Advertisement
Research Article| Volume 129, P19-31, April 2023

Immunogenicity of a fractional or full third dose of AZD1222 vaccine or BNT162b2 messenger RNA vaccine after two doses of CoronaVac vaccines against the Delta and Omicron variants

Open AccessPublished:January 19, 2023DOI:https://doi.org/10.1016/j.ijid.2023.01.022

      Highlights

      • No safety concern of the third dose boosting with either vaccine.
      • Half-dose immunogenicity of either vaccine was noninferior to the full dose.
      • Both doses elicited high immunogenicity and were best with longer boost intervals.
      • The immunogenicity against Omicron was better with the longer boost interval.

      Abstract

      Objectives

      The study aimed to compare the immunogenicity and safety of fractional (half) third doses of heterologous COVID-19 vaccines (AZD1222 or BNT162b2) to full doses after the two-dose CoronaVac and when boosting after three different extended intervals.

      Methods

      At 60-<90, 90-<120, or 120-180 days intervals after the two-dose CoronaVac, participants were randomized to full-dose or half-dose AZD1222 or BNT162b2, followed up at day 28, 60, and 90. Vaccination-induced immune responses to Ancestral, Delta, and Omicron BA.1 strains were evaluated by antispike, pseudovirus, and microneutralization and T cell assays. Descriptive statistics and noninferiority cut-offs were reported as geometric mean concentration or titer and concentration or titer ratios comparing baseline to day 28 and day 90 and different intervals.

      Results

      No safety concerns were detected. All assays and intervals showed noninferior immunogenicity between full doses and half doses. However, full-dose vaccines and/or longer 120-180-day intervals substantially improved the immunogenicity (measured by antispike or measured by pseudotyped virus neutralizing titers 50; P <0.001). Seroconversion rates were over 90% against the SARS-CoV-2 strains by all assays. Immunogenicity waned more quickly with half doses than full doses but remained high against the Ancestral or Delta strains. Against Omicron, the day 28 immunogenicity increased with longer intervals than shorter intervals for full-dose vaccines.

      Conclusion

      Immune responses after day 28 when boosting at longer intervals after the two-dose CoronaVac was optimal. Half doses met the noninferiority criteria compared with the full dose by all the immune assays assessed.

      Keywords

      Introduction

      The approved COVID-19 vaccines in Thailand include the messenger RNA (mRNA)-based vaccines, BNT162b2 (‘PF’; Pfizer Inc, New York, USA; BioNTech Manufacturing GmbH, Mainz, Germany) and mRNA-1273 (Moderna Inc, Cambridge, USA); the inactivated virus vaccine, CoronaVac (Sinovac Biotech, Beijing, China); and the adenovirus vector vaccine, AZD1222 (‘AZ’; Oxford-AstraZeneca, UK). However, variants of concern, such as Delta (B.1.617.2) and Omicron, especially BA.1 (B.1.1.529), are driving infection surges amid waning vaccine effectiveness (VE) [
      • Patel MK
      • Bergeri I
      • Bresee JS
      • Cowling BJ
      • Crowcroft NS
      • Fahmy K
      • et al.
      Evaluation of post-introduction COVID-19 vaccine effectiveness: summary of interim guidance of the World Health Organization.
      ]. Although Thailand has administered over 136 million vaccine doses, Omicron has increased deaths slightly among elderly, bedridden individuals, highlighting a need to sustain vaccinations. Thailand used CoronaVac widely but its immunogenicity and VE are inconsistent or limited [
      • Khoury DS
      • Cromer D
      • Reynaldi A
      • Schlub TE
      • Wheatley AK
      • Juno JA
      • et al.
      Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection.
      ]. In China [
      • Zhang Y
      • Zeng G
      • Pan H
      • Li C
      • Hu Y
      • Chu K
      • et al.
      Safety, tolerability, and immunogenicity of an inactivated SARS-CoV-2 vaccine in healthy adults aged 18–59 years: a randomised, double-blind, placebo-controlled, phase 1/2 clinical trial.
      ] and Chile [
      • Jara A
      • Undurraga EA
      • González C
      • Paredes F
      • Fontecilla T
      • Jara G
      • et al.
      Effectiveness of an inactivated SARS-CoV-2 vaccine in Chile.
      ], the two-dose CoronaVac (known as ‘primary series’) neutralizing antibody (nAb) titers were lowest against Delta, resulting in breakthrough infections, severe COVID-19 disease, and deaths. Although nAb titers from CoronaVac primary series waned after 3-4 months, nAb were increased more substantially when boosters were administered at 8 months than at 2 months after the primary series [
      • Zeng G
      • Wu Q
      • Pan H
      • Li M
      • Yang J
      • Wang L
      • et al.
      Immunogenicity and safety of a third dose of CoronaVac, and immune persistence of a two-dose schedule, in healthy adults: interim results from two single-centre, double-blind, randomised, placebo-controlled phase 2 clinical trials.
      ].
      In 2021, Thailand initiated heterologous regimens with AZD1222, BNT162b2, or mRNA-1273 but was limited by mRNA vaccine shortages. Although COVID-19 vaccines are widely available in many Western countries, vaccine inequity and supply constraints remain problematic in many low- and middle-income countries. Thus, different vaccination strategies should be explored, including extended dosing intervals (or dose stretching), dose reductions or fractional dosing, and vaccine platform combinations. Fractional dosing conserves supplies while increasing coverage without compromising immunogenicity and can help resource-limited countries extend supplies and reduce mortality [
      • Cowling BJ
      • Lim WW
      • Cobey S.
      Fractionation of COVID-19 vaccine doses could extend limited supplies and reduce mortality.
      ,

      World Health Organization. Interim statement on dose-sparing strategies for COVID-19 vaccines (fractionated vaccine doses), https://www.who.int/news/item/10-08-2021-interim-statement-on-dose-sparing-strategies-for-covid-19-vaccines-(fractionated-vaccine-doses); 2021 [accessed 16 August 2022].

      . Fractional dosing may also broaden the population coverage and facilitate herd immunity, as has been proven in polio. In the COVID-19 pandemic, 6 months after the vaccination with a fractional one-fourth dose of primary mRNA-1273 regimens, nAb responses were half as robust as full doses, but the VE was over 80% of that of full-dose vaccinations [
      • Mateus J
      • Dan JM
      • Zhang Z
      • Rydyznski Moderbacher C
      • Lammers M
      • Goodwin B
      • et al.
      Low-dose mRNA-1273 COVID-19 vaccine generates durable memory enhanced by cross-reactive T cells.
      ]. During the AZD1222 trials in UK recipients inadvertently primed with half doses and subsequently administered full-dose boosters, VE was 90% (95% confidence interval [CI]: 67-97) [
      • Voysey M
      • Clemens SAC
      • Madhi SA
      • Weckx LY
      • Folegatti PM
      • Aley PK
      • et al.
      Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK.
      ]. Thai adults boosted with 30 µg BNT162b2 and 15 µg BNT162b2 8-12 weeks after the two-dose CoronaVac or AZD1222 had high antireceptor-binding domain (anti-RBD) immunoglobulin (Ig)G concentrations [

      Angkasekwinai N, Niyomnaitham S, Sewatanon J, Phumiamorn S, Sukapirom K, Senawong S, et al. The immunogenicity and reactogenicity of four COVID-19 booster vaccinations against SARS-CoV-2 variants of concern (Delta, Beta, and Omicron) following CoronaVac or CHADOX1-S primary series. medRxiv. 04 February 2022. https://www.medrxiv.org/content/10.1101/2021.11.29.21266947v3; [accessed 20 August 2022].

      ], nAb titers against all variants, and T cell responses [
      • Angkasekwinai N
      • Sewatanon J
      • Niyomnaitham S
      • Phumiamorn S
      • Sukapirom K
      • Sapsutthipas S
      • et al.
      Comparison of safety and immunogenicity of CoronaVac and ChAdOx1 against the SARS-CoV-2 circulating variants of concern. Alpha: Delta, Beta) in Thai healthcare workers.
      ,
      • Niyomnaitham S
      • Quan Toh Z
      • Wongprompitak P
      • Jansarikit L
      • Srisutthisamphan K
      • Sapsutthipas S
      • et al.
      Immunogenicity and reactogenicity against the SARS-CoV-2 variants following heterologous primary series involving CoronaVac, ChAdox1 nCov-19 and BNT162b2 plus BNT162b2 booster vaccination: an open-label randomized study in healthy Thai adults.
      . Reducing vaccine doses does not necessarily reduce VE, and lower doses could quickly increase the immunity of at-risk populations when vaccine supply is low or during serious disease outbreaks and epidemics. To this end, Moderna has now halved the dose of its mRNA-1273 vaccine from 100 µg to 50 µg, without compromising immunogenicity [
      • Mateus J
      • Dan JM
      • Zhang Z
      • Rydyznski Moderbacher C
      • Lammers M
      • Goodwin B
      • et al.
      Low-dose mRNA-1273 COVID-19 vaccine generates durable memory enhanced by cross-reactive T cells.
      ,

      World Health Organization. The Moderna COVID-19 (mRNA-1273) vaccine: what you need to know, who.int/news-room/feature-stories/detail/the-moderna-covid-19-mrna-1273-vaccine-what-you-need-to-know; 2022 (accessed 16 August 2022).

