Antibiotic resistance: myths and misunderstandings

A pig flying at the Minnesota state fair. Picture by TCS.

I’ve been involved in a few discussions of late on science-based sites around yon web on antibiotic resistance and agriculture–specifically, the campaign to get fast food giant Subway to stop using meat raised on antibiotics, and a graphic by CommonGround using Animal Health Institute data, suggesting that agricultural animals aren’t an important source of resistant bacteria. Discussing these topics has shown me there’s a lot of misunderstanding of issues in antibiotic resistance, even among those who consider themselves pretty science-savvy.

I think this is partly an issue of, perhaps, hating to agree with one’s “enemy.” Vani Hari, the “Food Babe,” recently also plugged the Subway campaign, perhaps making skeptics now skeptical of the issue of antibiotics and agriculture? Believe me, I am the farthest thing from a “Food Babe” fan and have criticized her many times on my Facebook page, but unlike her ill-advised and unscientific campaigns against things like fake pumpkin flavoring in coffee or “yoga mat” chemicals in Subway bread, this is one issue that actually has scientific support–stopped clocks and all that. Nevertheless, I think some people get bogged down in a lot of exaggeration or misinformation on the topic.

So, some thoughts. Please note that in many cases, my comments will be an over-simplification of a more complex problem, but I’ll try to include nuance when I can (without completely clouding the issue).

First–why is antibiotic resistance an issue?

Since the development of penicillin, we have been in an ongoing “war” with the bacteria that make us ill. Almost as quickly as antibiotics are used, bacteria are capable of developing or acquiring resistance to them. These resistance genes are often present on transmissible pieces of DNA–plasmids, transposons, phage–which allow them to move between bacterial cells, even those of completely different species, and spread that resistance. So, once it emerges, resistance is very difficult to keep under control. As such, much better to work to prevent this emergence, and to provide conditions where resistant bacteria don’t encounter selection pressures to maintain resistance genes (1).

In our 75-ish years of using antibiotics to treat infections, we’ve increasingly found ourselves losing this war. As bacterial species have evolved resistance to our drugs, we keep coming back with either brand-new drugs in different classes of antibiotics, or we’ve made slight tweaks to existing drugs so that they can escape the mechanisms bacteria use to get around them. And they’re killing us. In the US alone, antibiotic-resistant infections cause about 2 million infections per year, and about 23,000 deaths due to these infections–plus tens of thousands of additional deaths from diseases that are complicated by antibiotic-resistant infections. They cost at least $20 billion per year.

But we’re running out of these drugs. And where do the vast majority come from in any case? Other microbes–fungi, other bacterial species–so in some cases, that means there are also pre-existing resistance mechanisms to even new drugs, just waiting to spread. It’s so bad right now that even the WHO has sounded the alarm, warning of the potential for a “post-antibiotic era.”

This is some serious shit.

Where does resistance come from?

Resistant bacteria can be bred anytime an antibiotic is used. As such, researchers in the field tend to focus on two large areas: use of antibiotics in human medicine, and in animal husbandry. Human medicine is probably pretty obvious: humans get drugs to treat infections in hospital and outpatient settings, and in some cases, to protect against infection if a person is exposed to an organism–think of all the prophylactic doses of ciprofloxacin given out after the 2001 anthrax attacks, for example.

In human medicine, there is still much debate about 1) the proper dosing of many types of antibiotics–what is the optimal length of time to take them to ensure a cure, but also reduce the chance of incubating resistant organisms? This is an active area of research; and 2) when it is proper to prescribe antibiotics, period. For instance, ear infections. These cause many sleepless nights for parents, a lot of time off work and school, and many trips to clinics to get checked out. But do all kids who have an ear infection need antibiotics? Probably not. A recent study found that “watchful waiting” as an alternative to immediate prescription of antibiotics worked about as well as drug treatment for nonsevere ear infections in children–one data point among many that antibiotics are probably over-used in human medicine, and particularly for children. So this is one big area of interest and research (among many in human health) when it comes to trying to curb antibiotic use and employ the best practices of “judicious use” of antibiotics.

