Systematic review and meta-analysis of the factors affecting waning of post-vaccination neutralizing antibody responses against SARS-CoV-2

Through a systematic literature review and meta-analysis, we found neutralizing antibodies declined after COVID-19 vaccination from months one to six ranging from 3.7-fold to 5.9-fold when evaluating post-primary or first booster against either the Index strain or Omicron BA.1. Waning rates were generally similar after primary or first booster regimens, and between infection-naïve and previously infected cohorts. Declines of neutralizing antibodies cross-reactive to Omicron BA.1 were greater than declines of Index-specific responses, both in infection-naïve and hybrid-immune cohorts, though this difference was not statistically significant. Only three studies evaluated a second booster; two reported no significant differences in waning kinetics between first and second booster^16,17 and one reported slightly enhanced antibody durability after the second booster, but the second booster cohort was small (n = 7)¹⁸. Because waning was similar after primary and first booster doses, degree of waning with subsequent doses is also expected to be similar. However, data to assess long-term waning, such as 12 months after the last vaccine dose, were unavailable, complicated both by needing to wait that long and by study subjects getting revaccinated before that time. As duration between vaccinations increases, this may be addressed in future studies. These waning rates could be used to predict waning for future relevant scenarios and adapt vaccination strategies accordingly.

Declines in neutralization titers were observed in all but three infection-naïve cohorts evaluated. One cohort that received a heterologous boost regimen (inactivated prime followed by a vector boost) was followed for 3 months and no change in titers was observed⁷. However, overall titers were low throughout. Two additional cohorts, one vaccinated with two doses of vector vaccine (out of five available vector-immunized cohorts) and one with a heterologous regimen (mRNA prime followed by a vector boost) had titers that increased against the Index strain through month 3 (longer follow-up was not conducted)¹⁰. Titers against Omicron BA.1 declined in these cohorts indicating that no undetected breakthrough infection occurred driving these titer increases. Interestingly, these three exceptional cohorts all received a vector vaccine as the last dose and hence it can be speculated that vector-mediated immunization might cause more durable antibody responses early after immunization/booster. On the other hand, studies with longer-term follow-up support overall comparable rates of waning across vaccine platforms beyond three months after the last dose, which might be explained by full clearance of the vector and any benefit it might add. More studies are needed to address this important observation and explore the potential role vector vaccines could play in enhancing durable immune responses. While we and others have shown that vector vaccines are generally less immunogenic compared to mRNA vaccines, heterologous regimens combining mRNA and vector vaccines have been shown to elicit immune responses comparable to mRNA vaccinations alone^11,19. Hence, boosters with vector vaccines in mRNA-primed (or already mRNA-boosted) participants could elicit the highest and most durable immune responses. This has been shown by Lyke et al. who observed that titers were more stable in subjects who received an mRNA prime followed by a vector boost than subjects who received three doses of mRNA vaccine¹⁰. We identified four additional cohorts (from three studies) evaluating hybrid-immune subjects that did not show a significant decline of titers in the observed period, and two of these cohorts were followed for more than six months post-last vaccine dose^6,14,20. However, two of the studies did not investigate breakthrough infections after the last vaccine dose^6,20. Breakthrough infections in even a small proportion of the subjects can have a large impact on the overall GMT because the impact of these few infections on the overall GMT of a group can be large. Therefore, these studies should be considered with caution. Still, one study on hybrid-immune subjects that ruled out breakthrough infections after the last vaccine dose showed stable titers over a three-month period¹⁴ which supports other observations that hybrid-immunity might have the potential to stabilize antibody titers at least temporarily¹⁸.

Neutralizing antibody titers can support and complement clinical vaccine effectiveness data as they correlate well with protection against infection and mild disease. Even if neutralizing antibodies fail to hinder initial infection and symptomatic disease, they will limit initial viral load and thus mitigate disease progression, so they correlate also with protection against severe disease. This can be seen by the see-saw pattern of COVID-19 vaccine effectiveness against severe disease which was similar to titers peaking in the first weeks after each dose and falling thereafter until the next dose³. However, the observed larger declines in titers against Omicron relative to the Index strain may correlate less well to clinical vaccine effectiveness against severe disease, which shows less waning than VE against symptomatic disease and infection. This supports that protection is aided by additional factors such as cellular immunity, which has gained increasing recognition for its importance for protection against severe disease²¹. While no precise correlates of protection are defined for neutralizing antibodies, an understanding of overall titers and waning rates will allow us to predict how fast protection against infection and mild disease will decline and whether this might differ by vaccine type, regimen, infection history or characteristics such as sex, age, or comorbidities to inform vaccine policy, including the time interval when additional vaccine doses should be given.

