Covid-19 Vaccines

Oxford AZ results: how do they change where we are on vaccines?

Little did I think when I posted last night on the differences between the DNA and RNA vaccines* (and in a parallel post I tried to explain these differences in scientific terms)**, that first thing this morning (UK time) Astra Zeneca (AZ) would announce in a press release the results of their DNA vaccine!  In this post I briefly answer the questions that their new data raise.



What are the headline results from the AZ Trial?

  • The AZ trial actually combined the results of several smaller trials which were all embedded within an umbrella of the single large investigation
  • These smaller trials (consider them ‘sub-studies’) used different dosage regimes:
    • Size of dose of the vaccine injected
    • Difference in the interval between doses 
    • Differences in the age of the participants
  • Although the numbers in some of these sub-studies are quite small, AZ decided the time had come that enough cases (131) had emerged that they could produce some preliminary results
  • The limited results presented were as follows:
    • In the total population whose results were analysed, the vaccine was 70% effective
    • The interest in the media this morning was that there was a difference when the study participants were divided into two groups:
      • 2,471 had a half the amount of the vaccine in their first dose and the full amount in the second dose – that regime had a 90% effectiveness
      • 8,895 had the full amount in both doses – that regime had a 62% effectiveness
  • The press release also said there were no severe cases of infection in those who had the active vaccine.  However, we are not given how many severe cases of infection occurred in the placebo group; so not easy to reach a conclusion on severity.

Which of these figures should we take?

  • At one level, AZ are proposing that the 90% effectiveness is proof of the success of the half/full dose regime
  • I would urge caution here for a number of reasons
    • The numbers are still quite small and indeed, even if the results from the half/full dose regime are better, we need much larger numbers to get more precise estimates of this percent benefit (we also need larger numbers for the other vaccines as well to get more precise estimates of these)
    • My understanding of the reason for having a half dose starter was not that the investigators thought this would be more effective but rather:
      • It might be less likely to give side effects
      • It would use less virus and hence be cheaper, and the  vaccine produced could be given to a larger number of people 
    • Thus, taking a formal statistical view there is always a worry when a result comes up from a ‘sub-group’ analysis which was not expected and is then presented as the headline result.  The approach in ‘normal’ times would be to check this unexpected result in new data
  • What is clear in the past from a number of trials of influenza vaccine, is that in a pandemic using half the expected effective dose can give the same level of protection as the full dose
  • So my conclusion is to note the 90% effectiveness in the half/full dose but, at this stage, take a cautious approach and use the 70% overall effect as being the key result to take forward

Does this mean the AZ vaccine is less effective than Pfizer or Moderna?

  • Firstly a 70% effectiveness, if that is how the AZ vaccine turns out, is still a success – would all conventional vaccines had that success!
  • Please read my posts yesterday on DNA and RNA vaccines where I show that the approach is different and that theoretically RNA vaccines (eg Pfizer and Moderna) might indeed be more effective
  • Against this, AZ vaccine follows perhaps a more tested approach, and will be much cheaper to produce and distribute
  • Thus in a world when only the AZ vaccine existed, I would consider that – unless contradicted as more data emerged –  this vaccine would be approved for widespread use

What about how easy will it be to achieve herd immunity with a vaccine with 70% effectiveness compared to one that is 95% effective?

  • Regular readers of this blog will know that I had anticipated this question emerging when the AZ data came out. The analysis I undertook therefore on this topic in my post you might find interesting to reread:

  • Some of you will remember this figure which showed the data where I tried to address how differences in vaccine effectiveness influence herd immunity (the yellow bars can now be read as the overall AZ vaccine result)


  • The results from the AZ DNA vaccine are good news of course
  • It would be wrong at this stage to say that this approach to making a Covid-19 vaccine is equally as effective as an RNA approach
  • More follow up data will provide a more definitive answer as to the relative benefits of the two approaches
  • With the limited evidence thus far available, and taking account of the different technologies, it is not a surprise that RNA vaccines might be more effective
  • All countries will now have a choice in terms of vaccine procurement but if (unlikely I know) that everything else (costs/availability/logistics) being equal it would make sense based on my previous post that the higher the effectiveness, the attainment of herd immunity will be so much easier- especially in the face of challenges in achieving high take up 

If you would like to receive email notifications of new posts on this blog just click this link and enter your email at the bottom of the home page:

6 replies on “Oxford AZ results: how do they change where we are on vaccines?”


Would I be right in thinking that herd immunity here will protect me whilst I am in the UK but if I travel (e.g. to Israel) man RNA vaccine will give me better chance of protection than a traditional DNA vaccine?


From: Making sense of CoVid-19 with Silman Reply to: Making sense of CoVid-19 with Silman Date: Monday, 23 November 2020 at 11:23 To: Mike Brown Subject: [New post] Oxford AZ results: how do they change where we are on vaccines?

alansilman posted: ” Little did I think when I posted last night on the differences between the DNA and RNA vaccines* (and in a parallel post I tried to explain these differences between a DNA vaccine)**, that first thing this morning (UK time) Astra Zeneca (AZ) would an”


I think that it may take time for herd immunity to build up in the UK, as per my pst last week and this will depend on how many people have the vaccine. I would be reluctant as things stand to rely on herd immunity until we know that the virus has been eliminated. The same applies to Israel. Whether an RNA or DNA vaccine is preferable we don’t know. As I said both are looking good, RNA vaccines look from these early data to be better and may therefore achieve herd immunity with a smaller take up


As a person with low immunity due to having to take anti rejection drugs and being unable to take “live vaccines”, what options would I have when it comes to deciding which vaccines would be safe for me to take?


Obviously I cannot give individual clinical advice

There are two aspects to having low immunity and vaccines

The first is safety, and RNA vaccines, not having a live virus, would not pose a risk as far as I can tell that you would ‘catch’ anything

The second is whether the vaccine would work, this would depend on the how immunosuppressed you were, and with what drugs. Vaccines can work in some people. The (second and new) AZ vaccine I mentioned in my post last night would be fir people like you as it provides thready made antibody. It doesn’t last though


Perhaps you have already covered this but there don’t appear to be meaningful data for a group for which AZ is recommended. There were too few in the over 55 group to derive outcome measures and it could be argued that even the decreased severity of failures within the treated group is based on small numbers. Do we therefore have a maybe safe vaccine with no evidence of efficacy over 55 which might be approved for the whole population?
Mike Robinson


Absolutely you are right that despite having a wide age range for recruitment, in reality there were too few people in older age groups for robust age specific analysis. But this is not unusual and we will have to wait for post licensing experience in the real world for answers. My guess is that the response will be good enough in mostelderly people without any other major problem


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s