Background

The vaccination programme for COVID-19 in the United Kingdom (UK) was started on 8 December 2020. Vaccination was in order of priority groups determined by the Joint Committee on Vaccination and Immunisation (JCVI) expert advisory group [1]: initially to people aged 80+ years, health and care workers and care home residents, people aged 70–79 and those extremely clinically vulnerable, followed by remaining adults in order of decreasing age or at increased risk. Three COVID-19 vaccines, approved by the Medicines and Healthcare products Regulatory Agency (MHRA) and requiring two doses to produce maximum protection, are currently in use in the United Kingdom (UK) [2]. To allow a higher percentage of the population to receive one vaccine dose quicker, second doses initially followed a 12-week interval from the first. This interval has since been shortened. These vaccines and their date of first administration and current second dose schedule are as follows: the Pfizer-BioNTech BNT162b2 mRNA COVID-19 vaccine (BNT162b2; first administered 8 Dec 2020, second dose at least 21 days later); the Oxford-AstraZeneca ChAdOx1 nCoV-19 vaccine (ChAdOx1; first administered 4 Jan 2021, second dose between 4 and 12 weeks later); and the Moderna mRNA-1273 vaccine (mRNA-1273; first administered 7 April 2021, second dose 28 days later). As of November 1st, 2021 78.7% of individuals aged over 16 years in England had been fully vaccinated (i.e., ≥14 days have passed since the receipt of their second dose of a COVID-19 vaccine) [3].

The vaccines against COVID-19 are considered to be highly effective and there is strong evidence that the UK’s COVID-19 vaccination programme has reduced infection and severe outcomes in vaccinated individuals [4,5,6]. However, breakthrough infections after vaccination against SARS-CoV-2 are increasingly reported [7] and there are concerns that the effectiveness of the vaccines may fade over time, due to the vaccines being less effective against the delta variant, and/or the waning of the immune response [8]. In addition, as no vaccine is 100% effective, COVID-19 vaccine breakthrough is likely (i.e., a small number of people will still get sick, be hospitalised, or die from COVID-19, despite being fully vaccinated). Describing individuals who have experienced a COVID-19 vaccine breakthrough could provide the first indication that the COVID-19 vaccine is less effective in certain groups and could be hugely important in helping to determine who may be at greatest risk and therefore might benefit most from booster doses.

To that end we used the new secure data analytics platform, OpenSAFELY [9] (built by our group on behalf of NHS England to support analysis of important questions related to COVID-19) to describe breakthrough COVID-19 amongst fully vaccinated individuals in England; and to describe how breakthrough COVID-19 varied between priority groups and by clinical and demographic characteristics.

Methods

Data sources

This OpenSAFELY study was conducted using the OpenSAFELY-TPP database which contains records for approximately 24 million people currently registered with GP surgeries using TPP SystmOne software (approximately 40% of the English population).

Study population

The base population consisted of all individuals aged 16 years or over with records indicating that they had received two COVID-19 vaccination doses since 8 December 2020 (the earliest vaccination date in England) and who were still alive and registered 2 weeks after their second dose. Individuals were excluded if they tested positive for SARS-CoV-2 or had been hospitalised or died due to COVID-19 within the 2 weeks after their second dose. Follow-up started 14 days after an individual’s second dose of the COVID-19 vaccination (the point by which individuals were classified as being fully vaccinated) and individuals were followed up until 1st November 2021 or until the first occurrence of the outcome of interest, death, and practice de-registration.

Outcomes

Four outcomes were assessed: positive SARS-CoV-2 swab test, via SGSS and based on swab date (we did not distinguish between symptomatic and asymptomatic infection for this outcome, nor PCR and lateral flow testing); COVID-19-related hospital admissions via Hospital Episode Statistics (HES) in-patient records (using both primary and non-primary diagnosis codes); COVID-19-related critical care admissions via HES and COVID-19-related death, via linked death registry records, which included individuals who died within 28 days of positive SARS-CoV-2 test or who had COVID-19 mentioned on the death certificate as one of the causes. COVID-19 hospitalisation variables were derived using two ICD-10 COVID-19 diagnosis codes; U071, U072 [10]. Outcomes were only included if they occurred 14 days or more after an individual's second dose of a COVID-19 vaccination.

