The Lancet 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
Reported in The lancet.com 06 March 2021
- Merryn Voysey, DPhil *
- Sue Ann Costa Clemens, MD *
- Shabir A Madhi, PhD *
- Lily Y Weckx, MD *
- Pedro M Folegatti, MD *
- Parvinder K Aley, DPhil
- et al.
- Show all authors
- Show footnotes
The ChAdOx1 nCoV-19 (AZD1222) vaccine has been approved for emergency use by the UK regulatory authority, Medicines and Healthcare products Regulatory Agency, with a regimen of two standard doses given with an interval of 4–12 weeks. The planned roll-out in the UK will involve vaccinating people in high-risk categories with their first dose immediately, and delivering the second dose 12 weeks later. Here, we provide both a further prespecified pooled analysis of trials of ChAdOx1 nCoV-19 and exploratory analyses of the impact on immunogenicity and efficacy of extending the interval between priming and booster doses. In addition, we show the immunogenicity and protection afforded by the first dose, before a booster dose has been offered.
We present data from three single-blind randomised controlled trials—one phase 1/2 study in the UK (COV001), one phase 2/3 study in the UK (COV002), and a phase 3 study in Brazil (COV003)—and one double-blind phase 1/2 study in South Africa (COV005). As previously described, individuals 18 years and older were randomly assigned 1:1 to receive two standard doses of ChAdOx1 nCoV-19 (5 × 1010 viral particles) or a control vaccine or saline placebo. In the UK trial, a subset of participants received a lower dose (2·2 × 1010 viral particles) of the ChAdOx1 nCoV-19 for the first dose. The primary outcome was virologically confirmed symptomatic COVID-19 disease, defined as a nucleic acid amplification test (NAAT)-positive swab combined with at least one qualifying symptom (fever ≥37·8°C, cough, shortness of breath, or anosmia or ageusia) more than 14 days after the second dose. Secondary efficacy analyses included cases occuring at least 22 days after the first dose. Antibody responses measured by immunoassay and by pseudovirus neutralisation were exploratory outcomes. All cases of COVID-19 with a NAAT-positive swab were adjudicated for inclusion in the analysis by a masked independent endpoint review committee. The primary analysis included all participants who were SARS-CoV-2 N protein seronegative at baseline, had had at least 14 days of follow-up after the second dose, and had no evidence of previous SARS-CoV-2 infection from NAAT swabs. Safety was assessed in all participants who received at least one dose. The four trials are registered at ISRCTN89951424 (COV003) and ClinicalTrials.gov, NCT04324606 (COV001), NCT04400838 (COV002), and NCT04444674 (COV005).
Between April 23 and Dec 6, 2020, 24 422 participants were recruited and vaccinated across the four studies, of whom 17 178 were included in the primary analysis (8597 receiving ChAdOx1 nCoV-19 and 8581 receiving control vaccine). The data cutoff for these analyses was Dec 7, 2020. 332 NAAT-positive infections met the primary endpoint of symptomatic infection more than 14 days after the second dose. Overall vaccine efficacy more than 14 days after the second dose was 66·7% (95% CI 57·4–74·0), with 84 (1·0%) cases in the 8597 participants in the ChAdOx1 nCoV-19 group and 248 (2·9%) in the 8581 participants in the control group. There were no hospital admissions for COVID-19 in the ChAdOx1 nCoV-19 group after the initial 21-day exclusion period, and 15 in the control group. 108 (0·9%) of 12 282 participants in the ChAdOx1 nCoV-19 group and 127 (1·1%) of 11 962 participants in the control group had serious adverse events. There were seven deaths considered unrelated to vaccination (two in the ChAdOx1 nCov-19 group and five in the control group), including one COVID-19-related death in one participant in the control group. Exploratory analyses showed that vaccine efficacy after a single standard dose of vaccine from day 22 to day 90 after vaccination was 76·0% (59·3–85·9). Our modelling analysis indicated that protection did not wane during this initial 3-month period. Similarly, antibody levels were maintained during this period with minimal waning by day 90 (geometric mean ratio [GMR] 0·66 [95% CI 0·59–0·74]). In the participants who received two standard doses, after the second dose, efficacy was higher in those with a longer prime-boost interval (vaccine efficacy 81·3% [95% CI 60·3–91·2] at ≥12 weeks) than in those with a short interval (vaccine efficacy 55·1% [33·0–69·9] at <6 weeks). These observations are supported by immunogenicity data that showed binding antibody responses more than two-fold higher after an interval of 12 or more weeks compared with an interval of less than 6 weeks in those who were aged 18–55 years (GMR 2·32 [2·01–2·68]).
