Missing Microbes: How the Overuse of Antibiotics Is Fueling Our Modern Plagues (27 page)

And about ten years ago
C. diff
infections became more severe; more people died. What was going on? Analyses show that the strains had changed. A small segment of DNA just upstream from the toxin gene had been deleted. As a result, those strains spewed out more toxin, with all of their damaging effects.

Even more remarkable to me is that several different strains have different deletions, but all of them lead to greater toxin production. To a biologist, this means that extremely strong pressures are operating on
C. diff,
selecting hypertoxigenic strains over normal toxin producers. That several clones have mutated in parallel at the same time points to some common change in the environment. These same highly toxic clones are present in Europe and in North America, suggesting that hospital environments common to developed countries could be a factor. Indeed, hospitals are dangerous places.

What we did not foresee is how quickly
C. diff
infections could spread into the community. People like Peggy Lillis, who was not hospitalized, are becoming ill, and some are dying. Like a lion escaped from the zoo,
C. diff
has escaped the confines of the hospital and is now loose in the community. And the same clones, as passengers in someone’s body aboard a jet plane, have crossed the oceans and set up shop in new communities—no passport required. In the United States, at least 250,000 people are hospitalized each year for
C. diff
infections that they acquired there or at home, and 14,000 die as a result.

The same thing has happened with MRSA, the infection due to antibiotic-resistant
Staph
that felled the two football players whose stories I told earlier. Twenty years ago, MRSA was found almost exclusively in hospitals, causing infections like the one suffered by the professional football player after his knee surgery. But now people with no exposure to hospitals, like the young high school player, are becoming infected. More virulent MRSA strains are appearing. That the two crises—
C. diff
and MRSA—have such similar characteristics and have arisen at more or less the same time tells us that our human microbial ecology is undergoing dramatic changes.

These are chilling stories, but sadly they are a harbinger of worse to come. The spread of these pathogens outside their “natural” reservoir, the hospital, to the larger community and across oceans represents a grave threat to our health. Finding ways to stop the spread of these lethal microbes must be a top priority.

The Centers for Disease Control and Prevention issued a landmark report in September 2013 that gave the first overall picture of drug-resistant bacteria in the United States. It ranked eighteen microbes according to their threat level and named three as “urgent.” At the top of the list is a relatively new group of microbes called CRE, short for carbapenem-resistant enterobacteriaceae, which kill a high number of those infected and are resistant to essentially every antibiotic thrown at them. Moreover, CRE have the ability to spread resistance genes to other microbes by having microbial “sex” with them. CRE already have been identified in health facilities in forty-four states.
C. diff
and drug-resistant gonorrhea came in second and third on the list. MRSA was ranked as “serious,” with eighty thousand infections a year and eleven thousand deaths.

Dr. Tom Frieden, who heads the center, warned that “antimicrobial resistance is happening in every community, in every health care facility, and in medical practices throughout the country. At least 2 million people per year in the U.S. get infections that are resistant to antibiotics, and 23,000 die. This is what happens,” he said, “when microbes outsmart our best antibiotics.” Saying that we face “catastrophic consequences” from overusing antibiotics, he added, “The medicine cabinet may be empty for patients with life-threatening infections in coming months and years.”

*   *   *

We have a cabin in the Rocky Mountains. It sits on a ridge within a wide valley ringed by tall peaks. These are high mountains with snowcapped summits nine months a year and patches of ice still present in the depths of the summer. The trees make the mountains green until at the highest altitudes they peter out, and the summits are barren. It is a timeless, rugged, and magnificent landscape.

Until recently, the forests have been thick—too thick—replete with trees of all ages: majestic pines reaching straight to the sky like two-hundred-foot arrow shafts surrounded by fir, blue spruce, and groves of aspens. Everywhere you could see the baby trees coming up on their flanks, their branches almost tender with bright green, soft needles.

But about ten years ago, a pine beetle invaded our valley. It was probably always there but was held in check by intensely cold winters. Now, as the climate has warmed, the beetle has come back with a vengeance and is eating its way through the forest, ravaging entire mountainsides. Ninety percent of the trees are dead, waiting for a fire to turn them to ash.

What is happening to this Colorado landscape is a compelling metaphor for my missing-microbe hypothesis. Like the pine beetle, human pathogens surround us all the time, but their spread depends on certain conditions. How easily can they be transmitted from individual to individual? What is the host density and how susceptible are the hosts to attack? How healthy is the community? And what happens when the ecology changes, not in a forest, but inside a person? What happens when humans lose biodiversity? And what if that loss includes “keystone” species that keep ecosystems stable?

In the early 1950s, decades before
C. diff
was identified as causing antibiotic-associated diarrhea, Marjorie Bohnhoff and C. Phillip Miller conducted a series of experiments to determine the role of the
normal flora
—the term back then for our resident microbes—in fending off disease-causing bacteria. They believed it would be protective, and they tested their hypothesis by feeding mice
Salmonella enteriditis,
a species of
Salmonella
that causes disease in both mice and humans. When they gave the strain to normal mice, it took about one hundred thousand organisms to infect half of them. But if they first gave the mice a single oral dose of an antibiotic—streptomycin—and then several days later gave them
Salmonella
, it took only about three organisms to infect them. This isn’t a 10 or 20 percent difference; it’s a thirty thousand–fold difference. Welcome to the world of bacteria.

Miller and colleagues continued the work, showing that the effect wasn’t limited to streptomycin. Other antibiotics, including penicillin, did the same. Even if the last antibiotic dose had been weeks earlier, the animals still could be infected by fewer microbes. In the sixty years since these experiments, many other researchers have confirmed and extended the findings. At least in mice, exposure to any of a number of different antibiotics increases susceptibility to an infection that sometimes is lethal. But is the same phenomenon true for humans?

