Evolutionary Applications

Evolutionary Applications ISSN 1752-4571

REVIEW AND SYNTHESIS

Antibiotics in agriculture and the risk to human health: how

worried should we be?
Qiuzhi Chang,1 Weike Wang,1 Gili Regev-Yochay,2,3 Marc Lipsitch1 and William P. Hanage1
1 Department of Epidemiology, Harvard School of Public Health, Boston, MA, USA
2 Infectious Disease Unit, Sheba Medical Center, Ramat-Gan, Israel
3 The Sackler School of Medicine, Tel-Aviv University, Tel-Aviv, Israel

Keywords Abstract
agriculture, antibiotic resistance,
Enterococcus, food, resistance transfer, The use of antibiotics in agriculture is routinely described as a major contributor
Staphylococcus aureus. to the clinical problem of resistant disease in human medicine. While a link is
plausible, there are no data conclusively showing the magnitude of the threat
Correspondence emerging from agriculture. Here, we define the potential mechanisms by which
William P. Hanage, Department of
agricultural antibiotic use could lead to human disease and use case studies to
Epidemiology, Harvard School of Public
critically assess the potential risk from each. The three mechanisms considered
Health, 677 Huntington Avenue, Boston, MA
02115, USA. are as follows 1: direct infection with resistant bacteria from an animal source, 2:
Tel.: +1 617 432 1282; breaches in the species barrier followed by sustained transmission in humans of
fax: +1 617 432 3564; resistant strains arising in livestock, and 3: transfer of resistance genes from agri-
e-mail: [email protected] culture into human pathogens. Of these, mechanism 1 is the most readily esti-
mated, while significant is small in comparison with the overall burden of
Received: 9 March 2014
resistant disease. Several cases of mechanism 2 are known, and we discuss the
Accepted: 14 May 2014
likely livestock origins of resistant clones of Staphylococcus aureus and Enterococ-
doi:10.1111/eva.12185 cus faecium, but while it is easy to show relatedness the direction of transmission
is hard to assess in robust fashion. More difficult yet to study is the contribution
of mechanism 3, which may be the most important of all.

The significance of agricultural antibiotics in the emergence the USA annually (Hollis and Ahmed 2013). Much of this
and spread of clinical antibiotic resistance is a matter of is not in veterinary medicine, but in the form of continu-
ongoing debate and controversy, with one prominent com- ous subtherapeutic application of antibiotics for growth
mentary asserting (albeit without support or citation) that promotion and disease prevention in intensively farmed
‘farming practices are largely to blame for the rise of antibi- animals (Mellon et al. 2001). Unsurprisingly, antibiotics
otic-resistant strains’ (Kennedy 2013). Antibiotic-resistant used in this context have been associated with a high fre-
infections have been conservatively estimated to cause quency of resistant bacteria in the gut flora of chickens,
some 23 000 deaths each year in the USA alone (Centers swine, and other food-producing animals (Witte 1998;
for Disease Control and Prevention 2013) and have been Aarestrup 1999). Without appropriate regulation, it is
described in apocalyptic terms by public health authorities. thought that a large diverse reservoir for resistant bacteria
If agriculture is a contributor to the spread of resistance, and resistance genes could facilitate the emergence and
immediate action is necessary to limit this source of spread of resistant pathogens to humans, and even the
human, as well as animal, morbidity, and mortality. How- ongoing transmission of such resistant organisms within
ever, as we shall argue, the magnitude of the threat arising the human population.
from the agricultural setting is uncertain for multiple The maintenance and increase in the prevalence of drug-
reasons. resistant organisms and resistance genes have been linked
Antibiotic use in humans has been shown to select anti- to the selective pressure of antibiotic use in both clinical
biotic-resistant strains, and the same should be expected in and agricultural settings (Levy 1997). How we should
livestock, which have been reported to receive over 13 mil- expect drug use in different environments to affect the tra-
lion kilograms, or approximately 80% of all antibiotics, in jectory of resistant organisms depends on the resulting

240 © 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd. This is an open access article under the terms of the Creative
Commons Attribution License, which permits use, distribution and reproduction in any medium, provided
the original work is properly cited.
Chang et al. Agriculture and antibiotic resistance