The results of this primary analysis of two doses of ChAdOx1 nCoV-19 were consistent with those seen in the interim analysis of the trials and confirm that the vaccine is efficacious, with results varying by dose interval in exploratory analyses. A 3-month dose interval might have advantages over a programme with a short dose interval for roll-out of a pandemic vaccine to protect the largest number of individuals in the population as early as possible when supplies are scarce, while also improving protection after receiving a second dose.
UK Research and Innovation, National Institutes of Health Research (NIHR), The Coalition for Epidemic Preparedness Innovations, the Bill & Melinda Gates Foundation, the Lemann Foundation, Rede D’Or, the Brava and Telles Foundation, NIHR Oxford Biomedical Research Centre, Thames Valley and South Midland’s NIHR Clinical Research Network, and AstraZeneca.
The widespread morbidity and mortality associated with the 2020 COVID-19 pandemic precipitated the most extensive and rapid global vaccine development programme in history,1 culminating in the development of several vaccines reaching phase 3 efficacy milestones and receiving emergency use authorisation by the end of that year.2, 3, 4 Widespread vaccination programmes have commenced in several countries as new vaccines are licensed for emergency use by regulators in each setting, with a focus primarily on high-risk groups such as the elderly, those with comorbidities, or front-line workers.Vaccine supply is likely to be scarce, at least initially, and so policy makers must decide how best to deliver available doses to achieve greatest public health benefit, and different approaches have been taken in different settings. In the UK, second doses of both available vaccines (a viral vector and mRNA vaccine) are being delivered with an interval of up to 12 weeks,5, 6 and this regimen is also being considered by several other countries.7, 8 By contrast, WHO has recently recommended a maximum 6 week interval between the two doses of the same mRNA vaccine.9The ChAdOx1 nCoV-19 vaccine (AZD1222) is a chimpanzee adenoviral vectored vaccine with full length SARS-CoV-2 spike insert, developed at the University of Oxford (Oxford, UK). The safety and immunogenicity of the vaccine were assessed in four randomised controlled trials in the UK, Brazil, and South Africa, and results in cohorts of healthy adults and in adults aged 70 years or older have been published.4, 10, 11, 12, 13 Efficacy of two doses of the vaccine in the interim analysis of 131 cases (data cutoff Nov 4, 2020), which pooled data from Brazil and the UK, was 70·4% (95·8% CI 54·8–80·6) overall.4 ChAdOx1 nCoV-19 was authorised for emergency use in the UK on Dec 30, 2020,14 on the basis of the interim analysis data,4 based on a regimen of two standard doses administered 4–12 weeks apart for adults aged 18 years and older, and has since been authorised for use in many other countries.Research in contextEvidence before this studyThe ChAdOx1 nCoV-19 (AZD1222) vaccine was approved for emergency use authorisation in the UK on the basis of interim efficacy results from 131 cases of primary symptomatic COVID-19, with efficacy based on two of the four trials of the vaccine. The planned roll-out of the vaccine in the UK involves the administration of two doses 12 weeks apart, a policy that has received substantial comment.Added value of this studyThis report provides updated primary efficacy results after a further month of data collection. The interim report included 131 cases of primary symptomatic COVID-19. The latest results with additional follow-up include 332 cases of primary symptomatic COVID-19. Efficacy estimates now include data from all four studies of the vaccine from three countries, whereas the interim analysis included only two studies in efficacy assessments because of the small number of cases in the smaller studies. In addition to the primary efficacy assessment, post-hoc exploratory analyses have been added, including a breakdown of efficacy by prime-boost interval, and the efficacy of a single dose of vaccine.Implications of all the available evidenceThe primary analysis supports the findings reported in the interim analysis that the vaccine is efficacious and safe. Exploratory analyses show that higher vaccine efficacy is obtained with a longer prime-boost interval, and that a single dose of vaccine is efficacious in the first 90 days, providing further evidence for current policy.The University of Oxford-sponsored studies were initially planned as single-dose studies but were amended to incorporate a second dose after review of the phase 1 immunogenicity data, which showed a substantial increase in neutralising antibody with a second dose of vaccine.12 After initially providing consent to participate in a single-dose study, some participants chose not to receive the second dose, providing a self-selected cohort of single-dose recipients. Additionally, because of the time required to manufacture the second dose, there were delays in administration of the second dose for a large number of trial participants who received the two-dose schedule. These two situations provide an opportunity to explore the immunogenicity and efficacy of a single dose of vaccine, and the effect of an extended interval before delivery of the second dose. In addition, data from an additional month of follow-up are now available for inclusion in the analysis, providing greater precision in estimates because of the larger number of cases for analysis in comparison with the previous report.4
