
==== Front
Vaccine X
Vaccine X
Vaccine: X
2590-1362
Elsevier

S2590-1362(24)00120-7
10.1016/j.jvacx.2024.100547
100547
Regular paper
Science-based exit from stringent countermeasures against COVID-19: Mortality prediction using immune landscape between 2021 and 2022 in Japan
Kayano Taishi ab
Sasanami Misaki a
Nishiura Hiroshi nishiura.hiroshi.5r@kyoto-u.ac.jp
ab⁎
a Kyoto University School of Public Health, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
b Center for Health Security, Kyoto University Graduate School of Medicine, Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan
⁎ Corresponding author at: School of Public Health, Kyoto University Yoshida-Konoe-cho, Sakyo-ku, Kyoto 606-8501, Japan. nishiura.hiroshi.5r@kyoto-u.ac.jp
12 8 2024
10 2024
12 8 2024
20 1005476 12 2023
9 8 2024
11 8 2024
© 2024 The Author(s)
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Highlights

• Immune fraction protected against either COVID-19-related illness or death in individuals aged ≥ 60 with a first booster dose was estimated to be less than 75 % at the end of 2022.

• Without a second booster vaccination program (fourth dose), approximately 155,000 potential deaths would be expected among people aged ≥ 60 years if all individuals were infected at the very end of 2022.

• A second booster could play a crucial role in reducing mortality by 60%, or by 33% with coverage identical or half that of the third dose.

• Understanding the mortality impact quantitively could provide insights to assist in decision-making processes.

Background

Stringent public health and social measures against COVID-19 infection were implemented to avoid an overwhelming hospital caseload and excessive number of deaths, especially among elderly people. We analyzed population-level immunity and predicted mortality, calculated as the potential number of deaths on a given calendar date in Japan, to develop a science-based exit strategy from stringent control measures.

Methods

Immune proportions were inferred by age group using vaccination coverage data and the estimated number of naturally infected individuals. Immunity against symptomatic illness and death were estimated separately, allowing for inference of the immune fraction that was protected against either COVID-19-related symptomatic infection or death. By multiplying the infection fatality risk by age group for the immune fraction, the potential number of deaths was obtained.

Results

Accounting for a second and third dose of messenger RNA vaccine in the present-day population, approximately 155,000 potential deaths would be expected among people aged ≥ 60 years if all individuals were infected at the very end of 2022. A fourth dose (i.e., second booster) with a coverage identical to that of the third dose could reduce mortality by 60%. In all examined settings, the largest number of deaths occurred among people aged 80 years and older.

Conclusions

Our estimates can help policymakers understand the mortality impact of the COVID-19 epidemic in a quantitative manner and the critical importance of timely immunization so as to assist in decision making.

Keywords

Remaining burden
Potential deaths
COVID-19
Statistical model
Vaccination
==== Body
pmcIntroduction

More than 2 and a half years have passed since the start of the pandemic caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1], [2]. At this point, it is vital to understand the likelihood that a forthcoming exit strategy from pandemic countermeasures would be successful. Coronavirus disease 2019 (COVID-19) involves substantial pre-symptomatic transmission and a large number of mild and asymptomatic infections, thus precluding containment via case isolation and contact tracing alone [3], [4], [5]. Therefore, before two rounds of vaccination had been completed, many countries had implemented public health and social measures (PHSM). The types of PHSM ranged from widespread mask wearing to strict lockdown measures to reduce COVID-19 transmission. The development of messenger RNA (mRNA) vaccines against SARS-CoV-2 substantially changed PHSM policies worldwide. In the presence of vaccine-induced immunity in the population, lockdown measures have become increasingly less common. Moreover, many nations with high immunization coverage began to cease strict border control policies.

Monitoring the landscape among populations with immunity against SARS-CoV-2 has been challenging because vaccine-induced protection has been shown to wane as a function of the time since vaccination. A government report from the United Kingdom on the Delta variant (B.1.617) showed that vaccine effectiveness against symptomatic illness was 90% at 2–4 weeks after receiving a second dose, but the effectiveness had declined to less than 50% by 20 weeks [6]. The emergence of the Omicron variant (B.1.1.529) brought about further complications because the effectiveness of available vaccines was not as high as that against the Delta variant. Booster vaccination programs (i.e., third and fourth doses) have had an important role in maintaining population immunity and avoiding severe COVID-19 infections, especially among older people [7], [8], [9]. In the presence of population immunity, many industrialized nations decided to gradually lift PHSM. Since then, the sequential emergence of Omicron subvariants has been ongoing. The newly dominant subvariants have tended to possess immune-evasion properties, leading to lowered immunity levels in the population. Thus, whereas a snapshot evaluation of the immune landscape was technically feasible prior to emergence of the Omicron variant [10], waning immunity and the emergence of subvariants with immune-escape capabilities have substantially complicated computation of the immune landscape.

Japan has maintained a cumulative incidence of COVID-19 that has been relatively lower than those of other industrialized nations (e.g., less than 5% of the population had confirmed COVID-19 infection by the end of 2021). The nationwide vaccination program in Japan started on February 17, 2021. This effort was initially focused on immunization of health care workers, followed by people aged 65 years and older, and achieved a coverage rate exceeding 80% [11]. The first booster vaccination program for people who completed the primary series (i.e., two doses) and the second booster program (i.e., fourth dose) began on December 1, 2021 and May 25, 2022, respectively [12], with the aim to boost immune protection and prevent mortality among elderly people and those with underlying comorbidities. According to the Prime Minister’s Office of Japan, as of July 7, 2022, the vaccination coverages of booster and second booster doses were 62.1% and 3.8% (23.5% among eligible people aged 60 years and older), respectively (Fig. 1) [13], [14]. With respect to vaccine type, BNT162b2 (Pfizer/BioNTech) accounts for more than 80% of administered vaccines, followed by the mRNA-1273 (Moderna), ChAdOx1-S (Oxford/AstraZeneca), and NVX-CoV2373 (Novavax/Takeda) vaccines [13]. Although the vaccination coverage is high, Japan has a substantial number of older residents who are more vulnerable to COVID-19 infection. Future epidemic waves in this super-aged country could yield substantial mortality if reopening were not objectively planned.Fig. 1 Vaccine rollout from February 2021 to June 2022 in Japan. Vaccination coverage is shown from mid-February 2021 to mid-June 2022 by age group and dose. Colors represent variations by age group, and lines denote different vaccine doses (continuous line for primary series, dashed line for first booster, and lower right corner for second booster dose). The visibility of the 4th dose may be limited due to low coverage.

