Third vaccine doses could have indirectly averted >100,000 deaths in the USA by reducing transmission of the SARS-CoV-2 Delta variant ======================================================================================================================================== * Billy J. Gardner * A. Marm Kilpatrick ## Abstract Vaccination programs often focus on direct protection of individuals against disease and discount reductions in transmission, which can result in preventable disease and death. Initial clinical trials for most COVID-19 vaccines only measured direct protection, and dosing and vaccine selection decisions have, so far, ignored effects on transmission. Here we provide an empirical framework for incorporating indirect effects of COVID-19 vaccination on transmission in a continuous dynamic immune landscape. For the first time, we quantified relationships between neutralizing antibody titers and vaccine effectiveness for both susceptibility (VES) and infectiousness (VEI) and quantified changes in VE with waning and boosting of immunity. We used these relationships to quantify the impact that additional doses of mRNA vaccines (BNT162b2 and mRNA-1273) could have had in reducing transmission and deaths caused by the deadliest SARS-CoV-2 variant, Delta, in Autumn 2021. Neutralizing antibodies waned 8-fold in 2021 over the six months following initial vaccination with mRNA vaccines, which reduced VES 33-38% (from 75-81% to 47-54%) and VEI 62-65% (from 47-57% to 16-22%) against the Delta variant. Third doses increased neutralizing antibody titers 13-26-fold, which more than restored VE and reduced the relative risk of transmission 7-10-fold. Administering third doses by September 1, 2021 could have reduced the effective reproductive number Rt by 19%, stopped surges in transmission in many populations, and averted an estimated 111,491 deaths in the United States. Vaccination programs that incorporate effects on transmission in trial design, vaccination frequency, and vaccine choice are needed to address current and future public health challenges. **Significance** Vaccination programs for many diseases often focus primarily on people at risk of severe disease and sometimes discount indirect reductions in transmission. We developed a framework for estimating the effects of vaccination on transmission and used it to estimate the number of COVID-19 deaths that could have been averted by deploying third doses during the Delta variant surge to reduce transmission. We show that immunity against transmission had waned significantly in late 2021. However, widespread deployment of third doses could have restored protection, stopped surges of the Delta variant, and prevented more than 100,000 deaths in the United States. Vaccination programs should incorporate effects on transmission in initial trial design, vaccination frequency, and vaccine selection to address public health challenges. Keywords * COVID-19 * vaccine effectiveness * surrogates of protection * vaccination * waning immunity * herd immunity * modeling ## Introduction Vaccination is the most effective tool for reducing disease, hospitalizations, and deaths for many infectious diseases 1–5. Core components of vaccination programs include how to allocate vaccine doses within and between individuals, when to recommend new doses and, when multiple vaccines are available, which vaccines to recommend. In making these decisions and developing new vaccines, public health agencies often focus on the direct impact of vaccines in reducing disease in vaccinated individuals 6,7. For example, the United States Food and Drug Administration guidelines for randomized control trials for COVID-19 vaccines did not include any criteria regarding protection against infection or reductions in infectiousness 8, and, as a result, these weren’t measured in the US trials for mRNA vaccines made by Pfizer-BioNTech and Moderna. Similarly, vaccines are often only recommended for subsets of the population at high risk disease severity 9, despite the fact that in many cases, such as pneumococcal disease, vaccines have been shown to reduce transmission 10. Many vaccines can have population-wide benefits by reducing transmission either by reducing susceptibility or reducing infectiousness (i.e. transmission from breakthrough infections), or both, which indirectly protects additional individuals 11. In these cases, vaccinating individuals who are not at risk of severe disease can reduce severe disease through reductions in transmission 12. Deploying vaccines to reduce transmission can increase health equity by reducing infection in individuals that lack access to health care, including vaccination, which is a key global challenge 13. However, determining whether the benefits of deploying vaccines to reduce transmission is warranted requires estimating vaccine effectiveness (VE) for transmission and the potential number of deaths averted for a specific population, which, in turn, depends on the dynamic levels of population immunity for the pathogen in question. The emergence of the Delta variant (B.1.617.2) of SARS-CoV-2 in 2021 illustrates the potential benefit of using vaccination to reduce transmission. The Delta variant caused a surge of deaths globally, even in populations with high initial two-dose vaccination coverage 14. This was due, in part, to the higher infectiousness and severity of this virus variant 15,16, partial immune evasion 17, and, critically, waning vaccine immunity. Waning immunity was evident in both levels of neutralizing antibodies and studies of vaccine effectiveness, including for severe disease in older individuals 18,19. To address the surge in COVID-19 in autumn 2021, third vaccine doses were offered to older ages and at risk individuals, but were not initially offered to healthy younger individuals because VE for severe disease in these age groups remained relatively high 20. However, the potential impacts of offering third doses to all individuals to reduce transmission was not considered because of a lack of information on the benefits of indirect protection 21. Here we estimate the potential impact of third doses of mRNA vaccines on population-level transmission of the Delta variant in late 2021 and the number of deaths that could have been averted. We first estimated how VES and VEI change over time due to waning and boosting of immunity. Previous work has demonstrated strong correlations between a surrogate measure of protection, neutralizing antibody titers, and VE for symptomatic and severe disease 22–27, and one study used these relationships to examine the impacts of vaccination for averting disease and deaths26. Here, we extend this approach by developing the first relationships, for any pathogen, between neutralizing antibody titers and VES and VEI, which are required to estimate the impact of vaccination on transmission. We then quantified relationships between neutralizing antibody titers and time since vaccination, and following boosting with a third dose of mRNA vaccines. We integrated these two relationships to estimate the potential impacts of deploying additional vaccine doses on the reproductive number of the virus, *Rt,* under multiple scenarios. Finally, we estimated the number of Delta infections and deaths that could have been averted in the United States in Autumn 2021 by deploying third doses to reduce transmission. ## Results We found very strong relationships between variant- and vaccine-neutralizing antibody titer ratios (NATRtot) and VE using 23 estimates of VES (susceptibility) and 14 estimates of VEI (infectiousness) across three variants and four vaccines (Figure 1A,B; Table S1). VES was higher than VEI across the range of NATRtot observed, but the slope of VEI was higher (VE increased more rapidly as NATRtot values increased) (Figure 1; Table S1). ![Fig. 1.