Fighting antimicrobial resistance

J Andrew Hudson of Jorvik Food and Environmental Microbiology summarises a review paper, published in 2017 in Trends in Food Science and Technology, which examines the development of antimicrobial resistant (AMR) bacteria from the food production perspective.

Antimicrobial resistant (AMR) bacteria are an obvious concern, with predictions that 10m people a year will be dying from AMR infections by 2050. However, how (or how much) is the use of antimicrobials in the food chain contributing to the problem? The extent to which the problem can be attributed to use in food production has been described as a ‘prominent and contentious issue’.

The first impression when investigating this area is the massive amount of literature that exists; tens of thousands of papers emerge from even a fairly superficial literature search. This raises the question as to whether it is possible to conduct a systematic review without dividing the topic up into smaller areas, which, when combined, may not then provide a sufficiently unified overview. Different technological approaches might be needed, such as model development or text mining, to create the necessary information links.

Development of AMR

The mechanism responsible for the development of AMR bacteria involves the antimicrobial exerting a selective pressure that selects for a resistant sub-population of the target species. This resistance may have been gained through a simple mutation or the acquisition of resistance genes on mobile genetic elements from other bacteria in the environment. Such resistance traits, and resistance to other bactericidal substances, such as heavy metals, can become co-located on the same genetic element so providing the bacterium with resistance to a suite of antimicrobial agents, a phenomenon known as co-resistance. The transfer of genetic material occurs between different species and so foodborne commensals need to be considered as part of the picture as well as the pathogens themselves. The Gram-positive pathogen/lactic acid bacterium/ enterococcal system found in fermented foods, provides a complex example of a situation where genes might flow between multiple taxa. This flow of genes can occur in the animal, environment, food or the human gut.

What seems to be clear is that treating livestock does result in the emergence of AMR bacteria that may also be of animal health concern.

Use of antimicrobials

The use of an antimicrobial agent in food production may result in resistance emerging to a structurally similar, but different, antimicrobial used for human clinical purposes. For example, chlortetracycline used to treat calves was reported to result in the appearance of tetracycline-resistant Campylobacter.

The overall use of antimicrobials within the EU seems to be quite dynamic. For example, in the Netherlands their use increased after the Antimicrobial Growth Promoter (AGP) ban, which came into force in 2006, but has reduced markedly since then according to data for 2014-2015. There are also big differences (up to ten-fold) in use on a like-for-like basis between Member States for reasons which are not obvious. For the rest of the world, an upward trajectory is predicted, with some countries almost doubling usage by 2030 as a result of increased intensification.

Antimicrobials are used in livestock production for a number of reasons: the treatment of diseased animals, treatment of part of a group (metaphylaxis) or use during high periods of susceptibility (prophylaxis). Although the use of antimicrobials as AGPs is no longer permitted, it is not clear whether the ban has had much of an effect on clinical AMR infections, but it may have saved farmers money as the beneficial effects of growth promoters are similarly unclear. One study from Denmark suggested that the feed conversion rate in broilers increased slightly when AGPs were withdrawn. It is possible that growth promoters are only of value when animal husbandry is less than optimum; in a good system it seems that they may not achieve anything for the farmer, but the data are far from unequivocal. Economic analysis in developing economies also suggests that AGPs provide ‘negligible’ effects to the overall economy.

What seems to be clear is that treating livestock does result in the emergence of AMR bacteria that may also be of animal health concern. Resistant bacteria have been detected in ruminant animals in many studies, with calves more frequently harbouring these organisms than adult animals, possibly as a result of greater antimicrobial use in young animals. In contrast, low prevalence rates of AMR bacteria are measured in game meat, for example in deer in both Spain and the USA, presumably because of the lack of exposure to antimicrobials by wild animals. However, it is interesting that differences in resistant bacteria detected in conventional and organic production systems are not consistent. Some studies are unable to determine any differences between the two, and this applies to foods of both animal and plant origin.

Much of the information relates to livestock and animal products. However, antimicrobials may be used in the cultivation of fruit and vegetables, as may manures from animals treated with antimicrobials that may contain the antimicrobials themselves or AMR bacteria. Finnish data suggested that the AMR profiles of Enterobacteriaceae in vegetables were different from those in the human faecal flora. In America, one study found that vegetarians were more frequently infected by AMR E. coli than newly hospitalised patients and so we should keep an open mind about which food exposures may be of most concern.

There are several reports suggesting that agricultural use does result in human infections.

