Missing Microbes: How the Overuse of Antibiotics Is Fueling Our Modern Plagues (25 page)

Rardi suggested that Genia start a gluten-free diet, to see if it helped. She did, and almost immediately her symptoms went away. For the first time in months, she didn’t have abdominal pain. Unfortunately, people with celiac disease face a serious problem: gluten is everywhere. One night, the sharp pain came back. She realized that she had eaten soy sauce at a restaurant, which is when she discovered that soy sauce often contains gluten.

After that she avoided gluten and felt great for a month. Then one day as she was traveling on the interstate highway back to Boston and stopped for fast food, she ordered french fries. Again, within an hour, abdominal pain struck her as severely as it had when they thought her appendix might burst. Later she learned that those french fries contained gluten.

Nowadays Genia is scrupulous about what she eats and has been almost entirely free of symptoms since the french fries episode. As a scientist, I can’t prove that Genia has celiac disease, but her symptoms are certainly consistent, and a more recent blood test showed for the first time that she has elevated levels of gluten antibodies. What piques my interest is all of those courses of amoxicillin that we gave her as a child, and then the courses of metronidazole for her “Giardia” infections. It suggests to me that she had major early-life disruptions in her resident microbes that led to her asthma and mango allergies. Later on, metronidazole was the coup de grâce, eliminating some population of microbes in her gut that instructed her immune cells to suppress certain allergies, including gluten.

Not long ago, a group of colleagues whose work focuses on celiac disease asked me to help them analyze some data collected in Sweden. Drs. Karl Marild and Jonas Ludvigsson had obtained records on thousands of people diagnosed with celiac disease, people with conditions similar to celiac disease (I’ll call it almost-celiac, perhaps like Genia’s case), and healthy controls. Plus they were able to access nationwide pharmacy records.

The main finding: people who recently had developed celiac disease were about 40 percent more likely to have been prescribed antibiotics in the preceding months compared to those who didn’t. This was true for people who definitely had celiac disease, for those who probably had celiac disease (60–90 percent more likely), for men, for women, for people of all ages, and for essentially every antibiotic examined individually. For those who were prescribed more courses, the risk was stronger. As with the diabetes studies, this kind of consistency was very important; it was not just a single finding. Most interesting to me is that metronidazole (the same drug that Genia was given repeatedly), which has major effects on gut bacteria, had the greatest association with celiac disease. People who were prescribed it ran more than twice the risk of getting celiac disease compared to those who were not recently prescribed any antibiotics.

To be sure, these studies show only an association between antibiotic use and celiac disease. We don’t even know whether those who were prescribed antibiotics took their medication, but we assume that people generally did, whether or not they later developed celiac disease. We also can’t prove what is called the direction of causality. One possibility is direct causality: exposure to antibiotics brings out the tendency for celiac disease. Another possibility is reverse causality: people already have celiac disease, and their doctors give them antibiotics to treat their symptoms without realizing what they have. At this point, we cannot distinguish which is correct, but clearly the first hypothesis is consistent with our work and with Genia’s case.

I was invited to join another analysis of celiac disease by the same group of experts who now were led by Dr. Ben Lebwohl at Columbia University. The question was: Is there any relationship between the presence of
H. pylori
in the stomach and celiac disease? We know that
H. pylori
is disappearing just as celiac disease is becoming more common. Could this suggest a protective relationship?
H. pylori
, when present, is almost always acquired very early in life before celiac disease develops. We also know that
H. pylori
helps suppress immune and allergic responses through its recruitment of T-reg cells, the cells that tone down and turn off immune reactions. So could the extinction of
H. pylori
be contributing to this modern plague as well?