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/13/2021.10.25.21265500/F1.medium.gif) [Fig. 1.](http://medrxiv.org/content/early/2024/11/13/2021.10.25.21265500/F1) Fig. 1. Vaccine Effectiveness (VE) plotted against variant- and vaccine-specific neutralizing antibody titer ratios (NATRtot) for (A) susceptibility, VES, (B) infectiousness, VEI, and (C) transmission VET. Each point (and 95% CI) represents a single empirical estimate of VE for a single vaccine (or infection) & virus variant. Points are jittered slightly along the x-axis to facilitate presentation. Black lines and ribbons in (A) and (B) show the fitted model and 95% CIs. The line and 95% CI ribbon in (C) is estimated from lines in (A) and (B): VET = 1-(1-VES)*(1-VEI); see Methods for details. In the eight months after vaccination, neutralizing antibody titers waned approximately 8-fold for vaccine-derived immunity for both BNT162b2 and mRNA-1273 vaccines and 3-fold for infection derived immunity, with most waning occurring in the first 3-5 months (Figure 2A, B; Table S2). Rates of waning were faster for the BNT162b2 vaccine than the mRNA-1273 vaccine and infection-derived immunity (Figure 2B; Table S2). Initial neutralizing antibody titers were highest for vaccination with mRNA-1273, followed by BNT162b2, and lowest following infection (Figure 2A; Table S2). ![Figure 2.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/13/2021.10.25.21265500/F2.medium.gif) [Figure 2.](http://medrxiv.org/content/early/2024/11/13/2021.10.25.21265500/F2) Figure 2. Waning of neutralizing antibody titers and VE for susceptibility (VES) and infectiousness (VEI) over time and following boosting with a third dose. Lines show fitted models and ribbons show ± 1 SE (Table S1). A) Neutralizing antibody titers relative to convalescent sera with wild-type (WT) virus, adjusted for the 2.2-fold lower neutralization titers with the Delta variant relative to wild-type virus. B) The same data as in panel A but rescaled relative to the initial peak value (i.e. all lines start at a value of 1) to show relative rates of waning and levels of stabilization (Table S1). C) VE for susceptibility (VES) over time using the patterns of waning in Panel A and the relationships between antibody titers and VE for susceptibility (Figure 1A). D) VE for infectiousness (VEI) over time using the patterns of waning in Panel A and the relationships between antibody titers and VE for infectiousness (Figure 1B). Points in the upper righthand corner of panels C and D, indicated by the black arrows, show the effect of third doses in boosting VE for the BNT162b2 and mRNA-1273 vaccines. We used the relationships between neutralizing antibody titers and VE (Figure 1) and patterns of waning of antibody titers (Figure 2A, B) to estimate the waning of VES and VEI over time (Figure 2C, D). As expected from levels of antibody titers, VES and VEI were highest for two-dose vaccination with mRNA-1273, and BNT162b2 vaccines, and then infection-derived immunity (Figure 2C, D). The relatively fast waning rate of neutralizing antibodies for the BNT162b2 vaccine suggests that VE for this vaccine was initially much higher than infection-derived immunity, but the two were similar after 3-4 months (Figure 2C, D). VE for the mRNA-1273 vaccine remained higher than both the BNT162b2 vaccine and previous infection as protection waned (Figure 2C, D). A third dose of the BNT162b2 vaccine boosted antibody titers 26-fold relative to levels after 8 months of waning, or 25.9/8.1 = 3.2-fold higher than one week after dose two 28. A third dose of the mRNA-1273 vaccine boosted antibody titers 13-fold relative to 8-month waned levels, or 1.5-fold higher than shortly after the second dose. The fitted relationship between VES and neutralizing antibody titer ratios (Figure 1) suggested that third doses of the mRNA vaccines would increase VES 1.6-1.8-fold from waned values of 47-54% to boosted values of 85-86% and would boost VEI 3-4-fold from 16-22% to 65-68% (Table S3; Figure 2C,D; compare blue and purple points in upper right of panels to lines of the same color below). Combined, boosting waned individuals with a third vaccine dose increased VE for transmission (VET = 1-(1-VES)*(1-VEI)) from 55.4% to 95.4% for BNT162b2 and from 64.3% to 94.5% for mRNA-1273. In doing so, boosting reduced the risk of transmission (1-VET) by ∼10-fold or 89.7% for BNT162b2 and ∼7-fold or 84.5% for mRNA-1273 vaccinated individuals compared to waned immunity. Next, we examined the impacts of third doses on population-level transmission in the United States via reductions in the reproductive rate of the pathogen, Rt. On September 1, 2021, the average time since vaccination for vaccinated individuals (53.2% of the population) was 4 months (118.3 days) (Figure S1) resulting in average antibody titers waning to less than 22% of peak values. If all doubly vaccinated individuals in the USA received a third dose of an mRNA vaccine that took effect on 1 September 2021 this would have reduced *Rt* by 18.4% (95% CI: 15.0% to 21.8%) from 1.17 to 0.96 (Figure 3D). This would have stopped the strong surge in cases, hospitalizations, and deaths that occurred over the next few months, assuming contact rates and other behaviors didn’t change following boosting (see Methods for further details). Notably, using the same number of doses to doubly vaccinate unvaccinated individuals (assuming they were willing to be vaccinated) rather than boost previously vaccinated individuals would have been similarly effective in reducing transmission (Rt would decrease from 1.17 to 0.94 versus 1.17 to 0.96 for using doses for boosting vaccinated individuals) (Figure 3D, compare right end of red line to red point labeled “R = 3.2, 79.8% vacc., no 3rd doses” on left side). We also explored several other scenarios with varying vaccination and infection rates and contact rates, which showed that the percent reduction in Rt was driven primarily by the fraction of the population vaccinated, which determined the potential boosting fraction (Supplemental Results Text, Figure S2). ![Figure 3.