Global Estimates of Lives Saved by COVID-19 Vaccination 2020-2024 PDF

medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity....

medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Global estimates of lives and life-years saved by COVID-19 vaccination during 2020-2024 John P.A. Ioannidis (1-3), Angelo Maria Pezzullo (3,4), Antonio Cristiano (3,4), Stefania Boccia (4,5) (1) Department of Medicine, Stanford University School of Medicine, Stanford, CA, USA (2) Department of Epidemiology and Population Health, Stanford University School of Medicine, Stanford, CA, USA (3) Meta-Research Innovation Center at Stanford (METRICS), Stanford University, Stanford, CA, USA (4) Section of Hygiene, Department of Life Sciences and Public Health, Università Cattolica del Sacro Cuore, Rome, Italy (5) Department of Woman and Child Health and Public Health, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy E-mail: [email protected] Correspondence to: John P.A. Ioannidis, MD, DSc, [email protected] Funding: None Acknowledgment: The work of John Ioannidis is supported by an unrestricted gift from Sue and Bob O’Donnell to Stanford University. Competing interest statement: None. 1 NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice. medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Contributions: J.P.A.I. had the original idea, developed the new methods, performed analyses, and wrote the manuscript. A.M.P. and A.C. discussed concepts, collected background data and previous studies, interpreted analyses, and commented on the manuscript. S.B. discussed concepts, interpreted analyses, and commented on the manuscript. All authors have approved the final version. Keywords: COVID-19, vaccines, life expectancy, mortality, deaths Conflicts of interest: none Data: All key data are in the manuscript. Word count: 2500; abstract: 275; 46 references; supplement (appendix 1, appendix 2, 26 supplementary references). 2 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. ABSTRACT Estimating global lives and life-years saved is important to put into perspective the benefits of COVID-19 vaccination. Prior studies have focused mainly on the pre-Omicron period or only on specific regions, lack crucial life-year calculations, and often depend on strong modeling assumptions with unaccounted uncertainty. We aimed to calculate the lives and life-years saved by COVID-19 vaccination worldwide from the onset of the vaccination campaigns and until October 2024. We considered different strata according to age; community-dwelling and long-term care residence status; pre-Omicron and Omicron periods; and vaccination before and after a SARS-CoV-2 infection. In the main analysis, 2.533 million deaths were averted. Eighty-two percent were among people vaccinated before any infection, 57% were in the Omicron period, and 90% pertained to people 60 years and above. Sensitivity analyses suggested 1.4 to 4.0 million lives saved. Some sensitivity analyses showed preponderance of the benefit during the pre-Omicron period. We estimated 14.8 million life-years saved (sensitivity range, 7.4-23.6 million life-years). Most life-years saved (76%) were in people over 60 years old, but long-term care residents contributed only 2% of the total. Children and adolescents (0.01% of lives saved and 0.1% of life-years saved) and young adults 20-29 years old (0.07% of lives saved and 0.3% of life-years saved) had very small contributions to the total benefit. Based on a number of assumptions, these estimates are substantially more conservative than previous calculations focusing mostly on the first year of vaccination, but they still undeniably demonstrate a major overall benefit from COVID-19 vaccination during 2020-2024. The vast majority of benefit in lives and life-years saved was secured for a portion of the elderly minority of the global population. 3 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. INTRODUCTION The development and wide implementation of COVID-19 vaccines are widely considered major successes for biomedical research and public health (1,2). It is important to estimate the number of lives saved by COVID-19 vaccination worldwide since their introduction. Previous efforts to estimate deaths averted by COVID-19 vaccines used epidemic modeling or counterfactuals from surveillance data (3-6). Models may give unreliable results, depending on assumptions (7,8). Moreover, most previous publications addressed the pre-Omicron period (3,4). The few that have included Omicron period data (5,6) focused on specific regions and have not calculated probable life-years saved. Life-year estimates are pivotal in decision-making. Here we estimate both lives and life-years saved by COVID-19 vaccination worldwide until October 2024, separating different age strata, community-dwelling and long- term care populations, pre-Omicron and Omicron periods, and populations vaccinated before or after SARS-CoV-2 infection. METHODS Outline of calculations We consider strata based on age and long-term care residence status. Furthermore, we separate the period until November 2021 (pre-Omicron) and the subsequent period (Omicron); and people vaccinated before any SARS-CoV-2 infection from those vaccinated after previous infection. Stratifications are important because the infection fatality rate (IFR) varies markedly across strata. The number of lives saved in each stratum i is the product of the number who would have died absent vaccination, and vaccine effectiveness (VEi) for mortality. The number of people who would have died is the product of the total stratum population Ni, the proportion who would have been infected PI*i (absent vaccination), and the respective IFRi: 4 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Li = Ni x PI*i x IFRi x VEi. Total lives saved are L = Σ Li = Σ (Ni x PI*i x IFRi x VEi.) Similarly, life-years saved LYi are proportional to Li, the stratum-specific life expectancy (LE) LEi, and to a factor fi that denotes how LE of those who died may have differed versus general population LE; f takes smaller values when those who die are in worse health than the respective same-stratum general population. Thus: LYi = Li x LEi, x fi Total life-years saved are LY= Σ LYi = Σ (Li x LEi, x fi) We first calculate the benefits for people vaccinated before any SARS-CoV-2 infection. For those first vaccinated after having at least one SARS-CoV-2 infection, we then assume that PI*i x IFRi x VEi is lower by R-fold (mostly because of lower IFR in re-infection and lower PI*). Values used and sensitivity analyses For details on values used, justification (with supporting references), and sensitivity analyses’ ranges, see Appendix 1: Supplementary Methods. In brief, we use the 2021 world population pyramid age strata 0-19, 20-29, 30-39, 40-49, 50-59, 60-69, and 70 years and higher, dividing