Among flu viruses, H3N2 is the one you should fear the most. It lands the most patients in hospitals. It kills the most people. Oh, and bad news: The flu shot has real trouble fighting it. Last year's seasonal flu vaccine was particularly weak against H3N2. In fact, that keeps happening, year after year—and no one is really sure why.
A typical flu vaccine is actually three, or sometimes even four, vaccines rolled in one, each protecting against a different strain. Flu experts have long suspected the H3N2 component was not up to snuff, and a meta-analysis of 60 past studies of flu vaccine effectiveness presented at an infectious disease conference in San Diego this week puts some numbers to it: The seasonal flu vaccine was, on average, only 38 percent effective against H3N2. That's terrible.
(Still: Get a flu shot. After all, 38 percent is a whole lot better than zero, and the effectiveness of the seasonal jab against other common strains is pretty good in this meta-analysis: 63 percent for influenza B and 65 percent for H1N1. So get the shot. Got it? Good.)
But that does make H3N2 something of a mystery. “The longer I study the flu,” says Edward Belongia, an epidemiologist at Marshfield Clinic who did the meta-analysis, “the less I think I understand it.”
If you want to figure out whether something works in medicine, you run a clinical trial. But clinical trials are expensive and slow—when it comes to the flu vaccine, by the time you get good results, flu season is already over and you’re onto next year’s version. Too late! Too bad!
A decade ago, statistics upended the world of flu monitoring. Epidemiologists in Canada figured out how to calculate flu vaccine effectiveness in real time using existing data sets: Plug the number of patients with flu-like symptoms who end up testing positive or negative for flu, along with whether they got vaccinated or not, into a formula and voila, out comes your vaccine effectiveness. Today epidemiologists call this a test-negative study design. “It’s the new gold standard for looking at flu vaccine effectiveness,” says Belongia. This design also gave epidemiologists, finally, the statistical power to routinely calculate the effectiveness against specific subtypes of flu—like H3N2.
“For a long time we weren’t able to see differences,” says Danuta Skowronski, an epidemiologist at the British Columbia Centre for Disease Control who helped develop the test-negative design. “We can’t fix the problem if we can’t see.”
Now that flu experts see the problem, they're thinking about chicken eggs.
Let’s back up for a second. The key to why H3N2 is so hard to pin down is the protein hemagglutinin. The virus makes it, and the human immune system uses it to recognize the pathogen as an invader.
But the flu virus's protein can mutate so that it escapes the immune system’s notice. And even though another common flu virus, H1N1, picks up genetic mutations at the same rate as H3N2, H3N2 can change the shape of its hemagglutinin faster. No one knows why.
These mutations add up to what epidemiologists call drift—and last year was an especially drastic case of drift for H3N2. Epidemiologists knew by late spring that the strain they chose for vaccines back in February did not match the one making people sick.
You read that right. Public health experts at the World Health Organization choose strains for the fall’s flu vaccines back in February. They have to do that so early, in part, because vaccine makers grow the viruses they use to make flu shots inside (bringing it back around) chicken eggs—hundreds of millions every year.
But not every virus will grow in eggs. The flu viruses that make you and me sick really want to live in human respiratory cells. Bird egg cells? Yuck. So every year, a handful of labs inject a chosen flu strain into eggs alongside another strain that loves chicken egg cells—it's especially adapted to grow inside eggs. The two viruses combine and, as flu viruses are wont to do, mix their genetic material. The researchers then pick one of these “reassortant strains” that looks like the chosen human-loving strain on the outside but the egg-adapted strain on the inside. This is the "seed strain" they inject into the eggs to reproduce and eventually make vaccines.
Remember how H3N2 is especially prone to drift? Well, creating the seed strain speeds up that process. In 2014, Skowioronski and her colleagues published a paper linking reduced effectiveness of the the previous winter’s H3N2 vaccine to the egg adaptation process. Other yet-unknown factors may play a role, says Skowioronski. “If our paper did anything, it opened our eyes to the fact it’s not just circulating viruses that mutate.”
People know how to make vaccines without using eggs. The US has approved one made in mammalian cell cultures, as well as as a recombinant vaccine for which manufacturers grow only the virus protein and not the whole virus. Most processes can be faster than using eggs. But no one has tried making these at the scale of hundreds of millions of doses yet. “We need to aggressively pursue and evaluate alternative techs that don’t require growing viruses in eggs,” says Belongia.
Vaccine makers have spent decades perfecting the egg process, and eggs aren’t going to disappear overnight. In fact, even cell-based vaccines still begin with egg-adapted strains for vestigial regulatory reasons. Getting to the bottom of the H3N2 mystery might add to the case for leaving eggs behind.