      .
      Prolonged or different booster intervals may also enhance immunogenicity [
      • Flaxman A
      • Marchevsky NG
      • Jenkin D
      • Aboagye J
      • Aley PK
      • Angus B
      • et al.
      Reactogenicity and immunogenicity after a late second dose or a third dose of ChAdOx1 nCoV-19 in the UK: a substudy of two randomised controlled trials (COV001 and COV002).
      ,
      • Voysey M
      • Clemens SAC
      • Madhi SA
      • Weckx LY
      • Folegatti PM
      • Aley PK
      • et al.
      Single-dose administration and the influence of the timing of the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine: a pooled analysis of four randomised trials.
      . In one study, third dose boosting at 44-45-week intervals significantly increased antibody levels versus boosting at 15-25-week or 8-12-week intervals. Alternatively, SARS-CoV-2 immune responses can also be restored using different booster intervals. A third dose of CoronaVac [
      • Zeng G
      • Wu Q
      • Pan H
      • Li M
      • Yang J
      • Wang L
      • et al.
      Immunogenicity and safety of a third dose of CoronaVac, and immune persistence of a two-dose schedule, in healthy adults: interim results from two single-centre, double-blind, randomised, placebo-controlled phase 2 clinical trials.
      ] administered 8 months after the second dose increased antibody levels more than when administered at 2 months, whereas the antibody responses were two-fold higher with a booster dose of AZD1222 administered at 12-week or longer intervals than 6-weeks or shorter intervals [
      • Voysey M
      • Clemens SAC
      • Madhi SA
      • Weckx LY
      • Folegatti PM
      • Aley PK
      • et al.
      Single-dose administration and the influence of the timing of the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine: a pooled analysis of four randomised trials.
      ].
      In at-risk populations in the United Kingdom, third doses of AZD1222 led to higher antibody levels that correlated with high efficacy and T cell responses after a prolonged, dose-stretched interval between doses than at shorter intervals [
      • Flaxman A
      • Marchevsky NG
      • Jenkin D
      • Aboagye J
      • Aley PK
      • Angus B
      • et al.
      Reactogenicity and immunogenicity after a late second dose or a third dose of ChAdOx1 nCoV-19 in the UK: a substudy of two randomised controlled trials (COV001 and COV002).
      ]. In fact, the odds ratios for symptomatic disease after shorter intervals between second and booster (third) doses [
      • Andrews N
      • Stowe J
      • Kirsebom F
      • Toffa S
      • Sachdeva R
      • Gower C
      • et al.
      Effectiveness of COVID-19 booster vaccines against COVID-19-related symptoms, hospitalization and death in England.
      ] was higher than after longer intervals, along with VE estimates of 93.2% (95% CI, 92.8-93.6) versus 95.6% (95% CI, 94.9-96.1), respectively. A Chinese trial of the ZF2001 vaccine [
      • Zhao X
      • Zheng A
      • Li D
      • Zhang R
      • Sun H
      • Wang Q
      • et al.
      Neutralisation of ZF2001-elicited antisera to SARS-CoV-2 variants.
      ] found that it may be feasible to combine these strategies because the ZF2001 neutralizing activity and resilience to all tested variants was higher with extended prime-boost (third dose) intervals than with shorter intervals. Extended intervals may allow antibodies to mature for longer, thus enhancing this regimen. In fact, Omicron-neutralizing antibodies were detected in only 56% of short-interval vaccine recipients versus all (100%) prolonged-interval vaccine recipients [
      • Zhao X
      • Li D
      • Ruan W
      • Chen Z
      • Zhang R
      • Zheng A
      • et al.
      Effects of a prolonged booster interval on neutralization of Omicron variant.
      ], 69% of whom also demonstrated Omicron-neutralizing antibodies at 4-6 months after the booster. Although data remain limited on the most optimal prime-boost interval, Israeli studies [
      • Bar-On YM
      • Goldberg Y
      • Mandel M
      • Bodenheimer O
      • Amir O
      • Freedman L
      • et al.
      Protection by a fourth dose of BNT162b2 against Omicron in Israel.
      ,
      • Regev-Yochay G
      • Gonen T
      • Gilboa M
      • Mandelboim M
      • Indenbaum V
      • Amit S
      • et al.
      Efficacy of a fourth dose of Covid-19 mRNA vaccine against Omicron.
      noted a restoration of antibody levels and enhanced immunogenic protection against severe disease when a second booster (fourth dose) was given 4 months or longer after a first booster, with no new safety concerns [
      • Munro APS
      • Feng S
      • Janani L
      • Cornelius V
      • Aley PK
      • Babbage G
      • et al.
      Safety, immunogenicity, and reactogenicity of BNT162b2 and mRNA-1273 COVID-19 vaccines given as fourth-dose boosters following two doses of ChAdOx1 nCoV-19 or BNT162b2 and a third dose of BNT162b2 (COV-BOOST): a multicentre, blinded, phase 2, randomised trial.
      ]. The World Health Organization now recommends an interval of 4-6 months after the primary series [

      World Health Organization. Interim statement on the use of additional booster doses of emergency use listed mRNA vaccines against COVID-19, https://www.who.int/news/item/17-05-2022-interim-statement-on-the-use-of-additional-booster-doses-of-emergency-use-listed-mrna-vaccines-against-covid-19; 2022 [accessed 16 August 2022].

      ], especially in the face of Omicron dominance, as well as heterologous or homologous schedule boosters (third or more).
      To better understand the applicability of these strategies, we compared the immune responses and safety of fractional (half) third doses of heterologous COVID-19 vaccines (AZD1222 or BNT162b2) with full doses after the CoronaVac primary series and when boosting after three different extended intervals.

      Methods

      Study design and participants

      This prospective, multicenter, randomized, observer-blinded phase II study enrolled 1320 healthy adults aged 20 years or older (Figure 1). Eligible participants had 60-<90 days, 90-<120 days, or 120-180 days ‘intervals’ after receipt of CoronaVac primary series (21-28 days apart); with no history of fever or symptoms within 7 days of enrollment; and were not pregnant. Written informed consents were obtained before all study procedures. Exclusion criteria were a history of COVID-19 infection within 3 months of enrollment, contraindication to ChAdOx1 or BNT162b2, and confirmed or suspected immunosuppressive or immunodeficient state. Briefly, participants were divided into two cohorts comprising 660 individuals each and subsequently into three interval-stratified subgroups. Each subgroup was randomly assigned in a 1: 1 ratio to either half-dose AZD1222 (AZHD) or BNT162b2 (PFHD) or full-dose AZD1222 (AZFD) or BNT162b2 (PFFD) vaccine. Participants were monitored after the vaccination for immediate adverse events (AEs), and recorded solicited AEs for up to 7 days. All participants completed five postvaccination follow-up visits at day 28, 60, and 90, and were assessed for safety at day 7, 28, 60, and 90. Blood was sampled at baseline and days 28, 60, and 90 to evaluate humoral immunity, and at baseline and day 28 from half of the participants per subgroup to evaluate T cell-mediated immunity. Patients were discontinued before study completion due to withdrawal (participant-initiated, for any reason), loss to follow-up, or sponsor-initiated study termination for administrative or other reasons.
      Figure 1
      Figure 1CONSORT diagram depicting trial design and vaccine administration groups.

      AEs

      Safety analyses included all randomized participants with at least one vaccination dose. All discontinued participants’ diary cards were collected and any AEs and concomitant, ongoing medications at the last visit were recorded. Newly reported AEs since the last visit were recorded, and AEs were systematically collected at all visits. Diary cards were reviewed at day 7 follow-ups and ongoing solicited AEs were recorded and monitored. AEs of special interest (AESI), medically-assisted AEs, and serious AEs (SAE) were collected through day 90.