selective pressure and on constraints to microbial adapta- Risk modeling may be used to estimate the burden of
tion. Humans are at risk for exposure to new resistant disease resulting from mechanism 1, although to date it has
pathogens or resistance determinants from animals largely been restricted to infection through food. In the
through direct contact, ingestion of contaminated food or case of direct infection of humans (mechanisms 1 and 2),
water, and contact with infected humans. A schematic rep- whether or not it is accompanied by onward transmission
resentation of the possible links between the agricultural (mechanism 2), molecular epidemiology can in principle
setting and human populations is shown in Fig. 1. The identify a link between agricultural bacteria, and those
major influence on the increase in resistant organisms infecting humans, if appropriate samples are available.
transmitted within each environment is expected to be However, with the passage of time, this may become more
antibiotic use in that environment. However, we can also difficult for organisms that have crossed the species barrier
consider at least three mechanisms by which antibiotic (mechanism 2), as lineages in humans and animals diverge
resistance in agriculture can lead to a threat to humans from the common ancestor of the strain that crossed the
(Lipsitch et al. 2002). species barrier. In the case of resistance genes (mechanism
1 A human is infected by a resistant pathogen of agricul- 3), this is particularly so, while we might be able to state
tural origin through contact with livestock, or through that the genetic elements or mutations causing resistance in
ingestion of bacteria from contaminated meat or water, different settings are extremely closely related, we will be
without ongoing transmission of the pathogen between unable to identify the direction of transfer with confidence.
humans. The central question then becomes: How likely are antibi-
2 A human is infected or colonized with a resistant otic-resistant strains to arise in humans from agricultural
microbe through any of these means, followed by ongo- antibiotic use as a result of bacterial adaptations, and how
ing transmission among humans, with some of these likely are these resistant strains to be a cause of significant
humans becoming ill. This scenario constitutes a break clinical disease? To lend a more concrete perspective on
in the ‘species barrier’ by a microbe that may be directly this complex issue, we shall ground our discussion in sali-
pathogenic to humans or may be a commensal with the ent case studies that highlight the dominant pathways
ability to cause opportunistic infections. through which such antibiotics may impede human health.
3 Resistance genes arising in the agricultural setting are
introduced into human pathogens by horizontal gene
Sporadic human infection from contaminated food
transfer. The resulting resistant lineages are then selected
or direct contact
by antibiotic use in humans.
Humans and other animals can acquire resistant patho-
gens and commensal organisms simply by ingesting them.
Contaminated meat and other cross-contaminated foods
(A) cause millions of cases of gastrointestinal illnesses such as
salmonellosis and campylobacteriosis each year in the
USA alone (Scallan et al. 2011). The threat that antibiotic
use in food-producing animals poses to human health via
this route has been estimated using microbial risk assess-
(B)
ment models (McEwen 2012). Using an exposure-based
model, one study assessed how many cases of Campylobac-
ter jejuni infection complicated by fluoroquinolone treat-
(C) ment failure (i.e., resistant cases) could arise from
contaminated ground beef. The study estimated 12 cases
in the USA after one year of fluoroquinolone use in cattle,
rising to 44 cases and one death after 10 years (Anderson
Figure 1 Schematic illustration of possible links between antibiotic use et al. 2001). The comparatively small burden estimated by
in agriculture and human disease. The prevalence of resistant bacteria this analysis is in contrast with the results of another
in agriculture is influenced by antibiotic use in that setting. The impact model using an outcome-attribution approach, which pre-
of infection depends crucially on the capacity for sustained human to dicted an 410 926 excess days of illness annually in the
human (H2H) transmission. Arrows linking the two populations repre-
USA due to fluoroquinolone-resistant Campylobacter
sent: a) direct transmission of bacteria not adapted to transmission in
humans via the food chain (e.g Campylobacter, Salmonella) or direct
infections, attributed to fluoroquinolone use in animals
contact with animals; b) direct transmission of organisms already (Travers and Barza 2002). Subsequently, the United States
adapted to transmission in humans; c) transfer of resistance genes from Food and Drug Administration (FDA) conducted an
the agricultural setting into pathogens transmitting among humans. assessment of the human health impact of fluoroquino-

© 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5 241
Agriculture and antibiotic resistance Chang et al.

lone-resistant Campylobacter associated with the consump- as a result of a lower burden of disease in food animals.
tion of chicken, which was later used to support the with- This study has been criticized on several grounds, including
drawal of fluoroquinolone use in poultry in the USA. The that it overlooks the contribution of commensal-contami-
