US study finds 36% flu vaccine protection, 25% against H3N2

An eagerly awaited estimate today of flu vaccine effectiveness (VE) in the United States so far this season confirmed that protection against the H3N2 strain is low and not much different than a similar report from Canada earlier this month.

The latest findings flesh out a vaccine protection gap that is part of what's fueling a record-high flu season in the United States and bolster the case that researchers and public health experts are making for the development of much improved flu vaccines.

With H3N2 as the dominant strain again this season, scientists have grappled with complex challenges regarding declining vaccine protection believed to be related to problems with egg-based production and wide genetic diversity in circulating H3N2 strains.

Research involved 5 study sites

To provide annual snapshots of flu VE during and at the end of each season, a team of researchers from the US Centers for Disease Control and Prevention (CDC) and their collaborators analyze children and adults seen for acute respiratory infections at five study sites connected to outpatient clinics. The investigators reported their interim findings—based on clinic visits and flu testing from Nov 2, 2017, through Feb 3, 2018—today in the latest issue of Morbidity and Mortality Weekly Report (MMWR).

Of 4,562 people who had respiratory illnesses, 38% tested positive for flu, with more than 80% of the infections caused by H3N2. The percentage of those vaccinated against flu varied by study location and demographic group, ranging from 45% to 59%. Of people with lab-confirmed flu, 43% had been vaccinated.

Based on their test-negative study, the overall adjusted VE against all seasonal flu subtypes was 36% (95% confidence interval [CI], 27% to 44%). Against the closely watched H3N2 strain, VE for all ages was 25% (CI, 13% to 36%), statistically significant but not much different from the statistically not-significant 17% VE against H3N2 reported from Canada 2 weeks ago.

Against the 2009 H1N1 virus, VE was higher at 67% (CI, 54% to 76%), and against influenza B, VE was 42% (CI, 25% to 56%). Nearly all (98%) of the influenza B viruses found in the study group belonged to the Yamagata lineage, which is not included in trivalent (three-strain) vaccines but is in quadrivalent (four-strain) formulations.

The researchers concluded that their interim estimate underscores ongoing challenges with the vaccine's H3N2 component that have persisted since the 2011-12 season. They added that the US findings of 25% VE against H3N2 suggest at least some protection against H3N2, and at a higher level than Canada's data and the 10% VE found in Australia—another finding that wasn't statistically significant. The US scientists also said VE against H3N2 this year was similar to the final 2016-17 estimate of 32%.

When the researchers looked at protection by age-group, they found that overall and H3N2-specific VE trended higher in children age 6 months through 8 years old.

In a nutshell, they said the vaccine provided substantial protection against 2009 H1N1 and moderate protection against influenza B. And the team said the findings emphasize the importance of antiviral treatment for any patients with suspected or confirmed flu who have severe or worsening symptoms or are at risk for flu complications, regardless of vaccination status or test results.

A separate US flu season update today from the CDC in the same MMWR issue said the proportion of 2009 H1N1 and influenza B infections are currently on the rise, and with several weeks more of flu activity expected, the vaccine is still poised to prevent significant illness, and CDC officials continue to recommend being immunized.

A need to probe other drivers of flu VE

Danuta Skowronski, MD, lead author of Canada's recent study, said the US findings for H3N2 are what experts have been expecting and aren't much different from what Canadian experts found.

"We wouldn't expect the overall H3N2 estimates to be identical," she said, noting differences in the proportion of participants who were children between the two studies. In both studies, VE estimates in working-age adults were only about 10% to 20%. "Overall," she added, "the estimates in Canada and the US are in the same ballpark—which is low."

The profile of circulating H3N2 viruses doesn't appear to be different between the two countries, Skowronski said. In both, more than 90% belonged to the same clade: 3C.2a.

"In combination, these findings signal something unique and important related to low VE in working-age adults especially that we need to understand better," she said. One factor in need of further study involves birth cohort effects triggered by different prime-boost exposures, also known as "imprinting," she said.