Priority groups for vaccination

Individuals were classified into seven priority groups (Table 1), based on the Joint Committee on Vaccination and Immunisation (JCVI) priority groups [11] using SNOMED-CT codelists and logic defined in the national COVID-19 Vaccination Uptake Reporting Specification developed by PRIMIS [12]. Individuals were assigned only to their highest priority group and not included again as part of any other priority group. In line with the national reporting specification, most criteria were ascertained using the latest available data at the time of analysis, with the exception of age which was calculated as at 31 March 2021 as recommended by Public Health England.

Table 1 Priority groups for vaccination
Full size table

Key demographic and clinical characteristics

Key clinical and demographic groups which were considered to have a higher possibility of experiencing a COVID-19 vaccine breakthrough were identified from previous studies [13,14,15,16,17]. This included age, sex, body mass index (BMI; kg/m2), smoking status, deprivation (measured by the index of multiple deprivation (IMD) as quintiles, ethnicity, NHS region of patient’s general practice, asplenia, asthma, blood pressure, cancer, chronic kidney disease, diabetes mellitus, dialysis, heart disease, haematological malignancy, immunocompromised, learning disability, liver disease, neurological diseases, respiratory disease, severe mental illness and transplant. Other variables considered were time since being fully vaccinated, time between vaccinations and any evidence of a prior SARS-CoV-2 infection.

Codelists and implementationg

Information on all variables were obtained from primary care records by searching TPP SystmOne records for specific coded data. Detailed information on compilation and sources for every individual codelist is available at https://codelists.opensafely.org/ and the lists are available for inspection and re-use by the broader research community.

Missing data

Individuals with missing ethnicity, IMD and region were included as “Unknown”.

Statistical methods

Simple descriptive statistics were used to summarise the counts and rates of COVID-19 vaccine breakthrough. Rates for each outcome were estimated by dividing the count by person-years, with 95% confidence intervals. Counts and rates were stratified by initial priority groups for all outcomes and by selected clinical and demographic groups.

Software and reproducibility

Data management and analysis was performed using the OpenSAFELY software libraries, Python 3 and R version 4.0.2. All code for the OpenSAFELY platform for data management, analysis and secure code execution is shared for review and re-use under open licenses at GitHub.com/OpenSAFELY. All code for data management and analysis for this paper is shared for scientific review and re-use under open licenses on GitHub (https://github.com/opensafely/covid-19-vaccine-breakthrough).

Patient and public involvement

Any patient or member of the public is invited to contact us at https://opensafely.org/ regarding this study or the broader OpenSAFELY project.

Results

Study population

Out of approximately 24 million patients, 15,501,550 individuals were identified as being fully vaccinated against COVID-19 and included in this study (Fig. 1). The median follow-up time was 149 days (interquartile range: 107–179 days). Since being fully vaccinated, 8,370,837 (54%) of the base population had at least one record for a (positive/negative) SARS-CoV-2 test recorded, with a positivity rate of 3.17%. Testing behaviours varied between priority groups with individuals in care homes testing more regularly than other groups; 91% of care home residents had 3+ SARS-CoV-2 tests since being fully vaccinated vs 41–73% in other groups. The total number of COVID-19 vaccine breakthrough cases was 590,279 (<4%). Table 2 shows the number and rate of individuals fully vaccinated broken down by initial JCVI priority groups, along with the number and rate (per 1000 patient years) of each outcome.

Fig. 1
figure 1

Inclusion/exclusion flowchart

Full size image
Table 2 Number of individuals fully vaccinated (2 doses + 2 weeks) in initial priority groups in OpenSAFELY-TPP, and number and rate with each outcome
Full size table

Positive SARS-CoV-2 test

In fully vaccinated individuals, the median number of days to a positive test for SARS-CoV-2 was 99 (IQR 63–134 days) with a total of 579,780 individuals testing positive for SARS-CoV-2 (about 1 in 25 or <4%). For every 1000 years of patient follow-up time, the corresponding incidence rate was ​98.06 (95% CI 97.93–98.19). Within the initial JCVI priority groups, positive SARS-CoV-2 test rates were highest in the CEV group (107.84, 95% CI 107.27-108.42) and lowest in those over 80 years of age (26.69, 95% CI 26.48-26.90). The overall cumulative incidence of positive SARS-CoV-2 tests at 275 days from the study start date was <0.1% (Fig. 2).