To guide the population of Japan through an exit from stringent countermeasures, evidence is needed regarding the size of the remaining COVID-19-related burden, especially the potential number of deaths in the future. In the present study, we aimed to estimate the immune fraction of the population, especially those with immune protection against COVID-19 from either symptomatic illness or death as a function of time and age group, such that the magnitude of the remaining burden (or the expected number of deaths) in the population can be estimated. Our estimates can guide not only epidemic exit policies but also objective planning of future immunization programs in Japan.

Material and Methods

Vaccination coverage data

Vaccination data were obtained from the Vaccination Record System (VRS), the nationwide monitoring system of the COVID-19 vaccination program in Japan, from February 17, 2021 to June 12, 2022. These data are reported by local governments to the central government via the VRS and include identifying information of vaccinees. An anonymized dataset was provided by the Ministry of Health, and Labour and Welfare. The data contained information on the number of vaccinated people, including health care workers (from February 17, 2021) and eligible individuals from the general population (from April 12, 2021), stratified by age group and the number of doses over time (Fig. 1). In the present study, we divided the data according to four age groups: 0–39, 40–59, 60–79 and ≥ 80 years. For the primary series vaccination (i.e., completed the second dose), we assumed that vaccination coverage would not change since the last date of empirical data (June 12, 2022) to estimate the potential number of deaths during the study period. Regarding the first booster dose (third dose), as the vaccination program had not yet been completed (i.e. still underway during the estimation), we estimated the number of vaccinated individuals by fitting the logistic curve described in our earlier study [15] and further technical details are available in the Supplementary Material. This approach enabled us to capture the dynamics of the protected proportion being accumulated at the population level and to project the future coverage (Figure S1). Because the proportion vaccinated with the second booster dose (fourth dose) among the eligible population was very small as of mid-June 2022, we considered the following two hypothetical settings. The first setting was that all individuals who received the first booster shot would receive the second booster dose at exactly the same rate as the previous dose since the start of the program on May 25, 2022; in other words, we obtained those distributions by shifting the coverage distributions of the first booster program starting December 1, 2021 by 176 days to the right, among people aged 60 years and older. The second setting was that the vaccination coverage of the second booster dose was expected to be one-half that of the first setting. Finally, we obtained the estimated immune fraction of the population owing to each vaccination over time, assuming that vaccinated individuals had developed protective immunity by exactly 14 days following the latest vaccination. This implies that the vaccination does not immediately elicit immunity during the first 14-day period, resulting in an abrupt increase in vaccine-induced immunity thereafter.

Calculation

We estimated the protected proportion of the population, considering changes in the immune landscape over time, as discussed in our earlier study [15]. Each vaccination program had approximately 6-month intervals between administrations, and the empirically reported length of vaccine-induced protection against symptomatic illness has been estimated to be 3–4 months or less, on average [6], [16], [17], [18], [19], [20]. Thus, to calculate the total number of people with vaccine-induced protection against symptomatic infection, we added the estimated immune fraction from each vaccination round. That is, the total immune proportion in age group a at time t was calculated as:(1) pa,tinfection=Ja,t2+Ja,t3+Ja,tinfection1Na

where Ja,tk is the immune fraction in age group a at time t with vaccination status k, i.e., primary series (k=2), booster dose (k=3) or natural infection (k=infection); Na is the number of people in age group a in October 2021 [6], [16], [21]. The speed of waning of vaccine-induced protection against death was slower than that against symptomatic infection; for example, the protective effect was expected to be more than 80% against the Delta variant at ≥ 20 weeks since the primary vaccination series, even during the time period with Omicron variant [17], [20], [22]. The total proportion immune against death in age group a at time t, accounting for the first booster vaccination, was described as:(2) pa,tdeath=Va,t2-Va,t3Va,t2Ma,t2+Ma,t3+Ma,tinfection1Na

where Ma,tk is the immune fraction protected against COVID-19-related death in age group a at time t with vaccination status k; Va,tk denotes the cumulative number of immunized individuals in age group a at time t with vaccination status k. Protected fractions in settings considering the second booster dose (k=4) were estimated in the same manner as the first booster dose. In the present study, we assumed that the vaccine efficacy against symptomatic infection and death was defined as that against Omicron subvariants BA.1 and BA.2 [6]. Moreover, in accordance with published evidence regarding ascertainment bias based on a seroepidemiologic dataset, the actual number of infected individuals was four times greater than that of confirmed cases [23], [24].

To explore the impact of decreased vaccine-induced immunity, accounting for the circulation of Omicron subvariants BA.4 and BA.5, our previous estimates of vaccine effectiveness were used for the comparison of immune proportions between subvariants of Omicron [22]. For instance, vaccine efficacy would exceed 50% against BA.1 and BA.2 two months after vaccination, but it would be less than 40% against BA.4 and BA.5 in later time. Furthermore, we varied protected fraction at the population level by changing the proportional reduction in vaccine effectiveness against symptomatic infection and also against death in order to address the emergence of antigenically distinct new variants and/or subvariants. Other scenario settings were explored in sensitivity analyses (see the Supplementary Material).

To estimate the remaining burden in the population, i.e., the potential number of deaths, in age group a at time t, the following equation was used:(3) Ba,t=Na(1-ha,t)fa

where 1-ha,t represents the unconditional risk of death related to COVID-19 among the unexposed population in age group a at time t, and fa represents the infection fatality risk (IFR) in age group a. The parameter h, (i.e., the fraction protected against death among people who avoided COVID-19-related symptomatic infection or death) was decomposed as:(4) ha,t=pa,tinfection+1-pa,tinfectionpa,tdeath.

