S. Anatum, both from the same patient. As the resistant isolates had molecular fingerprints identical to those of susceptible isolates of the same species isolated earlier from the same patient, it was concluded that the acquisition of the same plasmid by two different bacteria had probably occurred in the gut (Su et al., 2003). With similar type of evidence, Yan and co-workers reported on a S. Hadar with AmpC-type resistance apparently acquired from an E. coli from the same patient (Yan et al., 2005).

In summary, bacteria of animal origin carrying resistance genes encoding ESBL or AmpC can be present in food. Moreover, at that stage cross-contamination between food items and human contamination through food processing might also play a role (Anonymous, 2008). Transfer of such genes to bacteria causing disease in humans can occur in the intestine. The present extent of exposure via food is difficult to determine. Any further expansion of the occurrence of ESBL or AmpC resistance among animal bacteria is likely to have an influence on the occurrence in food, and thereby on human exposure.

Impact of infections with cephalosporin-resistant bacteria on human and animal health

Human health

Gastroenteritis is the most common clinical manifestation of Salmonella infections, but severe cases with systemic manifestations occur. In those cases, antimicrobial treatment is often recommended. Serious infections are most common in children and elderly (Arlet et al., 2006). Among the first line empiric treatments for adults are the fluoroquinolones. In very young patients and when fluoroquinolone resistance is present, third generation cephalosporins are the drugs of choice.

In a study of S. Newport infections, no significant differences in clinical signs, hospitalization, duration of illness or other outcomes between patients infected with susceptible isolates and isolates of MDR-AmpC resistance phenotype could be demonstrated (Devasia et al., 2005). The lack of demonstrable impact of the multiresistance phenotype was probably a consequence of empiric treatment mostly done with fluoroquinolones, to which the isolates were susceptible. Infections of humans with Salmonella resistant to both cephalosporins and ciprofloxacin have been described (Cheung et al., 2005; Chiu et al., 2004; Ko et al., 2005; Yan et al., 2005), in some cases in association with fatalities. The emergence of multiresistant Salmonella with additional resistance to cephalosporins and fluoroquinolones seriously limits the therapeutic options available.

The recent and rapid emergence of resistance of CTX-M type ESBLs in Enterobacteriaceae from human infection in Europe is of major public health concern (Canton & Coque, 2006; Livermore et al., 2007). The increasing frequency of community-acquired infections and in particular of infections with E. coli resistant to third and fourth generation cephalosporins is of concern. Prevalence data were reviewed by Rodriguez-Bano and Navarro (2008). For example, in Seville, the percentage of E. coli that was ESBL producers increased from 0.3% in 1995–1998 to 4.8% in 2002. Approximately half of these isolates were from nonhospitalized patients. The percentage of ESBL-producing E. coli among isolates from community-acquired urinary tract infections ranged from 1.5 to 3.3% in studies from Seville, Barcelona and the Gaza strip. In a multicentre study from Spain, 3.7% of E. coli isolated in 2003 from community-acquired intra-abdominal infections was ESBL-producers. Two Spanish studies illustrate that the recorded increase in infections was matched by an increase in faecal carriage in outpatients and healthy volunteers (Mirelis et al., 2003; Valverde et al., 2004).

Many patients with community-acquired infections have a history of hospitalization and have co-morbidities, but cases of uncomplicated cystitis also occur. The routes of spread of genes encoding CTX-M outside hospitals are still not clear, but the epidemiological pattern indicates that reservoirs may exist in the community (Livermore et al., 2007). Considering the emergence of CTX-M producing Salmonella and E. coli in animals as discussed above, it has been hypothesized that animals may be one of several potential reservoirs and food a potential vector (Livermore et al., 2007).

The therapeutic options for infections with bacteria resistant to third generation cephalosporins are limited. In particular, this is true for community-acquired infections where oral therapy is preferred (Canton & Coque, 2006). Theoretically, one option is fluoroquinolones, but as CTX-M type resistance is frequently linked to other resistance determinants such as plasmid-encoded quinolone-resistance, there is a high likelihood that such antimicrobials is not an effective alternative.

Animal health

For almost all of the indications for which ceftiofur or cefquinome are authorized for systemic therapy in food-producing animals, alternatives are available. For example, for streptococcal infections, cephalosporins have generally no advantage above benzylpenicillin in terms of antimicrobial efficacy or safety. In cattle, the only indication in which third or fourth generation cephalosporins could be the sole alternative is severe clinical mastitis with life-threatening sepsis caused by Enterobacteriaceae such as E. coli or Klebsiella. Cephalosporins are poorly distributed to the milk compartment and their systemic use would be rational only in septic mastitis. The few antimicrobials that have shown some beneficial effect in therapy of severe coliform mastitis are fluoroquinolones and third and fourth generation cephalosporins (Rantala et al., 2002; Erskine et al., 2002; Shpigel et al., 1997; Poutrel et al., 2008). In horses, the only indication where cephalosporins can be regarded as critically important is neonatal sepsis in foals. In the treatment of this condition, penicillin–aminoglycoside or penicillin–trimethoprim–sulphonamide combinations are listed as ‘first choice’ in standard textbooks (Giguère, 2006; Weese et al., 2008). In many countries, resistance to both gentamicin and trimethoprim–sulphonamides in the Gram-negative target pathogens exist. In such cases, third or fourth generation cephalosporins could be the only effective alternatives.

In conclusion, in most cases the direct impact of infections resistant to cephalosporins on animal health is low. The emergence of resistance mediated by genes encoding ESBLs or AmpC among Salmonella and E. coli is frequently linked to resistance to other antimicrobial agents. A further increase of cephalosporin resistance can indirectly impact on animal health by increasing the prevalence of multiresistance, thereby severely reducing the number of effective alternatives for treatment.