To find out, the Columbia group, working with a team of pathologists at a large national reference laboratory in Texas, investigated more than 136,000 people who had undergone upper gastrointestinal endoscopy for a variety of reasons. As part of their routine analyses, the pathologists looked for
H. pylori
in stomach biopsies and characterized inflammation in the duodenum. Since signs of celiac disease can be detected microscopically in the duodenum, Ben and his colleagues realized that they had a way to link celiac disease to the presence of
H. pylori
in the stomach. Would people who had
H. pylori
be more likely to have celiac disease or less likely, or would there be no association?

Because the study was performed in the United States, the overall rate of
H. pylori
was very low. Nevertheless the rate in patients who had signs of celiac disease was only 4.4 percent as opposed to 8.8 percent in those without. Because of the large number of subjects, we asked whether the same relationship was true of the subjects in the thirty-seven individual states from which the specimens had come, and indeed it was in every single state studied. The reciprocal relationship held in both men and women as well as in each age group studied. These consistencies suggest biological significance.

It is plausible that celiac disease is increasing because the microbes that protect against allergic responses are disappearing. Both stomach bacteria (
H. pylori
) and intestinal bacteria (those susceptible to metronidazole and/or other antibiotics) may provide some protection against celiac disease. People who have
H. pylori
in their stomachs can still develop celiac disease; it just may be less likely. Moreover, people born by C-section also face an increased risk. With this knowledge, we might one day be able to find the individually protective organisms and either prevent their loss or give them back to prevent celiac disease or to treat it.

Another condition to consider in the light of missing microbial diversity is what is called inflammatory bowel disease (IBD), a group of chronic, relapsing disorders of the intestine. IBD manifests in two main types, ulcerative colitis and Crohn’s disease, which partially overlap but have different pathology.

Ulcerative colitis affects just the colon and is most often limited to the most superficial layers of the bowel wall. People often bleed from their rectum, have severe diarrhea, lose weight, and become anemic. The disease can ruin their lives. And to make matters worse, the longer they have it, the greater their risk of cancer. One of my closest friends had ulcerative colitis until about ten years ago, when he had his colon removed. It was the right thing to do. His disease was out of control, and after more than thirty years of having an inflamed colon, his risk of colon cancer was much too high. He now has a bag, called a colostomy, in place of a colon. Although not ideal, he has added a new repertoire to his jokes. But he doesn’t have those bouts of illness anymore, and he can hike up mountains faster than I can.

Crohn’s disease can involve the entire GI tract, where patches of inflammation appear and where the resulting scarring, called fibrosis, leads to intestinal obstructions. Although ulcerative colitis has been recognized for a long time, Crohn’s disease was first described in 1932 by Dr. Burrill Crohn, a New York physician. Is Crohn’s disease new to the twentieth century, or had it been missed previously? We don’t know the answer to that, nor do we know what causes it. But we do know that the incidence has been rising in recent years all over the developed world and even in developing countries that are becoming more industrialized.

It’s clear that intestinal microbes are involved in IBD because nearly every mouse model of the disease requires their presence for colitis to develop. The illnesses wax and wane, and antibiotics help people get through their crises. Still, this doesn’t tell us whether the bacteria are of primary importance or secondary, but they clearly are involved at some level. The more important question is: Why is the incidence of IBD increasing?

In 2011 a group of Danish investigators reported on their examination of pharmacy records for all 577,627 children born in Denmark as singletons (not as twins) between 1995 and 2003 and their assessment of their risk of developing IBD at an early age. The children were followed for an average of nearly six years, representing more than 3 million person-years of follow-up. A study of this scale provides an opportunity to observe uncommon events.

One hundred and seventeen children developed IBD. They had their first contact with the medical system—in a clinic, an emergency room, or a hospital—when they were on average three years and five months old. These are very early cases of IBD; the peak generally comes later in life, and, predictably, there will be many more afflicted persons. Still, the researchers could identify what these children were exposed to before they got sick and seek correlations. Compared to healthy children, those who developed early IBD were 84 percent more likely to have received antibiotics. Furthermore, children who had taken antibiotics had more than triple the risk of developing Crohn’s disease than those who were antibiotic-free. The more often they took antibiotics, the higher the risk. The investigators calculated that each course of antibiotics was associated with an 18 percent increased risk of developing Crohn’s disease. The children who had received seven or more antibiotic courses had a risk more than seven times greater than that of those who received none.