Link between use of antimicrobials and infections

The nub of the issue lies in whether the use of antimicrobials results in human infections of AMR bacteria and, if so, to what extent. Meat, for example, may carry some AMR bacteria, but normal cooking should kill bacteria on the food rendering them harmless, whether resistant or otherwise. However, campylobacteriosis remains a prominent disease and the not particularly thermally-resistant causal organism is delivered into households on foods which are subsequently cooked, so cross contamination needs to be considered as a transmission route in addition to direct consumption. In general, studies show a decrease in AMR bacteria, in terms of prevalence and diversity, along the food chain, which is to be expected in a well-controlled system. As an example, methicillin-resistant Staphylococcus aureus (MRSA) were present in 61.9% of nasal swabs from slaughter pigs in Canada while the prevalence in retail pork was 1.2%. Since the detection of the same genetic type of an AMR bacterium in food and human cases does not necessarily show cause and effect, other data need to be obtained to explore the issue. There are several reports suggesting that agricultural use does result in human infections, for example colistin resistant Escherichia coli were more frequent in the Chinese agricultural environment than in clinical isolates, implying an agriculture to clinical flow. Again, the picture is not clear as AMR salmonellae were detected in German pig production even though the relevant antimicrobial is not permitted for such use.

A number of outbreaks of foodborne disease involving AMR bacteria have been reported, for example Salmonella Typhimurium DT104 has caused fatal infections associated with pork. When AMR bacterial infections of humans occur, the clinical consequences are more severe than would be the case in infection by an antimicrobial sensitive organism, with higher mortality rates and longer illness durations.

Even when the use of antimicrobials ceases, resistance can remain although, yet again, there is conflicting evidence. For example, the reduction in the use of avoparcin in broiler production is related to the reduction of vancomycin-resistant enterococci in humans, but a ban on the use of enrofloxacin in poultry in the USA did not reduce ciprofloxacin resistance in human or poultry Campylobacter isolates, although this could be the result of co-selection.

Arguably the best understood system is that of MRSA and its association with pigs. Here the evidence would suggest that the consumption of pork is not a major source of MRSA infections, which is perhaps not surprising as the types of staphylococci on food animals tend not to be those that cause foodborne disease. Contact with pigs can be a source of AMR infections and humans can introduce MRSA to pigs and to other humans. Humans may also contaminate food through poor food handling; a fatal nosocomial outbreak of MRSA infections resulted when a colonised worker contaminated food that was subsequently eaten by hospital patients.

Environment

Another plausible link between agricultural use of antimicrobials and AMR bacterial infections in humans is through the environment. Antimicrobials administered to farm animals largely pass through unmetabolised so that they are excreted into the environment and, as well as resulting in AMR bacteria, may end up in water, faeces, compost and other sources.

In China, studies have readily detected antimicrobials in farming environments, effluents and rivers. An examination of AMR genes themselves demonstrated an appearance in the environment at around the same time as their appearance in humans. Such examination of the epidemiology of AMR genes has been extended to meat processing plants.

Behaviour

Farmer and veterinarian behaviour with respect to the prescription and use of antimicrobials are a major issue. Classical agricultural economics would suggest that farmers only make such decisions based on financial criteria, but the evidence suggests that this is not so. Understanding the balance between financial and non-financial drivers is key to formulating policy. Illegal use is especially difficult to explore because of the obvious reticence of illegal users to supply information.

Various mitigation and intervention strategies have been proposed but to understand which of these are effective, robust surveillance systems need to be in place to measure the effects resulting from their implementation. Some existing programmes have been criticised for being disjointed. Data need to be obtained in a consistent manner to allow inter-study comparability and the formulation of policy. The problem is not national, or even within large geographical/ political blocs, such as the EU, but it is international because of the trade in food and mobility of people potentially harbouring novel AMR bacteria and so the global situation needs to be considered.

Conclusions

• There remain many data gaps that will need to be filled to enable the modelling effort that is needed.

• There are virtually no data on the concentration of AMR bacteria in foods and these are needed for exposure assessment.

• Data to show the directions and quantity of AMR bacteria/ gene flow are needed.

• In many foods the controls which are applied to foodborne pathogens should minimise the presence of AMR bacteria. However, some foods are eaten raw and so could present a greater exposure.

• It is clear that generalisations are hard to make - what happens in one system might not necessarily happen in another seemingly similar system.

• The risks and benefits of the use of antimicrobials in food production need proper evaluation to enable informed policy development. There is a clear need to use antimicrobials in food production to prevent/treat disease, to prevent production losses and to ensure animal welfare.

An extended summary prepared by J. Andrew Hudson

Jorvik Food and Environmental Microbiology, based on the article: John A. Hudson, Lynn J. Frewer, Glyn Jones, Paul A. Brereton, Mark J. Whittingham and Gavin Stewart (2017) The Agri-food Chain and Antimicrobial Resistance: A Review, Trends in Food Science and Technology, Volume 69, Part A, November 2017, pages 131-147.

Email andrewhudson@ foodsafetyteam.org

Web sciencedirect.com/ science/article/abs/pii/ S0924224417302285