To calculate the reduction in morbidity and mortality that could be produced by the hepatitis A vaccine candidates, the total vaccine preventable illness value for each IME perspective is multiplied by the predicted efficacy of the vaccine(s). For both hepatitis A virus vaccines (live attenuated virus and subunit), the predicted efficacy is 0.90. The potential reduction in morbidity and mortality available with either vaccine is 162 using the committee median perspective and 159 using the age-neutral perspective. These values are not adjusted for vaccine adverse reactions or anticipated utilization.

Use of these unadjusted potential benefits numbers for comparing vaccines is described in Chapter 7.

The major stumbling block to HAV vaccine development was lack of a suitable animal model. This was overcome when Holmes et al. (1969) first unequivocally demonstrated the infection of marmosets with HAV. Subsequently, Provost et al. (1975) demonstrated that the virus derived from marmoset liver was readily inactivated by treatment with formaldehyde. This finding led to preparation of the first killed HAV vaccine. Tests of this vaccine in marmosets demonstrated that it could stimulate antibody and that the resulting antibody was protective (Provost and Hilleman, 1978).

The next advance in hepatitis A vaccine research came in the late 1970s, when several laboratories reported reliable propagation of the virus in cell culture. Since then, it has become apparent that the virus grows in a variety of cells, including the WI-38 and MRC-5 human diploid strains (Provost and Hilleman, 1979; Provost et al., 1982). Sequential passage of the virus in cell culture attentuated it for both marmosets and chimpanzees, yet it retained the ability to elicit antibodies (Provost et al., 1983). Clinical studies are now underway to find the optimal level of attenuation for a human vaccine.

An alternative vaccine might employ subunit antigens of the virus prepared by recombinant DNA technology in commercial yeast cells. One of these noninfective subunits might be ideal for inclusion in a complex vaccine against multiple agents. A possible combination would include an HAV subunit and agents of the herpesvirus family (e.g., herpes simplex, cytomegalovirus, and varicella-zoster).

A single injection of live virus vaccine would be expected to induce lifetime immunity. The non-infective antigens might have to be administered intermittently. Predictions on the prospects of vaccine development are shown in Table 5.1.

The health belief model parameters (perceptions of risk of illness, severity, vaccination benefits, and barriers) used to predict vaccine