Infection fatality risk (IFR)

For fa, the risk of death among individuals with SARS-CoV-2 infection has not been constant owing to the different (sub)variants and time-dependent immune landscape. Additionally, it was challenging to obtain sufficient information in real time on the risk of death by vaccination status and variants at the time of the present analysis. Therefore, we used two hypothetical assumptions for the IFR related to COVID-19: (i) a constant IFR estimated in the pre-vaccination period over time and (ii) a two-step IFR separating the risk of death from December 2021 onwards, i.e., after introduction of the Omicron variant in Japan. The former values were estimated globally [25], and the latter estimates from December 2021 were calculated from the publicly available dataset in Japan. The two-step IFR from December 2021 was calculated using data from locally infected individuals in Osaka, Japan’s third most populated prefecture. The estimate was derived from the seroepidemiologic survey conducted by the National Institute of Infectious Diseases, Japan [26], and confirmed COVID-19-related deaths were obtained from the Osaka Prefecture website, corresponding to the infected individuals [27]. We used seroprevalence in Osaka rather than in Japan because the age distribution of confirmed deaths related to COVID-19 were only available in Osaka. Further details of the method, including estimation of the IFR, are described in the Supplementary Material.

Uncertainty in potential deaths

The uncertainties in potential mortality were measured by randomly sampling the number of COVID-19-related deaths constituting the IFR. We assumed that the number of deaths in age group a was sufficiently characterized as a binomial distribution:(5) EDaBinomial(n=Ia,p=Da/Ia),

where Da and Ia represent the number of deaths and infected individuals in age group a, respectively. Regarding the two-step IFR starting from December 2021, we used both empirically confirmed deaths and estimated infections in Osaka [27]. When using constant IFR values, we estimated the number of infected individuals by dividing the total number of confirmed deaths in Osaka by the global estimate of IFR [25]. We calculated 95% confidence intervals (CIs) of the potential number of deaths for the IFR using the bootstrapping method, with resampling 1000 times. It is important to highlight that the calculated uncertainties in potential deaths were based on IFR values rather than considering serial dependence, implying that the uncertainty bound may potentially be conservative.

Results

The estimated immune landscape was characterized by variations in assumed vaccination programs, each setting comprising different numbers of vaccination doses and coverages (Fig. 2). In all results, the immune proportions derived from natural infection with SARS-Cov-2 were included up to the last date of the available data (i.e., June 12, 2022). When we accounted for only the first booster dose, two peaks in the immune fraction were observed, attributable to the second and third doses (Fig. 2A–C). However, there were three peaks among people aged ≥ 60 years in settings where the second booster shot was administered, and the distinct peaks were owing to second, third and fourth doses (Fig. 2D–I). Although less than 75% of people aged ≥ 60 years were protected either from symptomatic illness or death without the second booster dose at the end of 2022, approximately 85% of the same age group were protected with the second booster dose, assuming an identical coverage for the previous dose. If the ascertainment rate was lower, the overall immune landscape was elevated (see Figures S3 and S4).Fig. 2 Immune landscape in Japan by age group and different vaccination scenarios. (A)–(C): Immune landscape considering the first booster vaccination program only. (D)–(F): Immune landscape considering the second booster dose with coverage identical to that of the first booster dose. (G)–(I): Immune landscape considering the second booster dose with half the coverage of the first booster dose. (A), (D), and (G): Immune proportions against symptomatic COVID-19. (B), (E) and (H): Conditional immune proportions against death related to COVID-19 infection. (C), (F) and (I): Unconditional immune fractions protected against either symptomatic illness or death related to COVID-19 (i.e., unconditional risk of death).

Fig. 3 illustrates the potential number of deaths by age group in the hypothetical scenario with a fourth dose (second booster), assuming an identical coverage to the third dose from March 2021 up to mid-June 2022. In retrospect, the figure presents an estimate of the potential death toll in the past. Fig. 3A displays the potential number of deaths using the constant IFR. Fig. 3B illustrates the potential number of deaths using the two-step IFR estimated globally from March to November 2021 and locally from December 2021 to June 2022. In both panels, potential fatalities dropped abruptly in summer 2021 and spring 2022 because of the primary series and first booster vaccination. At the beginning of the vaccination program in mid-February 2021, the potential number of deaths was estimated to be as high as 2110 (95% CI: 0–6330) among people aged 0–39 years, 57,354 (95% CI: 19,116–105,135) among people aged 40–59 years, 434,757 (95% CI: 316,223–553,390) among people aged 60–79 years, and 991,123 (95% CI: 817,777–1,164,713) among people aged ≥ 80 years. Subsequently, the primary series vaccination was implemented, substantially and rapidly reducing the potential number of deaths; however, those estimates fluctuated mainly because of waning immunity against SARS-CoV-2. Thus, the potential number of deaths as of June 12, 2022 in the setting using a constant IFR was estimated to be 162 (95% CI: 0–485), 3023 (95% CI: 1008–5542), 43,167 (95% CI: 31,398–54,946), and 87,593 (95% CI: 72,273–102,934) among people aged 0–39, 40–59, 60–79, and ≥ 80 years, respectively (Fig. 3A). For the setting using a two-step IFR, the potential number of deaths as of June 12, 2022 was estimated to be 39 (95% CI: 0–116), 477 (95% CI: 159–874), 3972 (95% CI: 2889–5055), and 43,796 (95% CI: 36,127–51,832) among people aged 0–39, 40–59, 60–79, and ≥ 80 years, respectively (Fig. 3B). The potential deaths in the past under other settings are depicted in Figures S5 and S6.Fig. 3 Potential number of deaths in the past by age group under the second booster shot scenario, assuming coverage identical to that of the previous dose. (A) Potential number of deaths estimated using constant infection fatality risk (IFR) and (B) two-step IFR from March 2021 to June 12, 2022 by age group. Shaded areas represent 95% confidence intervals.

Estimates of the projected number of potential deaths by age group at certain time points from 2021 to 2022 were described in Table 1. In any scenario, people aged ≥ 80 years yielded the largest number of COVID-19-related deaths, followed by people aged 60–79, 40–59, and 0–39 years. Using the two-step IFR, the potential number of deaths among people aged ≥ 80 years on December 31, 2022 was estimated to be 142,590 (95% CI: 117,623–168,755) in the setting with a first booster only, 56,830 (95% CI: 46,880–67,258) with a second booster and coverage identical to that of the first booster dose, and 94,952 (95% CI: 78,327–112,376) in the setting with a second booster and only half the coverage of the first booster dose. We also conducted a sensitivity analysis to explore the impact of the ascertainment rate on the potential number of deaths (Tables S1 and S2). The mortality estimates varied greatly according to the date of evaluation (Table 1). The minimum level was achieved around July 4, 2022 with a total of approximately 60,000 deaths, still much higher than but relatively close to that of seasonal influenza with 20,000–30,000 deaths per year. On December 31, 2022, the potential number of deaths was approximately 160,000 deaths.Table 1 Potential number of deaths at different time points by age group, vaccination setting, and infection fatality risk.