model estimated that between 4960 and 14 370 patients in nating flora, that it takes no account of individuals at
1998 were prescribed fluoroquinolone for fluoroquino- greater risk of disease, and crucially assumes no difference
lone-resistant Campylobacter infections attributed to between disease due to sensitive and resistant strains. It is
chicken (Bartholomew et al. 2003). The inconsistency in reasonable to suggest that this is not the case due to the dif-
these results may arise from differences in the modeling ficulty in treating resistant disease and concomitant risk of
approaches applied, the outcome units used and the genu- sequelae (Claycamp 2006). Moreover, the authors of this
ine uncertainty surrounding these estimates, reflected by study are in receipt of considerable industry funding, a
broad confidence intervals (IOM 1989; McEwen 2012). potential conflict of interest, and as a result we must note
However, these estimates make it clear that a direct that while superficially plausible this conclusion is reliant
impact on human health through resistant pathogens upon questionable assumptions and is counter to the over-
acquired from livestock is plausible. whelming weight of expert opinion (Marshall and Levy
While contaminated food can cause infections in the 2011).
context of outbreaks, the causative organisms do not typi-
cally transmit between humans, and this may limit the
Ongoing transmission in humans of resistant
impact of resistance. A comprehensive review of publicly
strains originating in livestock
available risk assessments in the USA, which have mainly
considered human infection via foodborne transmission Following exposure of humans-to-resistant organisms
rather than direct contact, indicated that the majority of from agriculture, there may be further spread within the
models focused on morbidity and quality of life as end- human population. Varying degrees of onward transmis-
points rather than mortality. Notably, no risk assessment sion within humans have been documented for different
model has found more than 100 annual deaths in the USA clonal types of methicillin-resistant Staphylococcus, origi-
caused by antibiotic use in food animals (McEwen 2012). nally acquired from livestock, such as ST398 in the Neth-
Among one of the studies that examined mortality, the erlands, CC93 in Denmark, and ST 130 in Europe
Institute of Medicine estimated that 40 deaths due to sal- (Armand-Lefevre et al. 2005; Harrison et al. 2013; Spoor
monellosis occur each year in the USA due to subtherapeu- et al. 2013). ST398 carriage in farmers appears to be
tic use of penicillin and tetracycline, although as in the transient; although they are readily colonized, ST398 does
examples of fluoroquinolone resistance in C. jejuni above, not spread easily to family members and the community
considerable uncertainty surrounds this estimate (IOM (Graveland et al. 2011). This suggests constraint in terms
1989). While these models can paint a telling picture of the of clearing the species barrier to transmit efficiently in
clinical impact of antibiotics use in food-producing ani- humans (or possibly a cost to resistance, although it
mals, a lack of concrete data, especially regarding the preva- should be noted that this has not impeded the emergence
lence of resistant strains in food-producing animals, make of numerous community acquired MRSA clones from
reliable estimates about the impact in human medicine elu- nonlivestock sources). However, several cases of human
sive. However, given what is available, the relatively low infections with ST398 without any contact with livestock
numbers indicate that antibiotic use in animals may not be or pig farms have been reported (Van der Mee-Marquet
as significant a public health problem in terms of direct et al. 2011). It is unclear how these individuals have
transmission, as has been alleged by numbers such as acquired these strains, whether through human-to-human
CDC’s ‘lower bound’ estimate of 23 000 annual deaths in contact, or through other exposure routes such as contact
the USA caused by resistant infections (Centers for Disease with pets, contaminated food, or environmental sources.
Control and Prevention 2013). In addition, very few studies This reflects poor sampling and limited data in the global
have examined the extent to which restrictions of agricul- surveillance of pathogens that colonize both humans and
tural antibiotics use would reduce the incidence of human animals (Fan et al. 2009). While surveillance systems
antibiotic-resistant infections caused by this particular based on clinical infections are capable of monitoring
mechanism. It has been suggested that the benefits of anti- occurrences of resistant pathogenic zoonotic bacteria such
biotics in preventing livestock disease result in an overall as Salmonella or Campylobacter species, zoonotic transfers
reduced risk to humans, that is, the withdrawal of antibiot- of commensals including Staphylococcus aureus and
ics could lead to an increase in human disease, albeit sus- Enterococcus usually go unnoticed. This possibility for
ceptible (Cox and Popken 2006). This conclusion is animal-to-human transmission, in the case of MRSA
dependent on the hypothesis that antibiotic use in animals ST398, brings heightened concerns about livestock as
leads to fewer pathogenic bacteria entering the food chain, potential reservoirs of zoonotic infections that may with