She said the US VE against influenza B is also comparable to Canada's, despite greater use of quadrivalent vaccine in the United States. Skowronski said hints of cross-lineage protection in the influenza B strains are interesting and have been found before. They require deeper study, she added, given that the hemagglutinin (HA) head of the two lineages is so antigenically distinct from each other.

"Clearly, there are other important contributors to vaccine performance beyond antigenic match between the HA head of the vaccine and circulating strains," she said, adding that there may be other components of the influenza B virus that are driving cross-lineage protection such as antibody to the neuraminidase or to the stalk of the hemagglutinin—not just the head.

FDA head casts doubt on egg-based problems

In a statement today, Food and Drug Administration (FDA) Commissioner Scott Gottlieb, MD, said though the US interim findings for the vaccine are better than some might have predicted, there is still significant room for improvement. And he added that the FDA is committed to working with scientific and medical communities to better protect the public against flu and apply lessons learned to next season's vaccine.

He added that the FDA is collaborating with the CDC, the National Institutes of Health (NIH), and other federal agencies to address steps needed to ensure optimal protection. He said a preliminary analysis of CDC data and patient database information suggests that cell-based flu vaccines appear to have performed somewhat better at protecting against flu than the egg-based vaccines that make up most of the US flu vaccine supply.

"Scientists at the FDA, CDC, and NIH are working diligently to fully understand the basis for this finding, so that all of next year's vaccines can provide better protection in preventing the flu," he said. He added that clarifying why cell-based vaccines provided better protection against H3N2 this season may help improve production of a more effective H3N2 vaccine for next season.

He also said that FDA tests on vaccine lots casts doubt on egg-based mutations as the cause of lower observed efficacy, given that agency scientists have confirmed, as part of quality control tests, that the genetic sequence in the vaccine strains used by manufacturers were the same as those provided to them at the beginning of the production process.

"Thus for egg-based vaccines, the lower-than-desirable efficacy being observed is not likely related to the process of adapting the virus to grow well in eggs," Gottlieb said in the statement.

Though developing a universal flu vaccine would be ideal, the reality is that such a vaccine is several years away, he said. In the meanwhile, the FDA is working with federal partners and industry to improve the manufacturing of current vaccines and looking at the possibility of continuous manufacturing technologies for cell-based and recombinant vaccines, which Gottlieb said could allow a quicker response to changes in circulating flu strains and a step-up in production, if needed.

Early estimates from Spain

In a separate study today in Eurosurveillance, Spanish researchers reported findings from primary care clinics and hospitals this season in Navarre, located in the north of the country. In October and November, the trivalent vaccine was offered for free to target populations, which included people age 60 and older and those with underlying health conditions.

Of 1,268 patients seen for flulike illness, 52% had lab-confirmed flu. The team found an overall adjusted VE of 39% (CI, 20% to 54%). Against H3N2 VE was 29% (CI, 15% to 57%), against 2009 H1N1 VE was 59% (CI, -6 to 84, so not statistically significant), and against influenza B VE was 29% (CI, 15% to 57%).

They concluded that the trivalent vaccine afforded moderate protection early in the area's flu season, and that performance of the H3N2 component higher than earlier reports from Canada and Australia, hinting that vaccination history might be a confounding factor. They also said their results for influenza B were consistent with those from Canada, which suggests the possibility of cross-lineage protection.

Skowronski noted that it's difficult to compare the Spanish report with the recent findings from Canada and the United States, because the Spanish study combined outpatient and hospitalized populations and involved only 118 H3N2 cases, too few to for a reliable estimation. "These results should be interpreted very cautiously," she said.

See also:

Feb 16 MMWR report on interim US flu VE

Feb 16 MMWR update on US flu season

Feb 1 CIDRAP News story "Canadian data show low flu vaccine protection against H3N2"

Feb 15 FDA statement

Feb 15 Eurosurveillance study

Newsletter Sign-up

Get CIDRAP news and other free newsletters.

Sign up now»


Unrestricted financial support provided by

Bentson Foundation Gilead 
Grant support for ASP provided by


  Become an underwriter»