In contrast to glycoconjugate vaccines the immune response generated by polysaccharide vaccines is generally poorly immunogenic in infants, does not elicit a memory response, and produces antibodies of lower avidity with less bactericidal activity relative to IgG concentration . Additionally, several polysaccharide vaccines including polysaccharide C  have been shown to elicit hyporesponsiveness, i.e. diminished immune response after booster compared to primary vaccination. The protective effect of meningococcal polysaccharide vaccines is strongly age dependent. Early clinical studies have shown inferior protection from invasive disease after polysaccharide A vaccination in children < 4 years , and no protection in recipients of polysaccharide C below 24 months of age . While total antibody concentrations after polysaccharide C vaccination fails to explain higher susceptibility in immunized infants , the SBA assay, specifically a titer of ≥ 8 when measured with rabbit complement, is currently the most practical method for the prediction of immunity after vaccination . It has superseded correlates based on antibody concentrations alone, e.g. an earlier proposal suggesting protection at concentrations of IgG ≥ 2.0 μg/ml .
While short-term qualitative and quantitative characteristics of the immune response to meningococcal polysaccharide vaccines have been extensively studied, less information is available on long-term effects. Antibody levels against serogroup C rapidly wane in infants [18, 29] and seem to persist less than 4 years in children above the age of two . Persistence of antibody levels in adults seems to last considerably longer, but very few studies have been published detailing long-term protection in older age groups: Zangwill et al. described elevated SBA titers to serogroup C (measured with rabbit complement) in adult vaccinees even after ten years .
In the present survey we have used a convenience sample of 20 individuals vaccinated only once with polysaccharide A, C, W135, and Y to analyze the association of observed antibody levels with time since immunization. Our analyses suggest that average durations of protective SBA titers to serogroups A, C, W135, and Y exceed 115 months. Previously published data for serogroup C may serve to validate the long-term predictive ability of our exponential model of decay: Zangwill et al. describe a rise of SBA titers from a pre-vaccination level of 13.6 to 1111.9 after vaccination. Using parameters 1111.9 and −0.032 (Table 2) as parameters L0 and k, respectively, our model predicts a SBA titer of 23.9 after 120 months (calculated as exp(log(1111.9) - 0.032 · 120)). This value is slightly below the observed value of 70 , which, assuming our model is valid, would correspond to a rate constant −0.023 (= log(70/1111.9)/120) and thus to a deviation of 28% regarding parameter k. Despite this small error, our model reproduces the original observation that titers after 10 years are still higher than pre-vaccination levels. We have found no further studies following up SBA levels exceeding five years after polysaccharide vaccination.
Moreover, we detailed the substantial inter-individual variability of both SBA titers and IgG levels directed against reference strains of targeted serogroups. The estimated minimal duration for W135 (0 months) suggests that some vaccinees may not attain lasting protection against this serogroup at all. Also, our model of decay for titers against W135 forecasts that a substantial part (approximately 23%) of vaccinees loses protection against this serogroup by 5 years, the time recommended for revaccination [10, 13]. The lower level of protection against W135 compared to other serogroups was further suggested by the significantly lower proportion of protective titers (65%) in our sample. Also, the quantification of IgG revealed that GMC reacting with polysaccharide W135 was the lowest (2.1 μg/ml) among the serogroups analyzed. On the other hand, our decay models predict that minimal durations of protection against serogroups A, C, and Y appear to exceed 24 months and only minor proportions (≤ 5%) probably lose protection within 5 years. As absolute risk for acquiring meningococcal disease for laboratory workers is low , only few case reports exist that could corroborate or refute suggested minimal durations of protection. Of those, we found only one report that contained previous vaccination history with polysaccharide vaccine. It describes infection of a laboratory worker 67 months after application of meningococcal A + C polysaccharide vaccine (Sanofi Pasteur MSD) caused by a serogroup A meningococcal strain . As our model forecasts a probability of > 5% of missing protection after this interval, it is plausible that the laboratory worker was not immune at the time of exposure.