242 © 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5
Chang et al. Agriculture and antibiotic resistance

further evolution become adapted to circulation within outbreaks in the USA were mainly due to a particular viru-
the human population. lent and adhesive clone, CC17 (Willems et al. 2005). While
In general, there is still a marked gap in our knowledge a recent genomic analysis showed that the CC17 clone was
regarding the successful transfer of resistant bacteria from much more closely related to isolates collected from diverse
animals to humans, or vice versa. However, it is clear that animal sources than a small sample of human commensal
the species barrier has been breached multiple times in isolates (Lebreton et al. 2013), it is not possible to define
both directions, with human-adapted strains giving rise to the direction of transmission without the context of a lar-
animal-adapted lineages as well as vice versa (Weinert et al. ger sample. Importantly, none of the samples from the ani-
2012; Shepheard et al. 2013; Spoor et al. 2013). In the case mal population harbored either vanA or the alternative
of S. aureus, population genetic analyses have clearly shown resistance gene vanB, and so the specific relation of the
the existence of host-specific clonal lineages, implying that CC17 clonal complex to agricultural VRE is obscure.
its adaptive evolution has led to host restriction due to eco-
logical differences among different hosts (Fitzgerald 2012)
Horizontal transmission of resistance genes
MRSA ST398 derived from animals appears to be a case of
originating from livestock
mechanism 1 (Fig. 1), and uncertainty still remains about
its origin and its implications in public health. In contrast, Bacteria are known to be capable of transferring genes
CC97 appears to have contributed two distinct host-switch- between strains of the same species, and between species,
ing events from animals to humans, resulting in sustained via multiple mechanisms including but not limited to plas-
onward transmission and representing a clear example of mid transfer, phage transduction, and natural transforma-
mechanism 2 (Spoor et al. 2013). tion. Numerous examples are already known of resistance
Another example of mechanism 2 may be vancomycin- genes arising in one species, perhaps a relatively innocuous
resistant enterococci (VRE). The Enterococcus, which nor- commensal, but then being donated to a different, patho-
mally colonizes the gut, has acquired resistance to multiple genic species, with consequences for human health (Coffey
antibiotics over time, making the glycopeptide vancomycin et al. 1993; Dowson et al. 1993; Bowler et al. 1994). As a
one of the last therapeutic options. The epidemiology of result of such bacterial promiscuity, the most significant
VRE differs substantially between the USA and Europe. In role of antibiotic use in agriculture may be in facilitating
Europe, Enterococcus faecium carrying the vanA resistance the emergence of new resistance genes, which can then be
element for vancomycin resistance was commonly found in transferred into pathogens already adapted to transmission
the intestinal flora of farm animals as well as healthy peo- in humans. The argument here is that antibiotic use creates
ple, but carriage of VRE in farm animals and healthy peo- a breeding ground for the accumulation and movement of
ple was absent in the USA until 2008 (Bonten et al. 2001). resistance genes, and it is the existence of this reservoir that
This difference has been proposed to be due to the wide- poses the most serious threat to public health.
spread agricultural use of avoparcin, a glycopeptide used in As noted above, resistance elements in VRE found in ani-
Europe since the 1970s, but was never approved for use in mals and humans are similar, but it is not clear whether
the USA. Avoparcin, which confers cross-resistance to this represents transfer of the whole organism or the genes
vancomycin, has been shown to select for VRE in animals alone. Given the relatively high rates of recombination in
(Aarestrup et al. 1996). A large reservoir of VRE in animals the Enterococcus, resistance loci such as vanA and vanB can
presents many opportunities for human infection, and the be readily transmitted into other lineages or bacterial spe-
potential for resistant bacteria to colonize the human niche. cies, with potentially catastrophic effects (Willems et al.
Molecular epidemiologic studies have found that the VRE 2011). The conjugative transfer of high-level vancomycin
strains isolated from animals and humans are similar, as resistance to S. aureus has been demonstrated in vitro and
are the resistance elements (Woodford et al. 1998; Jensen in vivo (Noble et al. 1992). The possibility of such transfer
et al. 1999); hence it is clear that at least the potential for may not imply that when it occurs, it will lead to a large
transmission exists. clinical problem. To date, only 13 clinical VRSA isolates
The combined epidemiological and molecular data pro- have been reported in the USA, all related to intensive
vide indirect evidence that vancomycin resistance may have vancomycin use (Kos et al. 2012; Limbago et al. 2014).
arisen from agriculture and transferred to humans, but the This likely reflects biological constraints: An elegant study
evidence is not conclusive. Notably, although Europe pre- has shown that acquiring vancomycin resistance results in a
ceded the USA in the circulation of VRE in healthy com- fitness cost to the organism as estimated by growth rate
munity-dwellers, the clinical problem of VRE circulating in and that this is partially compensated by deletion of the
hospitals emerged first in the USA (Bonten et al. 2001) and mecA gene that confers methicillin resistance (Noto et al.
was associated with the increase in oral vancomycin use in 2008). The resulting return to methicillin susceptibility
US hospitals (Handwerger et al. 1993). Furthermore, VRE impedes onward transmission. An even more worrisome

