The glass-walled landscaping center on the road south of Nijmegen looks like a gardener’s dream of heaven. My fingers tingle as I thread my way through stands of soaring bamboo, drifts of asters, and lanes of rhododendrons, tempted to grab a trowel and forget what I’m here for.
My host has little patience for my garden dreams. Jacques Meis, a physician and microbiologist, is muscling past the greenery to a wall of agricultural chemicals at the back of the long store. Wrapped in green and gold and aspirational images, organized by type of problem and method of application, the compounds stacked on the shelves are meant to keep these gorgeous plants healthy once they leave the nursery. I squint at the plant names that have been rendered into Dutch. Meis reaches over my shoulder, chooses a spray to protect boxwood from mildew and another to chase black spot from roses, and rotates them sideways so we can see the ingredient lists. Tapping the boxes gently, he shows me what we’ve come here to see.
“Tebuconazool,” he reads out in Dutch, and then switches to English for my benefit. “Tebuconazole. It’s a fungicide. It will kill funguses on garden plants.”
Meis is a big man, almost a foot taller than me, broad-shouldered and bullet-headed with an exuberant laugh. Eyeing the boxes, though, he looks solemn. “This is the same compound that we use in medicine,” he says. He looks around at the brilliant flowers and bags of bulbs and taps the ingredient list again. “Agriculture has almost 300 compounds they can use on fungi,” he tells me. “In medicine, we have four. And they use the four that we also use, because they work so well.”
It seems implausible to me that a box on a garden-center shelf, available to anyone who cares to buy it, could have any significance for human health. For the past decade, though, Meis and a small cadre of Dutch scientists have been building a case that one of our most commonly used classes of agricultural chemicals is simultaneously a profound health hazard. We have missed the connection, they say, because we do not pay attention to things too close to notice: the crops in fields, the flowers in gardens, the soil under our feet.
Nijmegen lies close to the German border, tucked against a river and folded into an agricultural landscape. There is nowhere in the Netherlands that is remote from farming—the densely packed country produces more food for export than any other nation save the United States—and you cannot drive far out of town without finding yourself next to a pig farm or a field of forage crops or a tractor taking up both lanes. Many of the diseases that affect humans come from animals and wildlife and the landscapes they live in, and infectious-disease experts such as Meis spend a lot of their time drawing connections between their patients and the places where their illnesses might have come from.
About a decade ago, he and some of his colleagues faced a diagnostic puzzle. Some of their patients were sicker than they ought to be, but there was no clear connection to a cause.
Meis is a consultant at Canisius-Wilhelmina Hospital and has an appointment at Radboud University Medical Center; these are side-by-side institutions, a hospital and a medical school, that attract patients from the entire Dutch southeast. He works alongside Paul Verweij, another physician and microbiologist, whose Ph.D. Meis once supervised. Verweij is now the chief of medical microbiology at Radboud, heading a lab that tests the organisms infecting hospital patients in order to identify the best treatments.
That kind of testing is most important for people who are very ill, because their fragility doesn’t leave much room for redirecting treatment if something isn’t working. The most vulnerable patients are people whose immune systems can no longer protect them against infections, because their innate defenses have been undermined by cancer treatment, by the immunosuppressive drugs given for autoimmune syndromes and after organ transplants, or by diseases such as HIV. People who are immunocompromised are like living petri dishes, fertile ground for whatever pathogen wafts their way. In the early 2000s, a cluster of those patients at Radboud were infected by an organism called aspergillus fumigatus.
In nature, aspergillus fumigatus breaks down decaying plant matter, keeping the world from becoming one giant pile of dead leaves and decaying crops. It reproduces by puffing out spores that drift through the air until they land somewhere damp and warm enough to germinate. One of the places they are likely to land is in our lungs: It’s estimated that each of us inhales about 200 aspergillus fumigatus spores every day. A healthy immune system sweeps them away—but in someone with compromised immunity, the spores can root into the lining of the lung and grow. Aspergillus fumigatus likes very warm temperatures; it grows happily in the steamy interior of a compost heap, which just happens to be the same temperature as the inside of our bodies. So once it settles in the lungs, the fungus reproduces wildly, spills into the bloodstream, and is conveyed to other organs, where it grows and overwhelms them in turn.