the last stratum further in community-dwelling (97%) and long-term care residents (3%). We assume that 10% in 0-19 years old, 20% in 20-29 years old, and 46% in higher age strata (overall 30% [sensitivity range 25-35%, retaining same age ratios]) had received at least one dose before any infection pre-Omicron. We assume that during Omicron, the remaining 56% of global population who remained uninfected by November 2021 were infected at least once until October 2024. An additional 18% of the global 5 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. population were first vaccinated during Omicron with slightly less than a third (5%) receiving at least one dose before being infected. We assume that absent vaccination all people would have been infected during the Omicron period. For pre-Omicron, we assume PI*=20% for all age strata by November 2021 (sensitivity range, 10%-40%) and that 5% of the population were first infected in the pre- Omicron period after vaccination. For IFR in unvaccinated people pre-Omicron, we use estimates from a systematic review for non-elderly age strata; from meta-regressions for community-dwelling 70 years old and over; and from a meta-analysis of case-fatality rates and studies estimating asymptomatic infection rates for long-term care residents. Sensitivity range is informed by the same sources. Omicron IFR among unvaccinated is assumed to be one-third of pre- Omicron values. We assume VE=75% (sensitivity range, 40-85%) pre-Omicron and 50% (sensitivity range, 30-70%) during Omicron. For people vaccinated after at least one infection, we assume R=5 (sensitivity range, 2.5-10). For LE, the UN population division life table for 2021 (World, both sexes) is used taking the mid-point in each age bracket. For 70 years and above, LE at age 77 is considered for community-dwelling individuals and 2 years for long-term care facility residents. The main analysis considers f=0.5 for all strata (sensitivity range, 0.25-0.8). RESULTS Lives saved Table 1 shows the characteristics of the different strata used in the calculations (see Appendix 1: Supplementary Methods (9-19)). In the main analysis (Table 2), 2.533 million lives were estimated to have been saved. They were mostly among people who were vaccinated before any infection (2.079/2.533, 82%). There were slightly more lives saved in 6 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. the Omicron period (1.448/2.533, 57%). 89.6% of lives saved pertained to people 60 years and above. Children and adolescents accounted for only 0.01% of total lives saved and young adults 20-29 years old for another 0.07%. Sensitivity analyses Table 3 shows values ranging from 1.4 to 4.0 million lives saved in one-way sensitivity analyses. Benefits tended to be larger for the Omicron period, but not when R values were low or when pre-Omicron PI* was assumed to be large. The widest range for two-way sensitivity analyses was 1.0 to 6.0 million (considering lower and upper range for both VE and R). Life-years saved In the main analysis, there were 14.8 million life-years saved, with sensitivity analyses ranging between 7.4 and 23.6 million life-years (Table 4). People over 60 years old accounted for most life-years saved (75.9%), but with very little contribution from long-term care residents (2% of total). 40-59 years old people also contributed a sizeable 20.6% of total. Children and adolescents (0.1%) and young adults 20-29 years old (0.3%) had negligible contributions. DISCUSSION We estimate that COVID-19 vaccination during 2020-2024 saved 2.5 million lives for 15 million life-years. Sensitivity analyses suggested ranges between 1.4 and 4 million averted deaths with 7.4-24 million life-years saved. However, uncertainty is substantially wider with multiple factors considered concurrently in multiple-way sensitivity analyses. Lives saved during the Omicron period appeared slightly higher than those saved pre- Omicron. However, Omicron period benefits become the minority under large assumed decreases in PI* x IFR x VE. The Omicron death burden was very low compared to earlier COVID-19 waves. This is unlikely to reflect mostly higher vaccination benefits. Pre-Omicron 7 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. PI* carries large uncertainty since the extent of pre-Omicron viral spread depends on exposure load and measures taken. E.g., there was hardly any pre-Omicron viral circulation in China or New Zealand. Therefore, in these countries hardly any lives were directly saved by COVID-19 vaccination pre-Omicron; all benefit materialized in the Omicron period. We estimate that 9 of 10 deaths averted and 8 of 10 life-years saved were in people 60 years old and over. While COVID-19 devastated long-term facilities (20), the proportion of life-years saved by vaccination in long-term facilities was only 2% of the total, mostly because of the very low LE of their residents. This may nevertheless vary across countries and institutions, depending on resident populations features (e.g. palliative care versus relatively healthy retired elderly). The relative contribution of children, adolescents, and young adults to lives and life- years saved appears minimal. Assessment of absolute net benefits in these populations, if any, require careful consideration of potential additional benefits for non-lethal outcomes (e.g. hospitalizations and other symptomatic disease), as well as any deaths and other consequences from adverse effects (not included in our calculations) (21,22). Cost- effectiveness ratios should be considered carefully in these age strata to document whether vaccination was worthwhile (23). Our estimates include countries with very different pandemic and vaccination experiences. Of note, most non-high income countries (with the notable exception of China) had high proportions of their populations infected before vaccination (13). Given the prominent lack of global vaccine equity (24,25), probably only a minority of lives saved were in non-high income countries, even though they represent 84% of the global population. A modeling study estimated that 1.5 million lives could have been saved with universal vaccination against Omicron in low and low-middle income countries (26). Little of this benefit probably materialized. Inequity, inefficient vaccination campaigns, and vaccine 8 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. hesitancy may have eroded substantially the benefit that might have been derived under ideal circumstances (see Appendix 2: “Deaths averted under ideal circumstances” ). Previous studies estimating lives saved by COVID-19 vaccination have focused on more limited time periods or more restricted areas, countries, or regions. The most-cited study to-date (3) used modeling to estimate 14.4 million COVID-19 deaths and 19.8 million excess deaths