      Immunological analyses

      At baseline and days 28, 60, and 90, IgG antibodies to full-length prefusion spike protein of SARS-CoV-2 (anti-S RBD IgG) and antinucleocapsid proteins levels were quantified using a validated enzyme-linked immunosorbent assay (ELISA; Roche). Anti-S and antinucleocapsid RBD IgG results were summarized with 95% exact CI, geometric mean concentrations (GMCs), geometric mean fold rise (GMFR) from baseline, percentages of participants achieving IgG seroresponse, and reaching the 95% CI for a ≥4-fold increase from baseline. Vaccination-induced nAb inhibition of SARS-CoV-2 seroresponse was quantified by SARS-CoV-2 pseudovirus neutralization assay (PNA), and defined as over 50% and 68% inhibition of Delta and the Ancestral strain, respectively, from baseline and at 28 and 90 days. Samples with >90% inhibition or 100% inhibition were evaluated for 50% neutralizing titer (NT50) of geometric mean titers (GMTs) against Ancestral, Delta, or Omicron pseudovirus S proteins (pseudotyped virus neutralizing titers [PVNT]). PVNT results were summarized from baseline with 95% CI, GMT, GMFR, and percentage of subjects with ≥4-fold increase in NT50 seroresponse against SARS-CoV-2 pseudovirus from baseline and at 28 and 90 days. SARS-CoV-2 microneutralization (microNT) was assayed in prescreening samples stratified by PVNT against Delta, Ancestral, or Omicron. N-proteins were measured by ELISA. SARS-CoV-2 cellular responses (i.e., T cell-mediated immunity) were quantified using an interferon-γ release ELISA assay (Euroimmun, Lubeck, Germany).

      Outcomes

      The primary outcome of immunogenicity was assessed through IgG levels against anti-S RBD at baseline and 28, 60, and 90 days after the third dose/booster given at different intervals among participants with CoronaVac primary series. The immunogenicity or functional (neutralizing) humoral immune response elicited by each regimen was also assessed by the PNA at baseline and at 28 and 90 days after the third dose. Third dose vaccination safety and tolerability were evaluated at different intervals among participants with CoronaVac primary series. The secondary outcomes of functional (neutralizing) humoral immune response at baseline and 28 and 90 days were measured by the microNT assay in specimens from participants with ≥4-fold seroconversion, with NT50 GMT assessed after the third dose vaccinations. The exploratory outcome—immunogenicity—was evaluated through S-protein-specific T cell responses at baseline and 28 days after the third dose.

      Statistical analyses

      Briefly, 110 participants were estimated to be required per arm. Immune response was assessed using two-sided statistical testing, with a significance level of 0.05 and a 95% CI. Baseline demographics were summarized for intent-to-treat (ITT) populations using descriptive statistics. The immunogenicity GMFR was computed using estimates of the log difference of the paired samples. Anti-S RBD IgG antibody concentrations were summarized using GMCs at baseline and 28, 60, and 90 days after vaccination; GMFR from baseline; and percentage of subjects with IgG seroresponse (all with 95% CI). NT50 nAb titers against SARS-CoV-2 pseudovirus were summarized using GMT at baseline and 28 and 90 days after vaccination, GMFR from baseline, and percentage of subjects with NT50 seroresponse against SARS-CoV-2 pseudovirus at 28 and 90 days after vaccination (all with 95% CI). Safety was assessed using the Clopper-Pearson method. NT50 nAb titers (seroresponse) against SARS-CoV-2 microNT were summarized at baseline and 28 and 90 days after vaccination using GMT, GMFR from baseline, and percentage of subjects with NT50 seroresponse at 28 and 90 days after vaccination (all with 95% CI). The noninferiority of seroresponse rates was calculated between full- and half-dose (AZ and PF) groups. GMC ratios were concluded to be noninferior when the lower bound of the two-sided 95% CI comparing the vaccine groups was >0.67 and the point estimate (PE) was >0.8. The difference in percentage of individuals with ≥4-fold GMFR was concluded to be noninferior when the lower bound of the two-sided 95% CI for the difference in proportions between vaccine groups was greater than -10%. The difference in percentage of individuals with seroresponse was concluded to be noninferior when the lower bound of the two-sided 95% CI for the difference in proportions between vaccine groups was greater than -10%. To achieve 80% power to demonstrate noninferiority. it is estimated that 247 subjects per group (a fractional or full dose as the third dose) would be required.

