Our results mirror those from an earlier trial conducted in the United Kingdom, which found that children who received influenza A vaccine had a lower risk of A/England infection in 1972 but later appeared to lack cross‐protection against A/Port Chalmers in 1974, compared with other children who had received influenza B vaccine. Similarly, in the 1957 Cleveland Family Study cohort, adults with laboratory‐confirmed influenza during the period 1950–1957 had significantly lower rates of pandemic influenza A/H2N2 infection in the 1957–1958 pandemic. Cross‐protection against heterologous strains following infection has also been demonstrated in animal models for A/California/04/2009(H1N1) in guinea pigs and other influenza viruses in pigs, chickens, mice, and cotton rats. There are few data on the duration of cross‐protective immunity following infection, although 1 study conducted in UK boarding schools found that prior influenza A/USSR(H1N1) infection in 1978 protected against infection and clinical illness from an antigenic drift variant A/England/83(H1N1) 5 years later.
Although seasonal influenza infection does appear to confer cross‐protection against pandemic H1N1, our results are consistent with those of other studies that have found little cross‐reactive antibody response to pandemic A/H1N1 virus following seasonal influenza vaccination. We found statistically significant cross‐reactive antibody responses to pandemic influenza A/H1N1 following seasonal A/H1N1 infection, although few individuals had antibody titers 1:40. The cross‐protection observed in our study may be associated with mechanisms such as cell‐mediated immunity or nonneutralizing antibodies via antibody dependent cell cytotoxicity. It is recognized that inactivated influenza vaccines are poor at inducing efficient CD8+ T cell responses in humans. However, natural seasonal influenza infection can elicit cross‐reactive T cell responses against new virus subtypes, and the presence of cross‐reactive cytotoxic T cells inversely relates to the amount of virus shedding in infected individuals. Experimental studies in mice have suggested that, although natural infection with seasonal influenza virus provides partial protection against the development of severe disease after challenge with a pathogenic H5N1 virus, prior vaccination with inactivated seasonal vaccine not only fails to elicit such cross‐subtype protection but impairs the development of such T cell–mediated protection that arises from subsequent natural infection with seasonal influenza. These findings are analogous to the observations in our study. We investigated the possibility of nonsterilizing cross‐immunity by TIV against pandemic influenza, but we did not find any evidence for different ILI attack rates between study subjects who received TIV or placebo and had confirmed pandemic A/H1N1 infection.
Our study has several limitations. First, our pilot study has a small sample size and, in particular, was underpowered to detect indirect benefits of vaccination. Second, although 40% of the participants had serologically confirmed influenza infections during our study, we only obtained RT‐PCR confirmation of influenza infections in 15% because of a lack of timely identification of illnesses and possible under reporting. In other prospective cohort studies, 10%–16% of serologically confirmed infections were confirmed by RT‐PCR We had insufficient sample size to explore which seasonal strains conferred greater cross protection. In our main phase, during 2009–2010, we have increased the frequency and intensity of telephone follow‐up to detect ARIs sooner, facilitate more home visits, and allow virologic confirmation of a greater proportion of infections. Finally, we may have failed to detect some seasonal influenza infections in vaccinees, because increases in antibodies associated with influenza infection might be obscured by higher post‐vaccination titers, which may have decreased over time.
In conclusion, administration of TIV in children aged 6–15 years can protect against seasonal influenza infection, whereas our results show that seasonal influenza infection can confer cross‐protection against pandemic A/H1N1 infection. Therefore, by protecting against strain‐matched seasonal infection, administration of seasonal TIV might lead to increased vulnerability to antigenically different influenza strains. Alternative vaccines, such as live attenuated vaccines or adjuvanted vaccines, may confer greater cross‐protection against heterologous strains and avoid this potential disadvantage of TIV.