Group B streptococci are subdivided into five serotypes (Baker, 1977), but have four protective polysaccharide antigens (Ia, Ib, II, and III). Serotype Ic shares type Ia polysaccharide and contains Ibc protein. Serotype III is responsible for about two-thirds of all invasive infections, and is found in 95 percent of isolates from infections occurring beyond the first week of life. Animal protection studies and clinical data indicate that antibodies to the sialated (complete) type-specific antigens are protective, variability in both lethal dose and susceptibility to opsonization has been reported among GBS strains.
Adults have been immunized with GBS polysaccharides from serotypes Ia, II, and III (Baker et al., 1978; DeCueninck et al., 1982; Eisenstein, et al., 1983; Fischer et al., 1983). The sialated polysaccharide vaccines induce an immune response, defined as an increase in antibody concentration of more than 1 g/ml, in 65 to 100 percent of recipients, depending on serotype and pre-immunization levels. GBS II vaccine is the best immunogen and GBS Ia the worst, based on preliminary reports (Fischer et al., 1983). The antibody isotypes produced in response to GBS vaccines and their persistence are not yet clear, but might be predicted from work with other polysaccharide vaccines. The extent to which antibodies to GBS will cross the placenta is an important consideration because fetal protection will depend on this mechanism. The protective level of antibody in human sera is difficult to estimate because most pregnant women have very low levels of antibody to GBS and still have low risk of infection, and because the attack rate in colonized infants is only 1 to 2 percent. Estimates of the level of antibody required for protection have been generated from animal models (Klegerman et al., 1983), and antibody concentrations measured in infected infants or their mothers always have been below the protective levels estimated from the animal models. However, a variety of factors (inoculum size, site, clinical complications) may affect the level required for protection (Gray and Dillon, 1984; Wilkinson, 1978).
For these disease burden calculations, estimates of morbidity and mortality are confined to neonatal and maternal disease.
Early-onset disease incidence rates have been estimated by hospitals with large perinatal centers and high proportions of premature and