That outcome is called invasive aspergillosis, and it is diagnosed as often as 500,000 times a year worldwide, in up to 10 percent of immunocompromised patients. It is deadly—or was, until a tiny group of drugs called triazoles came on the market in the 1990s and 2000s. The triazoles, which have names such as fluconazole and voriconazole, worked against many types of fungal infections—a rare feat, because fungi are more like us biologically than bacteria are, and it is harder to make an antifungal that will kill just them and not us than it is to make an antibiotic. Invasive aspergillosis had been a death sentence, but the triazoles dialed the death rate down from 100 percent to 40 percent. In other words, three out of every five patients who would have died began to survive their infections instead.
And then that trend reversed. At Radboud, the death rate began to creep up again, to 88 percent. Running a review of the samples his department had processed, Verweij spotted the reason: The patients’ infections were resistant to the triazoles, possessing cellular defenses that protected them from the drugs’ attack.
By itself, resistance wasn’t remarkable. If a patient takes an antibiotic or antifungal drug for a long time, there is always a risk that the organism causing their infection will mutate to make itself less vulnerable. But the infections in these patients were resistant from the start of treatment. They had never received azoles before, yet their infections did not respond to even the very first dose of drug.
Down the street at Canisius, Meis was observing the same phenomenon. A healthy young woman came into the hospital with a bad case of the flu, seemed to get better, and then abruptly developed invasive aspergillosis. She had never before received azoles, yet her infection did not respond to them. In just a few days, she died.
Where Meis is big and boisterous, Verweij is thoughtful and careful. Almost everyone in the Netherlands seems to speak excellent English, but Verweij’s is even better than most; he lived in the United Kingdom as a child, and his word choices are idiomatic and precise. I asked him to explain the significance of the cases, when aspergillosis was already a recognized and deadly problem. “These were resistant to azoles from the first dose,” he said. “That was ... unusual.”
It was so unusual that he felt compelled to double-check. He dug through the lab’s freezers and pulled out aspergillus fumigatus samples taken from patients during the years since the triazoles had been on sale, more than 1,200 of them going back into the 1990s. Thirty-two of them were resistant, and, on analysis, they were all alike. They shared two mutations that had never been recorded in Aspergillus before: a single amino-acid change at one end of a single gene, and at the other end, a string of nucleotides that repeated as though the fungus’s genome had stuttered.
If each patient’s infection has gained resistance individually, by adapting to the drugs as they were administered, the pattern would have differed from one to another. A consistent identical pattern suggested one common source—something that every aspergillus had been exposed to, even though they had occurred in patients many miles apart.
Aspergillus lurks in the environment, so Verweij decided to start by looking there. There was a flower bed at the entrance to Radboud hospital that patients passed as they entered, a big concrete dish planted with tulips and begonias and other plants, depending on the time of year. He scooped soil from it and sent a student to the local garden center—the same one Meis would later take me to—to buy the things that the hospital landscape crew would have installed in it: plants, seeds, and bags of compost. The soil and garden goods contained Aspergillus, of course; that was expected. But 30 percent of the Aspergillus samples were resistant to azoles. They were all resistant thanks to the same mutations—and the mutations were identical to those in the Aspergillus harvested from patients.
As Verweij had feared, the Aspergillus in Dutch patients had not become resistant inside their bodies after treatment began. It had already been resistant when they inhaled the organism, on the hospital grounds or somewhere else in the wider world. Yet it was resistant to azoles, a prescription-drug class that should only have been available inside the hospital. Unpacking the conundrum of how that had happened would pit Verweij, Meis, and their colleagues against one of the largest industries in their country.
I was four years old the first time I went to the Netherlands. A 4-year-old doesn’t retain much, but I never forgot the narrow boats on the canals or the chocolates left on the hotel pillows. And I didn’t forget the flowers. It would have been August, and the tiny front yards in the Amsterdam neighborhoods were ablaze with crimson roses climbing up the drainpipes and snowy hydrangea hedges taller than I was. To me, they were the perfect gardens—personal, fragrant, and small enough to feel like a refuge.