averted across 185 countries in the first year of vaccination alone, with very limited uncertainty (13.7-15.9 million and 19.1-20.4 million for 95% credible intervals, respectively). These results vary markedly from our pre-Omicron period estimates. We did not estimate total excess deaths averted globally, as this is fraught with extreme uncertainties (27). However, for COVID-19 deaths, our results suggest over a log10-scale lower deaths averted by COVID-19 vaccination in that early period. Differences may be due to the unreliability of modeling in such complex circumstances (7) and high estimates of IFR (especially in elderly) and VE (using short-term estimates available at that time) assumed in the modeling (3). Another modeling study estimated 620 thousand averted deaths from vaccination in the pre-Omicron period, increasing to 2.1 million based on underreporting assumptions (4). Our pre-Omicron estimates lie between these two estimates. Another study (5) covered 34 countries/territories in Europe and estimated 1.6 million lives saved until March 2023 with 96% of lives saved among those 60 years or older and 60% during the Omicron period. The analyzed countries include approximately half of the global population of high-income countries. While we did not obtain estimates limited to these countries, our global estimates seem modestly more conservative. Differences may be due to implied IFR and VE estimates. However, we agree that the vast majority of lives saved were in the elderly with a slight preponderance of lives saved in the Omicron period. A study covering Latin America and the Caribbean until May 2022 estimated 1.18 million deaths averted (sensitivity 9 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. range, 0.61-2.61 million) with 78% among those 60 years and over and 62% in the Omicron period (28). Several caveats should be discussed. First, the full range of uncertainty is larger than the range that we observe in presented sensitivity analyses. If one were to consider all factors in sensitivity analyses, the range of estimates would spread further. Moreover, our IFR estimates are derived from national seroprevalence studies before vaccination. For unvaccinated individuals, IFR in the second year (the pre-Omicron period that matters for calculation of lives saved) may have been lower with some effective treatments (e.g. dexamethasone) becoming available, better organization of healthcare services, and more experience in managing severe COVID-19. Moreover, there is debate on whether Delta was more or less lethal than the dominant variants of 2020 (29-30). Second, for most factors considered, data informing their values come mostly from high-income countries. The picture is more uncertain in other countries. The two largest countries, China and India, have major uncertainty on estimates of COVID-19 disease burden (31,32), let alone vaccine benefits. Third, VE assumptions try to amalgamate many different vaccines (of variable effectiveness (33,34)), different doses, and different vaccination policies, along with waning effectiveness over time. Unavoidably these assumptions simplify very complex backgrounds. Fourth, life-year calculations are a contentious topic. A previous study that calculated adjusted LE in COVID-19 deaths based on comorbidities found small LE reduction versus the general population (35) but was limited by incomplete information on comorbidities and their severity. Thus, there was probably substantial underestimation of the LE difference between COVID-19 victims and the general population. Another study showed that if LE reduction is modeled through a standardized mortality ratio (SMR) for COVID-19 victims versus the general population, mean LE at COVID-19 death in developed countries decreased 10 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. from ~10-12 years to ~6-8 years with SMR=2 (34), close to what our main analysis anticipates for f=0.5. That same study also estimated only 3.3-4.4 average discounted quality adjusted life-years (36). The SMR approach corresponds to higher f in young ages and larger f in elderly deaths; however, the share of life-years saved accounted by the elderly would be only slightly decreased. In principle, if a disease/condition/event kills anyone regardless of health status, e.g. a nuclear bomb, then f=1; conversely, for a condition that appears exactly when a patient is dying from other co-existing ailments, f approaches infinity. The exact positioning of COVID-19 in that spectrum (37) and the relative share of over- and under-counting of COVID-19 deaths (38) are still debated with substantial consequences for estimated disease burden and vaccination benefits. Regardless, taking LE at age of death directly as a measure of anticipated life-years may lead to grossly misleading inferences (39). E.g. average LE at age of death for all death causes in western countries is ~9-12 years anyhow (39,40) – very close to the average unadjusted LE at age of death for COVID-19 deaths. Interestingly, if many people saved from COVID-19 death by vaccination had indeed limited LE, postponement of death would be short-lived. Such short-lived postponement may explain in part why substantial excess deaths were seen in several high-income countries in 2022-2023 (41) despite achieving high levels of vaccination. Finally, one may put COVID-19 vaccination benefits in perspective along with benefits from other available vaccinations. Comparisons should be cautious, given the different calculation methods used and acknowledging that mathematical models for other vaccinations may also not be fully reliable. However, one study (42) estimated that vaccination for 10 pathogens across 112 countries in 2000-2019 saved 50 million lives; another 47 million may be saved in 2020-2030. Disability-adjusted life years saved were 2700 million and 2300 million, respectively. If these calculations are sound, COVID-19 11 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. vaccination apparently saved during 2020-2024 fewer lives than measles or hepatitis B vaccination in the same period, but more than vaccination for each of the other 8 pathogens. However, life-years saved by COVID-19 vaccination for the same period were more than 30- fold lower than the life-years saved from measles vaccination, 10-fold lower than from hepatitis B vaccination and substantially lower also than the life-years saved from HPV, yellow fever, H. influenzae, S. pneumoniae, and rubella vaccination (42). Therefore, even though COVID-19 vaccines are undeniably a major achievement, their benefits do not necessarily match the benefits of several other widely used vaccines. Decrease in trust and increased hesitancy for these vaccines may be devastating (43,44). The COVID-19 pandemic and pandemic response created a more challenging