Study design and participants
Data from three single-blind randomised controlled trials, one phase 1/2 study in the UK (COV001), one phase 2/3 study in the UK (COV002), and a phase 3 study in Brazil (COV003), and one double-blind phase 1/2 study in South Africa (COV005) are included in this primary analysis because all four trials now meet the required criteria for inclusion of having at least five primary outcome cases. Full descriptions of the methods as well as safety, immunogenicity, and interim efficacy analyses of the four studies have been previously published in detail, including full study protocols.4, 12, 13COV001 (UK) enrolled healthy adults aged 18–55 years. COV002 (UK) and COV003 (Brazil) enrolled adults aged 18 years and older, with a focus on recruitment of health-care workers and others at increased exposure to SARS-CoV-2 infection. COV005 (South Africa) enrolled adults aged 18–65 years.In the UK, the COV001 and COV002 studies were approved by the South Central Berkshire Research Ethics Committee (COV001 reference 20/SC/0145, March 23, 2020; and COV002 reference 20/SC/0179; conditional approval April 8, full approval April 19, 2020). The COV003 study was approved by the Oxford Tropical Research Ethics Committee (OxTREC; reference 36–20, June 12, 2020) and by the Comissão Nacional de Ética em Pesquisa (June 3, 2020). The COV005 study was approved by OxTREC (reference 35-20, June 5, 2020), the University of Witwatersand Human Research Ethics committee (reference 200501, May 21, 2020), and the South African Health Products Regulatory Authority (reference 20200407, June 1, 2020).
Randomisation and masking
Briefly, participants in efficacy cohorts from the four trials were randomly assigned 1:1 with full allocation concealment to receive either ChAdOx1 nCoV-19 vaccine or a control product (MenACWY in the UK, MenACWY prime and saline boost in Brazil, and saline only in South Africa). One group of participants in the COV002 study in the UK received a low dose as their first dose followed by a standard dose, as discussed previously.4 Other participants received two standard doses.
Procedures have been described in full previously.4, 12, 13 At baseline, eligibility and medical history was assessed and informed consent was taken from all participants. A baseline serum sample was taken to assess SARS-CoV-2 serostatus.In all studies, participants were asked to contact the study site if they had symptoms of COVID-19 and were then invited to attend for clinical review and a swab. Additionally, in the UK, asymptomatic infections were measured by means of weekly self-administered nose and throat swabs using kits provided by the Department of Health and Social Care as previously described.4 Those who tested positive on a self-swab were not specifically contacted by the study site and they are classed as having unknown symptoms in the analysis unless specific information was obtained (eg, through participants calling their study site) that the participant was asymptomatic, or unless the participant had additionally reported symptoms and could be classified as symptomatic.
The primary outcome was virologically confirmed symptomatic COVID-19 disease, defined as a nucleic acid amplification test (NAAT)-positive swab combined with at least one qualifying symptom (fever ≥37·8°C, cough, shortness of breath, anosmia, or ageusia). The primary analysis was of cases occurring more than 14 days after the second dose, with a prespecified secondary analysis of cases occurring more than 21 days after the first dose.A secondary analysis of any NAAT-positive case included a combination of primary symptomatic cases, non-primary symptomatic cases (eg, those who had other symptoms not meeting the primary definition such as nausea or diarrhoea), asymptomatic cases, and those with unknown symptoms.
The study was powered to include an α-adjusted interim analysis triggered when at least 53 cases had accrued in participants who had received two standard-dose vaccines. This analysis was statistically significant and subsequent efficacy results are considered supportive of that analysis, with no further adjustment of α.For the primary analysis, which we present here updated with additional cases from an extra month of follow-up, participants enrolled in efficacy cohorts were included in the analysis according to the vaccine received. Events were included that occurred more than 14 days after the second dose, in participants who were seronegative to SARS-CoV-2 N protein at baseline and had at least 14 days of follow-up after the second dose and no evidence of SARS-CoV-2 infection from NAAT swabs before day 14. Vaccine efficacy was calculated as 1–the adjusted relative risk (ChAdOx1 nCoV-19 vs control groups) computed using a robust Poisson regression model. The model contained terms for study, treatment group, and age group at randomisation. The logarithm of the period at risk was used as an offset variable in the model to adjust for volunteers having different follow-up times during which the events occurred. Cumulative incidence of primary symptomatic COVID-19 is presented using the Kaplan-Meier method.We present additional exploratory analyses of single-dose efficacy, which have been added at the request of regulators and policy makers. These are considered as supportive analyses to the previously published interim efficacy analysis and were not prespecified. The effect of the timing