Another big area of use is agriculture (2). Just as in humans, antibiotics in ag can be used for treatment of sick animals, which is completely justifiable and accepted–but there are many divergences as well. For one, animals are often treated as a herd–if a certain threshold of animals in a population become ill, all will be treated in order to prevent an even worse outbreak of disease in a herd. Two, antibiotics can be, and frequently are, used prophylactically, before any disease is present–for example, at times when the producer historically has seen disease outbreaks in the herd, such as when animals are moved from one place to another (moving baby pigs from a nursery facility to a grower farm, as one example). Third, they can be used for growth promotion purposes–to make animals fatten up to market weight more quickly.  The latter is, by far, the most contentious use, and the “low hanging fruit” that is often targeted for elimination.

From practically the beginning of this practice, there were people who spoke out against it, suggesting it was a bad idea, and that the use of these antibiotics in agriculture could lead to resistance which could affect human health. A pair of publications by Stuart Levy et al. in 1976 demonstrated this was more than a theoretical concern, and that antibiotic-resistant E. coli were indeed generated on farms using antibiotics, and transferred to farmers working there. Since this time, literally thousands of publications on this topic have demonstrated the same thing, examining different exposures, antibiotics, and bacterial species. There’s no doubt, scientifically, that use of antibiotics in agriculture causes the evolution and spread of resistance into human populations.

Why care about antibiotic use in agriculture?

A quick clarification that’s a common point of confusion–I’m not discussing antibiotic *residues* in meat products as a result of antibiotic use in ag (see, for example, the infographic linked above). In theory, antibiotic residues should not be an issue, because all drugs have a withdrawal period that farmers are supposed to adhere to prior to sending animals off to slaughter. These guidelines were developed so that antibiotics will not show up in an animal’s meat or milk. The real issue of concern for public health are the resistant bacteria, which *can* be transmitted via these routes.

Agriculture comes up many times for a few reasons. First, because people have the potential to be exposed to antibiotic-resistant bacteria that originate on farms via food products that they eat or handle. Everybody eats, and even vegetarians aren’t completely protected from antibiotic use on farms (I’ll get into this below). So even if you’re far removed from farmland, you may be exposed to bacteria incubating there via your turkey dinner or hamburger.

Second, because the vast majority of antibiotic use, by weight, occurs on farms–and many of these are the very same antibiotics used in human medicine (penicillins, tetracyclines, macrolides). It’s historically been very difficult to get good numbers on this use, so you may have seen numbers as high as 80% of all antibiotic use in the U.S. occurs on farms. A better number is probably 70% (described here by Politifact), which excludes a type of antibiotic called ionophores–these aren’t used in human medicine (3). So a great deal of selection for resistance is taking place on farms, but has the potential to spread into households across the country–and almost certainly has. Recent studies have demonstrated also that resistant infections transmitted through food don’t always stay in your gut–they can also cause serious urinary tract infections and even sepsis. Studies from my lab and others (4) examining S. aureus have identified livestock as a reservoir for various types of this bacterium–including methicillin-resistant subtypes.

How does antibiotic resistance spread?

In sum–in a lot of different ways. Resistant bacteria, and/or their resistance genes, can enter our environment–our water, our air, our homes via meat products, our schools via asymptomatic colonization of students and teachers–just about anywhere bacteria can go, resistance genes will tag along. Kalliopi Monoyios created this schematic for the above-mentioned paper I wrote earlier this year on livestock-associated Staphyloccocus aureus and its spread, but it really holds for just about any antibiotic-resistant bacterium out there:

And as I noted above, once it’s out there, it’s hard to put the genie back in the bottle. And it can spread in such a multitude of different ways that it complicates tracking of these organisms, and makes it practically impossible to trace farm-origin bacteria back to their host animals. Instead, we have to rely on studies of meat, farmers, water, soil, air, and people living near farms in order to make connections back to these animals.

And this is where even vegetarians aren’t “safe” from these organisms. What happens to much of the manure generated on industrial farms? It’s used as fertilizer on crops, bringing resistant bacteria and resistance genes along with it, as well as into our air when manure is aerosolized (as it is in some, but not all, crop applications) and into our soil and water–and as noted below, antibiotics themselves can also be used in horticulture as well.

So isn’t something being done about this? Why are we bothering with this anymore?