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/13/2021.10.25.21265500/F3.medium.gif) [Figure 3.](http://medrxiv.org/content/early/2024/11/13/2021.10.25.21265500/F3) Figure 3. United States population vaccination and infection history over time, population-level vaccine effectiveness against susceptibility (VES) and infectiousness (VEI) over time, and the relationship between the pathogen effective reproductive number, *Rt* and third dose coverage, as of 1 September, 2021. A) Fraction of the US population that had received their primary vaccination series but had not been infected (Frac. Vacc.), had been infected but not vaccinated (Frac. Inf.), had been both vaccinated and infected (Frac. Hybrid), or had received their third-dose booster (Frac. Boost). B) and C) Lines ± 1 SE show population-level VE against susceptibility (B) and infectiousness (C) over time for three subpopulations (vaccinated with their primary series, previously infected, and both vaccinated and previously infected [hybrid]) and the entire population in the US (“Total”) which is a mixture of the five subpopulations, as shown in panel (A). Rectangles above panels (A-C) show which variant was dominant during the period (frequency ≥ 50%). D) Line ± 1 SE shows Rt plotted against the fraction of vaccinated individuals receiving third dose boosters for the United States with R=3.2, 53.2% vaccinated, and 67.4% previously infected as of September 1, 2021 (see Methods for additional details). The single red point shows the impact of using all third doses from the right end of the red line to doubly-vaccinate unvaccinated individuals, which would bring the vaccination coverage from 53.2% to 79.8%. The dashed horizontal line shows the threshold reproductive number Rt = 1, separating growing from shrinking epidemics. Between June 2021, when the Delta variant emerged, and January 2022, the Delta variant caused an estimated 273,801 deaths 14, with the average number of estimated daily cases sometimes exceeding 200,000 (Figure S3), and reported deaths sometimes exceeding 2,000/day (Figure 4). However, during this period the estimated Rt for the US population mostly fluctuated between 0.9 and 1.1 (Figure S3), suggesting that many thousands of deaths could have been prevented if Rt could have been reduced, even marginally. As we just illustrated, boosting the 53.2% of the population that had been vaccinated with a third dose by 1 September, 2021 would have reduced Rt by 18.4% (Figure 3D, red line). This would have reduced the number of Delta deaths by 111,525 (95% CI: 103,819 to 116,641) (Figure 4E, blue shaded region) by averting 37,174,963 infections (95% CI: 34,606,246 – 38,880,500) (Figure 4A, orange shaded region). Boosting with third doses could have averted 629.5 (95% CI: 586.0 to 658.3) deaths per million vaccine doses administered due to effects on transmission alone. ![Figure 4.](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/11/13/2021.10.25.21265500/F4.medium.gif) [Figure 4.](http://medrxiv.org/content/early/2024/11/13/2021.10.25.21265500/F4) Figure 4. Estimated daily infections (A-D) and reported or estimated deaths (E-H) with and without third dose boosting plotted against time in the United States when the Delta SARS-CoV-2 variant was dominant. Black lines in the left panels show the estimated number of infections each day, calculated by deconvolution using the observed daily deaths (dark blue lines in right panels) and the delays between infection and death, and using and an infection fatality ratio (IFR) of 0.003 29. Deaths from the Delta variant were estimated using reported deaths and variant fractions based on sequencing data (see Methods). Red lines in left panels show the estimated number of infections with third dose boosting, and light blue lines show the number of deaths resulting from these infections with the timing estimated by reconvolution given the same distributions. The shaded regions show the averted infections (orange) and deaths (blue) if all eligible vaccinees received third doses that took full effect on Sept. 1, Oct. 1, Nov. 1, or Dec. 1 as indicated on the panel title with the reduction in Rt for these dates shown in Figure S2. Boosting populations after Sept. 1 would have resulted in greater reductions in Rt as immunity in the vaccinated group continued to wane but would avert fewer deaths caused by the Delta variant. Boosting on October 1, November 1, or December 1 would have reduced Rt by 29.7%, 31.2%, and 30.0%, respectively (Figure S2), but would avert 73,574 (95% CI: 69,253 – 75,962), 36,424 (95% CI: 33,085 – 38,350) and 5,970 (95% CI: 5,039 – 6,617) Delta deaths (Figure 4; Table S5). ## Discussion The indirect effects of vaccination in reducing disease by reducing transmission (via reduced susceptibility or infectiousness) were rarely considered at any stage of COVID-19 vaccination programs, including initial trial design, initial vaccination, and subsequent boosting. We found that boosting population immunity with third vaccine doses could have substantially reduced transmission of the most deadly SARS-CoV-2 variant, Delta, stopped surges, and averted huge numbers of deaths in many populations, including more than 100,000 in the US. We estimated that 629 deaths could have been averted per million vaccine doses which is similar to or higher than the direct effects of COVID-19 vaccination in other studies (93 to 836 deaths per million doses administered 30–33). However, many countries initially offered third doses only to protect at-risk individuals and only later offered third doses to the general population 20. Likely as a result, third dose uptake by the general population in most countries was low. It may have been higher if the third dose boosters were immediately and strongly recommended, or mandated for the whole population 34. The approach we developed here provides an empirical framework for estimating the effect of vaccination on transmission in a dynamic immune landscape. The key components are the temporal dynamics of a surrogate of protection neutralizing antibody titers and relationships between neutralizing antibody titers and the two components of transmission: susceptibility and infectiousness. Previous studies have shown that neutralizing antibody titers were strongly correlated with VE for symptomatic disease and severe disease 22–24,26,27. Here, we significantly advanced this approach by showing that antibody titers were also strongly correlated with VES and VEI. This advance is critical for building accurate mathematical models of SARS-CoV-2 transmission as immunity wanes and is boosted with additional vaccine doses or infection. Previous models have assumed VE for transmission was the same as VE for symptomatic disease26, which underestimates susceptibility (the relative risk