But even studying neutralization titers takes time; very few longitudinal studies with data against Omicron subvariants other than for BA.1 were available at the time of this review, and increasing immune escape from post-vaccination neutralizing antibody responses resulting in large proportions of participants with undetectable titers makes them difficult to evaluate, especially in infection-naïve cohorts post-primary vaccination. Indeed, available studies for newer sub-variants provide contradictory results with some observing increased rates of waning against Omicron sub-variants compared to the Index variant^22,23, some finding similar rates of waning^18,24,25, and others reporting lower rates^26,27. More evidence is needed to determine if waning of post-vaccination neutralizing antibody titers cross-reactive to newer Omicron subvariants differs from those reactive to the Index variant or to Omicron BA.1. Such results may depend on whether the vaccine targets the emerging subvariants. We found only one study assessing waning of neutralizing antibodies after bivalent mRNA vaccination (Index plus BA.4/5 antigen), which found greater waning during the first three months against Omicron subvariants than against the Index variant²⁸, similar to our observations for monovalent Index-directed vaccines. We were also unable to assess variant-specific effects on hybrid-immunity since all hybrid-immune cohorts investigated involved pre-Omicron infections.

Our results confirm observations of superior immunogenicity of some vaccination strategies over others. We observed significant differences in overall titers by vaccine platform, with mRNA vaccines resulting in higher titers and inactivated vaccines the lowest. Importantly, waning rates were not significantly different between the platforms. These results support previous findings that both booster doses and hybrid-immunity significantly increase overall titers and titers against Omicron BA.1 are generally lower than against the Index strain^{6,20,29,30,31}. Importantly, these results provide evidence of a relatively constant rate of waning for the different groups included in the analysis; thus, individuals immunized with a less immunogenic primary regimen are likely to reach non-protective antibody titers faster. This effect becomes more significant when comparing primary regimen to hybrid-immune or boosted cohorts. These results may prove informative for booster strategies, especially when vaccine supply is low or if over-immunization should be avoided because of possible imprinting and a lack of variant-adapted vaccines.

A systematic assessment of study reliability revealed that 88% of included studies had medium, low, or unclear reliability scores reflecting primarily poor reporting quality of study methods and details. While this does not necessarily translate to biased or unreliable data, the overall low-reliability scores and small percentage of studies with a high-reliability score reflects that data on neutralizing antibodies are difficult to compare across studies^4,5. This finding is further reflected by the wide confidence intervals observed in our meta-regression results. However, we included all studies meeting inclusion criteria irrespective of their reliability score for sample size reasons. A sensitivity analysis did not find an association between study results and reliability score, i.e., poorer scores were not more likely to be outliers.

In summary, neutralizing antibody titers are an important correlate of protection against infection and will continue to be important in providing alert signals and to help guide future vaccination regimens. While absolute values of neutralization titers varied widely between studies, this evaluation across many cohorts provides confidence that large differences in waning rates after booster doses are unlikely between vaccines used widely to date. However, we could not evaluate the most recent conditions, and waning against Omicron sub-variants that emerged after BA.1 may be faster. Considering the observed substantial rate of waning of neutralizing antibodies as well as the continuous emergence of new viral variants with altered antigenic features, additional booster vaccinations will continue to be an important tool to protect against COVID-19 in the future. Because peak titers differed widely by vaccine platform, number of doses received and number of prior infections, titers will drop to non-functional levels sooner for some conditions than others. Therefore, decisions around the timing of booster doses should consider both infection and vaccination history of the individual as well as other important factors such as risk for severe outcomes and population-level immunity. Other vaccine performance evidence should be considered as well, including clinical effectiveness of vaccines, particularly waning against clinically relevant Omicron sub-variants in combination with variant-adapted vaccination, and cellular immunity, which is important for long-term protection and protection against severe outcomes.