IFR	Date of snapshot
(Y-M−D)	Age group (years)	
0–39	40–59	60–79	≥80	
First booster dose only	
Constant	2021–03-03	2110.1 (0–6330.3)	57,353.7 (19,115.5–105,135.1)	43,4756.7 (316,222.6–553,389.6)	991,123.2 (817,777.2–1,164,713)	
Constant	2021–12-31	792.5 (0–2377.4)	11,713.3 (3904–21,471.7)	103,263.5 (75,109.3–131,441.2)	239,349.3 (197,487.5–281,270.1)	
Constant	2022–07-04	188.4 (0–565.1)	3889.1 (1296.2–7129.1)	52,599.5 (38,258.5–66,952.4)	108,475.2 (89,503–127,474)	
Constant	2022–10-02	365.9 (0–1097.8)	8754.4 (2917.8–16,047.8)	92,514.5 (67,291–117,759.2)	198,297.8 (163,615.8–233,028.6)	
Constant	2022–12-31	554.5 (0–1663.6)	13,865.9 (4621.4–25,417.6)	130,606.7 (94,997.5–166,245.6)	285,184.5 (235,306.1–335,133)	
Two-step	2021–03-03	2110.1 (0–6330.3)	57,353.7 (19,115.5–105,135.1)	434,756.7 (316,222.6–553,389.6)	991,123.2 (817,777.2–1,164,713)	
Two-step	2021–12-31	189.3 (0–568)	1847 (615.7–3386.1)	9501.4 (6910.1–12,092.6)	119,672.9 (98,718.8–141,632.4)	
Two-step	2022–07-04	45 (0–135)	613.2 (204.4–1124.3)	4839.7 (3519.8–6159.6)	54,236.8 (44,740.2–64,189)	
Two-step	2022–10-02	87.4 (0–262.3)	1380.4 (460.1–2530.8)	8512.4 (6190.8–10,833.9)	99,147.4 (81,787.2–117,340.6)	
Two-step	2022–12-31	132.5 (0–397.5)	2186.4 (728.8–4008.4)	12,017.2 (8739.8–15,294.6)	142,590.1 (117,623.3–168,754.9)	
Second booster dose, assuming the same coverage as the first booster dose	
Constant	2021–03-03	2110.1 (0–6330.3)	57,353.7 (19,115.5–105,135.1)	434,756.7 (316,222.6–553,389.6)	991,123.2 (817,777.2–1,164,713)	
Constant	2021–12-31	792.5 (0–2377.4)	11,713.3 (3904–21,471.7)	103,263.5 (75,109.3–131,441.2)	239,349.3 (197,487.5–281,270.1)	
Constant	2022–07-04	188.4 (0–565.1)	3889.1 (1296.2–7129.1)	51,445.1 (37,418.9–65,483.1)	105,727.4 (87,235.8–124,244.9)	
Constant	2022–10-02	365.9 (0–1097.8)	8754.4 (2917.8–16,047.8)	19,731.3 (14,351.7–25,115.5)	34,835.6 (28,742.9–40,936.9)	
Constant	2022–12-31	554.5 (0–1663.6)	13,865.9 (4621.4–25,417.6)	53,812.9 (39,141.1–68,497)	113,662.2 (93,782.9–133,569.6)	
Two-step	2021–03-03	2110.1 (0–6330.3)	57,353.7 (19,115.5–105,135.1)	434,756.7 (316,222.6–553,389.6)	991,123.2 (817,777.2–1,164,713)	
Two-step	2021–12-31	189.3 (0–568)	1847 (615.7–3386.1)	9501.4 (6910.1–12,092.6)	119,672.9 (98,718.8–141,632.4)	
Two-step	2022–07-04	45 (0–135)	613.2 (204.4–1124.3)	4733.5 (3442.5–6024.5)	52,862.9 (43,606.9–62,563.1)	
Two-step	2022–10-02	87.4 (0–262.3)	1380.4 (460.1–2530.8)	1815.5 (1320.4–2310.6)	17,417.5 (14,367.8–20,613.6)	
Two-step	2022–12-31	132.5 (0–397.5)	2186.4 (728.8–4008.4)	4951.4 (3601–6301.7)	56,830.3 (46,879.6–67,258.4)	
Second booster dose, assuming the half coverage of the previous dose	
Constant	2021–03-03	2110.1 (0–6330.3)	57353.7 (19115.5–105135.1)	434,756.7 (316,222.6–553,389.6)	991,123.2 (817,777.2–1,164,713)	
Constant	2021–12-31	792.5 (0–2377.4)	11713.3 (3904–21471.7)	103,263.5 (75,109.3–131,441.2)	239,349.3 (197,487.5–281,270.1)	
Constant	2022–07-04	188.4 (0–565.1)	3889.1 (1296.2–7129.1)	52,021.5 (37,838.1–66,216.7)	107,099.1 (88,367.6–125,856.9)	
Constant	2022–10-02	365.9 (0–1097.8)	8754.4 (2917.8–16,047.8)	49,547.6 (36,038.7–63,067.8)	99,854.2 (82,389.8–117,343.1)	
Constant	2022–12-31	554.5 (0–1663.6)	13,865.9 (4621.4–25,417.6)	88,051.5 (64,044.7–112,078.3)	189,907.5 (156,693–223,168.8)	
Two-step	2021–03-03	2110.1 (0–6330.3)	57,353.7 (19,115.5–105,135.1)	434,756.7 (316,222.6–553,389.6)	991,123.2 (817,777.2–1,164,713)	
Two-step	2021–12-31	189.3 (0–568)	1847 (615.7–3386.1)	9501.4 (6910.1–12,092.6)	119,672.9 (98,718.8–141,632.4)	
Two-step	2022–07-04	45 (0–135)	613.2 (204.4–1124.3)	4786.5 (3481.1–6092)	53,548.7 (44,172.6–63,374.7)	
Two-step	2022–10-02	87.4 (0–262.3)	1380.4 (460.1–2530.8)	4558.9 (3315.6–5802.3)	49,926.3 (41,184.5–59,087.6)	
Two-step	2022–12-31	132.5 (0–397.5)	2186.4 (728.8–4008.4)	8101.7 (5892.1–10,311.2)	94,952.3 (78,326.7–112,375.8)	
Estimated number of potential deaths at specific time points under various vaccination scenarios, considering different hypothetical first booster and second booster programs. Infection fatality risk (IFR) is based on either constant over time (global estimates) or two-step from December 2021 (global and local estimates).