of infection) and underestimates transmission (Figure S4). A similar empirical framework for estimating the effect of vaccination on transmission could be developed for other pathogens, such as influenza virus, but it would require a surrogate of protection35 and multiple estimates of vaccine efficacy for susceptibility and infectiousness in population with different levels of immunity. We found that VES and VEI declined rapidly in the first 3-5 months post-vaccination in BNT162b2 recipients and plateaued after ∼6 months, with slower, waning in mRNA-1273 recipients, which may or may not be related to long lasting plasma cells 36,37. These rates of waning suggest that vaccinating individuals multiple times per year would have substantially increased VE for transmission against the Delta variant, and would likely increase VE for transmission for the currently circulating SARS-CoV-2 variants as well. Given the summer surges of COVID-19 hospitalizations that now appear to be a repeatable phenomenon, recommending two vaccine doses per year for COVID-19, especially in sensitive settings such as skilled nursing facilities, could provide substantial benefits. We found third vaccine doses could have reduced Delta deaths dramatically, but the impact depended on the timing when doses were administered. Rolling out third doses by September 1 would have averted an additional 37,951 Delta deaths in the US compared to waiting until October 1, which emphasizes the timeliness of vaccinations. Two related factors made a September 1 rollout especially impactful. First, infections were near their peak in September. Second, Rt was below 1.03 from September 1 until late November when the Omicron variant began to invade. The potential reduction in Rt due to third doses (18.4%) was sufficient to reduce Rt well below 1 which would have driven Delta infections and deaths down rapidly (Figure 4). Administering third doses later would have averted fewer deaths caused by the Delta variant (but may have averted additional infections and deaths caused by the Omicron variant). Our study has several limitations. First, our analyses use population averages for estimates of VES and VEI and ignore age-specific variation among individuals, as well as other factors. Age-specific patterns of contact, disease severity, and immunity can play a key role in shaping epidemics 38. We also assumed well-mixed populations in calculating reductions in the reproductive number Rt. A targeted vaccination approach would be more effective than that outlined here if highly connected individuals could be targeted for third doses 39. Third, our calculations for the reductions in Rt with third doses were based on neutralizing antibody titers one month after the administration of a third dose. Neutralizing antibody titers also decay following third doses, but the decay rate is much slower than after second doses 40, suggesting that substantial reductions in Rt from third doses would persist for many months, as we assumed here. Fourth, our analyses were focused on population characteristics in the US in late 2021 and vaccine effectiveness against the Delta variant because data were available for quantifying immunity in populations at this time. The current impact of vaccine doses on transmission of SARS-CoV-2 is more difficult to quantify with currently available data, due to a more complex immune landscape, with most people having been infected several times with multiple virus variants and vaccinated with multiple vaccine types, and the variant composition of SARS-CoV-2 viruses is sometimes diverse and changes frequently. Determining the current benefit of vaccination for reducing transmission would require estimates of the distribution of neutralizing antibody titers against currently circulating viruses in target populations, and measuring increases in antibody titers following vaccination. Finally, quantifying the impact of vaccination in reducing deaths due to reduced transmission does not take into account additional benefits of reduced infections including reduced absenteeism from work and school and reduced cases of long COVID. The total benefits of reduced transmission may outweigh the costs of administering additional vaccine doses in many populations. In summary, we showed that the indirect benefits of vaccination in reducing transmission can lead to a huge number of deaths averted. In addition, reductions in transmission increase health equity by reducing infection in individuals that lack access to vaccination. Quantifying vaccine impacts on transmission should be an integral part of developing new vaccines, and estimates of reductions in transmission should be used for planning vaccination strategies, including vaccine selection, targeting, and vaccination schedules. ## Methods ### Relative neutralizing antibody titers by variant, vaccine dose and with waning We collected data on neutralizing antibody titers from the literature and estimated the total neutralizing antibody titer ratio (NATRtot) for each vaccine, virus variant, and immune status 24. NATRtot is the ratio of neutralizing antibody titer for a given vaccine, variant and immune status, relative to the neutralizing antibody titer of convalescent sera after infection with the original wild-type virus 24,27. For example, two weeks after a second dose of the BNT162b2 vaccine NATRtot for the Delta variant is 1.08, because vaccination with BNT162b2 results in neutralizing antibody titers 2.37-fold higher than infection 24,27, and neutralizing antibody titers against the Delta variant are 2.20-fold lower than against wild-type virus; 2.37*(1/2.2) = 1.08 24. We also collected data from studies measuring neutralizing antibody titers over time in sets of individuals following infection or vaccination to quantify waning of neutralizing antibody titers, as well as increases in antibody titers following boosting with third doses28,41,42. We then fit models to the decay in antibody titers over time following vaccination with two mRNA vaccines (BNT162b2 and mRNA-1273), and infection (Table S1). We assumed antibody titers from hybrid immunity (infection followed by vaccination) waned at the same rate as following BNT 162b2 vaccination based on a study that measured both 42. ### VE for susceptibility and infectiousness We collected VE estimates for susceptibility and infectiousness for SARS-CoV-2 from the literature, including a systematic living review 43, and categorized each study by vaccine, time since vaccination, and variant type (wild-type/D614G, Alpha, Delta). We excluded estimates where the virus variant could not be determined. For estimates of VES, we only included studies where the endpoint was “all infections” (often measured using frequent testing or serology) and excluded studies where the endpoint was “any infection” because “any infection” studies do not capture all infections and include an unknown