Waning of first- and second-dose ChAdOx1 and BNT162b2 COVID-19 vaccinations across the UK

Summary in Plain English

As of October 2022, the main COVID-19 vaccines that have been administered in the UK are:

• Pfizer-BioNTech (BNT162b2 mRNA) – introduced in UK on 8 December 2020;
• Oxford-AstraZeneca (ChAdOx1 nCoV-19) – introduced in UK on 4 January 2021;
• Moderna (mRNA-1273) – introduced in UK on 7 April 2021.
• Novavax (NVX-CoV2373) – approved in UK on 3 February 2022

Why did we carry out this work?

All COVID-19 vaccines available in the UK have been shown to provide protection against COVID-19 related hospital admissions and deaths, both in clinical trials and in real world settings.

Unfortunately, there is growing evidence that this protection decreases (wanes) in the weeks and months following vaccination.

To our knowledge, all COVID-19 vaccine waning studies in the UK thus far have analysed data from a single nation (England, Northern Ireland, Scotland or Wales). This study is the first of its kind to do a pooled analysis using data from all 4 nations of the UK.

It is important that researchers from across the four UK nations work together on these types of analyses. This allows us to better understand how the effectiveness of COVID-19 vaccines changes over time, across the entire UK population.

What did we do?

We used health data from across the UK to carry out this study. These data are routinely collected from patients in England, Scotland, Wales and Northern Ireland in near-real time. Datasets are stored in each UK nation, in a secure Trusted Research Environment (TRE).

Click here to find out more about Trusted Research Environments (TREs) [Understanding Patient Data]

The datasets for Northern Ireland (1.9 million people) and Scotland (5.3 million) included relevant health information for 99-100% of the population. The dataset for England contained information for 10% of the population (5.4 million), whilst the dataset for Wales covered 80-87% of the population.

Click to find out more about the data asset we used in England

Click to find out more about the data asset we used in Northern Ireland

Click to find out more about the data asset we used in Scotland

Click to find out more about the data asset we used in Wales

From these datasets, we selected people who were aged 18 or older on 8 December 2020 and followed them until the end of the study period on 30 June 2021.

We used a ‘target trial’ study design’. This method mimics a clinical trial, but in a real-world setting. The idea is to match, at the date of vaccination, a vaccinated person with a person who is unvaccinated at that date and who has very similar characteristics to the vaccinated person.

In this study, we matched people who received a vaccine dose with people who didn’t that were in same age group lived in the same area, and had a similar likelihood to receive that dose.

We then compared the rate at which COVID-19 hospital admission or death happened in these two groups.

We calculated the effectiveness of first and second dose COVID-19 vaccines in preventing hospital admissions and deaths over time.

To ensure that all health data was kept safe and secure, we analysed each nation’s data in separate TREs, and then carried out a pooled analysis using novel techniques that only involved sharing a minimal amount of non-disclosing data between the different TREs. As we used a ‘target trial’ approach the final result that we obtained was the same as we would have obtained as if all of the data had started out in the same location.

What did we find?

Understanding the results

When looking at effectiveness of a vaccine:

• A value greater than 0% indicates that receiving that particular vaccine does was more effective at preventing serious outcomes of COVID-19 than if that dose was not given;
• A value of 0% means that receiving that dose of a vaccine was equally effective compared to if the dose was not given; and
• A value below 0% means that receiving that dose of a vaccine was less effective at preventing serious outcomes than if the dose was not given.

How vaccine effectiveness changed over time

Our results showed that between 60-80 days post-vaccination, the effectiveness of the first and second dose of the Oxford-AstraZeneca vaccine fell to 0% and then went negative. This means that from this point onwards, individuals who received the vaccine dose were more likely to have serious COVID-19 related events than those who did not.

A likely explanation for this is that people who received the vaccine dose changed their behavior compared to people who didn’t, putting themselves in situations where they are at greater risk of coming into contact with the virus, as they believed they were protected from it.

In contrast, the effectiveness of a second dose Pfizer BioNTech vaccine remained above 0% throughout the study period, reaching 46% at 98 days post-vaccination.

Why is this important?

Our study presents strong evidence for the waning of COVID-19 vaccine effectiveness. This suggests that the vaccines may best be viewed as a tool for ‘flattening the curve’ for serious COVID-19 outcomes, and preventing healthcare services from being overwhelmed.

We used a novel method that allowed us to conduct a UK-wide study without needing to share individual-level data between TREs. This study serves as proof of concept for further pooled studies in the future.

We hope this evidence may be used to help inform policies in relation to additional vaccine doses.