What next for AstraZeneca's 'intriguing' half-dose COVID-19 vaccine?

By Rachel Arthur

- Last updated on GMT

Pic:getty/ca-ssis
Pic:getty/ca-ssis

Related tags Astrazeneca COVID-19 vaccine

AstraZeneca and the University of Oxford's COVID-19 vaccine candidate threw up a surprise last month when a lower initial dose of the two-dose vaccine showed higher efficacy. More detailed results from this interim analysis, now peer-reviewed and published in The Lancet this week, shed some light on this 'intriguing' efficacy.

AstraZeneca and Oxford University released basic results from its Phase 3 interim analysis for its viral-vectored COVID-19 candidate AZD1222 last month: giving a pooled efficacy rate of 70%. This was drawn from two different regimens: a half dose followed by a full dose giving 90% efficacy; and two full doses giving 62% efficacy. 

And the vaccine drew headlines when AstraZeneca admitted the more effective half dose regimen had come about as a result of a mistake in dosage in a small batch of participants in the Phase 3 trial in the UK. 

How did the low-dose regimen come about?

Initial dosing in the UK COV002 trial was batch manufactured via a contract manufacturing organisation using chromatographic purification, reveals the full interim analysis in The Lancet​ this week. During quality control, differences were observed between the quantification methods – spectrophotometry and quantitative PR – prioritized by different manufacturing sites.

However, the Oxford University highlights this was an issue with measurement methods and not the accuracy of the manufacturing process.

“There were no problems with manufacturing the vaccine, however, due to slight differences in the materials used in the production process, the spectrophotometry over-estimated the dose, so that after adjustment a lower dose was produced. 

“This unplanned lower dosing regimen became apparent following our investigations into the different methods of dose estimation, and when we observed the lower vaccine dose was better tolerated by our volunteers than expected, this confirmed that the dose needed to be adjusted to the standard dose agreed in the trial protocol. 

“This was immediately discussed with the regulators, who agreed to allow our trial protocol to be updated to allow us to follow this scientifically interesting group of volunteers.”

Oxford University decided to continue to follow the subgroup and says it has established tests to ensure consistency in all future doses.

“When we realised that the dose was better tolerated by the volunteers in our trial, we discussed this with regulators and independent data safety monitoring board and agreed that in a time of a global pandemic, it would be valuable to global health and still scientifically appropriate to continue to follow this sub-group of volunteers. 

“The half dose group was unplanned, but we did know in advance that there was a discrepancy in the dose measurements and discussed this with the regulators. We chose the approach to dosing that was logical and safe, as there was never any risk of overdose. 

“Subsequently, we have worked with AstraZeneca to establish a suite of tests to allow consistency across all batches of vaccine.” 

The low dose conundrum: ‘Efficacy was intriguingly high’

Participants in the full dose regimen (SD/SD) received two full doses (~5x1010​ viral particles) of AZD1222; while the placebo group received the meningococcal vaccine MenACWY.

Participants who received an initial half dose (LD/SD) received a first dose (~2.5 x1010​ viral particles) followed by a full second dose  (~5x1010​ viral particles) of AZD1222.

The high efficacy of the initial low dose regimen was a surprise to researchers, although they do point to a number of possible explanations in The Lancet.

“Efficacy of 90% seen in those who received a low dose as prime in the UK was intriguingly high compared with the other findings in the study.

"Although there is a possibility that chance might play a part in such divergent results, a similar contrast in efficacy between the LD/SD and SD/SD recipients with asymptomatic infections provides support for the observation.

“Use of a low dose for priming could provide substantially more vaccine for distribution at a time of constrained supply, and these data imply that this would not compromise protection."

It notes, however, wide CIs around its estimates and says further analyses will follow in the ongoing trial.

“Similar results have been seen for other vaccines where a reduced number or type of priming dose in infancy can lead to higher responses to a booster vaccine,”​ says the study, referencing a 2015 study on serogroup C meningococcal conjugate vaccine in infants.

“Further work is needed to determine the mechanism of the increased efficacy with a LD/SD regimen, which might be due to higher levels of neutralising antibody, lower levels of anti-vector immunity with lower vector-derived antigen content of the first dose, or differential antibody functionality or cellular immunity, including altered avidity or immunodominance.”

Phase III interim analysis: Main points from The Lancet study

  • The vaccine is 'safe and effective at preventing symptomatic COVID-19 and protects against severe disease and hospitalisation'.
  • The interim analysis for efficacy was based on 11,636 participants accruing 131 symptomatic infections from Phase III UK and Brazil trials.
  • The vaccine is 70.4% (95.8% CI: 54.8% to 80.6%) effective at preventing symptomatic COVID-19 occurring more than 14 days after receiving two doses of the vaccine (based on pooled data).
  • A secondary efficacy endpoint of prevention of severe disease demonstrated no cases of severe infections or hospitalisations in the vaccine group.
  • When the vaccine was given as two full doses, vaccine efficacy was 62.1% (n=8,895; CI 41.0% to 75.7%), and 90.0% (n=2,741; CI 67.4% to 97.0%) in participants who received a half dose followed by a full dose
  • There were no hospitalisations or severe cases of COVID-19 more than 21 days after the first dose of the vaccine. Ten participants in the control group were hospitalised due to COVID-19, among whom two were assessed as severe, including one fatal case.

Full figures and analysis in The Lancet.

So which vaccine dosage will people get?

AstraZeneca and Oxford University are simply giving all information – for both dose regimens – to regulators (the vaccine is under rolling reviews with regulators such as Europe's EMA, the UK's MHRA, Health Canada and Switzerland's SwissMedic). It has not specifically applied for usage of one regimen or the other.

At this point, the LD/SD regimen only has a small amount of data available. Occuring in the COV002 UK study, the LD/SD regimen covered 2,741 people; compared to 8,895 people in the SD/SD regimen. The LD/SD regimen is a small cohort unique to the UK trial; whereas the SD/SD regimen is being carried out across wider Phase 3 trials the UK, Brazil and South Africa.

Enrolled over 11 days between May 31 and June 10, the LD/SD cohort was aged 18-55 years old. Therefore the efficacy of the LD/SD regimen is unknown in older people (enrolment of older age cohorts for SD/SD began in August and now includes participants aged 70+). 

And as highlighted in The Lancet​, further investigations are required to understand how this lower dose regimen works. With this in mind, a statement from Oxford University implies it still considering the SD/SD regimen as its main candidate: “We have a lot of data on our planned dosing regimen of two full doses and we will provide extensive information for regulators to consider.”

As for the future of the LD/SD regimen, an AstraZeneca spokesperson told this publication that this data will be included in information submitted to regulators: and the company “will decide in due course”​ as to what the future holds for this possible regimen.

In the meantime, it is using the pooled efficacy figure of 70.4%. While this figure does not correlate specifically to either regimen, Oxford University defends the use of the figure.

The 70.4% efficacy reported is a weighted average from a pooled analysis of our data that combined information from our trial sites around the UK and in Brazil. Before we looked at the data and began our analysis, we agreed the approach to this analysis with regulators. The pooling was considered appropriate as it improved the generalisability of the findings, by including different regions and populations and because the immune responses appeared similar across the different groups.” 

Related topics Bio Developments

Related news

Show more

Follow us

Webinars