The global rise in resistance to antibiotics is threatening the foundation of modern medicine. Just as the invention of penicillin and other antibiotics transformed health care in the 20th century, today, waning effectiveness means that at least 23,000 Americans die each year from drug-resistant bacteria. Soon, therapies we take for granted like elective surgery, cancer treatments, and the immunosuppressants that enable organ transplants could become too dangerous because of the risk of infection. Simple cuts could become fatal, just as they often were before antibiotics revolutionized medical care. In 2016 the Review on Antimicrobial Resistance—a report commissioned by then-British Prime Minister David Cameron—estimated that by 2050, 10 million lives would be lost each year globally due to antimicrobial-resistant infections, cumulatively costing 100 trillion US dollars.
Considering the urgency of the problem, we understand remarkably little about the causes of antimicrobial resistance. Rounds of finger-pointing, in which doctors blame the problem on overuse of antibiotics in agriculture and farmers blame overuse in medicine, have done little to shed light. Rapid scientific advances, though, are giving us hints on how to get to the root causes. They suggest we need a “One Health” approach: That is, we can only understand—and ultimately stop—antimicrobial resistance by looking at human, animal, and environmental health together. When these three areas of study fail to share and communicate, we can’t see the whole picture.
It’s complicated. The agricultural community blames medicine of causing antimicrobial resistance with its tendency to overprescribe. In many countries, after all, antibiotics are available over-the-counter without a prescription, essentially serving as substitutes for hygiene, sanitation, and medical care. The medical community in turn blames agriculture, particularly livestock producers, for the injudicious use of low-dose antibiotics as growth-promoting agents. It’s true that this has become common practice, though it started for a good reason. Animals need protein in order to build the muscle that becomes meat. Before World War II, US protein supplies for animal feed came from Japan (which provided fish meal) and Norway (which provided cod liver oil). Both of those sources disappeared during the war. Meat prices skyrocketed in the United States, in turn depriving poor people of an important source of protein and giving rise to a black market. The federal government was desperate to find a protein substitute for animal consumption that would help lower meat prices for humans. In 1949, scientists inadvertently discovered that feeding low levels of antibiotics to farm animals dramatically increased growth and reduced mortality rates in the animals studied, including chickens, pigs, and calves.
The results were so impressive that livestock producers around the world quickly embraced low-dose antibiotics. At the time, of course, there was much more concern about food security than antimicrobial resistance, which was still a largely a theoretical threat. Today the priorities have switched—although between climate change and the spread of zoonotic diseases like avian influenza and salmonella, food security should not be considered guaranteed.
Look at the whole picture. In fact, neither of the accusations are correct, or at least not entirely. When antimicrobial resistance is examined from a One Health perspective—taking into account human, animal, and environmental factors—interesting and unexpected findings emerge.
For example, in the late 1980s in Europe, there emerged a bacterium resistant to the antibiotic vancomycin, which was typically used to treat methicillin-resistant Staphylococcus aureus (MRSA) and Clostridium difficile infections in humans. Scientists dubbed the new strain VRE, for vancomycin-resistant Enterococcus faecium, and it afflicted both humans and animals. Vancomycin was chemically related to avoparcin, an antibiotic used to promote animal growth, so researchers initially implicated avoparcin as the cause of VRE.
Avoparcin had been in use in Europe since the 1970s, and Denmark relied on it heavily for food production, especially in pigs. After the emergence of VRE, the country banned avoparcin in 1995. Antimicrobial-resistance surveillance data showed an impressive 90 percent decrease in VRE in pigs and poultry in the twelve years after the ban. Among patients in Danish hospitals, though, the rate of VRE inexplicably increased.
Across Europe, the picture was fuzzier. The European Union as a whole banned avoparcin in 1997, but because reporting VRE rates in livestock was voluntary, not enough data was available to determine the effect of the ban on animal health. Post-ban hospital data showed just what you might expect among humans: Countries that used a lot of vancomycin on human patients had high rates of VRE in hospitals, and countries with low human vancomycin use had low VRE hospital rates. Unfortunately, pre-ban VRE rates in hospitalized patients were not available. However, if avoparcin had been the cause of VRE in humans, all countries should have experienced a decline in VRE rates after the ban, and that’s not what the data showed.
The incidence and distribution of VRE in the United States has been very different from that in Europe. The United States never approved avoparcin in the first place because of concerns about potential carcinogenicity, but it does allow its chemical cousin vancomycin to treat human ailments. VRE has been a serious problem in US hospitals. But according to data from the National Antimicrobial Resistance Monitoring System and the National Animal Health Monitoring System, rates of VRE in pigs and poultry are zero. In the United States, there is no evidence that VRE came from the farm animals.
More questions than answers. So what’s going on? Genomic data may offer some answers. Genes are pieces of DNA that make up part of an individual’s entire genome. (Imagine one strand of hair as a gene, and an entire head of hair as a genome.) Surveillance systems track antimicrobial-resistance genes—snippets of genomes—which is analogous to tracking people who have red-hair genes. We can make certain assumptions about redheads, for example that they possess the genes for red hair, but that doesn’t mean that all redheads share the same entire genome, or tell us what their whole genomes look like. Likewise, just because surveillance systems can monitor bacteria that carry genes for antimicrobial resistance doesn’t mean we know those bacteria’s entire genomes—and the entire genomes are what we really need to track and understand.
Before 2008, the cost of sequencing the genomes of whole organisms was prohibitively expensive. The price has since plummeted, allowing researchers to analyze the genomes of antibiotic-resistant bacteria like VRE. When they do, surprises appear. Genomic data suggests that VRE in European farm animals and hospitalized humans are distinct and separate populations, which might explain the different rates. Genomic studies also suggest that the VRE in hospitals might have a common evolutionary link with an Enterococcus faecium bacteria isolated in dogs. A VRE clone called VRE CC 17—CC means “clonal complex”—has been responsible for most of the hospital outbreaks. Vancomycin is not prescribed for dogs, but the antibiotic ampicillin is. The VRE CC 17 clone appears to have started out as ampicillin-resistant Enterococcus faecium CC 17, likely in dogs, and evolved into VRE CC 17 in humans. The VRE in farm animals came from a different lineage that has not been the primary culprit responsible for hospital outbreaks. In other words, companion animals could be playing a big role in the spread of antimicrobial resistance.
These studies tell us two things: First, antimicrobial-resistance surveillance systems must track entire genomes, not just resistance genes, in order to better understand how resistant microbes originate and spread. Second, these genomic surveillance systems must track companion animals as well as humans, which none currently do.
Further complicating matters, environmental metagenomic studies—which involve extracting DNA from soil samples—have found antibiotic resistance genes everywhere, including the Arctic and Antarctic, even in places where antibiotic-using humans have never set foot. In fact, antibiotic-resistance genes appear to be ancient and exist naturally in the environment.
In short, science demonstrates that no one is to “blame” for the rise of antimicrobial resistance. It is clear, though, that everyone overuses antibiotics, and that by doing so we have changed—and are changing—the planet’s bacteria. Data coming in from the medical, veterinary medical, and environmental disciplines, all of it contributing to a larger picture, shows that we will only get to the bottom of the problem with a One Health approach on a global scale.
This column was based on findings in the author’s recent book, One Health and the Politics of Antimicrobial Resistance, published by Johns Hopkins University Press.