fraction of asymptomatic infections. Studies of VEI, or transmission given infection, usually compared secondary attack rates in households by vaccination status of the index case. To fit a relationship between VE and NATR, we needed estimates of the number of infections in the vaccine and control groups. However, most studies did not report the number of infections; however, all reported a VE and 95% CI. We estimated the effective number of infected individuals in the control group (Ic) and the effective number of infected individuals in the vaccine group (Iv) for each study by determining the number of each needed to match the mean and 95% CI given in a study. We held the effective number of individuals in the control group (Nc) and vaccine group (Nv) constant at 1,000,000 because the 95% CI was invariant to variation in these values for observed incidence values. We used a maximum value of 1000 infections in the control group (Ic) to reduce the undue leverage of some very large studies 44. ### Relationships between VE and neutralizing antibody titers by vaccine and variant We modeled the relationship between VE and total neutralizing antibody titer ratios (NATRtot) using 24: ![Formula][1] We estimated the coefficients c and c1 separately for the two endpoints, susceptibility and infectiousness, by maximizing the likelihood of observing the data underlying the VE estimates: the number of infections in the control group, Ic, and vaccine group, Iv, for each study. The likelihood was the product of two binomial distributions with Nc and Nv individuals: ![Formula][2] Here bs is the baseline risk for the study period, which is the fraction of control individuals infected during the study. We maximized the likelihood across all studies which was the product of the likelihoods of the individual studies. We used the fitted relationships for the two endpoints, susceptibility and infectiousness, between VE and neutralizing antibody titer ratios, NATRtot (eq 1) to estimate VE for populations based on the estimated NATRtot given the timing of vaccination and infection and estimated rates of waning as described above. Finally we calculated the VE for transmission, VET = 1-(1-VES)*(1-VEI). We calculated 95% CIs for VET using independent bootstrap samples from the VES and VEI estimates and 95% CIs. ### Vaccine effectiveness over time We estimated VES and VEI in the US for each day from January 1, 2021 to December 13, 2021 for three subpopulations: previously infected and unvaccinated, twice vaccinated and no previous infection, and twice vaccinated and previously infected (hybrid immunity), the timing of prior infections and vaccinations (Figure S1), and the relationships between neutralizing antibody titers and VES and VEI. We estimated the timing of infections using reported deaths 48 and deconvolution, using the distributions for the delays between infection and death (i.e. infection to shedding, shedding to symptom onset, symptom onset to hospitalization, and hospitalization to death) (Table S6) using the deconvolve package in R 45. For each day, we estimated VES and VEI in each subpopulation against each circulating variant based on the timing of infections and vaccinations, accounting for the waning that occurred in each subpopulation on that date using the fitted model for the relationship between neutralizing antibodies and time since infection and vaccination. We used sequencing data from GISAID 29,46 (offset by 9 days to account for the average delay between infection and reporting date) in the US to get a variant-averaged estimate of VES and VEI in each subpopulation for each day using a weighted mean where weights were the observed variant frequency each day. We then estimated a population-wide VE for each day using estimates for the fraction of the population in five subpopulations: fully susceptible, third-dose boosted, and the three subpopulations described above. ### The impact of third doses on the reproductive number, Rt We used patterns of waning and boosting of neutralizing antibody titers to estimate the impact of providing additional doses of mRNA vaccines to increasing fractions of vaccinated individuals in the USA and other populations on the reproductive number, *Rt*, of the Delta variant of SARS-CoV-2, at four time points in autumn 2021; Sept 1, Oct 1, Nov 1, and Dec 1 (Table S4). *Rt* is the average number of secondary cases that each case infects and is equal to the basic reproductive number for a pathogen in a fully susceptible population, *R*, multiplied by the fraction of the population that is susceptible (i.e. not immune from vaccination or previous infection). We examined the effect of boosting vaccinated individuals with a third dose on Rt by considering five scenarios which were combinations of vaccination, previous infection and contact rates that span the range of circumstances in many countries in late 2021. The five scenarios (with fraction vaccinated and previously infected shown for 1 September 2021) and rationale were (Table S4): * 1) US/D (United States with distancing): R = 3.2, 53.2.% vaccinated, 67.4% previously infected: approximated fraction of US population previously infected with contact rates similar to summer-fall 2021 * 2) CA/D (California, USA, with distancing): R = 3.2, 58.7% vaccinated, 59.8% previously infected: approximated some populations with higher vaccination and lower fraction infected than scenario (1) (e.g. California) * 3) NZ/ND (New Zealand, no distancing): R =7.0, 24.3% vaccinated, 0.2% previously infected: approximated countries with high vaccination rates that effectively suppressed transmission and had reduced contact rates when cases were detected (e.g. New Zealand, Australia, Hong Kong, etc.) * 4) US/ND (United States, no distancing): R = 7.0, 53.2% vaccinated, 67.4% previously infected: a scenario to compare to scenario (1) to determine if boosting could limit transmission if behavior returned to pre-pandemic levels * 5) US/ND-100 (United States, no distancing): R0 = 7.0, 100% vaccinated, 67.4% previously infected: a hypothetical optimistic scenario to compare to scenario (1) to determine if vaccination with or without boosting could limit transmission if behavior returned to pre-pandemic levels We estimated the fraction of the population that was susceptible using data on the fraction of the population that had been vaccinated or infected or both. We first split each population into three groups based on vaccination status: fraction unvaccinated, fU, fraction vaccinated, fV, with two doses of either BNT162b2 or mRNA-1273, and fraction boosted with a third dose of the same vaccine used for the primary series, fB (fU+fV+fB=1). We calculated the fraction of people in the USA receiving BNT162b2 and mRNA-1273 from vaccination records 47. We estimated the timing of infections using deconvolution of deaths as described above and calculated the number of infections