      Results

      Between September 24 and October 14, 2021, 1243 of 1250 screened individuals were recruited (Figure 1), randomized (AZHD n = 312, AZFD n = 307, PFHD n = 316, and PFFD n = 308; Table 1), administered the planned single dose of the vaccine and included in the safety analysis. Eight participants (0.6%) discontinued (AZHD n = 2, AZFD n = 3, PFHD n = 3, and PFFD n = 0), most commonly due to ‘other: subject inconvenient to come into site’ (n = 4). One participant was excluded due to developing COVID-19 (onset day 80). After excluding equal numbers of participants (n = 5) in the AZD1222 and BNT162b2 groups, 309 AZHD, 305 AZFD, 312 PFHD, and 307 PFFD participants were included in the per-protocol (PP) set. Exclusions were due to receipt of fourth vaccines (n = 6), failed eligibility criteria (n = 2), and SARS-CoV-2 infection (n = 2). Based on full set data (ITT), the population had a mean age of 41 years and 57% were female. Participants were 60-<90 days (33%), 90-<120 days (34%), or 120-180 days (33%) postfinal CoronaVac dose.
      Table 1Participant demographics of (a) per-protocol population, including their (b) medical history, and (c) anti-SARS-CoV-2 nucleocapsid antibody levels at baseline (intent-to-treat analysis).
      a. Patient demographics
      ChAdOX1-SBNT162b2
      Half-dose (N = 312)Full-dose (N = 307)Half-dose (N = 316)Full-dose (N = 308)All subjects (N = 1243)
      Age (years)
      n (missing)312 (0)307 (0)316 (0)308 (0)1241 (0)
      Mean (SD)41.7 (9.6)41.0 (9.6)40.5 (9.8)41.9 (9.8)41.3 (9.7)
      Median42.041.041.042.041.0
      Age Categories, n (%)
      n (missing)312 (0)307(0)316(0)308(0)1243(0)
      20-59305 (97.8)301 (98.0)315 (99.7)304 (98.7)1225(98.6)
      >=607 (2.2)6 (2.0)1 (0.3)4 (1.3)18(1.4)
      Sex, n (%)
      n (missing)312 (0)307(0)316(0)308(0)1243(0)
      Male118 (37.8)139 (45.3)133 (42.1)141 (45.8)531 (42.7)
      Female194 (62.2)168 (54.7)183 (57.9)167 (54.2)712 (57.3)
      b. Medical history
      ChAdOX1-SBNT162b2All subjects (N = 1243) n(%)
      System organ class preferred termHalf-dose (N = 312) n(%)Full-dose (N = 307) n(%)Half-dose (N = 316) n(%)
      Full-dose (N = 308) n(%)
      Metabolism and nutrition disorders34(10.9)27(8.8)32(10.1)27(8.8)120(9.7)
      Vascular disorders25(8.0)27(8.8)24(7.6)28(9.1)104 (8.4)
      Respiratory, thoracic and mediastinal disorders9(2.9)12(3.9)12(3.8)10(3.2)43 (3.5)
      Nervous system disorders11(3.5)8(2.6)9(2.8)8(2.6)36(2.9)
      c. Baseline anti-SARS-CoV-2 nucleocapsid antibody
      Baseline anti-nucleocapsid
      n (missing)46 (264)47 (260)48 (268)44 (264)185 (1056)
      Mean (SD)12.03 (40.19)12.22 (31.56)8.48 (28.50)5.41 (16.04)9.58 (30.32)
      Median1.81.51.61.31.6
      Unsolicited AEs based on ITT occurred in 38% of participants, including vascular (hypertension and tachycardia), cardiac disorders, and general and administration site conditions (pyrexia). Solicited AE occurred in 84% of participants (Figure 2), including local and systemic reactions (injection site pain, musculoskeletal and connective tissue disorders [fatigue, myalgia, arthralgia]), and nervous system disorders [headache]). After AZHD, three participants experienced unsolicited SAEs—appendicitis, SARS-CoV-2 infection and hemorrhoids—but all were unrelated to the vaccine and all participants recovered. A total of 19 participants (five solicited and 24 unsolicited events) experienced medically attended AEs. Two participants had an AESI SARS-CoV-2 infection after AZHD or after PFHD and both recovered, but all AESIs were unsolicited and unrelated to the vaccine. One PFHD recipient withdrew due to moderate AEs. Neither dose of AZD1222 or BNT2162b2 was directly related to an AE or led to reports of thrombosis with thrombocytopenia syndrome or myocarditis.
      Figure 2
      Figure 2Solicited treatment-emergent adverse events (intent-to-treat analysis).
      Abbreviations: AZHD, ChAdOx1-S half-dose; AZFD, ChAdOx1-S full-dose; PFHD, BNT162b2 half-dose; PFFD, BNT162b2 full-dose.
      The COVID-19 infection cases and those who had received additional dose of COVID-19 vaccine were excluded from PP analysis (Figure 1). Together with given negligible difference (0.8%) in participants in the full and PP analysis, the immune responses to the PP set were reported (ITT analysis results are shown in the Supplementary).
      The noninferiority criteria were met (Figure 3) across all three intervals in the comparisons of full-dose to half-dose vaccinations in four immunogenicity assays and maintained at each interval. To summarize, more than 97% seroconversion rates (≥4-fold GMFR within vaccine types using anti-S IgG) were observed at 28, 60, and 90 days after vaccination, regardless of dose (Figure 3 and Table 2a). The GMC ratio for anti-S was also noninferior for AZHD versus AZFD and PFHD versus PFFD (Table 2b).
      Figure 3
      Figure 3Noninferiority comparisons for all immunogenicity assays by vaccine group. The forest plots of point estimate and 95% confidence interval of all comparisons were between either AZHD and AZFD; or between PFHD and PFFD and stratified by interval after booster (third) vaccination in the per-protocol analysis group for (a) anti-SARS-Cov-2 spike-RBD antibody (shown as GMC ratio), (b) pseudovirus neutralization antibody titer (shown as GMT ratio) against Ancestral SARS-CoV-2 (left) or Delta (B1.617.2) variant (right), and (c) microneutralization (shown as GMT ratio) against Ancestral SARS-CoV-2 (left) or Delta (B.1.617.2) (right).
      Abbreviations: AZHD, ChAdOx1-S half-dose; AZFD, ChAdOx1-S full-dose; GMC, geometric mean concentration; GMT, geometric mean titer; PFHD, BNT162b2 half-dose; PFFD, BNT162b2 full-dose; RBD, receptor-binding domain.
      Table 2aImmunogenicity according to overall antispike receptor-binding domain immunoglobulin G activity.
      Analysis visitStatisticsAZHDAZFDPFHDPFFD
      PE95% CIPE95% CIPE95% CIPE95% CI
      BaselineN310307316308
      GMC (U/ml)48.543.0, 54.743.738.5,49.645.740.3, 51.944.839.8, 50.4
      Day 28N309307315308
      GMC (U/ml)8237.07679.3, 8835.28973.68328.1, 9669.214073.913,236.4, 14,964.415,920.715,135.7, 16,746.4
      GMFR166.9147.8, 188.5205.3181.8, 231.9308.2274.7, 345.7355.5316.0, 400.0
      ≥4-fold GMFR (%)99.497.7, 99.999.097.2, 99.899.497.7, 99.999.798.2, 100.0
      Day 60N308307315307
      GMC (U/ml)5088.34718.0, 5487.85496.25074.3, 5953.18150.67584.3, 8759.39340.68772.6, 9945.5
      GMFR103.090.9, 116.7125.7111.0, 142.4177.8157.9, 200.3207.8183.9, 234.8
      ≥4-fold GMFR (%)99.097.2, 99.899.097.2, 99.899.097.2, 99.899.097.2, 99.8
      Day 90N307303312307
      GMC (U/ml)3435.33166.7, 3726.63651.33355.8, 3972.85097.24716.2, 5508.95936.15540.1, 6360.4
      GMFR69.260.9, 78.683.873.8, 95.3111.598.4, 126.4132.1116.3, 149.9
      ≥4-fold GMFR (%)98.496.2, 99.599.097.1, 99.899.097.2, 99.899.097.2, 99.8
      Abbreviations: AZHD, ChAdOx1-S half-dose; AZFD, ChAdOx1-S full-dose; CI, confidence interval; GMC, geometric means concentration; GMFR, geometric mean fold rise; PE, point estimate; PFHD, BNT162b2 half-dose; PFFD, BNT162b2 full-dose.
      Immunogenicity was measured in terms of GMC (U/ml), GMFR and percentage of participants with a more than four-fold rise in GMFR.
      Table 2bSummary of geometric mean concentration ratios of anti-SARS-CoV2 spike-RBD antibody by vaccine group.
      Geometric mean concentration ratio (95% confidence interval)
      Anti-SARS-CoV2 spike-RBD antibodyInterval duration(days)ChAdOx1-S half-dose / ChAdOx1-S full-doseBNT162b2 half-dose / BNT162b2 full-dose
      Baseline (D1)Overall0.9 (0.8-1.1)1.0 (0.8-1.2)
      60 to <900.7 (0.6-1.0)0.9 (0.7-1.2)
      90 to < 1201.0 (0.8-1.4)1.0 (0.7-1.4)
      120 to 1801.0 (0.7-1.3)1.0 (0.7-1.4)
      Overall0.92 (0.83-1.02)0.88 (0.82-0.96)
      28 days after the third vaccination60 to <901.01 (0.85-1.20)0.91 (0.79-1.05)
      90 to < 1200.92 (0.79-1.08)0.88 (0.77-1.01)
      120 to 1800.84 (0.71-1.00)0.86 (0.77-0.98)
      Overall0.93 (0.83-1.03)0.87 (0.79-0.96)
      60 days after the third vaccination60 to <901.04 (0.87-1.25)0.91 (0.78-1.07)
      90 to < 1200.94 (0.79-1.11)0.88 (0.75-1.04)
      120 to 1800.83 (0.69-0.99)0.82 (0.71-0.96)
      Overall0.94 (0.84-1.06)0.86 (0.77-0.95)
      90 days after the third vaccination60 to <901.09 (0.89-1.32)0.90 (0.75-1.07)
      90 to < 1200.95 (0.78-1.15)0.87 (0.73-1.03)
      120 to 1800.82 (0.68-1.00)0.81 (0.68 - 0.96)
      Abbreviation: RBD, receptor-binding domain.
      Overall geometric mean titer ratios of Anti-SARS-CoV2 spike-RBD antibody was stratified by vaccine group and interval after booster (third) vaccination in the per-protocol analysis group.
      Pseudovirus neutralizing antibodies demonstrated more than 90% seroconversion rates (≥4-fold rise). Moreover, the responses across the dose groups were of similar magnitudes between the two vaccine platforms at 28 days after vaccination and against the Ancestral, Delta, and Omicron strains (Table 3a, Figure 4). However, GMT and GMFR declined against Delta and Omicron as detailed in the next sections (Table 3a). Regardless, noninferiority in terms of nAb was met for AZHD versus AZFD and for PFHD versus PFFD (Table 3b). Importantly, a longer interval between CoronaVac primary series and the booster (third) dose (‘prime-boost’) improved the immunogenicity. Boosting at 120-180 days substantially improved GMC, as measured by anti-S IgG, and GMT, as measured by PVNT50, especially with the full doses of both platforms against each strain (P <0.001; Table 2a-2b and Figure 4A).
      Table 3aImmunogenicity according to pseudovirus neutralizing antibody titer against different variants of concern.
      Analysis visitStatisticsChAdOx1-S half-doseChAdOx1-S full-doseBNT162b2 half-doseBNT162b2 full-dose
      PE95% CIPE95% CIPE95% CIPE95% CI
      Ancestral SARS-CoV-2
      BaselineN1009710099
      GMT24.221.5-27.224.621.5-28.227.422.8-33.125.822.9-29.0
      Day 28N1009710099
      GMT653.2556.9-766.3712.2607.0-835.61219.31048.7-1417.61362.71207.6-1537.9
      GMFR27.022.5-32.428.924.2-34.744.436.5-54.152.945.0-62.1
      ≥4-fold GMFR (%)97.194.5-98.797.795.4-99.1100.098.8-100.0100.098.8-100.0
      Day 90N100979899
      GMT407.0335.6-493.6406.1333.9-493.9539.6438.9-663.4575.3494.1-669.8
      GMFR16.813.7-20.716.513.2-20.720.416.3-25.522.318.6-26.8
      ≥4-fold GMFR (%)91.988.2-94.794.791.6-97.098.195.9-99.398.796.7-99.6
      Delta (B.1.617.2)
      BaselineN190187213202
      GMT22.120.5-23.723.821.6-26.322.820.9-24.921.720.7-22.9
      Day 28N190187213202
      GMT468.3409.8-535.2530.6470.4-598.5801.5715.4-897.8856.1777.8-942.1
      GMFR21.218.3-24.622.319.3-25.835.130.8-40.039.435.4-43.8
      ≥4-fold GMFR (%)90.686.8-93.694.190.9-96.598.195.9-99.399.497.7-99.9
      Day 90N190186210202
      GMT175.1152.6-200.9184.1160.9-210.8221.9194.4-253.3225.4200.9-252.9
      GMFR7.96.9-9.27.76.6-9.09.88.5-11.210.49.2-11.7
      ≥4-fold GMFR (%)77.572.4-86.182.277.4-86.386.982.6-90.491.287.5-94.1
      Omicron (B.1.1.529)
      BaselineN238238
      GMT1.61.4-1.81.51.3-1.8
      Day 28N238238
      GMT118.997.3-145.2255.9222.9-293.7
      GMFR76.159.4-97.6167.1137.7-202.7
      ≥4-fold GMFR (%)90.385.9-93.897.194.0-98.8
      Abbreviations: CI, confidence interval; GMFR, geometric mean fold rise; GMT, geometric means titer; PE, point estimate.
      Immunogenicity was measured in terms of GMT, GMFR and percentage of participants with a more than four-fold rise in GMFR.
      Figure 4
      Figure 4Vaccine-induced immune responses. GMT of (A) pseudo and (B) microneutralization antibody assays are stratified by variant after participant completed primary CoronaVac vaccinations and before administration of boosters.
