Study adds more data on effects of consecutive-year flu shots

Family building snowman
Family building snowman

The study involved households with children over consecutive flu seasons., Dejan Ristovski / iStock

A new study from the University of Michigan has yielded more evidence that getting a flu shot 2 years in a row may result in lower vaccine effectiveness (VE) in the second year, and also that the effects of a flu shot may last more than one season.

The researchers, who followed 321 households through the 2012-13 flu season, found that individuals who were vaccinated only in that season enjoyed better protection than those who were vaccinated in both the 2011-12 and 2012-13 seasons, according to their report in the Journal of Infectious Diseases.

In addition, the investigators found some serologic evidence that was consistent with the flu VE findings: those who were vaccinated only in the current season had higher antibody titers against influenza A/H3N2 than those vaccinated in both seasons.

The research suggested that flu vaccination lowered the risk of flu 48% in adults and a similar percentage in older children, but it did not appear to protect younger children.

The study is one of several in recent years suggesting that those who get a flu shot may derive somewhat better protection during the season if they didn't get the shot the year before. Experts speculate that no change or only a small change in the vaccine strains from one year to the next may result in a weaker antibody response the second year.

But in an accompanying editorial, an expert who was not involved in the study asserts that the overall public health impact of flu vaccination must be considered over the long term, not 1 year at a time, and that the current recommendation to get a flu shot every year should stand while further studies are done.

Tracking illnesses in households

The investigators, led by Suzanne E. Ohmit, DrPH, as first author, recruited 321 households of at least four persons each, for a total of 1,426 participants, 833 of whom were children. From October 2012 to early May 2013, the participants were asked to report all respiratory illnesses and to come in for tests. Blood samples were taken from all volunteers age 13 and older up to three times: summer, in late fall before the start of flu season, and in the spring.

Of the 1,426 volunteers, 845 (59%) had proof of getting the 2012-13 flu vaccine, and 86% of those had received an inactivated vaccine.

During the surveillance period, 695 (49%) participants representing 240 households reported 1,227 respiratory illnesses, and 1,133 specimens were collected. Ten percent (116) of these tested positive for flu by polymerase chain reaction. H3N2 viruses accounted for 65 (56%) of the cases, type B for 47 (41%) of cases, and 2009 H1N1 for 3 (3%). There was one H3N2/type B coinfection.

A total of 111 volunteers (8%) in 76 (24%) of the households tested positive for flu (5 volunteers were infected twice). Overall, the infection risk was significantly lower in vaccinated than unvaccinated participants: 6.0% versus 10.3% (P = .003).

Overall effectiveness estimated at 32%

The overall VE estimate for the study was 32% (95% confidence interval [CI], -6%-56%), the report says. The estimate was better for adults, at 48% (95% CI, 1%-72%), and was similar but not significant for children 9 to 17 years old: 49% (95% CI, -16% to 78%). But there was no evidence of protection for children under 9 years old, with VE estimated at 4% (95% CI, -110%-49%).

To assess VE by 2-year vaccination status, the researchers compared the findings for those vaccinated in either year or both years (current and prior) with findings for those who didn't get the vaccine either year.

By that standard, VE was 88% (95% CI, 2%-98%) for those vaccinated only in 2012-13, 47% (95% CI, 11%-69%) for those vaccinated both years, and 43% (95% CI, -66%-80%) for those vaccinated only in 2011-12. Only 88 participants were vaccinated only in 2012-13, and just 1 of them tested positive for flu.

By flu strain, VE against H3N2 was 40% (95% CI, -4%-65%), and for type B it was just 7% (95% CI, -94%-55%). Effectiveness was 30% for preventing community-acquired flu, 37% for preventing household-acquired flu, and 43% for preventing medically attended flu, but all these estimates were nonsignificant, because the lower boundaries of the confidence interval were less than 0.

The researchers obtained preseason serum samples from 504 of the volunteers. They found that those who were vaccinated either year had higher hemagglutination inhibition (HAI) antibody titers to each flu strain than those who were not vaccinated at all.