© 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5 243
Agriculture and antibiotic resistance Chang et al.

case may be the recent emergence of the carbapenem-resis- additional animal-to-human transfer of resistant organisms
tant Enterobacteriaceae, with a highly transmissible resis- may have little impact on the prevalence of such organisms
tance gene for the New Delhi metallo-beta-lactamase in humans. Indeed, mathematical modeling studies have
(NDM). This gene was demonstrated to easily transfer argued that the greatest impact of agricultural antibiotic
from one species to another on different plasmids. Zoo- bans would be those for antibiotics causing resistance that
notic transmission of this pathogen from chickens has also is not yet clinically important in humans, precisely to pre-
been demonstrated (Wang et al. 2012). vent crossing of the species barrier (Lipsitch et al. 2002;
The transfer of resistance determinants from animal to Smith et al. 2002).
human through horizontal gene transfer is tremendously Even after antibiotic selection pressures have been
difficult to trace and quantify, and the role of animal removed or reduced through a ban, it is not clear that
reservoirs as the ultimate source of genes contributing to there will be a reduction in the presence of resistant
clinical resistance remains to be definitively proven. Putt- organisms in the environment. In the USA, no decline in
ing the problem in perspective, the phenomenon of the levels of ciprofloxacin resistance has been observed
resistance to naturally occurring antibiotics greatly pre- following the ban of fluoroquinolones in chickens (Price
cedes the development of agriculture. The metallo-beta- et al. 2007; Nannapaneni et al. 2009). While it is possible
lactamases (of which NDM is one especially worrisome that insufficient time has elapsed for trends to be detect-
example) have an extremely ancient origin, so ancient in able, it is also possible that fluoroquinolone-resistant
fact that no detectable sequence homology remains strains may remain in the environment in absence of
between different classes of these genes (Hall et al. 2004; antibiotic selective pressure (Humphrey et al. 2005). This
Bebrone 2007). Horizontal transfer is thought to have may be due to a minimal cost of resistance, or co-selec-
been important in the evolution of these enzymes, but tion for other adaptive features of the resistant organism.
whether that process has been accelerated by the use of A resistance determinant is inherited along with these
antibiotics in agriculture is not known. other adaptations, and so maintained along with them.
An example of such co-selection is loci that confer resis-
tance to heavy metals, which are often found in close
The impact of banning antibiotic use in agriculture
linkage with resistance loci (Baker-Austin et al. 2006).
The possibility of clinical resistance against our last line of Discharge of heavy metals into the environment will then
treatment as a result of antibiotic use in animals has led to select directly for the heavy metal resistance loci, and
large-scale antibiotic bans as precautionary measures. By indirectly for antibiotic resistance. An example is copper
2006, the European Union had banned all nonmedicinal resistance (encoded by the tcrB gene in glycopeptide-
antibiotics in animals. Antibiotic regulations have also resistant enterococci isolates from pigs (Hasman and Aa-
become more stringent in the USA. In 2005, the emergence restrup 2002). This genetic linkage was connected to the
of fluoroquinolone-resistant Campylobacter jejuni in the higher copper exposure in pigs through feed additives in
clinical setting in conjunction with fluoroquinolone Denmark. Unlike antibiotics, heavy metals are not
administration in animals prompted the FDA to ban fluor- degraded and can linger in the environment producing a
oquinolone use in poultry, although it remains unclear if long-term selection pressure (Stepanauskas et al. 2005).
the dramatic increase of fluoroquinolone-resistant strains While the extent to which co-selection of resistance
was due to fluoroquinolone use in livestock (Engberg et al. determinants will affect clinical resistance is unclear, the
2001; Gupta et al. 2004). Recently in December 2013, the greatest value in restricting antibiotic use, as is the case
FDA issued a voluntary policy that asks farms to cut rou- in human medicine, may not be in reversing resistance,
tine use of antibiotics and consult veterinarians before use. but in preventing further increases in prevalence and the
The ban in Europe created a large-scale natural experi- possibility that the relevant genes find themselves hori-
ment in which we could observe the effects of an agricul- zontally transferred into yet more pathogens.
tural antibiotics ban. After the avoparcin ban, some
reported that the prevalence of VRE in farm animals rap-
Discussion
idly declined (Pantosti et al. 1999; Aarestrup et al. 2001).
However, this did not translate directly to a decrease of The topic of agricultural antibiotic use is complex. As we
VRE in humans (Phillips 2007). If the ultimate origin of noted at the start, many believe that agricultural antibiotics
VRE truly lies in agriculture, then it is plausible that an have become a critical threat to human health. While the
avoparcin ban in the past could have prevented the emer- concern is not unwarranted, the extent of the problem may
gence of this resistance threat. Under this scenario, how- be exaggerated. There is no evidence that agriculture is ‘lar-
ever, once the species barrier was breached and ongoing gely to blame’ for the increase in resistant strains and we
transmission in humans established, reducing the risk of should not be distracted from finding adequate ways to