It’s flowers, of course, that people associate with the Netherlands. Put the country’s name into Google and the first images to surface are ranks of brilliant tulips in front of windmills and canals. Walk through Amsterdam’s international airport, Schiphol, a transit hub set inside a shopping mall, and it seems as though every other shop offers bulbs for sale. Tulip fever, the mania for the newest and showiest flowers that bankrupted the Dutch in the 17th century, was so intense that it has come down to us as the words for a mass delusion; it inspired economic treatises and novels and films, and set the pattern for every financial bubble to come.
The Netherlands grows three-quarters of the billions of flower bulbs sold worldwide each year: tulips, lilies, hyacinths, crocuses, daffodils, alliums, and more. The trade earns the country about $750 million annually, and supports a farm economy of more than 100 bulb-trading companies, roughly 1,500 bulb farmers, and uncountable chemical dealers and fuel dealers and equipment salesmen and seasonal workers and farm towns. But that trade does not come easily. Despite the mild climate, the Dutch flower economy—along with its fruit, vegetable, and grain economies—is under attack all the time, from plant pathogens. One-fifth of the world’s food crops die in the field each year because of fungal plant diseases, and another tenth rot in storage because fungi have infiltrated them.
The main defense against plant pathogens is fungicides. And as Meis and Verweij and their colleagues would come to learn, the most important fungicides are azoles.
“When they came on the market, they were so much better than any of the other compounds that agriculture had been using,” Marin Talbot Brewer told me. She is a plant pathologist and associate professor of mycology at the University of Georgia, and is involved in a research project on azole resistance. “They had more of a targeted effect on fungi, and less of an impact on the plants themselves. And farmers could use much less of them—maybe 100 grams per hectare, instead of multiple kilograms.”
Azoles debuted in agriculture in 1973. Their effect was what a pharmacologist would call “broad spectrum”: They worked against many plant diseases, so a farmer dealing with several things attacking his crops could spray with just one compound instead of having to choose several. Plant diseases seemed to be slow to adapt to them, and azoles seemed to continue to work even once plant diseases began to show resistance. But to keep them as effective as possible, manufacturers tuned up the azoles over several different generations. The most recent update, the second-generation triazoles, came on the market in the 1990s.
It might be difficult to believe that farms and physicians could use the same core compounds and never realize it—but pharmaceutical chemistry and agricultural chemistry are separate professional fields that attend different conferences, publish in different journals, and have no reason to talk to each other. Without medicine ever recognizing it, azoles came to account for one-fourth of all fungicides worldwide. They are used on cereals and seeds, tree fruits and soft fruits, vegetables and flowers, hops and beans. Because they kill fungi so effectively, their use has bled out of agriculture into a vast array of consumer goods, from paints to lumber to glue.
“I don’t think there’s an analog in any other arm of biology where you have one class of chemicals being almost globally uniformly deployed on an organism that has such aerosolizable properties and is so numerous,” says Matthew Fisher, an epidemiologist at Imperial College London whose studies also suggest a strong link between agricultural azoles and azole resistance in Aspergillus. “It really is an enormous global experiment.”
By the late 2000s, when the Dutch medical researchers began to suspect that agricultural azoles were playing a role in patients’ deaths, farms in the Netherlands were using 25 different azole compounds, 143 tons of them per year. That was 325 times the amount of azoles that Dutch hospitals were administering to patients. The compounds—which medicine considered a drug and agriculture considered a fungicide—were being sprayed on flowering plants, fruits, and vegetables, and flower bulbs were being dipped in azoles before they went into the ground and before they were packaged for market.
To test their suspicions, Verweij and his group bought samples of agricultural azoles, took them into the lab, and exposed Aspergillus to them. Five of the fungicides caused the fungus to produce mutations identical to the ones the researchers had identified in their patients. All five belonged to the newest generation of azoles. They had been approved for sale in the Netherlands between 1990 and 1996, just a few years before patients had begun to come down with the resistant strains. Among them was tebuconazole, the compound Meis spotted on the shelves of the nursery we had visited in Nijmegen.