In contrast to SBA titers, we could not detect a significant decay of IgG levels with time. This is most likely due to the small power of our analysis to detect decays with slopes barely differing from zero. Zangwill et al., using a sample of 40 adults, described a decrease of total anti-capsular antibody concentration against serogroups A and C 10 years after vaccination; nevertheless, they noted that it was smaller than that seen in SBA titers against serogroup C .
GMCs differed significantly across serogroups with GMC against serogroup A considerably higher (17.4 μg/ml) than that of other serogroups. It is unlikely, however, that this represents a vaccine effect, as unimmunized adults frequently show high baseline levels against this serogroup. Levels of IgG in unimmunized individuals were 1.5 μg/ml and even 17.5 μg/ml in a study encompassing unimmunized individuals from North America and Sudan, respectively . The reasons for high pre-vaccination concentrations of IgG and other immunoglobulin classes remain unclear, yet it is likely that several commensal bacteria including Escherichia coli and Bacillus pumilus give rise to cross-reacting antibody populations. Their role in bactericidal immunity, however, seems to be minor . Although we cannot confirm it for our sample given the lack of pre-vaccination samples, it is probable that observed high concentrations of IgG against serogroup A are a corollary of high pre-existing levels rather than exceptional immunogenicity of serogroup A polysaccharide within the administered vaccine.
Several authors have investigated the correlation of SBA titers and concentrations of IgG after meningococcal polysaccharide vaccination. Maslanka et al. found a positive correlation for all investigated age groups after vaccination with serogroup C polysaccharide, noting that correlation was lowest in 1 year olds with a correlation coefficient of 0.34 . Granoff et al. observed increased correlation of high-avidity antibodies with SBA titers after serogroup C polysaccharide immunization and concluded that low-avidity, probably non-functional antibodies decrease correlation . Also, moderate correlations between IgG and SBA titers of 0.56 and 0.37 were found in adults  and toddlers  after vaccination with serogroup A polysaccharide. Moreover, several groups reported no correlation between antibody concentrations and SBA titers against serogroup C in non-immunized individuals [33, 36]. We found significant associations between antibody concentrations and SBA titers only against serogroups C and W135. For A and Y, however, it seems that antibody concentrations contribute little to the explanation of protection. While in the case of serogroup A this may be due to high and varying levels of cross-reactive antibodies, the reason for a missing association in serogroup Y remains elusive. In contrast to A the correlation coefficient in Y is so low that it seems unlikely that our failure to determine a positive association is due to low power of our sample.
Finally, we report the considerable proportion of incomplete protection (30%) in a convenience sample of laboratory workers. As recommended by several authorities including the RKI , polysaccharide-conjugate vaccine should be used to reinforce immunization and to avoid development of hyporesponsiveness. We have shown for a small subset, that, as expected, revaccination with conjugate vaccine is indeed effective in restoring SBA titers.
Our survey has several limitations. Firstly, our sample size is small with 20 individuals tested, which is due to the low availability of adult individuals who have been immunized once with meningococcal polysaccharide vaccine. This entails that our power to detect differences in the kinetics of antibody responses against different serogroups is small. Secondly, our sample is cross-sectional in nature, which leads to impaired temporal resolution in the description of decay. A consequence is that we are unable to test whether more elaborate models would provide a better fit to observed data; e.g. Zangwill et al. describe a decrease of titers by 94% in the first two years, which is considerably higher than a decrease of 54% predicted by our exponential model. Finally, we cannot say whether all individuals showed adequate response to the vaccine, as we do not have pre-vaccination samples. Judging from estimated L0 values, however, most participants will have attained SBA titres ≥ 8 after vaccination. In spite of these limitations, we have been able to correctly reproduce major long-term observations.
In summary, we present data on the long-term persistence of antibodies after meningococcal polysaccharide vaccination in adults, which represents an under-researched topic.