Quantifying the impact of a broadly protective sarbecovirus vaccine in a future SARS-X pandemic

Abstract

COVID-19 has underscored the need for more timely access to vaccines during future pandemics. This has motivated development of broad-spectrum vaccines providing protection against entire viral families, which could be stockpiled and deployed rapidly following detection. Using mathematical modelling, we assess the utility of a broadly protective sarbecovirus vaccine during a hypothetical SARS-X outbreak, for a range of implementation strategies including ring-vaccination, spatial-targeting and mass vaccination of high-risk groups. Broadly protective sarbecovirus vaccine ring- or spatial strategies alone are insufficient to contain epidemics driven by a SARS-CoV-2-like virus, but when paired with rapid isolation and quarantine, can achieve containment of a SARS-CoV-1-like virus. Where suppression fails, broadly protective sarbecovirus vaccine utilisation still reduces the effective reproduction number and slows epidemic growth - buying valuable time for health-system response and virus-specific vaccine development. Vaccination of high-risk populations with the broadly protective sarbecovirus vaccine ahead of virus-specific vaccine availability could reduce mortality and enable shorter and less stringent non-pharmaceutical interventions to be imposed; results are sensitive to vaccine properties (e.g., efficacy), health system capabilities (e.g. rollout speed) and timeline to virus-specific vaccine availability. Our modelling suggests that broadly protective sarbecovirus vaccine delivery to those aged 60+ years could have averted 21-78 % of COVID-19 deaths during the pandemic's first year, depending on the size of the stockpile. Realising this potential impact will require investment in manufacturing, delivery capacity and equitable access ahead of future pandemics.

Fig. 1. Exploring the prospects for outbreak containment via ring vaccination strategies using a broadly protective sarbecovirus vaccine (BPSV).

A stochastic branching-process-based approach was used to explore the impact of a BPSV on outbreak containment efforts utilising ring or spatially targeted vaccination strategies, and the factors most critical to control. A Schematic illustrating the ring-vaccination framework. B The effective reproduction number (Reff, top panel) and % of outbreaks controlled (bottom panel) via BPSV ring-vaccination and its dependence on R0 (x axis), for a “SARS-CoV-1-Like” virus and where no additional control measures are implemented. Black indicates scenario without BPSV, coloured lines/bars indicate different assumptions around vaccine delay to protection (VDP). For the Reff plot, bars indicate the average Reff across 100 stochastic simulations, and the error bars indicate the 95% confidence interval of the mean of those simulations. C As for B but for a “SARS-CoV-2-like” virus. For % outbreaks controlled plot, this is the percentage of 100 stochastic simulations that are successfully controlled. D As for B, but assuming that some infected symptomatic individuals quarantine and isolate such that onward transmission is reduced by 65% following isolation. E As for C but with the addition of quarantine. F Sensitivity analysis exploring how the % of outbreaks contained varies with R0 (x axis) and the ratio of the generation time to the VDP (Tg/VDP), in situations without quarantine (top heatmap) and with quarantine (bottom heatmap). Orange rectangle indicates the value held constant for other sensitivity analyses. G As for F but for vaccine efficacy against infection. H As for F, but % of presymptomatic transmission.

Fig. 2. Exploring the prospects for outbreak containment via spatially-targeted vaccination strategies using a broadly protective sarbecovirus vaccine (BPSV).

A Schematic illustrating the spatially targeted vaccination framework. B The effective reproduction number (Reff, top panel) and % of outbreaks controlled (bottom panel) via BPSV ring-vaccination and its dependence on R0, for a “SARS-CoV-1-Like” virus and where no additional control measures are implemented. Black indicates scenario without BPSV, coloured lines/bars indicate different assumptions around the number of hospitalisations required to trigger the spatially-targeted vaccination campaign (surveillance threshold), all assuming a 7 day VPD. For R_Eff plot, bars indicate the average Reff across 100 stochastic simulations, and the error bars the 95% confidence interval of the mean of those simulations. C As for (B), but for a “SARS-CoV-2-like” virus. For % outbreaks controlled plot, this is the percentage of 100 stochastic simulations that are successfully controlled. D As for (B), but assuming that some infected symptomatic individuals quarantine and isolate, which reduces onward transmission by 65%. E As for (C), but with the addition of quarantine. F Sensitivity analysis exploring how the % of outbreaks contained varies with R0 (x axis) and the ratio of the vaccination campaign spatial radius to the average distance between infections, in situations without quarantine (top heatmap) and with quarantine (bottom heatmap). Orange rectangle indicates the value held constant for other sensitivity analyses. G As for F but for vaccine efficacy against infection. H As for F but for the surveillance threshold.

Fig. 3. The potential impact of BPSV mass-vaccination campaigns on disease burden during a future SARS-X pandemic.