Kind of, but it’s not enough. Scientists and advocates have been trying to do something about this topic since at least 1969, when the UK’s Swann report on the use of Antibiotics in Animal Husbandry and Veterinary Medicine was released. As noted here:

One of its recommendations was that the only antimicrobials that should be permitted as growth promotants in animals were those that were not depended on for therapy in humans or whose use was not likely to lead to resistance to antimicrobials that were important for treating humans.

And some baby steps have been made previously, restricting use of some important types of antibiotics. More recently in the U.S., Federal Guidelines 209 and 213 were adopted in order to reduce the use of what have been deemed “medically-important” antibiotics in the livestock industry. These are a good step forward, but truthfully are only baby steps. They apply only to the use of growth-promotant antibiotics (those for “production use” as noted in the documents), and not other uses including prophylaxis. There also is no mechanism for monitoring or policing individuals who may continue to use these in violation of the guidelines–they have “no teeth.” As such, there’s concern that use for growth promotion will merely be re-labeled as use for prophylaxis.

Further, even now, we still have no data on the breakdown of antibiotic use in different species. We know over 32 million pounds were used in livestock in 2013, but with no clue how much of that was in pigs versus cattle, etc.

We do know that animals can be raised using lower levels of antibiotics. The European Union has not allowed growth promotant antibiotics since 2006. You’ll read different reports of how successful that has been (or not); this NPR article has a balanced review. What’s pretty well agreed-upon is that, to make such a ban successful, you need good regulation and a change in farming practices. Neither of these will be in place in the U.S. when the new guidance mechanisms go into place next year–so will this really benefit public health? Uncertain. We need more.

So this brings me back to Subway (and McDonald’s, and Chipotle, and other giants that have pledged to reduce use of antibiotics in the animals they buy). Whatever large companies do, consumers are demonstrating that they hold cards to push this issue forward–much faster than the FDA has been able to do (remember, it took them 40 freaking years just to get these voluntary guidelines in place). Buying USDA-certified organic or meat labeled “raised without antibiotics” is no 100% guarantee that you’ll have antibiotic-resistant-bacteria-free meat products, unfortunately, because contamination can be introduced during slaughter, packing, or handling–but in on-farm studies of animals, farmers, and farm environment, studies have typically found reduced levels of antibiotic-resistant bacteria on organic/antibiotic-free farms than their “conventional” counterparts (one example here, looking at farms that were transitioning to organic poultry farming).

Nothing is perfect, and biology is messy. Sometimes reducing antibiotic use takes a long time to have an impact, because resistance genes aren’t always quickly lost from a population even when the antibiotics have been removed. Sometimes a change may be seen in the bacteria animals are carrying, but it takes longer for human bacterial populations to change. No one is expecting miracles, or a move to more animals raised antibiotic-free to be a cure-all. And it’s not possible to raise every animal as antibiotic-free in any case; sick animals need to be treated, and even on antibiotic-free farms, there is often some low level of antibiotic use for therapeutic purposes. (These treated animals are then supposed to be marked and cannot be sold as “antibiotic-free”). But reducing the levels of unnecessary antibiotics in animal husbandry, in conjunction with programs promoting judicious use of antibiotics in human health, is a necessary step. We’ve waited too long already to take it.


(1) Though we know that, in some cases, resistance genes can remain in a population even in the absence of direct selection pressures–or they may be on a cassette with other resistance genes, so by using any one of those selective agents, you’re selecting for maintenance of the entire cassette.

(2) I’ve chosen to focus on use in humans & animal husbandry, but antibiotics are also used in companion animal veterinary medicine and even for aquaculture and horticulture (such as for prevention of disease in fruit trees). The use in these fields is considerably smaller than in human medicine and livestock, but these are also active areas of research and investigation.

(3) This doesn’t necessarily mean they don’t lead to resistance, though. In theory, ionophores can act just like other antibiotics and co-select for resistance genes to other, human-use antibiotics, so their use may still contribute to the antibiotic resistance problem. Studies from my lab and others have shown that the use of zinc, for instance–an antimicrobial metal used as a dietary supplement on some pig farms, can co-select for antibiotic resistance. In our case, for methicillin-resistant S. aureus.