      GMT, geometric mean titers.
      Table 3bSummary of GMT ratios of PVNT by vaccine group.
      Timepoint of PVNT50 measurement against SARS-CoV-2Interval duration (days)GMT Ratio (95% confidence interval)
      Ancestral SARS-CoV-2Delta (B.1.617.2) SARS-CoV-2
      AZHD/AZFDPFHD/PFFDAZHD/AZFDPFHD/PFFD
      Overall0.73 (0.45-1.18)1.25 (0.74-2.10)0.79 (0.55-1.13)1.04 (0.75-1.45)
      Baseline (D1)60 to <901.12 (0.46-2.69)2.88 (1.12-7.36)1.55 (0.81-2.98)1.40 (0.73-2.69)
      90 to < 1200.33 (0.15-0.71)1.41 (0.58-3.46)0.54 (0.29-1.02)0.76 (0.44-1.34)
      120 to 1801.05 (0.44-2.49)0.47 (0.20-1.10)0.69 (0.39-1.24)1.11 (0.66-1.88)
      Overall0.92 (0.73-1.15)0.89 (0.74-1.08)0.88 (0.74-1.06)0.94 (0.81-1.09)
      28 days after the third vaccination60 to <901.22 (0.78-1.89)1.15 (0.8-1.64)1.16 (0.80-1.68)0.94 (0.71-1.23)
      90 to < 1200.76 (0.52-1.11)1.01 (0.73-1.41)0.80 (0.59-1.07)1.01 (0.79-1.30)
      120 to 1800.84 (0.59-1.20)0.62 (0.45-0.84)0.81 (0.62-1.06)0.87 (0.68-1.11)
      Overall1.00 (0.76-1.32)0.94 (0.73-1.21)0.95 (0.78, 1.15)0.98 (0.83, 1.17)
      90 days after the third vaccination60 to <901.40 (0.84-2.34)1.01 (0.62-1.64)1.25 (0.86-1.82)1.10 (0.80-1.51)
      90 to < 1200.71 (0.43-1.17)1.07 (0.68-1.70)0.87 (0.61-1.25)1.01 (0.75-1.35)
      120 to 1801.02 (0.69-1.53)0.76 (0.53-1.10)0.86 (0.65-1.13)0.88 (0.66-1.19)
      Abbreviations: AZHD, ChAdOx1-S half-dose; AZFD, ChAdOx1-S full-dose; GMT, geometric mean titer; PFHD, BNT162b2 half-dose; PFFD, BNT162b2 full-dose; PVNT, pseudovirus neutralizing antibody titers.
      Overall GMTs of PVNT50 was stratified by vaccine group and interval after booster (third) vaccination in the per-protocol analysis group.
      Per-protocol and intent-to-treat showed similar results.
      The GMFR (%) for AZHD was 166.9, 103, and 69.2 at days 28, 60, and 90, respectively, versus 205.3, 125.7, and 83.8 for AZFD at the same respective visits. The GMFR for PFHD was 308.2, 177.8, and 111.5 at days 28, 60, and 90, respectively, versus 355.5, 207.8, and 132.1 for PZFD at the same respective visits. Anti-S RBD IgG GMC PE peaked at day 28 for all doses and vaccine platforms (data not shown; AZHD PE: 8237.0, 95% CI: 7679.3-8835.2; AZFD PE: 8973.6, 95% CI: 8328.1-9669.2; PFHD PE: 14,073.9, 95% CI: 13,236.4-14,964.4; and PFFD PE: 15,920.7, 95% CI: 15,135.7-16,746.4). GMC PE was still high and comparable at day 90 between AZHD and AZFD (3435.3, 95% CI: 3166.7-3726.6, and 3651.3, 95% CI: 3355.8-3972.8) and between PFHD and PFFD (5097.2, 95% CI: 4716.2-5508.9, and 5936.1; 95% CI: 5540.1-6360.4).
      At day 28 and 90, the baseline PVNT50 GMT against SARS-CoV-2 Delta and Ancestral strains (Table 3a) was similar between vaccines. However, at day 90, AZHD and AZFD had similar GMT against the Ancestral strain. Regardless of strain, PVNT50 GMT peaked at day 28. At days 28 and 90, AZHD PVNT50 GMT against Delta was noninferior to AZFD, whereas PFHD was noninferior to PFFD (Table 3b). Similarly, AZHD PVNT50 GMT against the Ancestral strain was noninferior to AZFD, whereas PFHD was noninferior to PFFD. Against Delta, the PVNT50 GMT PE for AZHD was 468.3 (95% CI: 409.8-535.2) and 175.1 (95% CI: 152.6-200.9) versus 530.6 (95% CI: 470.4-598.5) and 184.1 (95% CI: 160.8-210.7) for AZFD. The PVNT50 GMT for PFHD was 801.5 (95% CI: 715.4-897.8) and 221.9 (95% CI: 194.4-253.3) versus 856.1 (95% CI: 777.8-942.1) and 225.4 (95% CI: 200.9-252.9) for PFFD. Against the Ancestral strain, PVNT50 GMT for AZHD was 653.2 (95% CI: 556.88-766.28) and 407.0 (95% CI: 335.6-493.6) at days 28 and 90, respectively, versus 712.2 (95% CI: 607.0-835.6) and 406.1 (95% CI: 333.9-493.9) for AZFD. The PVNT50 GMT for PFHD was 1219.3 (95% CI: 1048.7-1417.6) and 539.6 (95% CI: 438.97-663.4) versus 1362.7 (95% CI: 1207.68-1537.9) and 575.3 (95% CI: 494.1-669.8) for PFFD. Strain-specific PVNT50 GMT ratios (Table 3b, Figure 4A) also showed an improved immunogenicity with longer prime-boost intervals (P <0.001). Against Delta, the AZHD/AZFD GMT ratio was high and increased at days 28 and 90, but the PFHD/PFFD GMT ratio, although high, was relatively unchanged between day 28 (0.9 [0.8-1.1]) and 90 (1.0 [0.8-1.2]). Against the Ancestral strain, the GMT ratios for both AZHD/AZFD and PFHD/PFFD were high and increased at days 28 and 90. When PVNT50 was stratified by boosting interval (Figure 4A, a-d), greater differences were seen between vaccine types at the 60-<90-day interval, with all comparisons inferior for Delta. At the 90-<120-day interval, all comparisons were noninferior, whereas at the 120-180-days interval, most comparisons were noninferior. Overall, the Ancestral strain prompted higher seroresponse rates than the Delta strain. Also, seroresponses against Delta were noninferior at day 28 between dose levels of either vaccine, but the seroresponses were only noninferior between PFHD and PFFD at day 90. For the Ancestral strain, noninferiority was observed at all time points and within dose levels for each vaccine.
      Full-dose vaccine microNT and PVNT50 were consistent across postbaseline visits for the Delta and the Ancestral strain, and noninferiority was observed at all intervals when stratified by variant (Figure 4A and 4B, a-d). At day 28, many participants also had ≥4-fold increase in microNT titers to both the Delta and the Ancestral strain, with all doses and vaccines. The microNT NT50 against both the Delta and the Ancestral strains peaked at day 28 versus day 90 (Table 4a, 4b). Against Delta, the microNT NT50 GMT PE for AZHD was 172.7 (95% CI: 142.3-209.5) versus 196.8 (95% CI: 163.2-237.4) for AZFD at day 28 (P = 0.336), and 79.5 (95% CI: 65.6-96.2) versus 84.1 (95% CI: 69.0-102.5), respectively, at day 90 (P = 0.681). Similarly, the microNT NT50 GMT for PFHD was 331.3 (95% CI: 274.5-399.8) versus 311.2 (95% CI: 265.7-364.4) for PFFD at day 28 (P = 0.613) and 103.9 (95% CI: 84.9-127.2) versus 104.4 (95% CI: 87.5-124.5), respectively, at day 90 (P = 0.974). Similar trends were observed with the Ancestral strain: the microNT NT50 GMT for AZHD was 519.8 (95% CI: 435.7-620.2) versus 575.0 (95% CI: 487.3-678.4) for AZFD (P = 0.410) at day 28 and 278.6 (95% CI: 231.0-336.0) versus 308.8 (95% CI: 254.4-374.8), respectively, at day 90 (P = 0.450). Also, the microNT NT50 GMT for PFHD was 930.5 (95% CI: 780.1-1110.1) versus 927.6 (95% CI: 793.7-1084.0) for PFFD at day 28 (P = 0.978) and 348.35 (95% CI: 285.5-4254.0) versus 363.0 (95% CI: 306.5-429.9), respectively, at day 90 (P = 0.755). All intervals between doses of each vaccine were also noninferior between the Delta and the Ancestral strain.
      Table 4aImmunogenicity according to micro neutralizing antibody titer against different variants of concern.
      Analysis visitStatisticsChAdOx1-S half-doseChAdOx1-S full-doseBNT162b2 half-doseBNT162b2 full-dose
      PE95% CIPE95% CIPE95% CIPE95% CI
      Ancestral SARS-CoV-2
      BaselineN1009710099
      GMT23.519.9-27.722.118.9-25.922.218.6-26.521.918.4-26.2
      Day 28N1009710099
      GMT519.8435.7-620.2575.0487.3-678.4930.5780.1-1110.0927.6793.7-1084.0
      GMFR22.217.9-27.426.021.3-31.841.933.9-51.842.333.9-52.9
      ≥4-fold GMFR (%)98.093.0-99.8100.0096.3-100.099.094.6-100.099.094.5-100.0
      Day 90N100979899
      GMT278.6231.0-336.0308.8254.4-374.8348.4285.5-425.0363.0306.5-429.9
      GMFR11.99.4-15.014.011.1-17.715.712.4-19.816.613.0-21.1
      ≥4-fold GMFR (%)93.086.1-97.192.885.7-97.091.884.5-96.494.988.6-98.3
      Delta (B.1.617.2)
      BaselineN1009710099
      GMT12.111.0-13.413.011.5-14.812.911.4-14.711.710.6-12.8
      Day 28N1009710099
      GMT172.7142.3-209.5196.8163.2-237.4331.3274.5-399.8311.2265.7-364.4
      GMFR14.211.7-17.315.112.2-18.725.620.9-31.526.722.1-32.2
      ≥4-fold GMFR (%)95.088.7-98.492.885.7-97.196.090.1-98.9100.096.3-100.0
      Day 90N100979899
      GMT79.565.6-96.284.169.0-102.5103.984.9-127.2104.487.5-124.5
      GMFR6.55.4-8.06.55.1-8.18.06.5-9.99.07.4-10.9
      ≥4-fold GMFR (%)81.071.9-88.279.470.0-86.981.672.5-88.788.981.0-94.3
      Abbreviations: CI, confidence interval; GMFR, geometric mean fold rise; GMT, geometric means titer; PE, point estimate.
      Immunogenicity was measured in terms of GMT, GMFR and percentage of participants with a more than four-fold rise in GMFR.
      Per-protocol and intent-to-treat showed similar results.
      Table 4bSummary of GMT ratios of microNT by vaccine group.
      Timepoint of microNT measurement against SARS-CoV-2Interval duration (days)GMT Ratio (95% confidence interval)
      Ancestral SARS-CoV-2Delta (B.1.617.2) SARS-CoV-2
      AZHD/AZFDPFHD/PFFDAZHD/AZFDPFHD/PFFD
      Overall1.06 (0.84-1.33)1.01 (0.79-1.30)0.93 (0.79-1.10)1.11 (0.95-1.29)
      Baseline (D1)60 to <901.20 (0.79-1.81)1.17 (0.76-1.79)0.95 (0.67-1.34)1.22 (0.92-1.62)
      90 to < 1200.90 (0.60-1.34)0.92 (0.54-1.55)0.93 (0.72-1.20)0.96 (0.68-1.35)
      120 to 1801.11 (0.77-1.60)0.96 (0.71-1.29)0.92 (0.73-1.16)1.16 (0.99-1.36)
      Overall0.90 (0.71-1.15)1.00 (0.79-1.27)0.88 (0.67-1.15)1.06 (0.83-1.36)
      28 days after the third vaccination60 to <901.25 (0.78-2.02)0.97 (0.63-1.51)1.27 (0.75-2.12)1.11 (0.69-1.80)
      90 to < 1200.75 (0.50-1.13)1.21 (0.83-1.77)0.68 (0.43-1.07)1.02 (0.69-1.50)
      120 to 1800.79 (0.56-1.12)0.86 (0.59-1.27)0.79 (0.52-1.21)1.07 (0.71-1.59)
      Overall0.90 (0.69-1.18)0.96 (0.74-1.24)0.94 (0.72-1.24)1.00 (0.76-1.30)
      90 days after the third vaccination60 to <901.12 (0.66-1.91)0.99 (0.61-1.62)1.41 (0.86-2.31)1.05 (0.66-1.66)
      90 to < 1200.69 (0.44-1.08)1.18 (0.74-1.88)0.76 (0.46-1.23)1.13 (0.71-1.82)
      120 to 1800.96 (0.65-1.42)0.75 (0.53-1.05)0.79 (0.51-1.24)0.83 (0.53-1.28)
      Abbreviations: AZHD, ChAdOx1-S half-dose; AZFD, ChAdOx1-S full-dose; GMT, geometric mean titer; microNT, microneutralization antibody titers; PFHD, BNT162b2 half-dose; PFFD, BNT162b2 full-dose.
      Overall GMT ratios of microNT were stratified by vaccine group and interval after booster (third) vaccination in the per-protocol analysis group.
      Per-protocol and intent-to-treat showed similar results.
      Day 28 T cell levels (Table 5) were higher with BNT162b2 than AZD1222 (P <0.01) and higher with PFFD than PFHD (P = 0.022) but were similar, regardless of AZD1222 dose (P = 0.642) or interval duration. Thus, AZFD was noninferior to AZHD, and PFFD was noninferior to PFHD.
      Table 5SARS-CoV-2 reactive T cell concentration interferon-gamma assay.
      Analysis visitChAdOx1-S half-dose N = 222ChAdOx1-S full-dose N = 231BNT162b2 half-dose N = 223BNT162b2 full-dose N = 224
      MedianQ1-Q3MedianQ1-Q3MedianQ1-Q3MedianQ1-Q3
      Baseline257.5111.1-623.5228.0110.4-486.0238.198.7-553.1250.6114.7-572.7
      Day 281722.4838.3-3245.21620.0968.6-3305.12866.91391.6-6398.13487.71853.1-6835.7
      Results are presented in units of mIU/ml.
      Q1, 1st quartile, Q3, 3rd quartile.
      Against Omicron, the day-28 PVNT50 GMT increased with longer booster intervals (Figure 4A) for AZFD and PFFD. A 60-<90-day interval increased the PVNT50 GMT PE from 1.4 (95% CI: 1.1-1.8) with AZFD and 1.6 (95% CI: 1.3-2.0) with PFFD to 58.00 (95% CI: 37.7-89.3) and 195.3 (95% CI: 140.4-271.6), respectively. A 90-<120-day interval increased the PVNT50 GMT from 1.6 (95% CI: 1.3273-2.1) with AZFD and 1.46 (95% CI: 1.2-1.8) with PFFD to 1260 (95% CI: 93.9-169.0) and 221.6 (95% CI: 184.0-2676.0), respectively. The 120-180-day interval increased the PVNT50 GMT from 1.6 (95% CI: 1.3-2.18) with AZFD and 1.5 (95% CI: 1.2-1.9) with PFFD to 195.3 (95% CI: 145.0-263.0) and 370.6 (95% CI: 307.2-447.1), respectively.