by dividing deaths by an estimate of the infection fatality ratio (IFR) across of 2020-21 of 0.003 29; the results were not sensitive to moderate variation in this estimate. We also examined estimating infections using cases and an infection/case ratio of 4.0 48 which produced a lower estimate of total infections, and the reduction in Rt with third dose boosting and deaths averted were very similar to the death-deconvolution and IFR method. We estimated the fraction of unvaccinated or vaccinated individuals that had been previously infected using the ratios of cases in these two groups over time in the USA 49 (Figure S5). We used vaccination and infection data to calculate the fraction of people previously infected among unvaccinated, fPU, and vaccinated, fPV individuals (Table S4). We used rates of waning of antibody titers and data on the timing of vaccination and infection from the USA and New Zealand 47 for US and New Zealand scenarios, respectively, to estimate the distribution of NATRtot in populations for the four time points in autumn 2021. We used the distributions of NATRtot to estimate the relative susceptibility, 1-VES, (Figure 1A) and relative infectiousness, 1-VEI, (Figure 1B) for the population in each of four groups: unvaccinated and previously infected (VESP and VEIP), vaccinated twice and no previous infection (VESV and VEIV), vaccinated three times or boosted (VESB and VEIB), and vaccinated and infected or hybrid immunity (VESH and VEIH), all with appropriate adjustments for waning antibody titers (see below). We used these estimates of VE to calculate Rt for populations composed of different fractions of five groups: never infected unvaccinated (1-fPU)fU, previously infected unvaccinated fPfU, previously uninfected vaccinated with two doses (1-fPV)fV, previously infected and vaccinated with two doses fPVfV, and boosted with a third dose, regardless of previous infection, fB: ![Formula][3] Note that in equation (3) the VE values for each population are integrated across the distribution of NATRtot values for that group, and the distribution of NATRtot values in each group reflects the timing of vaccination and infection on a daily timestep. We used the scenario estimates of the fraction of population with different types of immunity (with timing matched to vaccination and infection and waning as described above) for susceptibility and infectiousness to estimate values of Rt under each scenario at the four time points in autumn 2021 (Sept. 1, Oct. 1, Nov. 1, and Dec. 1; Table S4). We calculated 95% CIs for predicted values of Rt that incorporated uncertainty in the relationship between time since vaccination or infection and NATRistatus, uncertainty in the relationships between neutralizing antibody titers and VES and VEI, as well as uncertainty in NATRvac and NATRvar (Figures 1 and S2). ### The impact of third doses on infections and deaths caused by the Delta variant We estimated the timing of the infections that gave rise to observed deaths and total infections in the United States in autumn 2021 using deconvolution as described above. We used SARS-CoV-2 sequencing data 29,46 to estimate the daily fraction of all infections that were caused by the Delta variant. We estimated the reduction in Delta infections if third-dose boosters were given to all fully vaccinated individuals at four time points in autumn 2021. We assumed the additional doses would take full effect simultaneously on Sept. 1, Oct. 1, Nov. 1, or Dec. 1. We estimated the number of Delta infections that would have occurred if third doses had been deployed by multiplying the number of estimated infections on each day by the reduction in the daily Rt due to third doses (e.g. 0.812 if Rt was reduced by 18.8%), raised to the power of the number of generations of infections (generation time 3.8d 50) since Sept 1 (or Oct. 1, Nov. 1 or Dec. 1). The calculated reduction in Rt was the difference between the number of actual third-doses given, and the number of doses deployed in the scenario. We estimated the number of deaths that would have been averted in each scenario by re-convoluting the reduced infections, multiplied by the IFR, using the delays described previously. We used uncertainty associated with our Rt estimates to calculate 95% CIs in the number of deaths averted. ## Supporting information Supplemental materials [[supplements/265500_file02.pdf]](pending:yes) ## Data Availability All data produced in the present work are contained in the manuscript, and supplemental material [https://github.com/marmkilpatrick/Vaccine-boosters](https://github.com/marmkilpatrick/Vaccine-boosters) ## Author Contributors BJG and AMK conceived the study, performed the analyses and wrote the paper. ## Declaration of competing interests All authors declare no competing interests. ## Data sharing All data used in the paper are contained in the supplemental material. R Code and data files to replicate the figures and analyses of this paper are available: [https://github.com/marmkilpatrick/Vaccine-boosters](https://github.com/marmkilpatrick/Vaccine-boosters) Upon publication, R Code will be uploaded to Zenodo and data files will be archived with Dryad. ## Funding We received funding from the National Science Foundation (DEB-1911853, 1115895, and 1717498) and the California Department of Health. ## Acknowledgements ## Footnotes * **Classification** Biological Sciences: Applied Biological Sciences * The manuscript has been revised and updated to refine the analyses and focus on impacts in the US in 2021. * Received October 25, 2021. * Revision received November 13, 2024. * Accepted November 13, 2024. * © 2024, Posted by Cold Spring Harbor Laboratory This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/) ## References 1. Allen A. Vaccine—the controversial story of medicine’s greatest lifesaver. New York, NY: W.W. Norton & Co., 2007. 2. Minor PD. Live attenuated vaccines: Historical successes and current challenges. Virology 2015; 479–480: 379–92. 3. Greenwood B. The contribution of vaccination to global health: past, present and future. Phil Trans R Soc B 2014; 369: 20130433. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1098/rstb.2013.0433&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=24821919&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 4. Hinman AR, Orenstein WA, Schuchat A. Vaccine-preventable diseases, immunizations, and MMWR--1961-2011. MMWR Suppl 2011; 60: 49–57. 5. Roush SW. Historical Comparisons of Morbidity and Mortality for Vaccine-Preventable Diseases in the United States. JAMA 2007; 298: 2155. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jama.298.18.2155&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=18000199&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000250869800021&link_type=ISI) 6. Hudgens MG, Gilbert PB, Self SG. Endpoints in vaccine trials. Stat Methods Med Res 2004; 13: 89–114. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1191/0962280204sm356ra&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15068256&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000220254900002&link_type=ISI) 7. Li X, Mukandavire C, Cucunubá ZM, et al. Estimating the health impact of vaccination against ten pathogens in 98 low-income and middle-income countries from 2000 to 2030: a modelling study. The Lancet 2021; 397: 398–408. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0140-6736(20)32657-X&link_type=DOI) 8. COVID-19: Developing Drugs and Biological Products for Treatment or Prevention. 2023. [https://www.fda.gov/regulatory-information/search-fda-guidance-documents/covid-19-developing-drugs-and-biological-products-treatment-or-prevention](https://www.fda.gov/regulatory-information/search-fda-guidance-documents/covid-19-developing-drugs-and-biological-products-treatment-or-prevention). 9. CDC. Child and Adolescent Immunization Schedule by Age. 2023. [https://www.cdc.gov/vaccines/schedules/hcp/imz/child-adolescent.html](https://www.cdc.gov/vaccines/schedules/hcp/imz/child-adolescent.html). 10. CDC. Direct and indirect effects of routine vaccination of children with 7-valent pneumococcal conjugate vaccine on incidence of invasive pneumococcal disease--United States, 1998-2003. 2005. [https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5436a1.htm](https://www.cdc.gov/mmwr/preview/mmwrhtml/mm5436a1.htm). 11. Anderson RM, May RM. Vaccination and herd immunity to infectious diseases. 1985. 12. Medlock J, Galvani AP. Optimizing Influenza Vaccine Distribution. Science 2009; 325: 1705–8. [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6Mzoic2NpIjtzOjU6InJlc2lkIjtzOjEzOiIzMjUvNTk0OC8xNzA1IjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMTMvMjAyMS4xMC4yNS4yMTI2NTUwMC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 13. Marmot M. Achieving health equity: from root causes to fair outcomes. The Lancet 2007; 370: 1153–63. 14. Walker J, Grubaugh ND, Gonsalves G, Pitzer V, Rizvi Z. One Million and Counting: Estimates of Deaths in the United States from Ancestral SARS-CoV-2 and Variants. medRxiv 2022; : 2022.05.31.22275835. 15. Campbell F, Archer B, Laurenson-Schafer H, et al. Increased transmissibility and global spread of SARS-CoV-2 variants of concern as at June 2021. Eurosurveillance 2021; 26. DOI:10.2807/1560-7917.ES.2021.26.24.2100509. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.2807/1560-7917.ES.2021.26.24.2100509&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34142653&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 16. Elliott P, Haw D, Wang H, et al. Exponential growth, high prevalence of SARS-CoV-2, and vaccine effectiveness associated with the Delta variant. Science 2021; 374: eabl9551. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1126/science.abl9551&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34726481&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 17. Planas D, Veyer D, Baidaliuk A, et al. Reduced sensitivity of SARS-CoV-2 variant Delta to antibody neutralization. Nature 2021; 596: 276–80. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1101/2021.05.26.445838&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34237773&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 18. Lustig Y, Zuckerman N, Nemet I, et al. Neutralising capacity against Delta (B.1.617.2) and other variants of concern following Comirnaty (BNT162b2, BioNTech/Pfizer) vaccination in health care workers, Israel. Eurosurveillance 2021; 26. DOI:10.2807/1560-7917.ES.2021.26.26.2100557. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.2807/1560-7917.ES.2021.26.26.2100557&link_type=DOI) 19. Wu K, Werner AP, Moliva JI, et al. mRNA-1273 vaccine induces neutralizing antibodies against spike mutants from global SARS-CoV-2 variants. bioRxiv 2021; published online Jan 25. DOI:10.1101/2021.01.25.427948. [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoiYmlvcnhpdiI7czo1OiJyZXNpZCI7czoxOToiMjAyMS4wMS4yNS40Mjc5NDh2MSI7czo0OiJhdG9tIjtzOjUwOiIvbWVkcnhpdi9lYXJseS8yMDI0LzExLzEzLzIwMjEuMTAuMjUuMjEyNjU1MDAuYXRvbSI7fXM6ODoiZnJhZ21lbnQiO3M6MDoiIjt9) 20. Bar-On YM, Goldberg Y, Mandel M, et al. Protection Across Age Groups of BNT162b2 Vaccine Booster against Covid-19. medRxiv 2021; published online Oct 7. DOI:10.1101/2021.10.07.21264626. [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoibWVkcnhpdiI7czo1OiJyZXNpZCI7czoyMToiMjAyMS4xMC4wNy4yMTI2NDYyNnYxIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMTMvMjAyMS4xMC4yNS4yMTI2NTUwMC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 21. Krause PR, Fleming TR, Peto R, et al. Considerations in boosting COVID-19 vaccine immune responses. The Lancet 2021; 398: 1377–80. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/s0140-6736(21)02046-8&link_type=DOI) 22. Cromer D, Steain M, Reynaldi A, et al. Neutralising antibody titres as predictors of protection against SARS-CoV-2 variants and the impact of boosting: a meta-analysis. The Lancet Microbe 2022; 3: e52–61. 23. Cromer D, Steain M, Reynaldi A, et al. Predicting vaccine effectiveness against severe COVID-19 over time and against variants: a meta-analysis. Nat Commun 2023; 14: 1633. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41467-023-37176-7&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36964146&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 24. Gardner BJ, Kilpatrick AM. Predicting Vaccine Effectiveness for Hospitalization and Symptomatic Disease for Novel SARS-CoV-2 Variants Using Neutralizing Antibody Titers. Viruses 2024; 16. DOI:10.3390/v16030479. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.3390/v16030479&link_type=DOI) 25. Hogan AB, Doohan P, Wu SL, et al. Estimating long-term vaccine effectiveness against SARS-CoV-2 variants: a model-based approach. Nat Commun 2023; 14: 4325. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37468463&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 26. Hogan AB, Wu SL, Toor J, et al. Long-term vaccination strategies to mitigate the impact of SARS-CoV-2 transmission: A modelling study. PLoS Med 2023; 20: e1004195. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38016000&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 27. Khoury DS, Cromer D, Reynaldi A, et al. Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection. Nat Med 2021; 27: 1205–11. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41591-021-01377-8&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34002089&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 28. Falsey AR, Frenck RW, Walsh EE, et al. SARS-CoV-2 Neutralization with BNT162b2 Vaccine Dose 3. N Engl J Med 2021; 385: 1627–9. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1056/NEJMC2113468&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34525276&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 29. Kilpatrick AM. Ecological and Evolutionary Insights About Emerging Infectious Diseases from the COVID-19 Pandemic. Annu Rev Ecol Evol Syst 2023; 54: 171–93. 