(4) See many more of my publications here, or a Nature profile about some of my work here.


Is there such a thing as an “evolution-proof” drug?

Eleven years ago, two scientists made a bet. One scientist wagered that a new type of antimicrobial agent, called antimicrobial peptides, would not elicit resistance from bacterial populations which were treated with the drugs. Antimicrobial peptides are short proteins (typically 15-50 amino acids in length) that are often positively charged. They are also a part of our body’s own innate immune system, and present in other species from bacteria to plants. It is thought that these peptides work primarily by disrupting the integrity of the bacterial cell, often by poking holes in them. Sometimes they work with the host to ramp up the immune response and overwhelm the invading microbe. Because the peptides are frequently targeted at the bacterial cell wall structure, it was thought that resistance to these drugs would require a fundamental change in membrane structure, making it an exceedingly rare event. Therefore, these antimicrobial peptides might make an excellent weapon in the fight against multiply drug-resistant bacteria.

Additionally, the remarkable diversity of these peptides, combined with the presence of multiple types of peptides with different mechanisms of action present at the infection site, rendered unlikely the evolution of resistance to these molecules (or so some reasoning went). However, evolutionary biologists have pointed out that therapeutic use of these peptides would differ from natural exposure: concentration would be significantly higher, and a larger number of microbes would be exposed. Additionally, resistance to these peptides has been detailed in a few instances. For example, resistance to antimicrobial peptides has been shown to be essential for virulence in Staphylococcus aureus and Salmonella species, but we didn’t *witness* that resistance develop–therefore, it might simply be that those species have physiological properties that render them naturally resistant to many of these peptides, and were never susceptible in the first place.

The doubter of resistance, and the bet instigator, was Michael Zasloff of Georgetown University, who wrote in a 2002 review of antimicrobial peptides:

Studies both in the laboratory and in the clinic confirm that emergence of resistance against antimicrobial peptides is less probable than observed for conventional antibiotics, and provides the impetus to develop antimicrobial peptides, both natural and laboratory conceived, into therapeutically useful agents.

Certainly in the short term, resistance may be unlikely to evolve for reasons described above. However, if these peptides are used over an extended period of time, could the mutations necessary to confer resistance accumulate? This was the question asked in a new study by Dr. Zasloff along with colleagues Gabriel Perron and Graham Bell. Following publication of his 2002 paper where he called evolution of resistance to these peptides “improbable,” Bell challenged Zasloff to test this theory. Zasloff took him up on the offer, and they published their results in Proceedings of the Royal Society

The result?

Zasloff had egg on his face. Resistance not only evolved, but it evolved independently in almost every instance they tested (using E. coli and Pseudomonas species), taking only 600-700 generations–a relative blip in microbial time. Oops.

Well, everything old is new again. A very similar claim has been making the rounds recently, originating from the press release for a new paper claiming to have found bacteria’s “Achilles’ heel,” advancing the claim that “Because new drugs will not need to enter the bacteria itself, we hope that the bacteria will not be able to develop drug resistance in future.”  A grand claim, but history suggests otherwise. It was argued that bacteria could not evolve resistance to bacteriophage, as the ancient interaction between viruses and their bacterial hosts certainly must have already exploited and overcome any available defense. Now a plethora of resistance mechanisms are known.

Alexander Fleming, who won the 1945 Nobel Prize in Physiology or Medicine, tried to sound the warning that the usefulness of antibiotics would be short-lived as bacteria adapted, but his warnings were (and still are?) largely ignored. There is no “magic bullet;” there are only temporary solutions, and we should have learned by now not to underestimate our bacterial companions.

Part of this post previously published here.

The impact of HIV on Drug-Resistant Tuberculosis

Second of five student guest posts by Nai-Chung N. Chang

Tuberculosis (TB) is a major disease burden in many areas of the world. As such, it was declared a global public health emergency in 1993 by the World Health Organization (WHO). It is a bacterial disease that is transmitted through the air when an infected individual coughs, sneezes, speaks, or sings. However, not all individuals who contract the disease will display symptoms. This separates the infected into two categories, latent and active. Latent individuals are non-infectious and will not transmit the disease, whereas active individuals are able to transmit the disease.