These numbers are sobering and are consistent with other data, such as a Canadian study that showed double the risk of asthma in children who received antibiotics in the first year of life. When was the last time you heard a doctor tell you that taking antibiotics could increase your child’s risk of developing IBD or asthma? The answer is never. But at a recent conference on the microbiome, one of the physician participants suggested that maybe we should require black-box warnings—the ones in bold print on the slip of paper included with prescription drugs that warn about serious risks—for all antibiotics.

I discussed our studies of asthma earlier. Other conditions linked with asthma include hay fever and eczema, or atopic dermatitis. Hay fever involves sensitivity to environmental allergens, such as pollens, cat dander, and roses. Also called allergic rhinitis, it leads to sneezing and sinus problems. Eczema can be seen as patches of red skin or dry, scaly skin. Although several sites are prominent in children—scalp, face, and chest—it may be present everywhere.

The prevalence of both hay fever and eczema has been rising dramatically in recent years, paralleling the increase in asthma. In fact, many children start with eczema and end up with asthma (the so-called march to asthma), or they have all three conditions simultaneously. Millions of children in the United States alone are affected by these modern plagues. As I discussed earlier, mounting evidence suggests that lack of
H. pylori
is contributing to the rise in these conditions in childhood, but there could be other similarly missing microbes as well.

*   *   *

Other diseases also appear to have increased. Allergies to nuts used to be a very rare condition, but now as many as one in fifty children has the condition. Between 1997 and 2008 the percentage of children with diagnosed peanut allergy more than tripled. Although most cases are mild, and overdiagnosis is likely, allergic reactions can be severe, sometimes leading to sudden death. Even a trace of peanut in a food can be sufficient to set off a severe attack. That’s why manufactured foods today carry a label announcing “produced in a facility in which nuts are present” or “certified nut-free.” Nut allergies are changing the lives of tens of thousands of children worldwide. Where are these allergies coming from? I don’t think that it is lack of pets.

Recently I’ve begun to think about a possible suspect. Recall that virtually all antibacterial agents exert similar effects promoting growth in farm animals; the animals get bigger whether they receive penicillins, tetracyclines, or macrolides. This led me to believe that all antibiotics produce more or less equivalent harmful collateral effects on our human resident bacteria. However, what if all antibiotics are not created equal? In fact, why should they be? We know that all major classes of antibiotics differ in terms of which bacteria they are good at killing. In our mouse experiments, we consistently find that we get stronger effects from tylosin, a macrolide, than from the penicillin drugs. We chose to study these two classes of drugs—the macrolides and the penicillins (the original members of a class of antibiotics called beta-lactams because of their chemical structures)—because together they represent more than 80 percent of all antibiotics prescribed for our children.

In the past, the macrolide of choice for kids was erythromycin. But compared to amoxicillin, a type of penicillin, it was less effective against some important pathogens and often led to side effects such as nausea and upset stomach. In 1991 two new macrolides were licensed in the United States: clarithromycin and azithromycin. Both were superior to erythromycin in all important ways and soon replaced it. Azithromycin is long-acting; even a few pills can have an effect for a week. Its manufacturer understood its great value and created the Z-pak, a small collection of drugs that can be dispensed as a single prescription and used for the entire antibiotic course. It is simple, effective, and has an easy-to-remember name.

In 1990, the year before it was licensed, the use of the Z-pak was essentially zero. By 2010, the last year for which I was able to find records, its use had climbed to nearly 60 million courses, and azithromycin now is the number one–selling antibiotic in the United States, even supplanting the pink liquid of amoxicillin. What that means is that every year about one in five of us in the United States is taking a course of azithromycin.