This isn’t the first time that medicine and agriculture have used the same formulas for different ends. In the 1950s, livestock producers began adding low doses of antibiotics to feeds to make animals grow faster and protect them from diseases. It took a few decades for medicine to realize that the antibiotics farmers were using were the exact same ones that doctors wanted to reserve for patients—and that the use of antibiotics on farms was creating resistant pathogens that would make humans ill. The Netherlands in particular was so alarmed by resistant outbreaks emerging from its agriculture that it created some of the world’s tightest government controls on livestock antibiotics.
The Dutch government recognized that antibiotics are scarce and precious; there are only 12 classes of antibiotics—formulas that share the same core molecule— and no new class has entered the market since the 1980s. Antifungal drugs are even more scarce than antibiotics: There are only three classes of antifungal drugs, including the azoles. Yet the fact that medicine and agriculture are using the same antifungals has never provoked the same official concern that agriculture’s use of antibiotics did.
Here’s one possible reason. While the bacteria that cause illness in humans—Salmonella, Campylobacter, E. coli, and the like—are also found in animals, aspergillus fumigatus doesn’t affect plants. When flower growers use azoles, they are aiming for other fungi. Aspergillus is an accidental bystander, economically insignificant, hanging out in soil and in foliage that flowers shed.
I wanted to understand how azole resistance could be developing in the flower fields, and one of Verweij’s colleagues volunteered to show me. Sijmen Schoustra is an assistant professor of plant sciences at Wageningen University, one town over from Nijmegen, and belongs to a team of researchers assembled by Verweij, Meis, and the Wageningen geneticist Bas Zwaan to explore the puzzle of azole resistance. We folded ourselves into a tiny car and Schoustra drove us to the opposite end of the country. We ended up on a bulb farm outside the town of De Stolpen, in a field that ended at the North Sea.
The field resembled a swatch of corduroy. It was made up of beds of plants with stiff leafy stems that came to my knees. They had been cut at the same spot with industrial precision, and the breaks between the beds showed how: They held the tire tracks of the cutting tractor that had slid over them to scalp them. At the front of each bed, a few brilliant blossoms lingered on plants that had evaded the blades: oriental lilies, flame and cream and oxblood.
“They shear the blossoms off to force the plant’s energy down into the bulbs,” Schoustra said. He gestured to the next field over, divided from where we were standing by an irrigation canal that ran toward the dike with geometric precision. Tawny foliage covered the ground, laid flat and drying in the sun. “Those are tulips. They’re ready for harvest. These have weeks to go.”
Andre Conijn, an agricultural-chemicals consultant who had guided us to the farm, crouched to examine the rows of lilies. The plants didn’t look healthy. Almost every one of them bore one leaf or several that were crumpled around a dark-edged, tear-shaped scar. When he tugged on a stem, it left the ground easily. The roots were stunted, and at the spot below the roots where the lily bulb should have swelled to fill his hand, the stem had rotted through.
“They have botrytis,” Schoustra said. “It’s a fungus. It will damage the plants, and make the bulbs have less bulk.”
Conijn spoke in Dutch, and Schoustra translated. The farmers had sprayed routinely to keep fungal diseases off the flowers, every week since the plants had broken through the ground. It was clear the sprays hadn’t worked, since the disease had attacked the field so uniformly. There were dead leaves and shorn flower buds lying in the tire tracks, and Conijn was worried that could have made the farm vulnerable; if the farmers had cleaned up the plant debris, the disease might not have found a foothold. But either way, the farm’s plants were infected now, and the farmers would no doubt be spraying more, to keep the disease from spreading.
I asked Schoustra what chemicals the farmers were using, and he checked with Conijn before replying.
“Azoles,” Schoustra confirmed. “They are spraying azoles.”
As we left the farm, we passed a compost pile 30 feet wide and easily 10 feet high, a wall of dead leaves and stalks and rotten bulbs, layered with straw to heat the pile and encourage decomposition. Conijn hefted a shovel and chopped into the mass. A charcoal crust fractured just below the surface, and as he lifted the shovel, a fine gray mist drifted away.
“Aspergillus,” Schoustra said. “This is just the right temperature for them.”