244 © 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5
Chang et al. Agriculture and antibiotic resistance

ensure appropriate antibiotic use in all settings, the most of antibiotic resistance, results from agricultural uses. To
important of which being clinical medicine. establish such a causal mechanism requires quantifying the
The debate about agricultural use of antibiotics is further relationship between quantity of antibiotics used, selection
complicated by its relation to other scientific, political, and exerted, and human health impact. We also have limited
economic issues. The desire for meat raised without antibi- knowledge of the consumption of antibiotics in different
otics is part of a larger consumer movement toward more animal species and similarly limited surveillance programs
‘natural’ and sustainable food sources (Silbergeld et al. to monitor and trace the emergence of resistance in animals
2008; Paarlberg 2010). While controversial in light of a (Perron et al. 2008). The limited data available make it
trend toward ‘evidence-based’ policymaking (Pugh 2002; hard to quantify the relationship between antibiotic use in
Phillips et al. 2004), this precautionary consumer attitude animals and the occurrence of clinical resistance. As we
may be desirable even if quantifiable harms of a particular have shown, while there is considerable evidence associat-
practice are limited. The most consequential impacts of ing antimicrobial use in agriculture with resistant patho-
agricultural antibiotic use, such as possibly leading to the gens in livestock and in the food supply, the evidence for
origin of loci conferring high-level resistance, are among human health risks directly attributable to agricultural anti-
the hardest to demonstrate conclusively, yet the absence of biotics runs the gamut from speculative to scant. There is
evidence does not mean that there is no effect (Marshall an urgent need for better studies that combine quality sur-
and Levy 2011). Pointing to agricultural sources for clinical veillance with good data on antibiotic usage in agriculture,
problems of antimicrobial resistance serves the interests of which is at present hard to come by, and any serious
vendors and prescribers of antibiotics for clinical medicine attempt to address this problem will require the agricul-
by implying that they bear a correspondingly lesser share of tural industry to be more forthcoming. In determining
the blame. On the other hand, the strong economic inter- whether regulations should be in place, we must weigh the
ests favoring continued use of antibiotics in agriculture need for scientific evidence of an inherently difficult-to-
have resulted in major funding for studies that have found measure phenomenon against the consequences of inac-
modest human health burdens from agricultural animal tion. Regulations of antibiotic use in agriculture will likely
use (http://www.cox-associates.com/health.htm, accessed do the most good if they are in place early enough to pre-
March 2, 2014), although the potential conflicts of interest vent the rise of antibiotic-resistant strains. Once these
are not always reported in the peer-reviewed publications strains have emerged, it might be only a matter of time
reporting these analyses (Cox and Popken 2006). before they cross the species barrier and adapt to living in
While overuse of antibiotics in any setting is a matter of humans, at which time there is very little regulation of agri-
concern, it remains important to determine what exactly culture can do to prevent their persistence in the clinical
constitutes ‘overuse.’ It is important that we simulta- setting. The greatest value of reducing agricultural antibi-
neously preserve effective antibiotics as long as possible, otic use now may be in maintaining a status quo that, while
but also that we continue to deploy them in the service of far from ideal, is greatly preferable to the alternative.
human and animal health. We could stop using antibiotics
altogether, and this would greatly reduce the selective pres-
Acknowledgements
sure they exert (although given the presence of naturally
occurring antibiotics, selective pressure would not be WPH and ML were supported by Award Number
removed entirely) but the consequences for public health U54GM088558 from the National Institute of General
would obviously be dire. A more pragmatic scenario is that Medical Sciences. The content is solely the responsibility of
the use of subtherapeutic amounts of agricultural antibiot- the authors and does not necessarily represent the official
ics as prophylaxis or in growth promotion be closely scruti- views of the National Institute of General Medical Sciences
nized. Low-dose antibiotics favor the emergence of or the National Institutes of Health. We acknowledge the
resistance, and this practice has also been condemned as a use of icons made by Freepik from flaticon.com in the
cover for poor standards of animal care (The Pew Charita- design of Fig. 1.
ble Trusts 2008).
In recent years, agricultural antibiotics have gained a tre-
Literature Cited
mendous amount of attention from the media. In deter-
mining how much antibiotic use is too much, we must Aarestrup, F. M. 1999. Association between the consumption of antimi-
turn to the things we do not know. From the proportion of crobial agents in animal husbandry and the occurrence of resistant
bacteria among food animals. International Journal of Antimicrobial
antibiotics by weight used in agriculture as opposed to
Agents 12:279–285.
human medicine (Hollis and Ahmed 2013), it does not fol-
Aarestrup, F. M., P. Ahrens, M. Madsen, L. V. Pallesen, R. L. Poulsen,
low that the majority of selective pressure on human and H. Westh 1996. Glycopeptide susceptibility among Danish
pathogens, let alone the majority of human health impact

© 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5 245
Agriculture and antibiotic resistance Chang et al.