Summary assessment

Resistance to third generation cephalosporins in e.g. K. pneumoniae and E. coli in human infections is increasing in Europe

Genes coding for CTX-M type enzymes have rapidly emerged and spread not only in hospitals but also in the community. Production of ESBLs in Enterobacteriaceae is often associated with resistance to other antimicrobials. The changing epidemiological pattern may be explained by many interacting factors. Many problems in human medicine can be correlated to use of cephalosporins and other antimicrobials in humans, but it is possible that spread from animal reservoirs via food or via the environment may contribute to the dissemination of resistance in the community. The potential role of community reservoirs of animal origin such as food of different origins, and of other potential reservoirs such as the environment needs further investigation.

Available data indicate that resistance to third generation cephalosporins is increasing in E. coli and Salmonella from animals in Europe

Many countries still report low or zero prevalence but others have noted pronounced increases. A wide array of genes encoding ESBLs has emerged and is now present in enteric bacteria from animals. The occurrence of resistance to cephalosporins among bacteria isolated from animals may have been underestimated in the past, both because of methodological weaknesses and use of insensitive interpretation criteria.

Work aiming to harmonize methodology and interpretation criteria is undertaken by EFSA. Better data will be available if monitoring of resistance to third generation cephalosporins in the zoonosis-monitoring programmes is conducted, in accordance with the standards developed by EFSA. Furthermore, an expansion of the monitoring to include commensal E. coli from animals and food from all Member States would provide data valuable for the assessment the reservoir of resistance genes.

The genes encoding resistance to third and fourth generation cephalosporins are transferrable and often linked to other resistance genes

A wide array of genes encoding ESBLs has emerged and is now present in enteric bacteria from animals. The genes encoding ESBLs are often encoded by plasmids and/or other transferable genetic elements, and are often linked to other resistance genes. Spread of resistance can occur both through dissemination of clones, and through horizontal transmission of, e.g. epidemic plasmids.

Data on the extent of use of third and fourth generation cephalosporins for animals in the EU is not presented in a way that allows exposure to be properly assessed

More information is needed on the influence of the use of cephalosporins in veterinary medicine on the evolution of new variants of ESBLs, and on potential differences between different doses and dosing regimens in this respect. It is important that the use of antimicrobials will be monitored in a way that allows for the use of different generations of cephalosporins per animal species to be followed.

Systemic use of third and fourth cephalosporins selects for resistance

The use of third and fourth generation cephalosporins can influence resistance in two ways: either by favouring the evolution of new variants of ESBL genes by selecting for emerging mutants or by selecting for genes that have been introduced from other sources into the exposed population. Excretion of the drug into the intestine after systemic administration is low, but data on exact concentrations are not easily available. A relation between use of ceftiofur and occurrence of resistance at herd level has been documented, showing that the concentrations are high enough to select for resistance.

Co-selection by other antimicrobials is likely to influence prevalence of resistance to third and fourth generation cephalosporins

The emergence of cephalosporin resistance has been documented in poultry production systems where no cephalosporins are authorized for use. Resistance may also persist in farms in the absence of systemic use of beta-lactams. The genes encoding ESBL or AmpC-type enzymes are frequently linked to genes conferring resistance to other, unrelated antimicrobials. It is, therefore likely that co-selection by antimicrobials other than cephalosporins have an important role. In particular, medication of large groups of animals (mass medication) with various antimicrobials in animal husbandry may contribute to the occurrence and dissemination of resistance in exposed populations.

The importance of use of noncephalosporin antimicrobials for selection and maintenance of cephalosporin resistant bacteria in animal populations needs to be documented further.

Humans are exposed to cephalosporin-resistant bacteria via food or via direct contact with infected animals or indirectly through the environment

Humans are exposed to cephalosporin-resistant Salmonella via food or via direct contact with infected animals, and this can result in clinical infections. Furthermore, humans may be exposed to animal bacteria with resistance genes coding for ESBLs or AmpC type enzymes via direct contact, via contaminated food or indirectly through the environment. These genes can be transferred to bacteria with potential to cause infections in humans. The extent of such exposure is difficult to determine. It is related to the prevalence of such genes in bacteria colonizing animals, but will also be influenced by other factors, e.g. those related to food hygiene and consumption habits.

In human medicine, the options for effective treatments of infections that are resistant to third and/or fourth generation cephalosporins are limited

In the case of infections in humans with Salmonella, E. coli or other Enterobacteriaceae that are resistant to third and/or fourth generation cephalosporins, treatment alternatives are carbapenems, fluoroquinolones or aminoglycosides. The occurrence of co-resistance often seriously limits the options for effective treatment and some of the alternatives carry a high risk of adverse effects or are difficult to use in outpatient settings.

To conclude, resistance to third and fourth generation cephalosporins is rapidly increasing in humans. Available evidence indicates that resistance to third and fourth generation cephalosporins is also emerging in animal populations. Although there are many uncertainties, the potential consequences of a further increase of ESBL and AmpC-type resistance in bacteria-colonizing animals are serious. Measures to counter a further increase and spread of resistance in animals should therefore be considered.


  • 1

    The Scientific Advisory Group on Antimicrobials (SAGAM) was mandated to give advice to the CVMP on the need to exercise certain control on those classes of compounds of greater importance to human medicine e.g. fluoroquinolones and third and fourth generation cephalosporins. This document discusses cephalosporins with a focus on substances of the third and fourth generation and food-producing animals, excluding aquaculture.

  • 2

    A cephamycin.

  • 3

    According to international nonproprietory names (INN), the spelling is ‘meticillin’.