Reviewed by SAGE. meetings, July 2021

Executive summary

1. It is highly likely that vaccine induced immunity to SARS-CoV-2 infection, and potentially severe disease (but probably to a lesser extent) will wane over time.

2. This is likely to be first detected by vaccine failures in vulnerable cohorts (for example a high rate of infections in people vaccinated over time, including hospitalized cases).

3. It is therefore likely that there will be vaccination campaigns against SARS-CoV-2 for many years to come, but currently we do not know what will be the optimal required frequency for re-vaccination to protect the vulnerable from COVID disease.

4. Different vaccines may induce different sorts of immunity, which may have different duration. Absolute correlates of protection are yet to be determined, but the only ready measure that can be used at scale is of serum antibody.

5. Even if antibody levels after vaccination wane to allow significant numbers of infections, this might have an effect of boosting the primed immune system of vaccinees who would experience mild or asymptomatic infections. In turn this could result in a broader immune response than expected if vaccines induced a sustained sterilizing immune response. Measurements of immune responses in persons who are infected following vaccination will be important to understand this.

6. Current studies (PHE and National core studies) have assembled longitudinal cohorts but are only funded or planned to continue for 12-18 months, and waning may be important in the time frame of 2-5 years. There is a need to decide what further longitudinal studies are required, and to what extent they can build on existing funded studies which could be extended, or whether de novo studies will be needed.

7. We strongly recommend that funding is maintained to allow some of the current cohorts to be followed for longer times periods, and to included subsets who remain unboosted.

8. Maintaining focussed studies with cohorts of individuals in whom waning may happen sooner are particularly important since these groups may act as ‘canaries’ in signalling the time point beyond which waning vaccine immunity might become an issue. It will be important to relate immunological information to parallel studies of disease outcome, particularly those for severe disease.

9. These studies should continue to include a collaboration of public health, academia and clinical scientists working together with support for the appropriate data linkage, to measure VE in the real world and to understand mechanisms of protection and waning.

By Wendy Barclay, Paul Kellam, Paul Moss, Jamie Lopez-Bernal, Mary Ramsey and Maria Zambon.


Figure 1 Legend

The level of immunity is measure by the dilution of the sera that contains antibodies required to inhibit by 50% the infection of cells (neutralise) in the laboratory (IC50). After the 1st dose of the vaccine people make different levels of antibodies (blue dots), yielding a population distribution of neutralising antibodies levels (green distribution, with mean level represent by black dot). After the 2nd dose of vaccine, neutralising antibody levels are further boosted to give elevate individual responses (green dots) and a population distribution of neutralising antibodies levels (green distribution, with mean level represent by black dot). Over time from the 2nd vaccination, the levels of neutralising antibodies decline, more rapidly at first and then at a slower rate, but the dynamics are poorly understood at the individual or population level. The correlation between the level of immune response and vaccine efficacy is not known, but for other vaccines, detection of neutralising antibodies are a reasonable proxy for vaccine efficacy. The levels of neutralising antibodies required to neutralise the prototype ‘Wuhan’ like SARS-CoV-2 (grey dotted line) are different from the levels required to neutralise Variants of Concern (VOC; red dotted line), where mutations in the Spike gene decrease the potency of the neutralising antibody response, resulting in the need for a higher neutralising antibody level to achieve the same level of vaccine protection. As different people in the population have different levels of neutralising antibodies following two doses of vaccine, then the rate at which people become susceptible to infection by prototype ‘Wuhan’ like SARS-CoV-2 or a VOC is specific to a person within the vaccinated population. The average rate of change over time (black solid curve) and range (high-low, orange dashed curves) therefore yields three windows of concern: window 1, following the 1st vaccine dose, where most people are protected from infection by prototype ‘Wuhan’ like SARS-CoV-2 (above grey dashed line) but many are susceptible to infection by a VOC (below the red dashed line); window 2, following the 2nd vaccine dose, where most people are protected from infection by prototype ‘Wuhan’ like SARS-CoV-2 (above grey dashed line) and by a VOC (above the red dashed line), but where some people remain susceptible to infection by a VOC (below the red dashed line); window 3, immunity waning over time leading to many people becoming susceptible to infection especially to VOCs. It should be noted however, the VE to asymptomatic, symptomatic, severe, and fatal infection are all different endpoints, each of which requires separate interpretation within this framework, and will most likely involve different component of the total immune response including vaccine induced cell mediated immunity. Each vaccine platform and regimen will have a slightly different prime and boost response for detectable antibody levels. For comparison, the level of antibody immunity likely achieved from natural infection is indicated (orange dots and distribution), including an indication of the small number of people below the level of detection for neutralising antibodies.

In this paper we summarize relevant studies that:

• address the potential durability of vaccine-induced immunity to COVID; and

• tabulate ongoing or planned studies that will complement this knowledge.

Reprinted for educational purposes and social benefit, not for profit.