landscape on how to overcome general vaccine hesitancy (44-46). In conclusion, COVID-19 vaccination offered major benefits during 2020-2024. However, our estimates are more conservative than early modeling efforts to calculate lives saved based on the first year of vaccination alone and strong assumptions on IFR and VE (3). Moreover, from our estimates the vaccination benefits seem to have been largely limited to the elderly portion of the global population. Long-term outcomes in both vaccinated and unvaccinated people should also be examined with further follow-up. 12 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Table 1. Characteristics of strata considered in the calculations Age strata, years World Proportion vaccinated IFR in the pre-Omicron period population before infection in (sensitivity range) pre-Omicron period (sensitivity range) 0-19 2,664,996,463 0.1 (0.083-0.117) 0.000003 (0-0.00002) 20-29 1,209,691,398 0.2 (0.167-0.233) 0.00002 (0-0.00007) 30-39 1,173,183,969 0.46 (0.383-0.537) 0.00011 (0.00005-0.00032) 40-49 975,497,948 0.46 (0.383-0.537) 0.00035 (0.00011-0.00077) 50-59 849,924,808 0.46 (0.383-0.537) 0.00123 (0.00047-0.00220) 60-69 597,651,319 0.46 (0.383-0.567) 0.00506 (0.00208-0.00860) 70-, community 468,997,399 0.46 (0.383-0.567) 0.018 (0.013-0.023) 70-, long-term care 14,505,074 0.46 (0.383-0.567) 0.12 (0.10-0.25) All 7,954,498,378 0.30 (0.25-0.35) IFR: infection fatality rate. Population pyramid data are from (9), long-term care residents are assumed to be 3% of the to years and older stratum (10,11), proportion vaccinated before infection in pre-Omicron period makes assumptions that depend on (12-14), IFR are based on (15-19) and details on the justification of the assumptions and sensitivity analysis ranges can be found in the Supplementary Methods. 13 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Table 2. Lives saved by COVID-19 vaccination according to time period (pre-Omicron, Omicron) and whether vaccination was given to previously uninfected or previously infected people. Age strata Lives saved: Lives saved: Lives saved: Lives saved: Total lives saved pre-Omicron, pre-Omicron, Omicron, Omicron, (% of total saved) previously previously previously previously uninfected infected uninfected infected 0-19 112 17 133 36 299 (0.01) 20-29 677 104 806 220 1,808 (0.07) 30-39 8,311 1,274 9,894 2,704 22,183 (0.9) 40-49 21,988 3,371 26,176 7,155 58,690 (2.3) 50-59 67,324 10,323 80,148 21,907 179,702 (7.1) 60-69 194,753 29,862 231,849 63,372 519,836 (20.5) 70-, community 543,662 83,361 647,216 176,906 1,451,145 (57.3) 70-, long-term care 112,095 17,188 133,447 36,475 299,205 (11.8) All 948,922 145,501 1,129,669 308,776 2,532,869 (100) 14 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Table 3. Results of sensitivity analyses for lives saved by COVID-19 vaccination Previously Lives saved Lives saved Lives saved Lives saved Lives saved infected/ (millions) (millions) (millions) (millions) (millions) Period IFR D(≥1)pre-omicron PI* 10-40% VE 40-85% R 2.5-10 sensitivity 25-35% (pre-Omicron) (pre-), 30- range 70% (Omicron) No/Pre-omicron 0.603-1.455 0.791-1.107 0.474-1.898 0.506-1.075 0.949 Yes/Pre-omicron 0.092-0.223 0.177-0.114 0.073-0.291 0.078-0.165 0.146 No/Omicron 0.717-1.732 0.941-1.318 1.130 0.678-1.581 0.565-2.259 Yes/Omicron 0.196-0.474 0.346-0.271 0.309 0.185-0.432 0.154-0.617 Total 1.608-3.884 2.256-2.810 1.986-3.628 1.447-3.254 1.814-3.971 For the IFR sensitivity range, see Supplementary Methods. In the main analysis, in the pre- Omicron period, 30% and 23% of the global population are assumed to have been vaccinated (received at least one dose) before any infection and after previous infection, respectively; and for the Omicron period, 30% and 41% of the global population are assumed to have received at least one dose before any infection and after being infected, respectively. D(≥1)pre-omicron: proportion of global population that received at least one dose in the pre- Omicron period before any infection; PI*: proportion of pre-Omicron period vaccinated people who would have been infected in the absence of vaccination; VE: vaccine effectiveness for death; R: ratio for Omicron versus pre-Omicron product of PI* x IFR x VE (see Methods). 15 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Table 4. Estimates of life-years saved from COVID-19 vaccination Age strata Total Lives Life Life-years saved, Life-years Life-years saved expectancy, f=0.5 (%) [main saved, f=0.25 saved, f=0.8 years analysis] 0-19 299 64 9,560 (0.1) 4,780 15,297 20-29 1,808 49.9 45,114 (0.3) 22,557 72,183 30-39 22,183 40.7 451,430 (3.0) 225,715 722,288 40-49 58,690 31.8 933,167 (6.3) 466,584 1,493,068 50-59 179,702 23.5 2,111,502 (14.2) 1,055,751 3,378,403 60-69 519,836 16.2 4,210,672 (28.6) 2,105,336 6,737,076 70-, community 1,451,146 9.2 6,675,269 (45.3) 3,337,635 10,680,431 70-, long-term care 299,205 2 299,205 (2.0) 149,603 478,728 All 2,532,869 14,735,921 (100) 7,367,961 23,577,474 f: ratio of life expectancy in people dying from COVID-19 versus the total population in the same stratum. 16 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. REFERENCES 1. Fauci AS. The story behind COVID-19 vaccines. Science. 2021 Apr 9;372(6538):109. 2. Karikó K. Developing mRNA for Therapy. Keio J Med. 2022;71(1):31. 3. Watson OJ, Barnsley G, Toor J, Hogan AB, Winskill P, Ghani AC. Global impact of the first year of COVID-19 vaccination: a mathematical modelling study. Lancet Infect Dis. 2022 Sep;22(9):1293-1302. 4. Yang J, Vaghela S, Yarnoff B, et al. (2022). Estimated global public health and economic impact of COVID-19 vaccines in the pre-omicron era using real-world empirical data. Expert Rev Vaccines 2023; 22: 54–65. 5. Meslé MMI, Brown J, Mook P, Katz MA, Hagan J, Pastore R, Benka B, Redlberger- Fritz M, Bossuyt N, Stouten V, Vernemmen C, Constantinou E, Maly M, Kynčl J, Sanca O, Krause TG, Vestergaard LS, Leino T, Poukka E, Gkolfinopoulou K, Mellou K, Tsintziloni M, Molnár Z, Aspelund G, Thordardottir M, Domegan L, Kelly E, O'Donell J, Urdiales AM, Riccardo F, Sacco C, Bumšteinas V, Liausediene R, Mossong J, Vergison A, Borg ML, Melillo T, Kocinski D, Pollozhani E, Meijerink H, Costa D, Gomes JP, Leite PP, Druc A, Gutu V, Mita V, Lazar M, Popescu R, Popovici O, Musilová M, Mrzel M, Socan M, Učakar V, Limia A, Mazagatos C, Olmedo C, Dabrera G, Kall M, Sinnathamby M, McGowan G, McMenamin J, Morrison K, Nitzan D, Widdowson MA, Smallwood C, Pebody R; WHO European Respiratory Surveillance Network. Estimated number of lives directly saved by COVID-19 vaccination programmes in the WHO European Region from December, 2020, to March, 2023: a retrospective surveillance study. Lancet Respir Med. 2024 Sep;12(9):714-727. 