It came as a surprise this June when the Centers for Disease Control and Prevention recommended against using the nasal flu vaccine for the 2016-2017 flu season, citing a lack of evidence that it works.
But things aren't so simple. In fact, the conflicting evidence about the live nasal flu vaccine offers an excellent case study on how complex the task of analyzing flu vaccine data and making recommendations really is.
"Sometimes the public wants a very simple message, and unfortunately life's not like that," Mark Loeb, the new study's lead author and director of the division of infectious diseases at McMaster University in Ontario, tells Shots. "Things change as the evidence grows and we understand more. Unfortunately, that's how science and clinical medicine work. The challenge is to be able to help the public understand the shades of gray here."
Loeb's study focused on 52 Hutterite communities in rural Canada, because their relative isolation makes for an environment with fewer confounding factors. Researchers randomly assigned 1,186 children, ages 3 to 15, to receive either the live attenuated nasal flu vaccine or the inactivated flu vaccine by injection.
During the three flu seasons from 2012 to 2015, 5.3 percent of the children in the live nasal vaccine group and 5.2 percent of the children in the inactivated vaccine group had lab-confirmed influenza, revealing that the two vaccines were equally effective.
The study lacked a control group of children who did not get a vaccine, but previous studies showed a flu infection rate of about 10 percent among unvaccinated children and of 4 to 5 percent in vaccinated children in these communities, Loeb says. Rates of antibiotic prescriptions, ER visits, respiratory illnesses, hospital admissions and school or work absences were also similar across the two groups.
These findings, published Monday in Annals of Internal Medicine, parallel the early evidence for the live nasal vaccine in the U.S. which led ACIP to preferentially recommend it for children for the 2014-2015 season. CDC data consistently showed the live nasal vaccine to be very effective in children until 2013, when the vaccine went from including three strains (trivalent) to including four strains (quadrivalent).
And therein lies the rub: The new Canadian study used the trivalent vaccine, while ACIP analyzed data using the quadrivalent vaccine, and among U.S. children.
"Many of us felt very strongly that the [live attenuated influenza vaccine] was a better vaccine than the inactivated for children, and the data supported that," says Pedro Piedra, a professor of virology and microbiology at Baylor College of Medicine and one of the investigators involved in the nasal vaccine clinical trials in the late 1980s. "But something happened when it became a quadrivalent vaccine."
The CDC decision in June relied primarily on data from the U.S. Flu Vaccine Effectiveness Network, which assesses the health of people who have an acute respiratory illness with cough for at least a week during flu season and are at least 6 months old. Researchers compare the odds of vaccination among those testing positive or negative for flu to determine vaccine effectiveness.
From November 2015 through April 2016, CDC data from more than 7,500 individuals, including nearly 2,300 children, showed the inactivated flu shot was 49 percent effective against any flu strain that season. But the live nasal flu vaccine showed almost no effectiveness at all that season or during the two previous ones. The inactivated flu shot outperformed the nasal vaccine for all flu strains in the CDC study, as well as in a smaller Department of Defense study.
Experts disagree on how much, if at all, ACIP should consider this evidence when they meet again Oct. 19-20. The flu working group communicates throughout the year between meetings, but recommendations are made only at the thrice-yearly meetings of the full committee. Would they even consider changing the 2016-2017 recommendations well after flu season has started?
For one thing, comparing the trivalent and quadrivalent vaccines is like "comparing oranges and tangerines," says Kawsar R. Talaat, an assistant scientist at the Center For Immunization Research at Johns Hopkins Bloomberg School of Public Health. In fact, comparing any two flu vaccines involves a "quagmire of issues," partly because every year, vaccines act a little differently, Talaat tells Shots.
"It was not an easy decision, but ACIP did what was in the public's interest given the U.S. epidemiological data," she said. Her supervisor, Ruth Karron, is head of the ACIP flu vaccine working group. "The flu vaccine is different every single year, so they do their best to make the recommendation for that particular season based on the data that's available." She adds: "Given three years' worth of data showing this vaccine didn't work, do they wait to get more data? You're spending time and effort, and you're vaccinating children with a vaccine that is basically no better than giving them salt water."
At the same time, the Canadian study is a randomized controlled trial, which should produce higher-quality evidence than observational data collected by the CDC. These Canadian data also corroborate findings presented to ACIP from a small United Kingdom study, a large Finnish study and a moderately sized study from Astra Zeneca, the pharmaceutical company that manufactures FluMist, showing the quadrivalent nasal vaccine's effectiveness in children ranges from 46 to 58 percent.
"Personally, I could argue it both ways," says Litjen Tan, chief strategy officer for the Immunization Action Coalition. He says a Food and Drug Administration statement following the CDC recommendation noted that the vaccine's benefits outweighed its potential risks. The FDA sharply rebuked the ACIP decision, Tan says, by suggesting it didn't consider the "totality of the evidence" in making the decision.
But while ACIP looks at global data, it must prioritize data from the U.S., Piedra says.
"Data outside the U.S. cannot be used to drive your decision at this time," he says. "Something is different between what we see in Canada and what we see in the U.S., and we need to be able to understand that because a public health body like the CDC would be criticized pretty severely if they recommended a vaccine that has very little evidence of benefit."
Several phenomena may contribute to the incongruity in the data. Going from the trivalent to the quadrivalent vaccines, for example, could have introduced more vaccine interference. Each vaccine strain competes to infect enough cells so that the virus can replicate and induce an immune response in those cells. More strains means more competition among the strains to infect cells, so multistrain flu vaccines include higher doses of each virus.
But that alone would not explain lower U.S. effectiveness, because the U.K., Finland and Astra Zeneca studies showed higher effectiveness with the quadrivalent vaccine. Another factor could be that U.S. children have been vaccinated regularly enough since the first universal flu vaccine recommendation in 2007 that they've built up an underlying immunity. That underlying immunity could interfere with responses to new doses, progressively lessening the vaccine's apparent effectiveness each year. (This is more likely with live nasal vaccines than with the shot because they induce different types of immune responses.) Bias in the study or other as-yet-unidentified factors could play a role too.
"Like all things in life, it is never very clean, and there may be more than one reason we may not see an optimal immune response," Piedra says.
All the experts agree that better flu vaccines are needed, but influenza is a tricky virus to combat, and current inactivated vaccines are effective. With a public health concern as serious as the flu, which kills thousands of people in the U.S. every year, public health officials can't afford to send the wrong message to the public.
"What happens is that public health makes a message that's too simple and the public doesn't buy it, and that leads to vaccine hesitancy," Loeb says. Public health officials must therefore balance complexity and accessibility in describing evidence on the flu vaccine.
The question now is how much consideration ACIP gives this new evidence at their October meeting.
"Now they're in a rock and a hard place," Tan says. "In light of this data, are they going to come back and revisit this recommendation before flu season starts — even though it's already starting?"
If the live nasal vaccine isn't recommended for this flu season, that means the CDC will have far fewer U.S. data next year on the live vaccine's effectiveness, Tan says. "If the CDC says we're only to trust the data from the CDC, where's the data going to come from?"