Note: 95% confidence intervals are shown in parentheses.

To account for potential scenarios involving the emergence of new variants of concern (VOC) and/or subvariants with antigenic distinctions, we estimated the potential number of deaths by varying each immune fraction against symptomatic infection and COVID-19-related mortality. For the setting with second booster vaccination, assuming coverage identical to that of the third dose and incorporating two-step IFR, we calculated the cumulative potential number of deaths every 2 weeks, adjusting both the protected fractions, starting from August 2022 (Fig. 4). For instance, if the fraction with protection against symptomatic infection and death increased from 0.3 to 0.5, we anticipated approximately 300,000 potential deaths by mid-December 2022. Moreover, in the same setting, the potential number of deaths was estimated to be lower in September 2022 compared with other settings; subsequently, the estimates were expected to increase, mainly owing to waning immunity of the second booster dose. The projected numbers of potential deaths when varying the proportional reductions in vaccine-induced protection in other IFR scenarios are shown in Figure S7.Fig. 4 Potential number of deaths, varying the immune proportions from August to December 2022. Total number of potential deaths every 2 weeks was estimated based on the second booster vaccination scenario, assuming coverage identical to that of the third dose and using two-step infection fatality risk. Immune fractions against COVID-19-related symptomatic infection and death varied from 0.3 to 1 and from 0.5 to 1, respectively, with a focus on the middle of October 2022.

Lastly, we explored the impact of the immune landscape derived from vaccine efficacy against symptomatic infection with Omicron subvariants BA.4 and BA.5 (Fig. 5). The extent of severity associated with these subvariants had not been fully elucidated during the primary analysis in real time. Therefore, we calculated multiple sets of immune fractions by varying protections against symptomatic illness or death (while focusing on altering the risk of symptomatic infection). The vaccine-induced protection against death was assumed to be the same as that calculated for subvariants BA.1 and BA.2. The largest gap between BA.1/2 and BA.4/5 in terms of the fraction with protection against symptomatic infection was approximately 20% among those aged ≥ 80 years in September 2022. Furthermore, when calculating the unconditional risk of death, the gap in the protection level between BA.1/2 and BA.4/5 was limited and less than 5% and 1% in September and December 2022, respectively.Fig. 5 Comparison of immune landscape, accounting for Omicron subvariants. (A) Immune proportion against symptomatic infection and (B) protected against either symptomatic infection or death (i.e., unconditional risk of death) considering the second booster scenario and assuming vaccination coverage identical to that of the first booster dose, by different Omicron subvariants (December 1, 2021 to December 31, 2022). Each color indicates the age group. Line type represents the Omicron subvariant: BA.1 and BA.2 (main analysis in the present study) or BA.4 and BA.5.

Discussion

Inferring the immune fraction by accounting for vaccine doses and naturally infected individuals and assessing the remaining disease burden in the population associated with SARS-CoV-2 are critical steps to objectively understanding the timing and anticipating the likely impact of exit strategies from stringent countermeasures against COVID-19. In the present study, we estimated the potential number of deaths as a function of calendar time, considering changes in the immune landscape under different vaccination scenarios. At the end of 2022, the immune fraction protected against either COVID-19-related symptomatic illness or death among people aged ≥ 60 years who received a first booster dose was estimated to be less than 75%. Additionally, with a second booster dose that achieved the same coverage as that of the first booster dose, the protected proportion was projected to surpass 80%. Without the second booster vaccination program (i.e., fourth dose), approximately 155,000 deaths would occur among people aged ≥ 60 years, in the event that all susceptible individuals were infected on a single day (December 31, 2022) with two-step IFR. Nevertheless, the second booster could play a key role in reducing mortality by 60%, or by 33% with identical or half the coverage as that of the third dose were achieved. Our approach enabled us to compute the likely disaster size (e.g., age distribution of fatal cases) in the event of a large epidemic. Such estimates can help policymakers understand the likely mortality impact of an epidemic in a quantitative manner and the critical importance of timely immunization to assist in making value judgements.

There are two noteworthy takeaway findings of our study. First, exploring the age-dependent immune status allowed us to quantify the likely magnitude of deaths, if all susceptible individuals were infected with SARS-CoV-2 at a given point in time. Understanding the likely mortality impact is useful for two reasons. The first is associated with quantitative risk assessment, analyzing the epidemiologic risk of a future epidemic. The second reason is to guide policymakers in decision making regarding exit strategies. Comparing estimates against seasonal influenza and analyzing the economic harm of PHSM, the estimates could help in deciding whether non-specific behavioral restrictions should be discontinued. In fact, increasingly more countries have started to refrain from implementing strict PHSMs as they expand vaccine rollout, including primary series and booster doses [28].

The second takeaway is that visualizing the potential disaster size helps in preparing for potential outcomes. Using our proposed method, different vaccination scenarios with different timings were computationally explored over the course of time. We found that the mortality impact was very sensitive to different vaccination scenarios. Although the vaccine-induced risk reduction was estimated to be 10% by administering a fourth dose to older individuals in addition to the first booster setting by the end of 2022, its mortality impact at the population level cannot be underestimated. Thus, first and second booster programs could further mitigate the potential impact of an epidemic. As of July 15, 2022, the Japanese government had designated individuals aged ≥ 60 years and those aged 18–59 years with underlying health conditions as priority vaccine recipients in the second booster program; our analyses adhered to this protocol [29]. However, a further reduction in the remaining burden could be achieved if the scope of the vaccination program were extended to include other target groups, such as health care and nursing care workers (eventually, during summer–autumn 2022, eligibility was expanded to include these individuals and younger age groups).