Enterococcus faecium and Enterococcus faecalis isolates of animal and Graveland, H., J. A. Wagenaar, K. Bergs, H. Heesterbeek, and D. Hee-
human origin and PCR identification of genes within the vana cluster. derik 2011. Persistence of livestock associated MRSA CC398 in
Antimicrobial Agents and Chemotherapy 40:1938–1940. humans is dependent on intensity of animal contact. PLoS ONE 6:
Aarestrup, F. M., A. M. Seyfarth, H. D. Emborg, K. Pedersen, R. S. Hen- e16830.
driksen, and F. Bager 2001. Effect of abolishment of the use of antimi- Gupta, A., J. M. Nelson, T. J. Barrett, R. V. Tauxe, S. P. Rossiter, C. R.
crobial agents for growth promotion on occurrence of antimicrobial Friedman, K. W. Joyce et al. 2004. Antimicrobial resistance among
resistance in fecal Enterococci from food animals in Denmark. Anti- Campylobacter strains, United States, 1997-2001. Emerging Infectious
microbial Agents and Chemotherapy 45:2054–2059. Diseases 10:1102–1109.
Anderson, S. A., R. W. Yeaton Woo, and L. M. Crawford 2001. Risk Hall, B. G., S. J. Salipante, and M. Barlow 2004. Independent origins of
assessment of the impact on human health of resistant Campylobacter subgroup Bl + B2 and subgroup B3 metallo-beta-lactamases. Journal
jejuni from fluoroquinolone use in beef cattle. Food Control 12:13–25. of Molecular Evolution 59:133–141.
Armand-Lefevre, L., R. Ruimy, and A. Andremont 2005. Clonal compar- Handwerger, S., B. Raucher, D. Altarac, J. Monka, S. Marchione, K. V.
ison of Staphylococcus aureus isolates from healthy pig farmers, human Singh, B. E. Murray et al. 1993. Nosocomial outbreak due to Entero-
controls, and pigs. Emerging Infectious Diseases 11:711–714. coccus faecium highly resistant to vancomycin, penicillin, and gentami-
Baker-Austin, C., M. S. Wright, R. Stepanauskas, and J. V. McArthur cin. Clinical Infectious Diseases 16:750–755.
2006. Co-selection of antibiotic and metal resistance. Trends in Harrison, E. M., G. K. Paterson, M. T. G. Holden, J. Larsen, M. Stegger,
Microbiology 14:176–182. A. R. Larsen, A. Petersen et al. 2013. Whole genome sequencing iden-
Bartholomew, M. J., K. Hollinger, and D. Vose 2003. Characterizing the tifies zoonotic transmission of MRSA isolates with the novel mecA
risk of antimicrobial use in food animals: fluoroquinolone-resistant homologue mecC. EMBO Molecular Medicine 5:509–515.
campylobacter from consumption of chicken. In M. E. Torrence, and Hasman, H., and F. M. Aarestrup 2002. tcrB, a gene conferring transfer-
R. E. Isaacson, eds. Microbial Food Safety in Animal Agriculture: Cur- able copper resistance in Enterococcus faecium: occurrence, transfer-
rent Topics pp. 293–301. Iowa State Press, Ames, IA. ability, and linkage to macrolide and glycopeptide resistance.
Bebrone, C. 2007. Metallo-beta-lactamases (classification, activity, Antimicrobial Agents and Chemotherapy 46:1410–1416.
genetic organization, structure, zinc coordination) and their super- Hollis, A., and Z. Ahmed 2013. Preserving antibiotics, rationally. The
family. Biochemical Pharmacology 74:1686–1701. New England Journal of Medicine 369:2474–2476.
Bonten, M. J., R. Willems, and R. A. Weinstein 2001. Vancomycin-resis- Humphrey, T. J., F. Jørgensen, J. A. Frost, H. Wadda, G. Domingue, N.
tant Enterococci: why are they here, and where do they come from? C. Elviss, D. J. Griggs et al. 2005. Prevalence and subtypes of cipro-
The Lancet Infectious Diseases 1:314–325. floxacin-resistant Campylobacter spp. in commercial poultry flocks
Bowler, L. D., Q. Y. Zhang, J. Y. Riou, and B. G. Spratt 1994. Interspecies before, during, and after treatment with fluoroquinolones. Antimicro-
recombination between the penA genes of Neisseria meningitidis and bial Agents and Chemotherapy 49:690–698.
commensal Neisseria species during the emergence of penicillin resis- IOM, Institute of Medicine. 1988. Human Health Risks with the Sub-
tance in N. meningitidis: natural events and laboratory simulation. therapeutic Use of Penicillin Or Tetracyclines in Animal Feed. Com-
Journal of Bacteriology 176:333–337. mittee on Human Health Risk Assessment of Using Subtherapeutic
Centers for Disease Control and Prevention. 2013. Antibiotic Resistance Antibiotics in Animal Feeds. National Academy Press, Washington,
Threats in the United States, 2013. http://www.cdc.gov/drugresis- DC.
tance/threat-report-2013/ (accessed on 6 March 2014). Jensen, L. B., A. M. Hammerum, R. L. Poulsen, and H. Westh 1999.
Claycamp, H. G. 2006. Rapid benefit-risk assessments: no escape from Vancomycin-resistant Enterococcus faecium strains with highly similar
expert judgments in risk management. Risk Analysis 26:147–156; pulsed-field gel electrophoresis patterns containing similar Tn1546-
discussion 157–161. like elements isolated from a hospitalized patient and pigs in Den-
Coffey, T. J., C. G. Dowson, M. Daniels, and B. G. Spratt 1993. Horizon- mark. Antimicrobial Agents and Chemotherapy 43:724–725.
tal spread of an altered penicillin-binding protein 2B gene between Kennedy, D. 2013. Time to deal with antibiotics. Science 342:777.
Streptococcus pneumoniae and Streptococcus oralis. FEMS Microbiology Kos, V. N., C. A. Desjardins, A. Griggs, G. Cerqueira, A. Van Tonder, M.
Letters 110:335–339. T. G. Holden, P. Godfrey et al. 2012. Comparative genomics of vanco-
Cox, L. A., and D. A. Popken 2006. Quantifying potential human health mycin-resistant Staphylococcus aureus strains and their positions
impacts of animal antibiotic use: Enrofloxacin and Macrolides in within the clade most commonly associated with methicillin-resistant
chickens. Risk Analysis 26:135–146. S. aureus hospital-acquired infection in the United States. mBio 3:
Dowson, C. G., T. J. Coffey, C. Kell, and R. A. Whiley 1993. Evolution of e00112–12.
penicillin resistance in Streptococcus pneumoniae; the role of Strepto- Lebreton, F., W. van Schaik, A. M. McGuire, P. Godfrey, A. Griggs, V.
coccus mitis in the formation of a low affinity PBP2B in S. pneumoniae. Mazumdar, J. Corander et al. 2013. Emergence of epidemic multi-
Molecular Microbiology 9:635–643. drug-resistant Enterococcus faecium from animal and commensal
Engberg, J., F. M. Aarestrup, D. E. Taylor, P. Gerner-Smidt, and I. Na- strains. mBio 4:e00534–13.
chamkin 2001. Quinolone and macrolide resistance in Campylobacter Levy, S. B. 1997. Antibiotic resistance: an ecological imbalance. Ciba
jejuni and C. Coli: resistance mechanisms and trends in human iso- Foundation Symposium 207:1–9; discussion 9–14.
lates. Emerging Infectious Diseases 7:24–34. Limbago, B. M., A. J. Kallen, W. Zhu, P. Eggers, L. K. McDougal, and V.
Fan, J., M. Shu, G. Zhang, W. Zhou, Y. Jiang, Y. Zhu, G. Chen et al. S. Albrecht 2014. Report of the 13th vancomycin-resistant Staphylo-
2009. Biogeography and virulence of Staphylococcus aureus. PLoS coccus aureus isolate from the United States. Journal of Clinical Micro-
ONE 4:e6216. biology 52:998–1002.
Fitzgerald, J. R. 2012. Livestock-associated Staphylococcus aureus: origin, Lipsitch, M., R. S. Singer, and B. R. Levin 2002. Antibiotics in agricul-
evolution and public health threat. Trends in Microbiology 20:192– ture: when is it time to close the barn door? Proceedings of the
198.