TB is a significant concern in patients diagnosed with HIV, since individuals diagnosed with HIV and latent forms of TB infection is more likely to develop the disease, then the HIV negative individuals. In addition, in people living with HIV, TB is one of the leading causes of death. (CDC, 2012) The fact that latent forms of the disease are capable of becoming fully active forms given the right stimulus represents a high risk to individuals living in poor conditions, which is widely present in developing nations. It is of even greater concern to individuals who have immune system diseases, such as HIV. Individuals with latent TB infection depend on robust immune system responses to prevent the infection from going into active form. HIV and similar diseases targets and weakens immune systems so that the response to infections becomes weaker, providing increased risk of TB infections and the activation of latent forms.

TB is a major concern not only because of its status as a global epidemic. While there are many forms of prevention and treatment for the disease, such as antibiotics and vaccine, these treatments are not overly effective in combating and controlling the spread of the disease. TB is widespread and has a high chance of becoming resistant to any treatment that it is exposed to, especially antibiotics and other chemotherapeutic drugs such as isoniazid. Several of these strains already exist and each has varying levels of resistance, including Multidrug-Resistant (MDR) and Extensively Drug-Resistant (XDR). MDR is a strain that is resistant to two of the most often used and potent TB drugs, isoniazid and rifampin; whereas XDR is MDR strains that have developed resistance to any fluoroquinolone and at least one of three second-line drugs such as kanamycin or capreomycin. Also, the vaccine that has been developed for preventing TB is not overly protective, and sometimes fails to protect against infection. (CDC, 2012) The vaccine is not designed to prevent the infection of TB; instead, it is aimed towards boosting and speeding up the immune system response to any new infection so that the infected individual remains in latent forms. (Russell, et al., 2010)

The increasing trends in the resistance of TB to various treatments is a serious concern as it have major impacts in controlling the spread of the disease in many regions. This condition worsens with MDR and XDR TB. With regular, normal strains of TB, latent and early infections could be combated and controlled by a successful chemotherapeutic treatment even in patients with immune system diseases. However, with MDR and XDR TB, the strains are able to fully develop in an individual with weakened immune system, as evident in areas where incidence of TB and HIV is high, such as South Africa. (O’Donnell, et al., 2013) For cases with MDR and XDR strains, the weakened immune systems are not potent enough to prevent infections or keep them in latent form. Additionally, the active forms of these strains are resistant to common, and in some cases, advanced treatments.

With the increasing development of drug-resistant TB, the most effective way to combat TB is not only through vaccines and treatments. Instead, strict public health policy is needed to properly maintain control and combat the spread of TB. With a well-structured public health system, we can ensure that the long treatment of TB is complete, since most of the increase in the resistance to treatment often results from issues during treatment. Events such as patient non-compliance to the treatment and inadequate health-care supervision can all result in the development of new strains of the bacteria that have developed resistance to the treatments that was used. (Russell, et al., 2010) Also, a well-structured public health system can maintain better supply and quality of drugs throughout the treatment process, as well as the prevention and detection of possible new drug resistant strains. More importantly, it can maintain better surveillance and ensure patient compliance during the treatment process, which would help in reducing the development of drug resistant strains. The surveillance systems can also target comorbid diseases such as HIV to reduce risk factors for activating latent forms of the disease in patients with HIV and similar diseases.

CDC, 2012. Tuberculosis (TB). [Online]
Available at:
[Accessed 13 2 2013].
O’Donnell, M. R. et al., 2013. Treatment Outcomes for Extensively Drug-Resistant Tuberculosis and HIV Co-infection. Emerging Infectious Disease [Internet], 19(3).
Russell, D. G., Barry 3rd, C. E. & Flynn, J. L., 2010. Tuberculosis: What We Don’t Know Can, and Does, Hurt Us. Science, 328(5980), pp. 852-856.

MRSA in pork products: does the “antibiotic-free” label make a difference?