From a technical perspective, having a proper age-dependent IFR was key to computing the likely number of deaths with occurrence of a major epidemic. We used two different published estimates of the IFR, referred to as global and local estimates in Figure S2. A published study with a similar scope to ours used time-varying IFR [10]. The constant or near constant IFR over the course of time might be an oversimplified assumption. However, Japan had a very low cumulative incidence with spread of the Omicron variant from December 2021, and most of the Japanese population remained unexposed to SARS-CoV-2 at that time [23], [30]. Thus, having a simple two-step IFR from December 2021 among naïve people was extremely useful for application to the population considered in our study. Moreover, the global estimate was in conflict with the actual epidemiologic picture in Asia. The incidence of COVID-19 in countries belonging to the Western Pacific region has been relatively low compared with that in European countries and the United States, as has the mortality impact [31], [32]. This might be partly owing to the fact that health care sectors and elderly care services were not seriously affected during the pre-vaccination period, and therefore hospitals might not have been overwhelmed in maintaining their services. Thus, in the present study, we did not use a more detailed time-dependent case fatality risk, as seen in specific prefectures with a substantial surge in the number of COVID-19 cases [33].

When the initial analysis of the present study was carried out, the Omicron BA.4 and BA.5 subvariants, known to have the ability to evade the existing immune response [34], [35], [36], had recently replaced the Delta variant and Omicron BA.1 and BA.2 subvariants, fueling another epidemic wave with a substantial number of COVID-19 cases [37], [38]. Prior to the Omicron period, the primary series vaccination program had prevented over 14 million deaths across 185 countries and territories during 2021, thanks to both indirect and direct vaccine effects [39]. In the case of Japan, the direct effect of the vaccination program was the prevention of 18,000 deaths from its initiation through November 2021 [40]. Despite such success against the Delta variant, the newly predominant Omicron BA.4 and BA.5 subvariants and possible emergence of new VOCs posed challenges in determining the best mid-term and long-term control strategies in the epidemic. This is because emerging VOCs tended to increase the risk of infection, even among previously immune individuals, potentially leading to death among high-risk individuals. In fact, Japan experienced the world's highest number of newly confirmed COVID-19 cases owing to infection with BA.5 at the end of July 2022, followed by the highest recorded number of daily cases in the country, reaching nearly 250,000 on August 3, 2022 [41]. According to our sensitivity analyses concerning the decrease in vaccine effectiveness at population level, the maximum total number of potential deaths could triple in settings where no reduction in immune proportion was considered. These characteristics, together with our study findings, give rise to the notion that downgrading the official handling of COVID-19 and associated monitoring should not be hastily decided during such periods.

Three technical limitations of this study must be discussed. First, the two-step IFR was estimated from an epidemiologic dataset in Osaka, with seroepidemiologic data and information on COVID-19-related deaths retrieved from publicly available open data. Thus, the corresponding IFR was not strictly derived for all of Japan owing to the absence of case fatality data. At minimum, the seroprevalence data were similar between Osaka and all of Japan [26]. Second, we did not account for transmission dynamics, and our remaining burden rests on an unrealistic assumption in which all naïve individuals were infected at the same time. That is, in the present study, we did not consider the averted number of cases owing to an indirect effect of vaccination [8], [42], [43]. Strictly speaking, the estimated potential number of deaths should be regarded as the upper bound of deaths. Although our approach was not based on mechanistic models such as SIRD model, by examining the 7th wave of COVID-19 patients [44], when the most of all stringent PHSM were lifted, we roughly calculated the total potential number of deaths at around 75 thousand, assuming an ascertainment bias of 0.25. This calculation was performed according to age groups by dividing the number of deaths by the number of infections, then multiplying by the population size and the susceptible fraction (1–0.26), which was determined based on a seroepidemiological survey conducted by the National Institute of Infectious Diseases in December 2022 [45]. When comparing this figure to our estimate of approximately 60 thousand on 2nd October (Table 1), the discrepancy between the two estimates does not appear significant. Third, abruptly changing immune protection owing to continuous VOC replacement was dealt with in an ad-hoc manner. In the present study, we investigated the potential impact on the remaining burden in relation to the risk of death during circulation of primarily the Omicron BA.1 and BA.2 subvariants, which did not strictly or immediately apply to replacement VOCs including BA.4, BA.5 [46], BA.2.75, and XBB. We imposed a simplifying assumption that the effectiveness against death for any vaccine dose was the same as estimates based on the Omicron variant, mainly BA.1 and BA.2 [6]. A similar issue applies to bivalent booster vaccines. We performed ad-hoc analysis using this alternative vaccine for the fourth shot. If the bivalent booster dose was administered as the second booster shot, the potential number of deaths among individuals aged ≥ 80 years at the end of 2022 was estimated to be 64,123 (95% CI: 52,895–75,888), which represents a 24% reduction compared with the setting assuming half the coverage of the third dose, along with two-step IFR (refer to the Supplementary Material for more information).

Despite the above simplification and caution against strict interpretation, we believe that our estimates of the remaining burden in the population provide key insights into population-wide risk assessment based on immunity status, potentially informing the appropriate timing for lifting of strict PHSM against COVID-19.

Conclusion

Using immune landscape, we estimated the potential number of deaths by age group both in the past and near future. The primary series and booster doses in Japan’s vaccination programs were shown to substantially mitigate the number of deaths. Mortality could be further reduced if a second booster (i.e., fourth dose) vaccination program were implemented and extended to health care and nursing care workers. The variability among the assumed settings provides policymakers with a quantitative understanding of the mortality impact of an epidemic, emphasizing the critical importance of timely immunization to assist in decision-making processes.

Author contributions

Taishi Kayano: Writing- Original draft preparation, Data curation, Methodology, Analysis. Misaki Sasanami: Data curation, Analysis. Hiroshi Nishiura: Methodology, Writing, Review, Editing, and Supervision.

Funding source

T.K. received funding from JSPS KAKENHI (21 K10495). H.N. received funding from Health and Labour Sciences Research Grants (21HB1002, 21HA2016 and 23HA2005), the Japan Agency for Medical Research and Development (JP23fk0108612 and JP 23fk0108685), JSPS KAKENHI (21H03198 and 22 K19670), the Environment Research and Technology Development Fund (JPMEERF20S11804) of the Environmental Restoration and Conservation Agency of Japan, Kao Health Science Research, the Daikin GAP Fund of Kyoto University, the Japan Science and Technology Agency SICORP program (JPMJSC20U3 and JPMJSC2105), and the RISTEX program for Science, Technology, and Innovation Policy (JPMJRS22B4).