In 2015, Thomas Rogers began to be worried about Aspergillus. Rogers is the chair of clinical microbiology at St. James Hospital in Dublin, the largest hospital in Ireland, and he had been called to consult on the cases of two patients who were recovering from stem-cell transplants that had been used to treat leukemia. Those transplants use chemotherapy or radiation to wipe out the sick person’s immune system, so that the stem cells infused from the donor will take and not be rejected. They are brutally debilitating, and leave the recipients with no defense against infection. Both of these patients developed azole-resistant invasive aspergillosis, the first such infections the hospital had ever seen, and died.
Rogers had been thinking about Aspergillus for a while—the hospital was constructing a new building next door, and he worried about spores kicking up from the excavation—and he had been talking to Verweij and Meis about their discoveries in the Netherlands. He wondered what risks the environment held for his patients, and he decided to try to check. His postdoc, Katie Dunne, went to a Dublin garden center and bought bags of tulip and narcissus bulbs shipped from the Netherlands, the kinds of bulbs that the hospital’s landscapers would have planted to brighten the entryway. Five of the six packages contained azole-resistant Aspergillus, bearing the mutations that Verweij and Meis had linked to environmental exposure—and the organisms recovered from the patients bore environment-related mutations as well.
In a letter to a medical journal, Rogers and Dunne urged hospitals to stop planting decorative flower displays as part of their landscaping. They warned that people with compromised immune systems should be careful about gardening, too.
That warning came too late. In the summer of 2014, a 66-year-old man who had spent his working life farming and retired to a house surrounded by barley fields, checked in to a hospital in France. He had been healthy except for a longtime case of rheumatoid arthritis, for which he periodically received an immune-suppressing drug. When he arrived at the hospital, doctors thought he had pneumonia, and gave him antibiotics. He died 17 days later of invasive aspergillosis that was resistant to azoles and possessed one of the mutations identified in the Netherlands.
Months later, investigators examined the man’s house, taking samples from the nearby fields, his vegetable garden, his workshop, and his bedroom. No one had been in the house regularly since his death—yet in the garden, on his tools and on his nightstand, they found the resistant Aspergillus that had killed him. He had carried or tracked it unknowingly into his home.
The farmer’s death established that Aspergillus made resistant by azoles used in farming could be a risk to people anywhere the compounds are used. That means the landscape of risk includes Colombia, the world’s second-largest flower grower after the Netherlands: Azole-resistant fungi and azole drug residues have been found not just on flower farms, but also in flower beds and public gardens in the capital, Bogota. It includes China, where azole-resisant Aspergillus was found in soil in almost 6 percent of flower and vegetable greenhouses. It includes India, where Aspergillus with the environmental mutations has been found in tea fields and rice paddies, as well as in flowerpots outside hospitals.
When Meis counted in 2016, Aspergillus made resistant by farm use of azoles had also been identified in Belgium, Spain, Denmark, Italy, Norway, Germany, the United Kingdom, Poland, Romania, and Austria; Turkey, Iran, and Kuwait; Tanzania; Pakistan; Japan and Taiwan; and Australia. The problem is so newly recognized that researchers cannot say whether the same resistance mutation arises afresh in every location where Aspergillus is exposed to azoles, or if the sturdy spores of the newly resistant fungus have spread across long distances, drifting on the wind. By either route, it puts in peril millions of people living with compromised immune systems. Just in the United States, that number is almost 9 million.
Until recently, the United States was exempt from concerns about azole-resistant Aspergillus. Azole use here has historically been low, a small percentage of what happens in Europe—though the numbers that record it, kept by the U.S. Geological Survey, may be incorrect because submitting them is voluntary. Nevertheless, tebuconazole use in the United States rose five times over from 1995 to 2015; the crops that accounted for the most use were wheat and corn. Meanwhile, a Centers for Disease Control (CDC) poll of state health laboratories found no cases of resistance as recently as 2014. But last month, the CDC disclosed that 10 patients treated in the United States have been diagnosed with triazole-resistant Aspergillus bearing the mutations associated with agriculture. In a report in its weekly bulletin, the agency analyzed the records of seven of them. Four of them had never received azoles before they were diagnosed with the resistant strain. Three of the patients died.
Also last month, the CDC made its first attempt to estimate how costly fungal infections may be, given how serious they are and how frequently they are misdiagnosed. The agency’s calculation: $7.2 billion in 2017, just from health-care costs, not counting lost work time or the expense of home care or the ongoing impact of disability if people survive. Aspergillosis, they said, costs the United States $1.2 billion each year.
