Skip to Main Content

If you’ve ever gotten a flu shot — and then, later that season, gotten the flu — you were more than likely, and rightly, miffed. Your doctor might have explained that the problem was possibly that the influenza vaccine wasn’t well matched that year to the strains people coughed and sneezed in your direction.

But increasingly influenza researchers are offering another explanation: The problem, at least partly, could be you.


A growing body of evidence suggests that sometimes our immune systems simply don’t follow the instructions a vaccine tries to give them — that is, make antibodies to fight a particular H3N2 or H1N1 virus. The reason? We all have flu baggage that shapes the way our immune systems respond to both infections and vaccines.

“We’ve all been trained on different influenza viruses,” explained Scott Hensley, an associate professor of microbiology at the University of Pennsylvania who studies the factors that influence response to flu vaccine. “If you vaccinate 100 people, guess what? They’re all going to respond differently. We think a large part of that is that we all have a different immunological imprint.”

The idea is that the first flu viruses your immune system encounters make indelible marks on it. A person born in 1970 whose first influenza A infection was caused by an H3N2 virus will always mount a better immune response to H3N2 viruses — or that component of the vaccine — than she will to an H1N1 virus or vaccine. But a person born in 1949, when H1N1 was the only influenza A virus circulating, will likely always have an underwhelming response to the H3N2 component of the flu shot.


The phenomenon is called “original antigenic sin” in the influenza literature. It’s also known as “imprinting.”

“The vaccine is inducing different responses in different people,” said Sarah Cobey, an associate professor of computational biology at the University of Chicago.

Dr. Danuta Skowronski, an influenza epidemiologist who has been digging into this issue, recalled that flu dogma used to be that the reason influenza was so hard on seniors was that seniors were, well, old. Immune senescence — the age-related slowing down of the immune system — was blamed for the fact that flu vaccine didn’t elicit a robust response in them.

Then about a decade ago, Skowronski and some colleagues conducted a review of a number of studies that tested antibody responses in people 60 and older a few months after they had received a flu shot. The studies showed quick waning of antibodies to the H3N2 component of the vaccine — but not the H1N1 portion.

“For me, that’s when I started thinking: There’s something else going on,” said Skowronski, who is with the British Columbia Center for Disease Control in Vancouver.

Shortly thereafter, the 2009 H1N1 pandemic struck. The pandemic virus was distantly related to H1N1 viruses that had circulated from 1918 to 1957, and then later from 1977 to 2009. Older adults — those whose first flu infections occurred before 1957 — fared pretty well in the pandemic. But adults in their 30s and early 40s, who were likely first infected with H3N2 viruses, were among those hardest hit by the new virus.

Why first infections are so critical with flu isn’t really understood. But the effect appears to be powerful — and it might help mitigate the risk we face if one of the dangerous bird flu viruses that have occasionally infected people triggered a human pandemic. We’re talking here about H5N1 or H7N9, both of which have taken high tolls when they have infected people.

To understand why, you need to know this: The 18 or so influenza A viruses break down into two large groups. The viruses within each group are more closely related to each other than they are to viruses in the opposite group.

H5N1 bird flu viruses are in Group 1, as are human H1N1 viruses. H7N9 bird flu viruses are in Group 2, with human H3N2 viruses.

Most people who have died from H5N1 infections were of an age where their first flu infections would have been with H3N2, an opposite group virus. Most people who died from H7N9 infections were of an age where their imprinting virus would have been H1N1, also an opposite group infection, a study published in Science in 2016 showed.

In other words, the patterns suggested that antibodies generated by an imprinting flu virus actually offered protection from severe disease when a person encountered a bird flu virus from within the same group.

Michael Worobey, a professor of evolutionary biology at the University of Arizona, was one of the authors of the Science study. The same group of researchers is now looking at seasonal flu data from Arizona to see if they can see the effects of imprinting there.

The phenomenon, he suggested, is probably playing a significant role in the underperformance of flu vaccine. “I think the imprinting has been ignored for too long and it may be as important,” he told STAT.

Hensley said the effect can even be seen within a flu subtype. In 2017, the experts who advise the World Health Organization on which viruses to include in flu vaccines recommended vaccine manufacturers switch to a new H1N1 strain. The one they had been using seemed to work fine for most people. But it wasn’t working well for a slice of the population — adults between the ages of about 30 and late middle age.

Hensley and his lab discovered that the vaccine target was making people who had their first flu exposures between 1977 and 1985 create antibodies to a version of H1N1 that was circulating back then — their imprinting virus. The decades-old H1N1 strains were too different from the 2009 version for the vaccine to work well in these people.

The phenomenon is also seen with influenza B viruses, but the pattern is even less clear there, Cobey said. There are two lineages of flu B viruses, called B/Yamagata and B/Victoria.

In annual vaccine efficacy studies Skowronski conducts, she has often reported decent flu B protection even when the virus in the vaccine isn’t the one causing most of the disease.

But a trial she conducted in small children showed that those whose first vaccine exposure was to a B/Yamagata produced more antibodies to it than to B/Victoria when they got vaccine containing the latter virus.

That raises an important question: Does vaccination imprint the immune system in the same way as infection? The Centers for Disease Control and Prevention recommends flu vaccine for all children aged 6 months and older. Flu vaccines now contain protection against two flu A viruses and the two B lineages. If a young child’s first exposure to flu is vaccine, not infection, will that child’s immune system imprint on all four viruses?

The short answer is no one knows. “That’s a really hot question,” Cobey said.

The National Institute of Allergy and Infectious Diseases wants to find out; last year it indicated it wanted to fund studies where researchers follow and chart the influenza exposures of infants and young children.

Dr. Alan Embry, chief of the respiratory diseases branch of NIAID’s division of microbiology and infectious diseases, said the proposals are being reviewed and grants should be issued in the first half of this year. The approved studies could last as long as seven years, he said.

Figuring out what happens with the earliest flu infections could point to ways to make more effective influenza vaccines. “It may inform designs of novel vaccines that elicit broader immunity from the beginning,” Embry said. “Or it may potentially provide insight into how imprinted [immune] memory can be rewired later in life.”