6. Steele MK, Couture A, Reed C, Iuliano D, Whitaker M, Fast H, Hall AJ, MacNeil A, Cadwell B, Marks KJ, Silk BJ. Estimated Number of COVID-19 Infections, 17 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Hospitalizations, and Deaths Prevented Among Vaccinated Persons in the US, December 2020 to September 2021. JAMA Netw Open. 2022 Jul 1;5(7):e2220385. 7. Ioannidis JPA, Cripps S, Tanner MA. Forecasting for COVID-19 has failed. Int J Forecast. 2022 Apr-Jun;38(2):423-438. 8. Jewell NP, Lewnard JA, Jewell BL. Predictive Mathematical Models of the COVID- 19 Pandemic: Underlying Principles and Value of Projections. JAMA. 2020 May 19;323(19):1893-1894. 9. https://www.populationpyramid.net/world/2021, last accessed October 2, 2024. 10. https://www.who.int/data/gho/indicator-metadata-registry/imr-details/percentage-of- older-people-receiving-long-term-care-at-a-residential-care-facility-and-at-home, last accessed October 2, 2024. 11. https://www.ecdc.europa.eu/en/publications-data/surveillance-COVID-19-long-term- care-facilities-EU-EEA, last accessed October 2, 2024. 12. https://www.kff.org/interactive/kff-global-covid-19-vaccine-coverage-tool-current- and-projected-coverage/, last accessed October 2, 2024. 13. COVID-19 Cumulative Infection Collaborators. Estimating global, regional, and national daily and cumulative infections with SARS-CoV-2 through Nov 14, 2021: a statistical analysis. Lancet. 2022 Jun 25;399(10344):2351-2380. 14. https://ourworldindata.org/grapher/covid-vaccine-by-age, last accessed October 2, 2024. 15. Pezzullo AM, Axfors C, Contopoulos-Ioannidis DG, Apostolatos A, Ioannidis JPA. Age-stratified infection fatality rate of COVID-19 in the non-elderly population. Environ Res. 2023 Jan 1;216(Pt 3):114655. 16. Axfors C, Ioannidis JPA. Infection fatality rate of COVID-19 in community-dwelling elderly populations. Eur J Epidemiol. 2022 Mar;37(3):235-249.22.17% 18 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 17. Zhang J, Yu Y, Petrovic M, Pei X, Tian Q-B, Zhang L, Zhang W-H. Impact of the COVID-19 pandemic and corresponding control measures on long-term care facilities: a systematic review and meta-analysis. Age Ageing 2023 Jan 8;52(1):afac308. 18. Borras-Bermejo B, Martínez-Gómez X, San Miguel MG, Esperalba J, Antón A, Martin E, Selvi M, Abadías MJ, Román A, Pumarola T, Campins M, Almirante B. Asymptomatic SARS-CoV-2 Infection in Nursing Homes, Barcelona, Spain, April 2020. Emerg Infect Dis. 2020 Sep;26(9):2281–3. 19. Patel MC, Chaisson LH, Borgetti S, Burdsall D, Chugh RK, Hoff CR, Murphy EB, Murskyj EA, Wilson S, Ramos J, Akker L, Bryars D, Thomas-Smith E, Bleasdale SC, Ezike NO. Asymptomatic SARS-CoV-2 Infection and COVID-19 Mortality During an Outbreak Investigation in a Skilled Nursing Facility. Clin Infect Dis. 2020 Dec 31;71(11):2920-2926. 20. https://ltccovid.org/country-reports-on-covid-19-and-long-term-care/, last accessed October 2, 2024. 21. Fraiman J, Erviti J, Jones M, Greenland S, Whelan P, Kaplan RM, Doshi P. Serious adverse events of special interest following mRNA COVID-19 vaccination in randomized trials in adults. Vaccine. 2022 Sep 22;40(40):5798-5805. 22. Bardosh K, Krug A, Jamrozik E, Lemmens T, Keshavjee S, Prasad V, Makary MA, Baral S, Høeg TB. COVID-19 vaccine boosters for young adults: a risk benefit assessment and ethical analysis of mandate policies at universities. J Med Ethics. 2024 Jan 23;50(2):126-138. 23. Pilz S, Ioannidis JPA. Does natural and hybrid immunity obviate the need for frequent vaccine boosters against SARS-CoV-2 in the endemic phase? Eur J Clin Invest. 2023 Feb;53(2):e13906. 19 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 24. Carlson C, Becker D, Happi C, O'Donoghue Z, de Oliveira T, Oyola SO, Poisot T, Seifert S, Phelan A. Save lives in the next pandemic: ensure vaccine equity now. Nature. 2024 Feb;626(8001):952-953. 25. Nunes C, McKee M, Howard N. The role of global health partnerships in vaccine equity: A scoping review. PLOS Glob Public Health. 2024 Feb 22;4(2):e0002834. 26. Savinkina A, Bilinski A, Fitzpatrick M, Paltiel AD, Rizvi Z, Salomon J, Thornhill T, Gonsalves G. Estimating deaths averted and cost per life saved by scaling up mRNA COVID-19 vaccination in low-income and lower-middle-income countries in the COVID-19 Omicron variant era: a modelling study. BMJ Open. 2022 Sep 13;12(9):e061752. 27. Ioannidis JPA, Zonta F, Levitt M. Flaws and uncertainties in pandemic global excess death calculations. Eur J Clin Invest. 2023 Aug;53(8):e14008. 28. Savinkina A, Weinberger DM, Toscano CM, De Oliveira LH. Estimated Deaths Averted in Adults by COVID-19 Vaccination in Select Latin American and Caribbean Countries. Open Forum Infect Dis. 2024 Sep 10;11(10):ofae528. 29. Lin L, Liu Y, Tang X, He D. The Disease Severity and Clinical Outcomes of the SARS-CoV-2 Variants of Concern. Front Public Health. 2021 Nov 30;9:775224. 30. Ong SWX, Chiew CJ, Ang LW, Mak TM, Cui L, Toh MPHS, Lim YD, Lee PH, Lee TH, Chia PY, Maurer-Stroh S, Lin RTP, Leo YS, Lee VJ, Lye DC, Young BE. Clinical and Virological Features of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Variants of Concern: A Retrospective Cohort Study Comparing B.1.1.7 (Alpha), B.1.351 (Beta), and B.1.617.2 (Delta). Clin Infect Dis. 2022 Aug 24;75(1):e1128-e1136. 31. Ioannidis JPA, Zonta F, Levitt M. Estimates of COVID-19 deaths in Mainland China after abandoning zero COVID policy. Eur J Clin Invest. 2023 Apr;53(4):e13956. 20 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 32. Jha P, Deshmukh Y, Tumbe C, Suraweera W, Bhowmick A, Sharma S, Novosad P, Fu SH, Newcombe L, Gelband H, Brown P. COVID mortality in India: National survey data and health facility deaths. Science. 2022 Feb 11;375(6581):667-671. 33. Lin DY, Gu Y, Wheeler B, Young H, Holloway S, Sunny SK, Moore Z, Zeng D. Effectiveness of Covid-19 Vaccines over a 9-Month Period in North Carolina. N Engl J Med. 2022 Mar 10;386(10):933-41. 34. Korang SK, von Rohden E, Veroniki AA, Ong G, Ngalamika O, Siddiqui F, Juul S, Nielsen EE, Feinberg JB, Petersen JJ, Legart C, Kokogho A, Maagaard M, Klingenberg S, Thabane L, Bardach A, Ciapponi A, Thomsen AR, Jakobsen JC, Gluud C. Vaccines to prevent COVID-19: A living systematic review with Trial Sequential Analysis and network meta-analysis of randomized clinical trials. PLoS One. 2022 Jan 21;17(1):e0260733. 35. Hanlon P, Chadwick F, Shah A, Wood R, Minton J, McCartney G, Fischbacher C, Mair FS, Husmeier D, Matthiopoulos J, McAllister DA. COVID-19 - exploring the implications of long-term condition type and extent of multimorbidity on years of life lost: a modelling study. Wellcome Open Res. 2021 Mar 1;5:75. 36. Briggs AH, Goldstein DA, Kirwin E, Meacock R, Pandya A, Vanness DJ, Wisløff T. Estimating (quality-adjusted) life-year losses associated with deaths: With application to COVID-19. Health Econ. 2021 Mar;30(3):699-707. 37. Rubo M, Czuppon P. How should we speak about years of life lost (YLL) values? Eur J Epidemiol. 2023 Mar;38(3):345-347. 