There is no mandatory registry in the United States of either azole application or diagnoses of azole-resistant illness. In 2016, CDC researchers went looking for azole-resistant Aspergillus in crop fields where they thought the compounds might have been used. They found it in a pile of crop debris in a set of experimental farm fields belonging to the University of Georgia, where the university’s agriculture school was testing varieties of peanuts. But there have been no cases of resistant Aspergillus among American farmworkers—at least none that anyone knows about. In a survey the CDC conducted in 2016, fewer than half of infectious-disease specialists were aware that azole-resistant Aspergillus exists, and only 14 percent had heard of a link to agricultural fungicides.
Invasive fungal infections “are not on the radar of clinicians,” according to Brendan Jackson, a medical epidemiologist who works on fungal diseases at the CDC. “And when they do occur, no one thinks to do a susceptibility test, to see if they are resistant.”
Doctors and researchers take their cues of what to worry about from national and international health authorities, and those bodies are just beginning to grapple with the health impact of farm fungicides. In 2013, the European Centre for Disease Prevention and Control, the EU equivalent of the CDC, pulled together a panel of 16 experts, including Verweij. Universally, they expressed alarm over the problem, predicting that aspergillosis will “rise relentlessly” and stress the health-care systems of even rich nations. But given how little was known about farm-related azole resistance, they could recommend only the beginnings of a research program: national surveillance to determine how many cases there are, better diagnostic tests to confirm infection, and experiments to confirm the environmental link.
The fungicide industry has set up its own monitoring body, called FRAC, but its chief concern understandably is whether its compounds still work against plant diseases; Aspergillus, the accidental bystander in farm soil and compost, is not on their list of concerns. In 2014, the Dutch Parliament passed a bill that asked the government’s health and environment ministries to consider banning the azoles from the Dutch market. The ministries replied that the science wasn’t yet solid enough to do so. They asked Verweij and Zwaan to launch a research program that could help formulate policies.
“We are doing as much as we can, but there is so much still to be done,” Zwaan told me. “At conferences, we still come up with very basic questions that just haven’t been addressed. I wonder if we will be really ashamed, maybe 10 years from now, if we look back and wonder why we didn’t do anything about this 10 years ago.”
The gardens of the Netherlands exist in my memory as bright and innocent and perfect, and I can still admire their beauty. But they are not the shelters I once imagined them to be.
Meis had told me that before I left the Netherlands, there was one more garden I ought to see. East of Schiphol Airport, he said, south of the center of Amsterdam, I would find the heart of the Dutch garden economy, an expanse of greenhouses and warehouses that form the portal where Dutch bulbs and flowers leave the country. Very early on a weekday, before the sun rose over the town of Aalsmeer, I stepped into Royal FloraHolland, the largest flower auction in the world.
FloraHolland (royal designates a firm that has been in business for more than 100 years) is a single building so large that the numbers describing it make no sense: It covers 1.3 million square meters, 320 acres, the area of 220 football fields. It is one unfathomably large room, but a gantry stretches across it at the level of a second story, for visitors to walk along without getting in the way of business. Suspended in the middle of the gantry is the auction itself, rooms of traders in bleacher seating, wearing headsets and stabbing keyboards, staring at wall-sized screens of flower lots while electronic clocks tick down.
Watching the traders was dull. Watching the flowers they were bidding on was different. Under the gantry beneath my feet, forklifts zoomed around the space, pushing pallets of flowers into a semblance of knots and labyrinths and breaking them apart again. Feeding the forklifts were lines of robotic carts that chugged through channels in the floor, bumping together like freight cars shuttling into a train yard. From above, they were a sea sucking at my ankles in a spume of colors: sheaves of roses, spikes of delphinium, more tulips than anyone could count. I checked the brochure they had given me at the desk. The auction processes 34.5 million flowers every day. It sends out 1.7 billion tulips a year.
I leaned over the railing to watch the blind carts roll past me, chartreuse and verdigris, carmine and violet and bone. The auction knew where they were going. It was impossible to say, beneath their beauty, what they might contain.