CRediT authorship contribution statement

Taishi Kayano: Writing – review & editing, Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Misaki Sasanami: Writing – review & editing, Investigation, Formal analysis, Data curation. Hiroshi Nishiura: Writing – review & editing, Supervision, Methodology, Investigation, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A Supplementary data

The following are the Supplementary data to this article:Supplementary Data 1

Data availability

The data are publicly available and the source is referenced.

Acknowledgements

We thank Analisa Avila, MPH, ELS, of Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.

Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.jvacx.2024.100547.
==== Refs
References

1 Li Q. Guan X. Wu P. Wang X. Zhou L. Tong Y. Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus-Infected Pneumonia N Engl J Med 382 13 2020 1199 1207 10.1056/NEJMoa2001316 31995857
2 Lu H. Stratton C.W. Tang Y.W. Outbreak of pneumonia of unknown etiology in Wuhan, China: The mystery and the miracle J Med Virol 92 4 2020 401 402 10.1002/JMV.25678 31950516
3 Carfì A. Bernabei R. Landi F. Gemelli Against COVID-19 Post-Acute Care Study Group. Persistent Symptoms in Patients After Acute COVID-19 JAMA 324 6 2020 603 605 10.1001/jama.2020.12603 32644129
4 Wu Z. McGoogan J.M. Characteristics of and Important Lessons From the Coronavirus Disease 2019 (COVID-19) Outbreak in China: Summary of a Report of 72 314 Cases From the Chinese Center for Disease Control and Prevention JAMA 323 13 2020 1239 1242 10.1001/JAMA.2020.2648 32091533
5 Kim M.K. Lee B. Choi Y.Y. Um J. Lee K.S. Sung H.K. Clinical Characteristics of 40 Patients Infected With the SARS-CoV-2 Omicron Variant in Korea J Korean Med Sci 37 3 2022 e31 35040299
6 UK Health Security Agency. COVID-19 vaccine surveillance report: week 27 2022. https://assets.publishing.service.gov.uk/media/62c6c2168fa8f54e855dfe29/Vaccine-surveillance-report-week-27.pdf (accessed July 11, 2022).
7 Ashworth M. Thunström L. Cherry T.L. Newbold S.C. Finnoff D.C. Emphasize personal health benefits to boost COVID-19 vaccination rates Proc Natl Acad Sci U S A 118 32 2021 10.1073/PNAS.2108225118 e2108225118
8 Gavish N. Yaari R. Huppert A. Katriel G. Population-level implications of the Israeli booster campaign to curtail COVID-19 resurgence Sci Transl Med 14 647 2022 9836 10.1126/scitranslmed.abn9836
9 Barda N. Dagan N. Cohen C. Hernán M.A. Lipsitch M. Kohane I.S. Effectiveness of a third dose of the BNT162b2 mRNA COVID-19 vaccine for preventing severe outcomes in Israel: an observational study Lancet 398 10316 2021 2093 2100 10.1016/S0140-6736(21)02249-2 34756184
10 Chapman L.A.C. Barnard R.C. Russell T.W. Abbott S. van Zandvoort K. Davies N.G. Unexposed populations and potential COVID-19 hospitalisations and deaths in European countries as per data up to 21 November 2021 Euro Surveill 27 1 2022 2101038 10.2807/1560-7917.ES.2022.27.1.2101038 34991776
11 Ministry of Health Labour and Welfare of Japan. COVID-19 Vaccines 2022. https://www.mhlw.go.jp/stf/covid-19/vaccine.html (accessed July 14, 2022).
12 Prime Ministry’s Office of Japan. Schedule of COVID-19 vaccines 2022. https://www.kantei.go.jp/jp/headline/kansensho/vaccine_supply.html (accessed July 14, 2022).
13 Prime Minister’s Office of Japan. Information on Vaccination of COVID-19 2022. https://www.kantei.go.jp/jp/headline/kansensho/vaccine.html (accessed July 11, 2022).
14 Digital Agency of Japan. Vaccination coverages of COVID-19 vaccines 2022. https://info.vrs.digital.go.jp/dashboard (accessed July 14, 2022).
15 Sasanami M. Fujimoto M. Kayano T. Hayashi K. Nishiura H. Projecting the COVID-19 immune landscape in Japan in the presence of waning immunity and booster vaccination J Theor Biol 559 2023 111384 10.1016/j.jtbi.2022.111384
16 Andrews N. Stowe J. Kirsebom F. Toffa S. Rickeard T. Gallagher E. Covid-19 Vaccine Effectiveness against the Omicron (B.1.1.529) Variant N Engl J Med 386 16 2022 1532 1546 10.1056/NEJMOA2119451 35249272
17 Lin D.-Y. Gu Y. Wheeler B. Young H. Holloway S. Sunny S.-K. Effectiveness of Covid-19 Vaccines over a 9-Month Period in North Carolina N Engl J Med 386 10 2022 933 941 10.1056/NEJMOA2117128 35020982
18 Skowronski D.M. Febriani Y. Ouakki M. Setayeshgar S. El Adam S. Zou M. Two-Dose Severe Acute Respiratory Syndrome Coronavirus 2 Vaccine Effectiveness With Mixed Schedules and Extended Dosing Intervals: Test-Negative Design Studies From British Columbia and Quebec Canada Clin Infect Dis 75 11 2022 1980 1992 10.1093/CID/CIAC290 35438175
19 Goldberg Y. Mandel M. Bar-On Y.M. Bodenheimer O. Freedman L. Haas E.J. Waning Immunity after the BNT162b2 Vaccine in Israel N Engl J Med 385 24 2021 e85 34706170
20 Nordström P. Ballin M. Nordström A. Risk of infection, hospitalisation, and death up to 9 months after a second dose of COVID-19 vaccine: a retrospective, total population cohort study in Sweden Lancet 399 10327 2022 814 823 10.1016/S0140-6736(22)00089-7 35131043
21 Statistics Bureau: Ministry of Internal Affairs and Communications Japan. Summary of Population Estimates 2022. https://www.stat.go.jp/data/jinsui/2.html#monthly (accessed July 12, 2022).
