Searching for a Killer:
Inside the CDC’s scramble to solve a mysterious vaping disease
But as the lab team raced to test a last batch of lung fluid samples, the tool needed for the chemical analysis suddenly crashed. Scientists feared their precious samples would be destroyed.
“My stomach dropped,” recalled Jim Pirkle, who oversees CDC’s chemical lab analyses. “Since we have limited volume [of fluid samples], it could mean we weren’t going to be able to analyze them.”
The equipment failure in early November was one of countless challenges confronted by the more than 100 CDC scientists who have been working around the clock on the agency’s Chamblee campus in Atlanta, scrambling to create new investigative tools and techniques as they seek to prevent additional injuries and deaths from the vaping related illness. The devastating lung illness was reported for the first time this summer among previously healthy young people in the Midwest, and then began popping up all over the country. Doctors thought it was linked to e-cigarettes, battery-powered devices that can look like flash drives or pens and mimic smoking by heating liquids containing nicotine or marijuana, among other substances. But the devices had been sold for more than a decade without reports of such a disease. Why was this happening now?
Over the past three months, teams of scientists working to solve the mystery developed extensive lab tests and built a new data-collection system on the fly. They even relied on 3-D printers to manufacture custom parts so vaping devices could be fit into CDC’s special smoking machines, enabling them to “puff” on the devices to test the aerosols that are released.
In many infectious disease outbreaks, disease detectives have some idea of the likely culprit , or at least a sense of how to approach the investigation. The challenge in this case was that no one had any idea what might be making so many people so sick.
Even more daunting, federal and state health officials were dealing with a vast number of possible suspects that included several hundred e-cigarette or vaping devices and thousands, if not more, of e-liquids, many of which contain more than one ingredient . Supply chains for the vaping products also appear to vary state to state.
“The level of difficulty on this one is probably a 9 out of 10,” said Pirkle, director of the laboratory science division at CDC’s National Center for Environmental Health. 1/5 Diagnosis of a new disease that presented itself in different ways was also confounding, said Peter Briss, medical director at CDC’s National Center for Chronic Disease Prevention and Health Promotion, who helped lead the investigation. Many patients had chest pain and shortness of breath . Others reported nausea and vomiting, or fever and chills.Even more daunting, federal and state health officials were dealing with a vast number of possible suspects that included several hundred e-cigarette or vaping devices and thousands, if not more, of e-liquids, many of which contain more than one ingredient . Supply chains for the vaping products also appear to vary state to state. “The level of difficulty on this one is probably a 9 out of 10,” said Pirkle, director of the laboratory science division at CDC’s National Center for Environmental Health.
Diagnosis of a new disease that presented itself in different ways was also confounding, said Peter Briss, medical director at CDC’s National Center for Chronic Disease Prevention and Health Promotion, who helped lead the investigation. Many patients had chest pain and shortness of breath . Others reported nausea and vomiting, or fever and chills.
Pirkle said that only two other investigations he participated in during his nearly four decades at CDC were harder. One involved a mysterious syndrome in the 1980s involving contaminated oil fraudulently sold as olive oil in Spain that and killed hundreds and sickened thousands. That took several years to solve, he said. The other involved developing a way to measure levels of the toxin dioxin in Vietnam War veterans who had been exposed to Agent Orange. That also took several years “to get a methodology that was sensitive enough,” he said.
“What we have done here, we have done in weeks,” he said, crediting scientists working 14- to 15-hour days.
Among the earliest challenges was gathering data. The illnesses first appeared in Wisconsin and Illinois in July. When CDC got involved in August, officials had no idea how big or widespread the outbreak would become. The agency relied on a scaled-down version of a system used to track food-borne outbreaks, said Macarena Garcia, chief data scientist for the investigation.
But as one state after another reported cases, she realized the system couldn’t be scaled up fast enough. Information from the states was also arriving in different formats, she said, making it hard for federal officials to quickly analyze the national picture.
Data managers had to write code for every single state to transfer the information into the CDC system, she said. Even the questions states had asked patients were not uniform. For example, some asked them whether they vaped daily, weekly or monthly. Others asked if they vaped two to three times a day, or more than 50 times a day. While juggling incoming data, Garcia’s team had to create a separate, parallel system to collect the information, which took several weeks, she said.
Lab tests turned out to be even more challenging.