of the second dose is explored in more detail.For the analysis of single-dose efficacy, randomised participants enrolled in efficacy cohorts were included in the analysis according to the vaccine they received as their first dose. Events were included if they occurred more than 21 days after the first dose. Participants were excluded if they had a NAAT-positive swab in the first 21 days after the first dose or had fewer than 22 days of follow-up. Participants who received a second dose were censored in the analysis at the time of their booster dose. Participants who did not receive a second dose are censored in the analysis at the data cutoff date.For exploratory analysis, the persistence of anti-spike IgG responses after a single dose was measured in the UK by standardised ELISA. Decay of antibody over time was modelled for low-dose and standard-dose recipients using a linear model of log-transformed antibody values. A non-linear generalised additive model was also used to assess the shape of the decay curve to establish whether linear modelling was appropriate. Both models gave similar outputs.Baseline serum samples were measured for nucleocapsid reactivity with the Roche Elecsys Anti-SARS-CoV-2 serology test (PPD Central Laboratories, Zaventem, Belgium and Highland Heights, KY, USA) and a multiplexed immunoassay (3-plex ECL based assay on the MSD platform, PPD Vaccines, Richmond, VA, USA) was used to measure the spike-specific response to ChAdOx1 nCoV-19 vaccination. Antibody neutralisation was measured with lentivirus-based pseudovirus particles expressing the SARS-CoV-2 spike protein as described.12For exploratory analyses of the effect of varying the timing of the second dose of vaccine, we fit separate efficacy models, using unadjusted log-binomial models, for each 20-day window starting with a window of 26–46 days (midpoint for plot 36 days) and increasing by 1 day for each model. Participants who received their second dose within the window were included in that model. Vaccine efficacy for each window was plotted with 95% CIs. Unadjusted models were used to achieve convergence across every model consistently and to remove bias from the potentially different effect of variation in the distribution of adjustment variables in different models. Participants were not randomly assigned to their dosing interval and these exploratory analyses should be interpreted with caution because it is not possible to exclude the possibility that any apparent trend is due to measured or unmeasured confounding factors.To explore the potential for waning of efficacy after the first dose before a booster dose was received, a similar approach was taken with separate efficacy models fitted to 21-day windows of the time from vaccination. Cases occurring outside the windows were censored.Potential differences in population baseline characteristics between those who received a second dose of vaccine and those who did not are explored descriptively, with comparisons made between groups using χ2 tests, Wilcoxon rank-sum tests, or Cochran-Armitage tests as appropriate.Safety was assessed in all participants who received at least one dose. Safety data were reviewed on an ongoing basis by the independent data monitoring safety board. All endpoints were adjudicated for inclusion in the analysis by an independent masked endpoint review committee.Data analysis was done using R, version 3.6.1 or later. Robust Poisson models were fitted using “proc genmod” function in SAS, version 9.4. The four trials are registered at ISRCTN89951424 (COV003) and ClinicalTrials.gov, NCT04324606 (COV001), NCT04400838 (COV002), and NCT04444674 (COV005).
Role of the funding source
AstraZeneca reviewed the data from the study and the final manuscript before submission, but the academic authors retained editorial control. All other funders of the study had no role in the study design, data collection, data analysis, data interpretation, or writing of the report.
Between April 23 and Dec 6, 2020, 24 422 participants were recruited and vaccinated across the four studies, of whom 17 178 were included in this primary efficacy analysis (8597 receiving ChAdOx1 nCoV-19 and 8581 receiving control vaccine). 8948 were from the UK trial, 6753 from the Brazil trial, and 1477 from the South Africa trial (appendix p 2). Here, we provide safety data on 100 958 person-months of follow-up after first dose and 49 945 person-months of follow-up after two doses. Baseline characteristics were similar for vaccine and control groups (appendix p 3). Duration of follow-up varied by prime-boost interval (appendix p 4). The day for data cutoff for cases to be included in this report was Dec 7, 2020.There were 332 cases of primary symptomatic COVID-19 occurring more than 14 days after a booster dose, 84 (1·0%) in the 8597 participants in the ChAdOx1 nCoV-19 group and 248 (2·9%) in the 8581 participants in the control group, with overall efficacy of 66·7% (95% CI 57·4–74·0; table 1). In the participants who received two standard doses, 74 (1·0%) cases occurred in the 7201 participants in the ChAdOx1 nCoV-19 group and 197 (2·7%) in the 7179 in the control group, with vaccine efficacy of 63·1% (51·8–71·7). 61 cases were recorded in the participants who received a low dose plus standard dose, ten (0·7%) of 1396 participants in the ChAdOx1 nCoV-19 group and 51 (3·6%) of 1402 in the control group, with vaccine efficacy of 80·7% (62·1–90·2).Table 1Efficacy of ChAdOx1 nCoV-19 more than 14 days after a second dose