The quest for a fully immunogenic vaccine against influenza H5N1 viruses has gone on for more than 10 years, since this family of potentially pandemic viruses emerged as a cause of human disease in Hong Kong in 1997. H5N1 has caused 381 human cases of influenza, with a mortality rate exceeding 60%. H5 strains have now been found in birds throughout much of the world (though not yet in the Americas), and human illness has occurred in 14 countries throughout Asia and in northern Africa.¹ The much-feared rapid spread through and between communities, however, has not occurred. Aside from small clusters of cases within families, each human case has been associated with close contact with poultry. The culling of poultry in the face of recognized bird disease has been a major defense strategy since the first outbreak.

Each human infection constitutes an opportunity for genetic modification of the virus through reassortment, mutation, or both — modifications that could enable the virus to overcome the remaining barrier to a pandemic by gaining the capacity for efficient person-to-person transmission. The fact that no epidemic has yet occurred has prompted questions about whether H5 viruses face some insurmountable barrier of viral fitness that renders them incapable of causing widespread illness in humans. Yet all agree that history is a powerful teacher and that future influenza pandemics caused by novel strains are highly probable.

So where do we stand with vaccines against emergent influenza strains? The preparation of a vaccine against H5 influenza has not proved as simple as adhering to the standard manufacturing techniques used to create yearly influenza vaccines. The H5 strains have had to be modified, since their virulence in chicken eggs causes rapid death of the embryo, precluding the generation of an acceptable antigen yield. Assays that are used to evaluate potential vaccine efficacy have also had to be modified, since horse erythrocytes have proved more sensitive than the commonly used chicken or turkey cells in measuring responses to the H5 hemagglutinin antigen (HA). Attempts have been made to standardize a neutralization assay in the hope that it would be a more sensitive reflection of the functional activity of the vaccine. The difficulties in producing and standardizing conventional H5 vaccines have prompted innovative and extensive examinations of options for improving all influenza vaccines.

During the pandemics of Asian influenza in 1957 and Hong Kong influenza in 1968, there were efforts to explore possible vaccines, but the first systematic attempt to develop vaccines against an influenza virus representing a pandemic threat was made in the face of the swine influenza of 1976. Both pediatric and adult trials were conducted, studying both whole-virus and subvirion vaccines. These trials showed that whole-virus vaccine was not only more immunogenic but also more reactogenic than subvirion vaccine; that with a new immunogen, two doses were needed; and that immunogenicity was roughly doubled with a 10-fold increase in the vaccine´s HA content.

These lessons, coupled with epidemiologic data on the seasonal impact of influenza, have formed the basis for the broadening use of influenza vaccine in the United States for yearly epidemic influenza. The relatively poor immunogenicity of the 1976 vaccines in recipients who had not previously been exposed to the vaccine´s influenza virus strain also presaged the difficulties of developing an effective vaccine against H5 influenza. This history suggests that H5 may not be a uniquely poor immunogen; rather, limited responses are to be expected whenever people are given an entirely new influenza vaccine.