246 © 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5
Chang et al. Agriculture and antibiotic resistance

National Academy of Sciences of the United States of America Silbergeld, E. K., J. Graham, and L. B. Price 2008. Industrial food animal
99:5752–5754. production, antimicrobial resistance, and human health. Annual
Marshall, B. M., and S. B. Levy 2011. Food animals and antimicrobials: Review of Public Health 29:151–169.
impacts on human health. Clinical Microbiology Reviews 24:718–733. Smith, D. L., A. D. Harris, J. A. Johnson, E. K. Silbergeld, and J. Glenn
McEwen, S. A. 2012. Quantitative human health risk assessments of anti- Morris 2002. Animal antibiotic use has an early but important impact
microbial use in animals and selection of resistance: a review of pub- on the emergence of antibiotic resistance in human commensal bacte-
licly available reports. Revue Scientifique et Technique (International ria. Proceedings of the National Academy of Sciences of the United
Office of Epizootics) 31:261–276. States of America 99:6434–6439.
Mellon, M., C. Benbrook, and K. L. Benbrook 2001. Hogging it: esti- Spoor, L. E., P. R. McAdam, L. A. Weinert, A. Rambaut, H. Hasman, F.
mates of antimicrobial abuse in livestock. Union of Concerned Scien- M. Aarestrup, A. M. Kearns et al. 2013. Livestock origin for a human
tists, UCS Publications, Cambridge, MA. pandemic clone of community-associated methicillin-resistant Staph-
Nannapaneni, R., I. Hanning, K. C. Wiggins, R. P. Story, S. C. Ricke, ylococcus aureus. mBio 4:00356–13.
and M. G. Johnson 2009. Ciprofloxacin-resistant Campylobacter per- Stepanauskas, R., T. C. Glenn, C. H. Jagoe, R. C. Tuckfield, A. H. Lindell,
sists in raw retail chicken after the fluoroquinolone ban. Food Addi- and J. V. McArthur 2005. Elevated microbial tolerance to metals and
tives & Contaminants Part A, Chemistry, Analysis, Control, Exposure antibiotics in metal-contaminated industrial environments. Environ-
& Risk Assessment 26:1348–1353. mental Science & Technology 39:3671–3678.
Noble, W. C., Z. Virani, and R. G. Cree 1992. Co-transfer of vancomycin The Pew Commission on Industrial Farm Animal Production, 2008. Put-
and other resistance genes from Enterococcus faecalis NCTC 12201 to ting Meat on the Table: Industrial Farm Animal Production in America.
Staphylococcus aureus. FEMS Microbiology Letters 72:195–198. The Pew Commission on Industrial Farm Animal Production. http://
Noto, M. J., P. M. Fox, and G. L. Archer 2008. Spontaneous deletion of www.pewtrusts.org/en/research-and-analysis/reports/2008/04/29/
the methicillin resistance determinant, mecA, partially compensates putting-meat-on-the-table-industrial-farm-animal-production-in-
for the fitness cost associated with high-level vancomycin resistance in america (accessed on 6 March 2014).
Staphylococcus aureus. Antimicrobial Agents and Chemotherapy Travers, K., and M. Barza 2002. Morbidity of infections caused by anti-
52:1221–1229. microbial-resistant bacteria. Clinical Infectious Diseases 34(Suppl 3):
Paarlberg, R. 2010. Food Politics: What Everyone Needs to Know. S131–S134.
Oxford University Press, New York, NY. Van der Mee-Marquet, N., P. Francßois, A.-S. Domelier-Valentin, F. Cou-
Pantosti, A., M. Del Grosso, S. Tagliabue, A. Macrı, and A. Caprioli lomb, C. Decreux, C. Hombrock-Allet, O. Lehiani et al. 2011. Emer-
1999. Decrease of vancomycin-resistant enterococci in poultry meat gence of unusual bloodstream infections associated with pig-borne-
after avoparcin ban. Lancet 354:741–742. like Staphylococcus aureus ST398 in France. Clinical Infectious Dis-
Perron, G. G., S. Quessy, and G. Bell 2008. A reservoir of drug-resistant eases 52:152–153.
pathogenic bacteria in asymptomatic hosts. PLoS ONE 3:e3749. Wang, Y., C. Wu, Q. Zhang, J. Qi, H. Liu, Y. Wang, T. He et al. 2012.
Phillips, I. 2007. Withdrawal of growth-promoting antibiotics in Europe Identification of New Delhi metallo-b-lactamase 1 in Acinetobacter
and its effects in relation to human health. International Journal of lwoffii of food animal origin. PLoS ONE 7:e37152.
Antimicrobial Agents 30:101–107. Weinert, L. A., J. J. Welch, M. A. Suchard, P. Lemey, A. Rambaut, and J.
Phillips, I., M. Casewell, T. Cox, B. De Groot, C. Friis, R. Jones, C. R. Fitzgerald 2012. Molecular dating of human-to-bovid host jumps
Nightingale et al. 2004. Does the use of antibiotics in food animals by Staphylococcus aureus reveals an association with the spread of
pose a risk to human health? A critical review of published data. domestication. Biology Letters 8:829–832.
The Journal of Antimicrobial Chemotherapy 53:28–52. Willems, R. J. L., J. Top, M. van Santen, D. A. Robinson, T. M. Coque,
Price, L. B., L. G. Lackey, R. Vailes, and E. Silbergeld 2007. The F. Baquero, H. Grundmann et al. 2005. Global spread of vancomycin-
persistence of fluoroquinolone-resistant Campylobacter in poultry resistant Enterococcus faecium from distinct nosocomial genetic com-
production. Environmental Health Perspectives 115:1035–1039. plex. Emerging Infectious Diseases 11:821–828.
Pugh, D. M. 2002. The EU precautionary bans of animal feed additive Willems, R. J. L., W. P. Hanage, D. E. Bessen, and E. J. Feil 2011. Popula-
antibiotics. Toxicology Letters 128:35–44. tion biology of gram-positive pathogens: high-risk clones for dissemi-
Scallan, E., R. M. Hoekstra, F. J. Angulo, R. T. Tauxe, M. A. Widdowson, nation of antibiotic resistance. FEMS Microbiology Reviews 35:872–
S. L. Roy, J. L. Jones, and P. M. Griffin 2011. Foodborne Illness 900.
Acquired in the United States - Major Pathogens. Emerging Infectious Witte, W. 1998. Medical consequences of antibiotic use in agriculture.
Diseases. 17:7–15. Science 279:996–997.
Shepheard, M. A., V. M. Fleming, T. R. Connor, J. Corander, E. J. Feil, Woodford, N., A.-M. A. Adebiyi, M.-F. I. Palepou, and B. D. Cookson
C. Fraser, and W. P. Hanage 2013. Historical zoonoses and other 1998. Diversity of vanA glycopeptide resistance elements in Entero-
changes in host tropism of Staphylococcus aureus, identified by cocci from humans and nonhuman sources. Antimicrobial Agents
phylogenetic analysis of a population dataset. PLoS ONE 8:e62369. and Chemotherapy 42:502–508.

© 2014 The Authors. Evolutionary Applications published by John Wiley & Sons Ltd 8 (2015) 240–5 247
Copyright of Evolutionary Applications is the property of Wiley-Blackwell and its content
may not be copied or emailed to multiple sites or posted to a listserv without the copyright
holder's express written permission. However, users may print, download, or email articles for
individual use.