We did not identify any statistically significant differences in attack rates of seasonal and pandemic influenza infection or clinical influenza among the household contacts of children who received TIV or placebo, although our study had limited power to detect such differences. Results were similar when stratified by winter and summer seasons.
Of 91 children and their household contacts who had laboratory‐confirmed (by serological testing or RT‐PCR) seasonal influenza infection during the follow up period, 7 (8%; 95% CI, 3%‐15%) had confirmed pandemic influenza infection by RT‐PCR or serological testing. This was a significantly lower pandemic strain attack rate than that among the 276 subjects who did not have laboratory‐confirmed seasonal influenza infection, of whom 56 (20%; 95% CI, 16%‐26%) had laboratory‐confirmed pandemic influenza infection. In an adjusted model, individuals who had laboratory‐confirmed seasonal influenza infection had a substantially and statistically significantly lower risk of pandemic influenza infection. Receipt of TIV did not statistically significantly affect the risk of pandemic influenza infection in the adjusted model. Adjusted estimates were similar in a sub‐analysis restricted to the children who received TIV or placebo although the protective effect of seasonal influenza infection was not statistically significant, although it did remain significant in analysis of their household contacts. In a sensitivity analysis, results were similar when an 8‐fold or greater increase in antibody titer was used to define infection (data not shown). Our sample size was insufficient to allow us to distinguish whether seasonal A/H1N1 or A/H3N2 infections were associated with greater cross‐protection against pandemic influenza, and both were associated with a similar nonsignificant protective effect against pandemic A/H1N1 infection. Among individuals infected with seasonal influenza between baseline and mid‐season, few had antibody titers 1:40 against pandemic A/H1N1 virus at mid‐season, although there was some evidence of cross‐reactive antibody to pandemic A/H1N1 virus following seasonal A/H1N1 infection.
From January through September 2009, the predominant circulating strains of influenza in Hong Kong matched the vaccine strains in our study, except for antigenically drifted A/Perth/16/2009‐like (H3N2) viruses, which circulated in the summer. Our results are consistent with the prior expectation that administration of TIV would be effective in preventing serologically confirmed seasonal influenza infection in school‐age children. Pandemic A/H1N1 was the predominant influenza strain in Hong Kong from mid‐August 2009 onwards. During the study period, we estimated that 31% of the children who received TIV and 12% of the children who received placebo were infected with the pandemic virus. Furthermore, subjects who had seasonal influenza infection during our study were found to have a significantly lower risk of subsequent infection with pandemic A/H1N1 virus, although we could not distinguish whether greater protection was associated with seasonal A/H1N1 or A/H3N2.