In 1985 there was a massive outbreak of
Salmonella
infections in Chicago. At least 160,000 people became ill and several died. What could cause an event that would affect so many in one locality? Usually there are two main culprits, water or milk. Chicago had a municipal water system that was tightly regulated and policed; it was not the likely suspect. Besides, some of the people who fell sick were never in the city; they lived in suburbs that had their own water systems.

Thus suspicions turned to milk, which careful investigation confirmed. In particular, drinking milk from one grocery chain, the ubiquitous “Supermarket A,” was implicated. Within days it became clear that milk from that chain was the source of the outbreak and that all of the milk was from their single large dairy. This dairy, a massive industrial facility with miles of pipes and huge vats, which I visited and inspected as an expert for the class of victims that brought a lawsuit, produced more than 1 million gallons of milk a week.

Most germane to our story is that the health department studied a group of fifty victims of the outbreak (cases) and fifty unaffected people (controls). They asked this simple question: Have you received any antibiotics in the month prior to becoming ill? They found that people who had taken antibiotics at any time during the month prior to the outbreak were five and a half times more likely to become ill than people who drank the milk but had not recently received any antibiotics.

Just as Bohnhoff and Miller showed in mice decades earlier, antibiotic exposure left people more susceptible to becoming ill from
Salmonella.
In the PAT experiments I described a few chapters ago, the mice were given their last doses of antibiotics on their fortieth day of life. But more than one hundred days later, we could still find evidence that their intestinal microbes were substantially perturbed.

It’s not likely that Chicagoans were warned by their doctors that taking antibiotics would increase their susceptibility to infections, specifically to
Salmonella
. Has any health-care professional ever told you that? But increased susceptibility to new infections is one of the hidden costs of antibiotic use.

We now are in a good position to address one of the main questions raised throughout this book: How do antibiotics exert long-term effects on our resident microbes? In an earlier era, we relied on “indicator” organisms to represent overall microbial populations. An indicator organism is one that is used to estimate the presence of other microbial populations. For example,
E. coli
in surface water is an indicator organism for broader fecal contamination.

In 2001, my colleague in Sweden and good friend Dr. Lars Engstrand invited me to join in a study of how antibiotics affect indicator bacteria found in the human gut and on human skin. We used common colonizing bacteria that are easy to grow in culture:
Enterococcus fecalis
for the GI tract and
Staphylococcus epidermidis
for the skin. We asked whether people receiving a macrolide antibiotic—in this case, clarithromycin—as part of a one-week regimen to eradicate
H. pylori
from the stomach would exhibit an increase in macrolide-resistant bacteria elsewhere inside and on their bodies.

Unfortunately, the experiment worked beautifully. Before the subjects received the antibiotic, they had very low numbers of macrolide-resistant
Enterococcus
and
Staph
epidermidis
, as did the control subjects, who were not treated at all. Things were different for the study subjects who received the antibiotic. Immediately following their treatment, the numbers of macrolide-resistant indicator organisms increased dramatically both in their feces and on their skin, but no such changes occurred in the untreated control subjects.

But our main question was how long would these blooms of antibiotic-resistant organisms last without any further macrolide exposure. The results were sobering. In the treated subjects but not in the controls, we found resistant
E. fecalis
three years later and resistant
S. epidermidis
four years later
,
which is when the respective studies ended, so we do not know how much longer the organisms would have persisted. I find it remarkable that a one-week course of an antibiotic can lead to persistence of resistant organisms more than three years later and in sites far away from the intended target of the antibiotic.

We also wanted to know whether the strains present at the beginning of the study were the same ones identified three years later, or if they had been replaced by new strains of the same species. Using DNA fingerprinting techniques, we found that in the beginning of the study each of the control subjects had a few different
Enterococcus
strains, which were mostly present three years later. However, in the treated group, the strains present before the treatment largely disappeared and were replaced by others. And over the course of the three-year study, strains with new fingerprints kept appearing. In other words, not only had we selected for resistance (which persisted), but we destabilized prior
Enterococcus
populations. We don’t know whether those new strains had been present the entire time as minor populations or whether they were newly acquired but, in any case, the week of antibiotic treatment had a long-term and totally unintended effect on the stability of the particular strains of our indicator organism.

From the type of study we conducted, we cannot tell whether such changes lead to illness. If there is an effect, I would predict that the risk in most people would be small under usual circumstances. But we don’t know the cumulative effect of many billions of doses of antibiotics given to hundreds of millions of people. Widespread treatments certainly enhance the pool of resistance genes, including those that can jump from our friendly bacteria to newly acquired pathogens. But the
Salmonella
experiments in mice, the Chicago outbreak, and the current epidemic of
C. diff
infections show us that antibiotic pretreatment increases susceptibility to pathogens. This is another hidden cost of changing our internal ecosystem.

*   *   *

It should be clear by now that even short-term antibiotic treatments can lead to long-term shifts in the microbes colonizing our bodies. A full recovery or bounce-back of healthy bacteria is in no way guaranteed, despite the long-held belief that such was the case. But that is not my only worry. I also fear that some of our residential organisms—what I think of as contingency species—may disappear altogether.

Recent research shows that people carry a small number of highly abundant species and a large number of much less common ones. For example, you may carry trillions of
Bacteroides
in your colon and only a thousand cells or fewer of other species. We are not sure how many rare, or contingency, species any of us has, but if you had only fifty or sixty cells of a particular type, it would be very difficult to detect them against the background of trillions of other bacteria.