38. Ioannidis JPA. Over- and under-estimation of COVID-19 deaths. Eur J Epidemiol. 2021 Jun;36(6):581-588. 39. Marshall RJ. Standard expected years of life lost as a measure of mortality: norms and reference to New Zealand data. Aust N Z J Public Health. 2004 Oct;28(5):452–7. 21 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 40. Wengler A, Rommel A, Plaß D, Gruhl H, Leddin J, Ziese T, Lippe EV; BURDEN 2020 Study Group. Years of Life Lost to Death—A Comprehensive Analysis of Mortality in Germany Conducted as Part of the BURDEN 2020 Project. Dtsch Arztebl Int. 2021 Mar 5;118(9):137-144. 41. Ioannidis JPA, Zonta F, Levitt M. Variability in excess deaths across countries with different vulnerability during 2020-2023. Proc Natl Acad Sci U S A. 2023 Dec 5;120(49):e2309557120. 42. Toor J, Echeverria-Londono S, Li X, Abbas K, Carter ED, Clapham HE, Clark A, de Villiers MJ, Eilertson K, Ferrari M, Gamkrelidze I, Hallett TB, Hinsley WR, Hogan D, Huber JH, Jackson ML, Jean K, Jit M, Karachaliou A, Klepac P, Kraay A, Lessler J, Li X, Lopman BA, Mengistu T, Metcalf CJE, Moore SM, Nayagam S, Papadopoulos T, Perkins TA, Portnoy A, Razavi H, Razavi-Shearer D, Resch S, Sanderson C, Sweet S, Tam Y, Tanvir H, Tran Minh Q, Trotter CL, Truelove SA, Vynnycky E, Walker N, Winter A, Woodruff K, Ferguson NM, Gaythorpe KA. Lives saved with vaccination for 10 pathogens across 112 countries in a pre-COVID-19 world. Elife. 2021 Jul 13;10:e67635. 43. Pandey A, Galvani AP. Exacerbation of measles mortality by vaccine hesitancy worldwide. Lancet Glob Health. 2023 Apr;11(4):e478-e479. 44. Durrheim DN, Andrus JK, Tabassum S, Bashour H, Githanga D, Pfaff G. A dangerous measles future looms beyond the COVID-19 pandemic. Nat Med. 2021 Mar;27(3):360-361. 45. Pertwee E, Simas C, Larson HJ. An epidemic of uncertainty: rumors, conspiracy theories and vaccine hesitancy. Nat Med. 2022 Mar;28(3):456-459. doi: 10.1038/s41591-022-01728-z. Epub 2022 Mar 10. PMID: 35273403. 22 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 46. Kaufman J, Rak A, Vasiliadis S, Brar N, Atif E, White J, Danchin M, Durrheim DN. The Case for Assessing the Drivers of Measles Vaccine Uptake. Vaccines (Basel). 2024 Jun 20;12(6):692. 23 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. SUPPLEMENTARY MATERIAL APPENDIX 1: SUPPLEMENTARY METHODS General principles The analysis compares the outcomes of global COVID-19 vaccination strategies to a counterfactual scenario of no vaccination. In calculating our estimates, we do not consider deaths and other consequences from adverse effects of SARS-CoV-2 vaccines, nor do we make any adjustment for the quality of life-years saved. Moreover, we do not attempt to calculate indirect effects of COVID-19 vaccination which may have modulated excess deaths through an impact on non-COVID-19 causes of death. Values used and sensitivity analyses Population: The world population pyramid in 2021 was used in the calculations (1) considering age strata 0-19, 20-29, 30-39, 40-49, 50-59, 60-69, and 70 years and higher. The last age stratum was further divided into those living in the community and those living in long-term care facilities, assuming that the latter represent a 3% fraction of that age stratum; data on the size of this fraction are available mostly from high-income countries and there is sparse use of long-term care facilities in other countries (2,3). Residents of long-term care facilities are 0.7% of the total population in European countries, and 0.4% in the USA (2,3), while our assumption translates to ~0.18% of the global population. Proportion vaccinated before/after infection and before/after Omicron: By mid- November 2021, 53% of the global population had received at least one dose and 41% had been fully vaccinated with substantial disparity per income group (4). A systematic review has estimated (5) that 44% of the global population had been infected by mid-November 2021. If vaccination and prior infection were independent, then D(≥1)pre-omicron=53%x(1- 0.44)=30% had received at least one dose before any infection (and similarly D(≥2)pre- omicron=0.41%x0.56=23% had been fully vaccinated before any infection) before the advent of 24 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Omicron. Vaccination and prior infection at a global level may be nearly independent events because the vast majority of infections were undetected, thus they would not have affected vaccination decisions. Moreover, most countries introduced campaigns that advised vaccination regardless of prior infection status. We also assumed that D(≥1)pre-omicron differed across age strata. Data are available mostly from some high income countries (6), showing substantially or far lower vaccination rates in children and young people in many countries. At global level, we assumed this age gradient may be more pronounced: 10% in those 0-19 years old, 20% in those 20-29 years old, and otherwise similar (46%) in higher age strata (overall 30%). In sensitivity analysis, we considered D(≥1)pre-omicron ranging from 25% to 35%, retaining the same ratios between age strata as in the main analysis. We assumed that after Omicron emerged, the remaining 56% of the global population who had not been infected by November 2021 were infected at least once until October 2024. People remaining uninfected are probably less than 5% in high-income countries with the most extensive vaccination programs (7) and probably overall very rare globally. Calculations assuming 2% of the global population not infected until October 2024 did not materially change results (not shown). Between December 2021 and October 2024, an additional 71%-53%=18% of the global population received at least one dose and an additional 65%-41%=24% reached full vaccination status (6). Given the massive, rapid onslaught of the Omicron waves and the relatively slowing of vaccination of previously unvaccinated individuals after November 2021, we considered that only an additional D(≥1)omicron=5% (slightly less than a third of the 18%) were first vaccinated with at least one dose before being infected and assumed no distortion of the relative vaccination rates across age strata. 