Dynamical modelling of BPSV mass-vaccination of priority groups (those aged 60+) following pathogen detection during a hypothetical SARS-X pandemic. A Illustrative figure of simulated scenarios and the timing of key events. B Time-varying reproduction number (Rt) profiles for the different non-pharmaceutical intervention (NPI) scenarios imposed in response to the epidemic that are considered for the analyses presented here. These Scenarios differ by assumed stringency (either no measures, a minimal mandate reducing transmission by 25% or stringent measures reducing Rt to 0.9), duration (either until the BPSV campaign is completed or the disease-specific vaccination campaign is completed) and the nature by which these NPIs are relaxed (either instantaneous or gradual). C BPSV impact on disease burden for each NPI scenario, assuming the VSV is available 100 days (top-panel) or 250 days (bottom-panel) following detection, for an R0 of 2.5. Uncoloured crosses indicate scenarios without BPSV (VSV only); points indicate scenarios where BPSV is available, coloured according to NPI scenario. Inset panels show deaths averted by the BPSV, coloured by NPI scenario. D BPSV impact on the need for NPIs for the same disease burden. For each NPI Scenario, the Pareto frontier was constructed for the VSV-only scenario, and used to calculate how many fewer NPI days can be imposed in the BPSV scenario whilst still limiting disease burden to the level observed in the corresponding VSV-only scenario.

Fig. 4. Retrospective evaluation of BPSV impact during the COVID-19 pandemic in selected countries.

Analyses using published model fits calibrated to excess mortality data retrospectively assessed the potential impact of BPSV availability on COVID-19 mortality during the SARS-CoV-2 pandemic, under various assumptions about the size a BPSV stockpile countries have access to. A Cumulative global COVID-19 deaths during the first year of the pandemic without (grey) the BPSV, and with the BPSV (coloured lines). Low coverage = BPSV stockpile size sufficient to vaccinate 40% of elderly population; Moderate coverage = 60%; High coverage = 80%. Variable coverage indicates size of stockpile varies according to the World Bank Income Group each country belongs to (LIC = 20%, LMIC = 40%, UMIC = 60%, HIC = 80%). B As for A but for daily COVID-19 deaths. C Modelled impact of the BPSV during the first year of the COVID-19 pandemic in different countries around the world, assuming stockpile size varies by World Bank Income Group (“Variable coverage”). Country colour indicates the percentage of COVID-19 deaths occurring in the first year of the pandemic that could have been averted if a BPSV had been available. Bars plot the mean % of deaths averted across 100 simulations, each utilising a single draw from the previously estimated posterior distribution of Rt for each country; error bars represent the 95% confidence interval for those 100 simulations. D Modelled impact of the BPSV during the first year of the COVID-19 pandemic in different countries around the world, assuming the variable coverage BPSV scenario—results plotted are the mean of 100 simulations, with country colour indicating the percentage of COVID-19 deaths occurring in the first year of the pandemic that could have been averted if a BPSV had been available. E BPSV impact on COVID-19 mortality in Italy during the first year of its COVID-19 epidemic. Grey line indicates model fit to COVID-19 excess mortality data (light grey points), and ribbon indicates the 95% CI across 100 model simulations using different posterior draws for the Rt trajectory. Coloured line indicates expected mortality when the BPSV is available, with the mean trajectory across 100 simulations plotted. Line colours reflect the assumption about the size of the BPSV stockpile. F As for E but for Iran instead of Italy. G As for F, but for Bangladesh instead of Italy.

Fig. 5. Dependence of BPSV impact on intrinsic vaccine properties, vaccination campaign dynamics and health system capabilities.

Sensitivity analyses exploring the sensitivity of BPSV impact to intrinsic BPSV properties and factors governing the speed, availability and coverage of the BPSV vaccination campaign. A Deaths averted by the BPSV (per 1000 population) and BPSV efficacy against severe disease. Results coloured according to NPI scenario considered (pink = minimal, orange = moderate, blue = stringent), for R0 = 2.5. Inset panels show Rt profile for each NPI scenario. Assumed virus-specific vaccine (VSV) development timeline was 250 days. B As for (A), but for BPSV efficacy against infection. C As for (A) but for BPSV immunity duration. D As for (A) but for BPSV stockpile size (and associated coverage of the target population that can be achieved). E As for A, but for the rate of vaccination during the BPSV campaign (and the associated time taken to vaccinate all eligible and willing individuals). F The delay (in days) between the first country in the world achieving 1% of its population vaccinated with COVID-19 vaccines and other countries achieving this same milestone. Individual coloured points are specific countries—data from Our World In Data. Empirical data are plotted as points, with the underlying boxplot displaying the median (central vertical line), interquartile range (box) and the 1.5× the interquartile range (extent of horizontal whiskers). G Impact of delays to BPSV access on deaths averted per 1000 population. Scenarios shown are for moderate NPIs and with continent-specific VSV access delays derived from (F).