22 Okada Y. Kayano T. Anzai A. Zhang T. Nishiura H. Okada Y. Protection against SARS-CoV-2 BA.4 and BA.5 subvariants via vaccination and natural infection: A modeling study Math Biosci Eng 20 2 2023 2530 2543 10.3934/mbe.2023118 36899545
23 Sanada T. Honda T. Yasui F. Yamaji K. Munakata T. Yamamoto N. Serologic Survey of IgG Against SARS-CoV-2 Among Hospital Visitors Without a History of SARS-CoV-2 Infection in Tokyo, 2020–2021 J Epidemiol 32 2 2022 105 111 10.2188/JEA.JE20210324 34776499
24 Zhang T. Nishiura H. Estimating infection fatality risk and ascertainment bias of COVID-19 in Osaka, Japan from February 2020 to January 2022 Sci Rep 13 1 2023 5540 10.1038/s41598-023-32639-9 37016060
25 O’Driscoll M. Ribeiro Dos Santos G. Wang L. Cummings D.A.T. Azman A.S. Paireau J. Age-specific mortality and immunity patterns of SARS-CoV-2 Nature 590 7844 2020 140 145 10.1038/s41586-020-2918-0 33137809
26 National Institute of Infectious Diseases. Report of sero-epidemiological investigation of COVID-19 2022. https://www.niid.go.jp/niid/ja/2019-ncov/2484-idsc/11118-covid19-79.html (accessed July 12, 2022).
27 COVID-19 Task Force Meeting Osaka Prefecture. Patients and Deaths of COVID-19, Osaka 2022. https://www.pref.osaka.lg.jp/iryo/osakakansensho/happyo.html (accessed July 12, 2022).
28 Zhang Y. Quigley A. Wang Q. MacIntyre C.R. Non-pharmaceutical interventions during the roll out of covid-19 vaccines BMJ 375 2021 2314 10.1136/BMJ.N2314
29 Ministry of Health, Labour and Welfare, Japan. Information on vaccination of fourth dose 2022. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/vaccine_fourth-dose.html (accessed July 19, 2022).
30 Japan Broadcasting Corporation. Special website for COVID-19: confirmed cases, hospitalizations and deaths 2022. https://www3.nhk.or.jp/news/special/coronavirus/data-all/ (accessed July 19, 2022).
31 COVID-19 Forecasting Team. Variation in the COVID-19 infection–fatality ratio by age, time, and geography during the pre-vaccine era: a systematic analysis. Lancet 2022;399(10334):1469–88. doi: 10.1016/S0140-6736(21)02867-1.
32 Shaw R, Kim Y kyun, Hua J. Governance, technology and citizen behavior in pandemic: Lessons from COVID-19 in East Asia. Prog Disaster Sci 2020;6:100090. doi: 10.1016/J.PDISAS.2020.100090.
33 Hayashi K. Nishiura H. Time-dependent risk of COVID-19 death with overwhelmed health-care capacity in Japan, 2020–2022 BMC Infect Dis 22 1 2022 933 10.1186/s12879-022-07929-8 36510193
34 Hachmann N.P. Miller J. Collier A.Y. Ventura J.D. Yu J. Rowe M. Neutralization Escape by SARS-CoV-2 Omicron Subvariants BA.2.12.1, BA.4, and BA.5 N Engl J Med 387 1 2022 86 88 10.1056/NEJMc2206576 35731894
35 Cao Y. Yisimayi A. Jian F. Song W. Xiao T. Wang L. BA.2.12.1, BA.4 and BA.5 escape antibodies elicited by Omicron infection Nature 608 7923 2022 593 602 10.1038/s41586-022-04980-y 35714668
36 Qu P. Faraone J. Evans J.P. Zou X. Zheng Y.-M. Carlin C. Neutralization of the SARS-CoV-2 Omicron BA.4/5 and BA.2.12.1 Subvariants N Engl J Med 386 26 2022 2526 2528 10.1056/NEJMc2206725 35704428
37 Tegally H. Moir M. Everatt J. Giovanetti M. Scheepers C. Wilkinson E. Emergence of SARS-CoV-2 Omicron lineages BA.4 and BA.5 in South Africa Nat Med 28 9 2022 1785 1790 10.1038/s41591-022-01911-2 35760080
38 Mohapatra R.K. Kandi V. Sarangi A.K. Verma S. Tuli H.S. Chakraborty S. The recently emerged BA.4 and BA.5 lineages of Omicron and their global health concerns amid the ongoing wave of COVID-19 pandemic – Correspondence Int J Surg 103 2022 106698 10.1016/J.IJSU.2022.106698 35690362
39 Watson O.J. Barnsley G. Toor J. Hogan A.B. Winskill P. Ghani A.C. Global impact of the first year of COVID-19 vaccination: a mathematical modelling study Lancet Infect Dis 22 9 2022 1293 1302 10.1016/S1473-3099(22)00320-6 35753318
40 Kayano T. Sasanami M. Kobayashi T. Ko Y.K. Otani K. Suzuki M. Number of averted COVID-19 cases and deaths attributable to reduced risk in vaccinated individuals in Japan Lancet Reg Health West Pac 28 2022 100571 10.1016/J.LANWPC.2022.100571
41 Kyodo News. Japan’s daily COVID-19 cases hit record high of nearly 250,000 2022. https://english.kyodonews.net/news/2022/08/a94fa9df7a97-japans-daily-covid-19-cases-hit-record-high-of-nearly-250000.html (accessed August 5, 2022).
42 Gallagher M.E. Sieben A.J. Nelson K.N. Kraay M.A.N. Orenstein W.A. Lopman B. Indirect benefits are a crucial consideration when evaluating SARS-CoV-2 vaccine candidates Nat Med 27 1 2021 4 5 10.1038/s41591-020-01172-x 33230343
43 Kayano T. Ko Y. Otani K. Kobayashi T. Suzuki M. Nishiura H. Evaluating the COVID-19 vaccination program in Japan, 2021 using the counterfactual reproduction number Sci Rep 13 1 2023 17762 10.1038/s41598-023-44942-6 37853098
44 Ministry of Health Labour and Welfare of Japan. Materials in COVID-19 Advisory Board 2022. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/0000121431_00395.html (accessed April 23, 2024).
45 National Institute of Infectious Diseases. Report of sero-epidemiological investigation of COVID-19 2023. https://www.niid.go.jp/niid/ja/2019-ncov/2484-idsc/12191-covid19-85.html (accessed April 23, 2024).
46 UK Health Security Agency. SARS-CoV-2 variants of concern and variants under investigation in England. Technical briefing 43. 2022. https://assets.publishing.service.gov.uk/media/631a0b8d8fa8f50207d07b57/Technical-Briefing-43-24June2022.pdf (accessed July 19, 2022).