25 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. Proportion of people among those who are vaccinated who would have been infected in the absence of vaccination: PI* is a very uncertain parameter, because it largely depends on the strictness of other policies and measures taken at public and individual levels (e.g. lockdowns, masks, other restrictive measures or personal restrictions of exposure). We may assume that almost all people would have been infected during the Omicron period anyhow, since aggressive restrictive measures would have been impossible to hold for long. However, in separating the pre-Omicron period up to November 2021, several countries did achieve to have only minimal or even negligible circulation of the virus in that period. For the main analysis, we assumed for all age strata that PI*=20% of those vaccinated would have been infected until November 2021 in the absence of vaccination, during a mean follow-up of half a year of less. Sensitivity analyses considered values ranging from 10% to 40%. We also assumed that a small percentage of the population (5%) were infected for the first time in the pre-Omicron period after having received at least 1 vaccine dose, based on high vaccine effectiveness against infection in the first month but modest decrease by 6 months (8). IFR: For IFR in unvaccinated people before the advent of Omicron, we used the median estimates obtained from a systematic review for non-elderly age strata and we used the interquartile range [IR] values in each age stratum for the sensitivity range (9). For the two youngest age strata the lower IQR value is 0% (due to no deaths observed in modest size population samples), but using instead the median or half the median for the lower sensitivity range yielded very similar total lives saved (not shown). For the 70 years old and over community-dwelling population we considered IFR=1.8% based on a previous meta- regression analysis that allows calculating expected IFR as a function of the proportion of people over 85 years old among those over 70 years (in the global population this proportion is 13.4%), and taking the median of estimates from two meta-regressions based on different seroprevalence study eligibility criteria (10). The estimates from the two meta-regressions 26 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. were considered in the sensitivity analyses (range, 1.3 to 2.3). We considered IFR=12% for residents of long-term care facilities based on another review and meta-analysis (11) showing case-fatality rate of 22.71% from a summary of 51 studies covering 2020 and assuming approximately half of infections globally were missed because of being asymptomatic (given asymptomatic rates of 39-70% among infected residents in different surveillance studies) (12,13). In sensitivity analyses, we considered a range of 10% to 25%. These estimates are derived from pre-vaccination data, but it is reasonable to assume a similar IFR for unvaccinated people during 2021. Conversely, for Omicron, we assumed that IFR was reduced to a third of the previous values in each stratum, as suggested by refs. (14,15). Vaccine effectiveness for death: We assumed VE=75% during the pre-Omicron period and 50% during the Omicron period. This is an aggregate estimate considering the large heterogeneity of vaccination experiences (different vaccines, some of which had probably lesser effectiveness than others (16,17)), waning effectiveness especially with long- term follow-up (18), and also different vaccination experiences including many people who received only one or two doses in the pre-Omicron period. and one, two, three, or more doses in the Omicron period with overall lower effectiveness in the Omicron period). Sensitivity analyses considered values between 40% and 85% in the pre-Omicron period and 30% to 70% in the Omicron period. R: For people vaccinated after at least one infection, we assumed R=5 (sensitivity range, 2.5 to 10), given the very low re-infection risk and lower IFR for them among previously infected in the pre-Omicron period (19), and the more common re-infections but with very low IFR in the Omicron period (20,21). Life expectancy: For life expectancy, the UN population division life table for 2021 for World, both sexes (22) was used and for each age stratum, the mid-point in the age 27 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. bracket was considered. For the 70 years and above, the life expectancy at age 77 was considered for the community-dwelling individuals and 2 years was used for long-term care facility residents based on epidemiological studies in such settings, e.g. (23,24). For the factor f values, the main analysis considered f=0.5 for all strata, assuming that those who succumb to COVID-19, in the absence of infection would have half the life expectancy of the general population given the high burden of comorbidities and higher background socioeconomic burden in people dying from COVID-19 (25,26). In sensitivity analyses, values between 0.25 and 0.8 were considered. Calculations were conducted in Excel. All key data used are in the manuscript and supplement and clarifications can be requested by the authors. APPENDIX 2: DEATHS AVERTED UNDER IDEAL CIRCUMSTANCES Calculating formally the number of deaths that could have been averted by COVID- 19 vaccination under ideal circumstances is highly speculative. The results may vary substantially under different assumptions about the feasible speed of vaccination campaigns and the eventual vaccination coverage achieved – perfect coverage is not a realistic expectation. However, the large majority of the benefit was derived when vaccinated before any SARS-CoV-2 infection and we assume in the main analysis that 35% of the global population overall and 51% of the elderly (that account for the vast majority of the benefit) were vaccinated before any infection (46% in the pre-Omicron phase and another 5% in the Omicron phase). Therefore, it is likely that close to half of the maximum achievable vaccination benefit in terms of lives saved materialized. The proportion is likely higher in high-income countries and lower in poor countries that suffered from the consequences of vaccine inequity. As an illustrative calculation, one may assume the scenario to have the entire global population vaccinated already during the pre-Omicron period. One may assume that even 28 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. under the most optimistic conditions, only 75% of the population would be vaccinated before any SARS-CoV-2 infection (as many had already been infected when vaccines became available and some more would be infected while deployment occurred, even if this were very fast). Then in the pre-Omicron period 75% are assumed to be vaccinated before infection and 25% are assumed to be vaccinated after infection; and in the Omicron period 70% have been vaccinated before infection and 30% have been vaccinated after infection. If vaccine choice and dosing were also optimized, we can then also assume the upper limits of the range of VE. Under these idealized circumstances, we estimate that 4,623,154 deaths would have been averted (1,881,231 in pre-Omicron/not infected, 125,415 in pre- Omicron/previously infected, 2,409,942 in Omicron/not infected, and 206,566 in Omicron/previously infected). The actual averted deaths (2,532,869) are therefore estimated to be 55% of the total that could have been averted under these idealized circumstances. References for supplementary methods 1. https://www.populationpyramid.net/world/2021, last accessed October 2, 2024. 