In early fall, New York’s Wadsworth Center laboratory and the Food and Drug Administration began analyzing vaping products collected from sick patients and found vitamin E acetate in those containing THC — the ingredient in marijuana that produces a high. But it was CDC’s analysis of lung fluid from sick patients that provided the first direct evidence of the oil as a common factor in their damaged lungs.
Vitamin E acetate, an oil from the vitamin, is in many foods and nutritional supplements and cosmetics. It is also an additive used to dilute THC in black-market vape cartridges . The oil does not usually cause harm when swallowed as a supplement or applied to the skin. Pirkle described it as an “enormously sticky” substance like honey that clings to the inner lining of the lungs. When inhaled, research suggests that it may disrupt normal lung function.
Developing reliable and accurate tests took time — about three weeks — because scientists wanted to ensure their methods were sensitive enough to detect the wide array of substances being investigated, said Eric Blank, chief program officer at the Association of Public Health Laboratories. Scientists are “looking for things you wouldn’t normally see in these [lung] specimens,” he added.
By early November — more than three months after the first cases had been reported in Wisconsin — a team of 25 scientists had found vitamin E acetate in lung fluid from 19 patients. That in itself was remarkable, Pirkle said. More remarkable, the scientists found no evidence of other chemicals or compounds on the list of possible suspects. The laborintensive and meticulous process involved more than 1,000 measurements just to rule out many plant oils as culprits, he said.
Still, “we needed to get results fast,” Pirkle recalled. “We couldn’t keep having people getting sick and dying.”
To buttress their case, the team sought to test 10 more samples. But the night before the analysis was to take place, a software failure crashed the chemical analysis tool, called a mass spectrometer.
When he first heard the news, Pirkle thought his team was playing a joke on him.
“No, it really crashed,’” he recalled being told. He was horrified. The scientists couldn’t afford to lose any samples — some of them so small they contained only about a dropper-and-a-half of liquid.
“It was tense,” Pirkle said. The CDC had already begun writing a report detailing the findings. Deadline was less than 12 hours away.
Technicians fixed the tool and the new samples provided even stronger confirmation. When the results came back positive for vitamin E acetate in those as well — now 29 out of 29 samples — team members were so tired that “we just went home,” Pirkle said.
Two days later, Anne Schuchat, CDC’s principal deputy director, announced that vitamin E acetate was “a very strong culprit of concern” in the illnesses. The evidence, while not yet definitive, has helped public health experts better understand how the puzzle pieces might fit together.
“Things are becoming more clear,” said Ruth Lynfield, Minnesota’s state epidemiologist, describing the lab findings as a “compelling association.”
The findings reinforced official health warnings against using e-cigarette and vaping products that contain THC, especially those bought off the street. But because a small population of sick patients insist that they used only nicotine-containing products, the CDC is maintaining its recommendation that consumers consider refraining from using all vaping and e-cigarette products, including those containing nicotine.
Animal studies still need to be done to determine the mechanism by which vitamin E acetate might cause the harm seen in patients’ lungs. The National Institutes of Health has already issued a solicitation for such animal studies.
Investigators are also trying to find out what other toxins might be flowing into people’s lungs as they vape. Federal health officials have said the outbreak may have more than one cause.
“It’s possible the harmful thing doesn’t even get produced until fluid is vaped” or heated, Pirkle said.
To understand what other substances are being inhaled, CDC began testing aerosols from vaping devices linked to sick patients. The agency is relying on 13 special smoking machines used in the past to study toxic chemicals in conventional cigarettes. But the smoking machines have round holes, whereas many vaping devices have rectangular mouthpieces. So CDC scientists used 3-D printers to create “custom interfaces” to fit on the mouthpieces and connect them to the machines.
“You’re literally putting a square peg in a round hole,” Pirkle said. “We had to construct something to make airtight connections.”
The machines puff at different voltages, and for different lengths of time, just like “all the things people do” when they’re vaping, he said. Scientists will analyze whether higher temperature and dosages make a difference. Filters will collect the aerosol and any lipidcontaining droplets.
“We’ll be looking to see if we get anything unusual,” Pirkle said. Most of the sick patients vaped multiple products, including those containing THC. “It could be that only one of the products causes disease and the others did not. But we don’t know which one, so we still have to analyze all of them.”
Scientists hope to have “a good chunk” of results within about six weeks, he said.