Further, they found that HAI titers against H3N2 were significantly higher for those vaccinated only in 2012-13 than for those vaccinated both years. But for influenza B, antibody titers were highest for those vaccinated in both years, though differences between the various vaccine exposures were not significant. For H3N2, neuraminidase antibodies were slightly higher in those vaccinated only in 2012-13 than in those vaccinated both years.

The authors comment that the preseason HAI and neuraminidase antibody titers against H3N2 "were consistent with the observed pattern of VE for each combination of current and prior season vaccination status." For type B, they add, "Residual protection for those vaccinated in the prior season only was indicated in both serologic and VE results."

Findings extend results from 2010-11

The researchers observe that their findings extend what they found in a similar household cohort study published last year covering the 2010-11 season. That study revealed lower VE than was found in contemporaneous studies in healthcare settings, "an apparent negative effect of prior-season vaccination on current season VE estimates, and no evidence that vaccination prevented household transmission" of flu.

The current report also comes on the heels of a much larger study that produced somewhat similar findings on flu VE for 2012-13. That investigation, an observational study from the US Flu VE Network, showed that those who missed that season's vaccine but had been vaccinated the year before still seemed to have some protection.

The researchers say their new findings are consistent with the hypothesis that "small or no change in the vaccine from year to year may reduce antibody response and possibly VE." They note that the type B component of the 2012-13 vaccine was changed from the Victoria to the Yamagata lineage, while the H3N2 component was updated because of signs of antigenic drift. The authors don't say this, but the type B change presumably represented a larger shift than the H3N2 change.

The authors also comment that putting volunteers who were vaccinated in the prior season into an unvaccinated comparison group may reduce overall estimated VE for the current season, because the prior-season vaccinees have some residual protection. They add, "Confirmatory studies of the effects of longer term vaccination and infection histories and their methodological implications are warranted."

For now, they state, "Our results support the recommendation for annual vaccination, given current vaccine options, as those unvaccinated in the current year appear to be at greater risk when considering both VE and serologic evidence."

Growing evidence of an effect

In an accompanying editorial, Kathleen M. Neuzil, MD, MPH, says evidence of an effect of prior-year vaccination on VE is growing. She is a clinical professor of allergy and infectious diseases and of global health in the University of Washington's Department of Global Health, and also directs vaccine access and delivery programs for the Seattle-based global health group PATH.

"This phenomenon of vaccination against influenza in the prior year being significantly associated with a modestly lower level of clinical protection in the current year has now been reported from multiple influenza seasons, using both the prospective household design as well as the test-negative case-control design," she writes.

She adds that a research team in Wisconsin recently used 5 years' worth of historical vaccination data to show that vaccine effectiveness was higher in those with no vaccination history than in those who had frequent annual vaccinations.

However, she suggests that in the long term, annual vaccination is likely to prevent more flu illnesses than less frequent vaccination would: "While looking at vaccination in a single year may illustrate an apparent negative effect compared to either year alone, the overall public health impact must be considered over ALL years, not just the current year."

For a hypothetical example, she used the new study's VE estimates for the different vaccine-exposure groups to estimate total flu cases per 100 persons over two flu seasons (2011-2013), assuming that if none of them were vaccinated, there would be 10 flu cases each year. With vaccination both years, her example shows 6.5 flu cases over the 2 years, versus 6.9 cases for 2011-12 vaccination only and 11.2 cases for 2012-13 vaccination only.

"Only through multi-year prospective studies, with evaluation of subclinical infections, will we be able to solve the enigma of vaccine-induced influenza immunity," Neuzil writes. "In the meantime, the current policy to administer influenza vaccine every year should be maintained."

Ohmit SE, Petrie JG, Malosh RE, et al. Influenza vaccine effectiveness in households with children during the 2012-2013 season: assessments of prior vaccination and serologic susceptibility. J Infect Dis 2014 (Early online publication Nov 21) [Abstract] [Full text]

Neuzil KM. How can we solve the enigma of influenza vaccine-induced protection? (Commentary) J Infect Dis 2014 (Early online publication Nov 21) [Full text]

See also:

Nov 20 CIDRAP News story on US Flu VE Network findings for 2012-13

Mar 1, 2013, CIDRAP News story on 2010-11 Michigan household cohort study

This week's top reads