Back in November, I blogged about one of our studies, examining methicillin-resistant Staphylococcus aureus (MRSA) in Iowa meat products. In that post, I mentioned that it was one of two studies we’d finished on the subject. Well, today the second study is out in PLoS ONE (freely available to all). In this study, we focused only on pork products, and included 395 samples from Iowa, Minnesota, and New Jersey. We also looked at not only conventional meats, but also “alternative” meat products. Most of the latter were products labeled “raised without antibiotics” or “raised without antibiotic growth promotants”–in the markets we tested, very few USDA-certified organic products were available unfrozen, and we were looking for fresh meat products.

In our previous paper, we found MRSA on 1.2% of 165 meat samples. In the current study, we found a higher prevalence–6.6% of 395 samples were contaminated with MRSA. (More about the differences in methods between our two studies later). Interestingly, we didn’t find a statistically significant difference in MRSA prevalence on conventional versus alternative pork products–a finding that surprised me, as it contradicts what we’ve found to date looking at the sources of this meat–conventional versus “alternative” pig farms. Other groups have also found differences on-farm versus on-meat: a 2011 study looking at feedlot cattle didn’t find any MRSA in animal samples, though the same group found MRSA in beef products. So, our disparate findings between farms and meat samples are not unheard-of. However, even though our sample size was larger than other U.S. studies to date, it was still fairly small overall–300 conventional and 95 alternative pork samples over a 4-week sampling period from 3 states, so larger multi-state studies are needed to further examine this angle.

It also suggests that we need processing plants and packing companies to work with us to determine where products are being contaminated–because while there may be arguments about the public health importance of MRSA on meats (or lack thereof), it’s very likely that if S. aureus are ending up on meat products, other pathogens are as well.

What does the molecular typing tell us, speaking of contamination source? We carried out analyses on all the MRSA and found that the most common type of MRSA was ST398, the “livestock” strain that we previously found on pig farms in the U.S. We also found two “human” types were common: USA300 (a “community-associated” strain) and USA100 (typically considered a “hospital-associated” strain). In the simplest analysis of these findings, these molecular types (a combination of “human” and “pig” strains) suggests that MRSA on raw pork products may be coming both from farms and from food handlers. However, in real life, it’s not quite so straightforward. USA100 types have also been found in live pigs. So has USA300. As such, the source of contamination and relative contributions of live pigs versus human meat handlers currently isn’t certain.

Within the MRSA strains, we found high levels of antibiotic resistance, similar to what was reported in the recent Waters et al. study. In ours, 76.9% were resistant to two or more antibiotics and 38.5% were resistant to three or more antibiotics tested. (I’ll note that we only had funding to test the MRSA–we weren’t able to do these tests on all the methiciin-susceptible strains).

Did MRSA prevalence increase in the period between our first study (spring 2009) and this one (late summer/fall 2010)? I doubt it. For this paper, we used a different sampling method, adding the samples to a sterile stomacher bag so that the entire sample was immersed in the culture medium; for the first paper we used external swabbing and so likely didn’t capture as many bacteria. This current study more likely represents the “true” MRSA prevalence. But–all isolates were only called as positive/negative, and we didn’t measure the number of bacteria on each piece of meat. So, there theoretically could have been just a few colonies of MRSA on the entire piece of meat, and that would have been called a positive sample, while another meat product covered with hundreds of MRSA would have been put in the same category. Therefore, more subtle differences may exist that we didn’t pick up in this study, but we will examine in other ongoing studies.

So–what’s the take-home here? Don’t assume that any meat product is contamination-free, and always use good food handling/cooking practices when dealing with raw meats. As far as the titular question, well, we’re still hashing that one out.


Hanson et al. Prevalence of Staphylococcus aureus and methicillin-resistant Staphylococcus aureus (MRSA) on retail meat in Iowa.

Waters et al. Multidrug-Resistant Staphylococcus aureus in US Meat and Poultry.

O’Brien et al. MRSA in conventional and alternative retail pork products.

Lin et al. Evidence of multiple virulence subtypes in nosocomial and community-associated MRSA genotypes in companion animals from the upper midwestern and northeastern United States.

Weese et al. The Prevalence of Methicillin-Resistant Staphylococcus aureus Colonization in Feedlot Cattle.

Weese et al. Detection and quantification of methicillin-resistant Staphylococcus aureus (MRSA) clones in retail meat products.