      Discussion

      This randomized, double-blind study compared the safety and immunogenicity induced by boosting with an additional (third) AZHD to AZFD and PFHD to PFFD and between three extended postprimary vaccination intervals. Comparisons were not conducted between vaccine platforms (AZD1222 vs BNT162b2). Instead, our intraplatform comparisons differentiated between dose, boosting intervals, and virus variants, with sufficient statistical power to detect noninferiority between different doses of the same vaccine. We did so to understand if and how to maintain high levels of immunogenicity when vaccine doses were limited.
      Although our safety assessments were limited by small cohort sizes, all vaccine doses were found to be very safe, with no vaccine-related or life-threatening AE occurring at any dose. Thrombosis with thrombocytopenia syndrome, although rare, has been documented with AZD1222 [

      European Medicines Agency. European Union risk management plan AstraZeneca. p. ChAdOx1-S (recombinant) (AZD1222), https://www.ema.europa.eu/en/documents/rmp-summary/vaxzevria-previously-covid-19-vaccine-astrazeneca-epar-risk-management-plan_en.pdf; 2021 [accessed 25 August 2022].

      ]. However, no thrombosis with thrombocytopenia syndrome occurred in our study with either AZHD or AZFD. Moreover, no previously reported BNT162b2 AESI or SAE occurred in our study [

      European Medicines Agency. BNT162b2 risk management plan, https://www.ema.europa.eu/en/documents/rmp-summary/comirnaty-epar-risk-management-plan_en.pdf; 2022 [accessed 25 August 2022].