30. Kayano T, Sasanami M, Kobayashi T, et al. Number of averted COVID-19 cases and deaths attributable to reduced risk in vaccinated individuals in Japan. The Lancet Regional Health - Western Pacific 2022; 28: 100571. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35971514&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 31. Haas EJ, McLaughlin JM, Khan F, et al. Infections, hospitalisations, and deaths averted via a nationwide vaccination campaign using the Pfizer–BioNTech BNT162b2 mRNA COVID-19 vaccine in Israel: a retrospective surveillance study. The Lancet Infectious Diseases 2022; 22: 357–66. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/s1473-3099(21)00566-1&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34562375&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 32. Santos CVBD, Noronha TGD, Werneck GL, Struchiner CJ, Villela DAM. Estimated COVID-19 severe cases and deaths averted in the first year of the vaccination campaign in Brazil: a retrospective observational study. The Lancet Regional Health - Americas 2023; 17: 100418. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36575682&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 33. Mendes D, Chapman R, Gal P, et al. Public health impact of booster vaccination against COVID-19 in the UK during Delta variant dominance in autumn 2021. Journal of Medical Economics 2022; 25: 1039–50. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36097853&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 34. Mello MM, Opel DJ, Benjamin RM, et al. Effectiveness of vaccination mandates in improving uptake of COVID-19 vaccines in the USA. The Lancet 2022; 400: 535–8. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/S0140-6736(22)00875-3&link_type=DOI) 35. Ward BJ, Pillet S, Charland N, Trepanier S, Couillard J, Landry N. The establishment of surrogates and correlates of protection: Useful tools for the licensure of effective influenza vaccines? Human Vaccines & Immunotherapeutics 2018; 14: 647–56. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29252098&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 36. Nguyen DC, Hentenaar IT, Morrison-Porter A, et al. SARS-CoV-2-specific plasma cells are not durably established in the bone marrow long-lived compartment after mRNA vaccination. Nat Med 2024; published online Sept 27. DOI:10.1038/s41591-024-03278-y. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41591-024-03278-y&link_type=DOI) 37. Schulz AR, Fiebig L, Hirseland H, et al. SARS-CoV-2 specific plasma cells acquire long-lived phenotypes in human bone marrow. eBioMedicine 2023; 95: 104735. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37556944&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 38. Yuan H-Y, Baguelin M, Kwok KO, Arinaminpathy N, Van Leeuwen E, Riley S. The impact of stratified immunity on the transmission dynamics of influenza. Epidemics 2017; 20: 84–93. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.epidem.2017.03.003&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28395850&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 39. Firth JA, Hellewell J, Klepac P, Kissler S, Kucharski AJ, Spurgin LG. Using a real-world network to model localized COVID-19 control strategies. Nat Med 2020; 26: 1616–22. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=http://www.n&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 40. Gilboa M, Regev-Yochay G, Mandelboim M, et al. Durability of Immune Response After COVID-19 Booster Vaccination and Association With COVID-19 Omicron Infection. JAMA Netw Open 2022; 5: e2231778. 41. Chu L, Vrbicky K, Montefiori D, et al. Immune response to SARS-CoV-2 after a booster of mRNA-1273: an open-label phase 2 trial. Nat Med 2022; 28: 1042–9. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35241844&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 42. Goel RR, Painter MM, Apostolidis SA, et al. mRNA vaccines induce durable immune memory to SARS-CoV-2 and variants of concern. Science 2021; 374: abm0829. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34648302&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 43. Higdon MM, Wahl B, Jones CB, et al. A Systematic Review of Coronavirus Disease 2019 Vaccine Efficacy and Effectiveness Against Severe Acute Respiratory Syndrome Coronavirus 2 Infection and Disease. Open Forum Infectious Diseases 2022; 9: ofac138. [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35611346&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 44. Lewnard JA, Patel MM, Jewell NP, et al. Theoretical Framework for Retrospective Studies of the Effectiveness of SARS-CoV-2 Vaccines. Epidemiology 2021; 32: 508–17. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1097/EDE.0000000000001366&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34001753&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 45. Lewnard JA, Liu VX, Jackson ML, et al. Incidence, clinical outcomes, and transmission dynamics of hospitalized 2019 coronavirus disease among 9,596,321 individuals residing in California and Washington, United States: a prospective cohort study. 2020; published online April 16. DOI:10.1101/2020.04.12.20062943. [Abstract/FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6NDoiQUJTVCI7czoxMToiam91cm5hbENvZGUiO3M6NzoibWVkcnhpdiI7czo1OiJyZXNpZCI7czoyMToiMjAyMC4wNC4xMi4yMDA2Mjk0M3YxIjtzOjQ6ImF0b20iO3M6NTA6Ii9tZWRyeGl2L2Vhcmx5LzIwMjQvMTEvMTMvMjAyMS4xMC4yNS4yMTI2NTUwMC5hdG9tIjt9czo4OiJmcmFnbWVudCI7czowOiIiO30=) 46. Shu Y, McCauley J. GISAID: Global initiative on sharing all influenza data – from vision to reality. Eurosurveillance 2017; 22. DOI:10.2807/1560-7917.ES.2017.22.13.30494. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.2807/1560-7917.ES.2017.22.13.30494&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28382917&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) 47. Mathieu E, Ritchie H, Rodés-Guirao L, et al. Coronavirus Pandemic (COVID-19). Our World in Data. 2020. 48. CDC. Estimated COVID-19 Burden. 2021. [https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/burden.html](https://www.cdc.gov/coronavirus/2019-ncov/cases-updates/burden.html). 49. COVID Data Tracker. CDC. 2021. [https://covid.cdc.gov/covid-data-tracker/](https://covid.cdc.gov/covid-data-tracker/). 50. Park SW, Sun K, Abbott S, et al. Inferring the differences in incubation-period and generation-interval distributions of the Delta and Omicron variants of SARS-CoV-2. Proceedings of the National Academy of Sciences 2023; 120: e2221887120. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1073/pnas.2221887120&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37216529&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F11%2F13%2F2021.10.25.21265500.atom) [1]: /embed/graphic-5.gif [2]: /embed/graphic-6.gif [3]: /embed/graphic-7.gif