Members of Lee-Ann Jaykus’s lab waited near the finish line of the Krispy Kreme Challenge. But the Jaykus lab from North Carolina State University wasn’t cheering on participants of the race, which involves scarfing a dozen Krispy Kreme doughnuts followed by a 5k. Instead, they were documenting the after-effects.

“About one in 10 racers throws up after they cross the finish line,” Jaykus, a food scientist at NCSU, said.

For her lab, this was when the magic happened. To study how norovirus, the virus notorious for causing massive outbreaks on cruise ships, could be aerosolized during vomiting, Jaykus wanted to build a machine that would simulate vomiting. The problem was that no one had a lot of detail on the biomechanics of vomiting, which meant the researchers had to observe the event as it happened—hence the crowd at the Krispy Kreme Challenge.

The final machine, which could upchuck a fake vomit solution of dilute instant Jell-O vanilla pudding on command, gave Jaykus the realistic answers she was hoping for. The study, published in PLOS ONE, found that significant amounts of norovirus was aerosolized during vomiting and can potentially float around in the air, which is important for understanding the spread of the virus and for cleaning up after it.

“It provides some of the first evidence that norovirus could actually be aerosolized during vomiting,” Jaykus said, “and that amount is enough to make several people sick.”

Although norovirus might be best known for causing outbreaks of diarrhea and vomiting on cruise ships and in hospitals and nursing homes, it’s actually the most common cause of foodborne disease outbreaks, says Aron Hall, leader of the Norovirus Epidemiology Program at the Centers for Disease Control and Prevention (CDC). The virus infects 20 million people each year in the United States, putting 75,000 people in the hospital and causing around 800 deaths.

“It’s a tremendous burden. Understanding norovirus’ spread is a key public health priority,” he said. “The virus is spread through a variety of different means, which is why it’s so hard to control.”

About 70 percent of people acquire the virus from direct person-to-person contact, usually from a person who hasn’t adequately washed their hands after using the bathroom or vomiting. Another 20 percent of cases come from contaminated food. But the cause of the remaining 10 percent of cases isn’t quite clear. Amy Kirby, a microbiologist at Emory University, thinks, however, that the aerosolization of the virus during vomiting likely plays a role.

She points out several cases, such as a norovirus outbreak that spread through an elementary school. Numerous children vomited in various classrooms in the school, and the researchers found that classrooms that had vomiting children had more kids get sick, and that the more kids who vomited in class, the more kids who got sick. Kirby also cited the case of a girls’ soccer team that got sick after one player vomited in a hotel bathroom. This girl drove separately with her parents, went straight to the hotel room, got sick, and went home the next morning without ever seeing her teammates. In the bathroom, however, was a bag of snacks that was distributed to the team. Although the girl never touched the snacks, the researchers found norovirus genetic material on the snacks, which ultimately sickened the rest of the team.

“While this might not be direct proof, outbreak investigations show that vomiting is an important method of transmission,” Kirby said.

Since Jaykus couldn’t very well get people to sign up for a study that involved a mass barfing session, she decided to do the next best thing and simulate it with a machine. With a team of mechanical engineers and inexpensive parts from the hardware store, such as a plastic sac for a “stomach” and PVC piping for the esophagus and mouth, Jaykus got to work. To get the machine as realistic as possible, they wanted to mimic the angle of the head and body as someone threw up, which is how they found themselves at the Krispy Kreme Challenge. For good measure, they also watched videos of people vomiting on YouTube. They found that they head was generally bent down, and that people coughed around three or four times after vomiting to clear their airways of any debris.

The team also swapped out infectious norovirus with a bacteria-infecting virus called MS2 that has a similar size and shape and, in place of real vomit, used a surrogate made out of watered-down instant vanilla pudding. The machine vomited into a plastic box, where sensors measured the amount of virus in the air. Jaykus and colleagues had the machine vomit at various forces, to simulate projectile and more normal vomiting, and with various amounts, everything from a freshly eaten meal to a dry heave. Overall, the researchers found that only about one in 5,000 virus particles were aerosolized during a typical vomiting event.

It doesn’t sound like much, Jaykus admits, but when you consider that around 340 million virus particles are in a typical amount of vomit that someone produces, it ends up being more than enough to potentially make someone sick.

“It was a very well-controlled and well-designed study,” Kirby said. “The results support the idea that transmission by aerosol occurs, but that other methods of transmission, like vomit directly on surfaces, is likely more important.”

And while it is possible to be infected by norovirus by inhaling aerosolized particles through your mouth, the most likely method of infection is when the aerosolized virus settles on surfaces like tables, doorknobs, or a bag of snacks. There, it can lurk for weeks or even months while remaining fully infectious.

Jaykus says that she is currently working to determine the size of an area contaminated by aerosolized virus so that epidemiologists can develop protocols to determine who is at highest risk of infection and how to best decontaminate an affected area. Whether her lab returns to the Krispy Kreme Challenge remains unclear.