      ].
      Importantly, AZHD or PFHD were immunologically noninferior to the corresponding full-dose in anti-S IgG assays, PNA, and microNT assays of humoral immune responses and S-protein-specific interferon-γ assays for T cell response. The noninferiority between doses of any vaccine type persisted at all intervals, even 90-120 days after the primary series of CoronaVac. Although seroconversion was high at all doses, it declined slightly with longer intervals in half-dose recipients and slower in full-dose recipients. Specifically, at day 90, the half-dose immune response waned more than the full-dose immune response; nevertheless, the half-dose immune response at day 28 remained robust and noninferior. Reassuringly, a high and comparable proportion of participants seroconverted after the different intervals, with increased seroconversion and T cell induction rates at longer intervals, even at 120 days after the primary series. Also, our full-dose vaccination induced similar immune responses to another Thai study that had a smaller cohort [

      Angkasekwinai N, Niyomnaitham S, Sewatanon J, Phumiamorn S, Sukapirom K, Senawong S, et al. The immunogenicity against variants of concern and reactogenicity of four COVID-19 booster vaccinations following CoronaVac or ChAdOx1 nCoV-19 primary series. medRxiv. 19 January 2022. https://www.medrxiv.org/content/10.1101/2021.11.29.21266947v2 [accessed 8 September 2022].

      ]. A vaccination strategy which allows lower doses can potentially mitigate supply constraints, especially in resource-limited settings facing an urgent need to ensure high vaccination coverage. We showed in a large adult population that AZHD boosting was noninferior to AZFD boosting and PFHD boosting was noninferior to PFFD boosting in terms of eliciting high immunogenicity. Half-dose vaccinations were not evaluated for Omicron because it is associated with immune escape and must be overcome by full doses [
      • Khan K
      • Karim F
      • Ganga Y
      • Bernstein M
      • Jule Z
      • Reedoy K
      • et al.
      Omicron BA.4/BA.5 escape neutralizing immunity by BA.1 infection.
      ]. Notably, virus neutralization declined gradually from the Ancestral strain to the Delta variant and more markedly from these strains to Omicron. Regardless, neutralization remained high, even with extended intervals between the primary and booster doses, and as recently shown, nAb titers may correlate with protection against infection [

      Khoury DS, Schlub TE, Cromer D, Steain M, Fong Y, Gilbert PB, et al. Correlates of protection, thresholds of protection, and immunobridging in SARS-CoV-2 infection. medRxiv. 06 June 2022. https://www.medrxiv.org/content/10.1101/2022.06.05.22275943v1 [accessed 08 September 2022].

      ]. Halving doses did not compromise immune responses or safety and could be an important strategy for ensuring population coverage, especially when boosting. Now, a main vaccine provider, Moderna Inc., is manufacturing 50-µg doses of mRNA-1273 and will produce bivalent boosters with 25 µg of each antigen [

      United States Food and Drug Administration. mRNA-1273.214 Moderna COVID-19 investigational bivalent vaccine (original + Omicron) Moderna, Inc. Vaccines and Related Biological Products Advisory Committee Meeting, https://www.fda.gov/media/159492/download; 2022 [accessed 08 September 2022].

      ].
      Low or inconsistent vaccine supplies can be circumvented with heterologous regimens. mRNA-1273, BNT162b2, and AZD1222 vaccine combinations are safe, well tolerated, and comparably or even more immunogenic than homologous regimens. A strategy that combines fractional dosing, dose stretching, and heterologous schedules can prevent vaccination campaign disruptions even if vaccine logistics or safety are problematic. Future investigations should include age-based stratifications and comparative durations of humoral immunity between fractional and full doses. However, with the current spread of Omicron worldwide, provision of AZFD or PFFD boosters is consistent with recent effectiveness data [
      • Suphanchaimat R
      • Nittayasoot N
      • Jiraphongsa C
      • Thammawijaya P
      • Bumrungwong P
      • Tulyathan A
      • et al.
      Real-world effectiveness of mix-and-match vaccine regimens against SARS-CoV-2 Delta variant in Thailand: a nationwide test-negative matched case-control study.
      ].

      Conclusion

      We found similar immune responses after day 28 when boosting using any of the three progressively longer intervals after the primary series, and that a third dose given at a longer interval was optimal. Halving doses does not compromise immunogenicity or safety and can broaden vaccine coverage when vaccine supply is a problem.

      Declaration of competing interest

      The authors have no competing interests to declare.

      Funding

      Funding was provided by the Program Management Unit for Competitiveness Enhancement, C17F640221 National research, National Higher Education, Science, Research and Innovation Policy Council, Thailand through Clinixir Ltd. The Program Management Unit for Competitiveness Enhancement funded the trial but had no other involvement in the study. Clinixir provided logistical and monitoring supports.

      Ethical approval

      The protocol (ClinicalTrials.gov Identifier: NCT05049226) has received approval from central and each institutional ethical review boards. All participants had signed written informed consent forms before enrolling into the study.

      Acknowledgments

      The authors wish to thank the staff of all seven medical schools: Siriraj Institute of Clinical Research, Faculty of Medicine, Mahidol University, Prof. Dr. Kulkanya Chokephaibulkit, Patimaporn Wongprompitak; The Academic Clinical Research Office, Faculty of Medicine, Khon Kaen University, Dr. Apichart So-ngern, Dr. Manchumad Manjavong, Dr. Worawat Chumpangern, and Dr. Kwanchanok Yimtae; Clinical Research Center, Faculty of Medicine, Prince of Songkla University, Dr. Sarunyou Chusri and Dr. Smonrapat Surasombatpattana; Clinical Research Center, Faculty of Medicine, Thammasart University; Maha Chakri Sirindhorn Clinical Research Center Under the Royal Patronage, Faculty of Medicine, Chulalongkorn University; Chakri Naruebodindra Medical Institute, Dr. Sirawat Srichatrapimuk and Prof. Dr. Somnuek Sungkanuparph; CM Clinical Trial Unit Faculty of Medicine, Chiang Mai University, Dr. Parichat Salee, Dr. Poramed Winichakoon, and Dr. Nattinee Laksananun; and Clinixir Co., Ltd., Dr. Kanokwan Pornprasit. The authors’ grateful thanks are also extended to National Center for Genetic Engineering and Biotechnology and Faculty of Science, Mahidol University. The authors would like to thank Shawna Tan of Medical Writers Asia for assisting in drafting manuscript. Finally, the authors would like to express their very great appreciation to Prof. Kanta Subbarao, Dr. Jean-Louis Excler, and Assoc. Prof. Jaranit Kaewkungwal who have served as the data and safety monitoring board.

      Author contributions

      All authors were involved in the study design; in the interpretation of data; in reviewing the manuscript; and in the decision to submit the manuscript for publication. All authors had full access to all of the data in the study and can take responsibility for the integrity of the data and the accuracy of the data analysis.

      Appendix. Supplementary materials

      References

        • Patel MK
        • Bergeri I
        • Bresee JS
        • Cowling BJ
        • Crowcroft NS
        • Fahmy K
        • et al.
        Evaluation of post-introduction COVID-19 vaccine effectiveness: summary of interim guidance of the World Health Organization.
        Vaccine. 2021; 39: 4013-4024https://doi.org/10.1016/j.vaccine.2021.05.099
        • Khoury DS
        • Cromer D
        • Reynaldi A
        • Schlub TE
        • Wheatley AK
        • Juno JA
        • et al.
        Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection.
        Nat Med. 2021; 27: 1205-1211https://doi.org/10.1038/s41591-021-01377-8
        • Zhang Y
        • Zeng G
        • Pan H
        • Li C
        • Hu Y
        • Chu K
        • et al.
        Safety, tolerability, and immunogenicity of an inactivated SARS-CoV-2 vaccine in healthy adults aged 18–59 years: a randomised, double-blind, placebo-controlled, phase 1/2 clinical trial.
        Lancet Infect Dis. 2021; 21: 181-192https://doi.org/10.1016/S1473-3099(20)30843-4
        • Jara A
        • Undurraga EA
        • González C
        • Paredes F
        • Fontecilla T
        • Jara G
        • et al.
        Effectiveness of an inactivated SARS-CoV-2 vaccine in Chile.
        N Engl J Med. 2021; 385: 875-884https://doi.org/10.1056/NEJMoa2107715
        • Zeng G
        • Wu Q
        • Pan H
        • Li M
        • Yang J
        • Wang L
        • et al.
        Immunogenicity and safety of a third dose of CoronaVac, and immune persistence of a two-dose schedule, in healthy adults: interim results from two single-centre, double-blind, randomised, placebo-controlled phase 2 clinical trials.
        Lancet Infect Dis. 2022; 22: 483-495https://doi.org/10.1016/S1473-3099(21)00681-2
        • Cowling BJ
        • Lim WW
        • Cobey S.
        Fractionation of COVID-19 vaccine doses could extend limited supplies and reduce mortality.
        Nat Med. 2021; 27: 1321-1323https://doi.org/10.1038/s41591-021-01440-4
      1. World Health Organization. Interim statement on dose-sparing strategies for COVID-19 vaccines (fractionated vaccine doses), https://www.who.int/news/item/10-08-2021-interim-statement-on-dose-sparing-strategies-for-covid-19-vaccines-(fractionated-vaccine-doses); 2021 [accessed 16 August 2022].