2. https://www.who.int/data/gho/indicator-metadata-registry/imr-details/percentage-of- older-people-receiving-long-term-care-at-a-residential-care-facility-and-at-home, last accessed October 2, 2024. 3. https://www.ecdc.europa.eu/en/publications-data/surveillance-COVID-19-long-term- care-facilities-EU-EEA, last accessed October 2, 2024. 4. https://www.kff.org/interactive/kff-global-covid-19-vaccine-coverage-tool-current- and-projected-coverage/, last accessed October 2, 2024. 5. COVID-19 Cumulative Infection Collaborators. Estimating global, regional, and national daily and cumulative infections with SARS-CoV-2 through Nov 14, 2021: a statistical analysis. Lancet. 2022 Jun 25;399(10344):2351-2380. 29 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 6. https://ourworldindata.org/grapher/covid-vaccine-by-age, last accessed October 2, 2024. 7. Riedmann U, Chalupka A, Richter L, Sprenger M, Rauch W, Schenk H, Krause R, Willeit P, Oberacher H, Høeg TB, Ioannidis JPA, Pilz S. Estimates of SARS-CoV-2 Infections and Population Immunity after the COVID-19 Pandemic in Austria: Analysis of National Wastewater Data. Available at SSRN: https://ssrn.com/abstract=4995470 or http://dx.doi.org/10.2139/ssrn.4995470 8. Ssentongo P, Ssentongo AE, Voleti N, Groff D, Sun A, Ba DM, Nunez J, Parent LJ, Chinchilli VM, Paules CI. SARS-CoV-2 vaccine effectiveness against infection, symptomatic and severe COVID-19: a systematic review and meta-analysis. BMC Infect Dis. 2022 May 7;22(1):439. 9. Pezzullo AM, Axfors C, Contopoulos-Ioannidis DG, Apostolatos A, Ioannidis JPA. Age-stratified infection fatality rate of COVID-19 in the non-elderly population. Environ Res. 2023 Jan 1;216(Pt 3):114655. 10. Axfors C, Ioannidis JPA. Infection fatality rate of COVID-19 in community-dwelling elderly populations. Eur J Epidemiol. 2022 Mar;37(3):235-249.22.17% 11. Zhang J, Yu Y, Petrovic M, Pei X, Tian Q-B, Zhang L, Zhang W-H. Impact of the COVID-19 pandemic and corresponding control measures on long-term care facilities: a systematic review and meta-analysis. Age Ageing 2023 Jan 8;52(1):afac308. 12. Borras-Bermejo B, Martínez-Gómez X, San Miguel MG, Esperalba J, Antón A, Martin E, Selvi M, Abadías MJ, Román A, Pumarola T, Campins M, Almirante B. Asymptomatic SARS-CoV-2 Infection in Nursing Homes, Barcelona, Spain, April 2020. Emerg Infect Dis. 2020 Sep;26(9):2281–3. 13. Patel MC, Chaisson LH, Borgetti S, Burdsall D, Chugh RK, Hoff CR, Murphy EB, Murskyj EA, Wilson S, Ramos J, Akker L, Bryars D, Thomas-Smith E, Bleasdale SC, Ezike 30 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. NO. Asymptomatic SARS-CoV-2 Infection and COVID-19 Mortality During an Outbreak Investigation in a Skilled Nursing Facility. Clin Infect Dis. 2020 Dec 31;71(11):2920-2926. 14. Ward IL, Bermingham C, Ayoubkhani D, Gethings OJ, Pouwels KB, Yates T, Khunti K, Hippisley-Cox J, Banerjee A, Walker AS, Nafilyan V. Risk of covid-19 related deaths for SARS-CoV-2 omicron (B.1.1.529) compared with delta (B.1.617.2): retrospective cohort study. BMJ. 2022 Aug 2;378:e070695. 15. Nyberg T, Ferguson NM, Nash SG, et al., COVID-19 Genomics UK (COG-UK) consortium. Comparative analysis of the risks of hospitalisation and death associated with SARS-CoV-2 omicron (B.1.1.529) and delta (B.1.617.2) variants in England: a cohort study. Lancet. 2022;399:1303-12. 16. Lin DY, Gu Y, Wheeler B, Young H, Holloway S, Sunny SK, Moore Z, Zeng D. Effectiveness of Covid-19 Vaccines over a 9-Month Period in North Carolina. N Engl J Med. 2022 Mar 10;386(10):933-41. 17. Korang SK, von Rohden E, Veroniki AA, Ong G, Ngalamika O, Siddiqui F, Juul S, Nielsen EE, Feinberg JB, Petersen JJ, Legart C, Kokogho A, Maagaard M, Klingenberg S, Thabane L, Bardach A, Ciapponi A, Thomsen AR, Jakobsen JC, Gluud C. Vaccines to prevent COVID-19: A living systematic review with Trial Sequential Analysis and network meta-analysis of randomized clinical trials. PLoS One. 2022 Jan 21;17(1):e0260733. 18. Chalupka A, Riedmann U, Richter L, Chakeri A, El-Khatib Z, Sprenger M, Theiler- Schwetz V, Trummer C, Willeit P, Schennach H, Benka B, Werber D, Høeg TB, Ioannidis JPA, Pilz S. Effectiveness of the First and Second Severe Acute Respiratory Syndrome Coronavirus 2 Vaccine Dose: A Nationwide Cohort Study From Austria on Hybrid Versus Natural Immunity. Open Forum Infect Dis. 2024 Sep 19;11(10):ofae547. 19. COVID-19 Forecasting Team. Past SARS-CoV-2 infection protection against re- infection: a systematic review and meta-analysis. Lancet. 2023 Mar 11;401(10379):833-842. 31 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 20. Medić S, Anastassopoulou C, Lozanov-Crvenković Z, Vuković V, Dragnić N, Petrović V, Ristić M, Pustahija T, Gojković Z, Tsakris A, Ioannidis JPA. Risk and severity of SARS-CoV-2 reinfections during 2020-2022 in Vojvodina, Serbia: A population-level observational study. Lancet Reg Health Eur. 2022 Sep;20:100453. 21. Erikstrup C, Laksafoss AD, Gladov J, Kaspersen KA, Mikkelsen S, Hindhede L, Boldsen JK, Jørgensen SW, Ethelberg S, Holm DK, Bruun MT, Nissen J, Schwinn M, Brodersen T, Mikkelsen C, Sækmose SG, Sørensen E, Harritshøj LH, Aagaard B, Dinh KM, Busch MP, Jørgensen CS, Krause TG, Ullum H, Ostrowski SR, Espenhain L, Pedersen OBV. Seroprevalence and infection fatality rate of the SARS-CoV-2. Omicron variant in Denmark: A nationwide serosurveillance study. Lancet Reg Health Eur. 2022 Oct;21:100479. 22. https://population.un.org/wpp/Download/Standard/Mortality/, last accessed October 2, 2024. 23. Collingridge Moore D, Payne S, Keegan T, Van den Block L, Deliens L, Gambassi G, Heikkila R, Kijowska V, Pasman HR, Pivodic L, Froggatt K. Length of stay in long-term care facilities: a comparison of residents in six European countries. Results of the PACE cross- sectional study. BMJ Open. 2020 Mar 8;10(3):e033881. 24. Kelly A, Conell-Price J, Covinsky K, Cenzer IS, Chang A, Boscardin WJ, Smith AK. Length of stay for older adults residing in nursing homes at the end of life. J Am Geriatr Soc. 2010 Sep;58(9):1701-6. 25. Singh AK, Gillies CL, Singh R, Singh A, Chudasama Y, Coles B, Seidu S, Zaccardi F, Davies MJ, Khunti K. Prevalence of co-morbidities and their association with mortality in patients with COVID-19: A systematic review and meta-analysis. Diabetes Obes Metab. 2020 Oct;22(10):1915-1924. 32 medRxiv preprint doi: https://doi.org/10.1101/2024.11.03.24316673; this version posted November 4, 2024. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. It is made available under a CC-BY-NC-ND 4.0 International license. 26. Khanijahani A, Iezadi S, Gholipour K, Azami-Aghdash S, Naghibi D. A systematic review of racial/ethnic and socioeconomic disparities in COVID-19. Int J Equity Health. 2021 Nov 24;20(1):248. 33