        • Mateus J
        • Dan JM
        • Zhang Z
        • Rydyznski Moderbacher C
        • Lammers M
        • Goodwin B
        • et al.
        Low-dose mRNA-1273 COVID-19 vaccine generates durable memory enhanced by cross-reactive T cells.
        Science. 2021; 374: eabj9853https://doi.org/10.1126/science.abj9853
        • Voysey M
        • Clemens SAC
        • Madhi SA
        • Weckx LY
        • Folegatti PM
        • Aley PK
        • et al.
        Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK.
        Lancet. 2021; 397: 99-111https://doi.org/10.1016/S0140-6736(20)32661-1
      2. Angkasekwinai N, Niyomnaitham S, Sewatanon J, Phumiamorn S, Sukapirom K, Senawong S, et al. The immunogenicity and reactogenicity of four COVID-19 booster vaccinations against SARS-CoV-2 variants of concern (Delta, Beta, and Omicron) following CoronaVac or CHADOX1-S primary series. medRxiv. 04 February 2022. https://www.medrxiv.org/content/10.1101/2021.11.29.21266947v3; [accessed 20 August 2022].

        • Angkasekwinai N
        • Sewatanon J
        • Niyomnaitham S
        • Phumiamorn S
        • Sukapirom K
        • Sapsutthipas S
        • et al.
        Comparison of safety and immunogenicity of CoronaVac and ChAdOx1 against the SARS-CoV-2 circulating variants of concern. Alpha: Delta, Beta) in Thai healthcare workers.
        Vaccine X. 2022; 10100153https://doi.org/10.1016/j.jvacx.2022.100153
        • Niyomnaitham S
        • Quan Toh Z
        • Wongprompitak P
        • Jansarikit L
        • Srisutthisamphan K
        • Sapsutthipas S
        • et al.
        Immunogenicity and reactogenicity against the SARS-CoV-2 variants following heterologous primary series involving CoronaVac, ChAdox1 nCov-19 and BNT162b2 plus BNT162b2 booster vaccination: an open-label randomized study in healthy Thai adults.
        Hum Vaccin Immunother. 2022; 182091865https://doi.org/10.1080/21645515.2022.2091865
      3. World Health Organization. The Moderna COVID-19 (mRNA-1273) vaccine: what you need to know, who.int/news-room/feature-stories/detail/the-moderna-covid-19-mrna-1273-vaccine-what-you-need-to-know; 2022 (accessed 16 August 2022).

        • Flaxman A
        • Marchevsky NG
        • Jenkin D
        • Aboagye J
        • Aley PK
        • Angus B
        • et al.
        Reactogenicity and immunogenicity after a late second dose or a third dose of ChAdOx1 nCoV-19 in the UK: a substudy of two randomised controlled trials (COV001 and COV002).
        Lancet. 2021; 398: 981-990https://doi.org/10.1016/S0140-6736(21)01699-8
        • Voysey M
        • Clemens SAC
        • Madhi SA
        • Weckx LY
        • Folegatti PM
        • Aley PK
        • et al.
        Single-dose administration and the influence of the timing of the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine: a pooled analysis of four randomised trials.
        Lancet. 2021; 397 (Erratum in: Lancet 2021;397:880): 881-891https://doi.org/10.1016/S0140-6736(21)00432-3
        • Andrews N
        • Stowe J
        • Kirsebom F
        • Toffa S
        • Sachdeva R
        • Gower C
        • et al.
        Effectiveness of COVID-19 booster vaccines against COVID-19-related symptoms, hospitalization and death in England.
        Nat Med. 2022; 28: 831-837https://doi.org/10.1038/s41591-022-01699-1
        • Zhao X
        • Zheng A
        • Li D
        • Zhang R
        • Sun H
        • Wang Q
        • et al.
        Neutralisation of ZF2001-elicited antisera to SARS-CoV-2 variants.
        Lancet Microbe. 2021; 2: e494https://doi.org/10.1016/S2666-5247(21)00217-2
        • Zhao X
        • Li D
        • Ruan W
        • Chen Z
        • Zhang R
        • Zheng A
        • et al.
        Effects of a prolonged booster interval on neutralization of Omicron variant.
        N Engl J Med. 2022; 386: 894-896https://doi.org/10.1056/NEJMc2119426
        • Bar-On YM
        • Goldberg Y
        • Mandel M
        • Bodenheimer O
        • Amir O
        • Freedman L
        • et al.
        Protection by a fourth dose of BNT162b2 against Omicron in Israel.
        N Engl J Med. 2022; 386: 1712-1720https://doi.org/10.1056/NEJMoa2201570
        • Regev-Yochay G
        • Gonen T
        • Gilboa M
        • Mandelboim M
        • Indenbaum V
        • Amit S
        • et al.
        Efficacy of a fourth dose of Covid-19 mRNA vaccine against Omicron.
        N Engl J Med. 2022; 386: 1377-1380https://doi.org/10.1056/NEJMc2202542
        • Munro APS
        • Feng S
        • Janani L
        • Cornelius V
        • Aley PK
        • Babbage G
        • et al.
        Safety, immunogenicity, and reactogenicity of BNT162b2 and mRNA-1273 COVID-19 vaccines given as fourth-dose boosters following two doses of ChAdOx1 nCoV-19 or BNT162b2 and a third dose of BNT162b2 (COV-BOOST): a multicentre, blinded, phase 2, randomised trial.
        Lancet Infect Dis. 2022; 22: 1131-1141https://doi.org/10.1016/S1473-3099(22)00271-7
      4. World Health Organization. Interim statement on the use of additional booster doses of emergency use listed mRNA vaccines against COVID-19, https://www.who.int/news/item/17-05-2022-interim-statement-on-the-use-of-additional-booster-doses-of-emergency-use-listed-mrna-vaccines-against-covid-19; 2022 [accessed 16 August 2022].

      5. European Medicines Agency. European Union risk management plan AstraZeneca. p. ChAdOx1-S (recombinant) (AZD1222), https://www.ema.europa.eu/en/documents/rmp-summary/vaxzevria-previously-covid-19-vaccine-astrazeneca-epar-risk-management-plan_en.pdf; 2021 [accessed 25 August 2022].

      6. European Medicines Agency. BNT162b2 risk management plan, https://www.ema.europa.eu/en/documents/rmp-summary/comirnaty-epar-risk-management-plan_en.pdf; 2022 [accessed 25 August 2022].

      7. Angkasekwinai N, Niyomnaitham S, Sewatanon J, Phumiamorn S, Sukapirom K, Senawong S, et al. The immunogenicity against variants of concern and reactogenicity of four COVID-19 booster vaccinations following CoronaVac or ChAdOx1 nCoV-19 primary series. medRxiv. 19 January 2022. https://www.medrxiv.org/content/10.1101/2021.11.29.21266947v2 [accessed 8 September 2022].

        • Khan K
        • Karim F
        • Ganga Y
        • Bernstein M
        • Jule Z
        • Reedoy K
        • et al.
        Omicron BA.4/BA.5 escape neutralizing immunity by BA.1 infection.
        Nat Commun. 2022; 13: 4686https://doi.org/10.1038/s41467-022-32396-9
      8. Khoury DS, Schlub TE, Cromer D, Steain M, Fong Y, Gilbert PB, et al. Correlates of protection, thresholds of protection, and immunobridging in SARS-CoV-2 infection. medRxiv. 06 June 2022. https://www.medrxiv.org/content/10.1101/2022.06.05.22275943v1 [accessed 08 September 2022].

      9. United States Food and Drug Administration. mRNA-1273.214 Moderna COVID-19 investigational bivalent vaccine (original + Omicron) Moderna, Inc. Vaccines and Related Biological Products Advisory Committee Meeting, https://www.fda.gov/media/159492/download; 2022 [accessed 08 September 2022].

        • Suphanchaimat R
        • Nittayasoot N
        • Jiraphongsa C
        • Thammawijaya P
        • Bumrungwong P
        • Tulyathan A
        • et al.
        Real-world effectiveness of mix-and-match vaccine regimens against SARS-CoV-2 Delta variant in Thailand: a nationwide test-negative matched case-control study.
        Vaccines. 2022; 10: 1080https://doi.org/10.3390/vaccines10071080