Fig. 2
figure 2

Kaplan-Meir plot for positive SARS-CoV-2 test over time, by priority group

Full size image

When broken down into clinical/demographic groups and comorbidities (Table 3), rates of individuals testing positive for SARS-CoV-2 were highest in 40–40 years old (179.96, 95% CI 179.50–180.42) and lowest in individuals over 80 years of age (28.90, 95% CI 28.69–29.11). Rates were higher in females than in males: 102.60 (95% CI 102.42–102.78) vs 92.76 ( 95% CI 92.58–92.95), respectively. Comorbidities with the highest rates of positive SARS-CoV-2 tests included kidney transplant, dialysis and immunocompromised: 143.51 (95% CI 136.46–150.93), 134.75 (95% CI 124.32–146.06), and 91.53 (95% CI 90.91–92.15), respectively. Rates of positive SARS-CoV-2 tests were lowest 4–8 weeks after being fully vaccinated and highest at 12+ weeks: 48.65 (95% CI 47.97–49.33) vs 100.3 (95% CI 100.16–100.43), respectively.

Table 3 Count and rates of breakthrough positive SARS-CoV-2 tests and hospitalisation and death in fully vaccinated individuals in OpenSAFELY-TPP, broken down by selected clinical and demographic groups
Full size table

COVID-19-related hospital admission

From within the fully vaccinated population, 28,580 had a COVID-19-related hospital admission (about 1 in 550 or 0.18%), with a median time to hospitalisation of 143 days (IQR 102–174). Of those who had a COVID-19-related hospital admission, 18,800 (66%) had a positive SARS-CoV-2 test in their records; 9170 (49%) occurred prior to an individual being hospitalised, 7030 (37%) occurred within the first 2 days of their hospital admission, 2365 (13%) occurred between 3 and 29 days after being hospitalised and 235 (1%) occurred 30 days or more after an individual had been hospitalised.

Rates of COVID-19-related hospital admissions increased with age: 1.09 (1.84–1.96) for 16–29 years old vs 12.75 (12.61–12.89) for those over 80 years, respectively. Rates were higher in those who were more deprived; most deprived IMD quintile versus least deprived IMD quintile was 7.32 (7.23–7.41) vs 3.55 (3.5–3.61), respectively. Comorbidities with the highest rates of COVID-19-related hospital admissions included kidney transplant, dialysis, and chronic kidney disease: 76.08 (95% CI 71.03–81.49), 70.73 (95% CI 63.34–78.99), and 49.49 (95% CI 45.33–54.02), respectively.

COVID-19-related critical care admission

Of the 28,580 COVID-19-related hospital admission, 1980 needed to be admitted to critical care (about 1 in 1000 of fully vaccinated people or 1 in 14 of hospitalised patients), with a median time to critical care admission of 136 days (IQR 94-166). For every 1000 years of patient follow-up time, the corresponding incidence rate was ​0.33 (95% CI 0.32–0.34). Within the initial JCVI priority groups, COVID-19-related critical care admissions rates were highest in the CEV group (1.73, 95% CI 1.66–1.80) and lowest in health care workers 0.10 (95% CI 0.08–0.12).

Rates of COVID-19-related critical care admissions were higher in males than females and in; 0.45 (95% CI 0.44–0.46) vs 0.23 (95% CI 0.22–0.24), respectively. Rates were higher in those who were obese compared to not obese; BMI >30 (kg/m2) versus BMI <30 (kg/m2) was 0.68 (0.66–0.71) vs 0.22 (0.21-0.23), respectively. Comorbidities with the highest rates of COVID-19-related critical care admissions included dialysis, kidney transplant and chronic kidney disease: 19.74 (95% CI 16.03–24.32), 17.86 (95% CI 15.5–20.57) and 5.31 (95% CI 4.06–6.93), respectively.

COVID-19-related death

Of those who were fully vaccinated, 6435 had a COVID-19-related death (about 1 in 2500 or 0.04%) with a median number of days to the death of 165 (IQR 129–188). While the majority (43%) of deaths occurred in the 80+ JCVI priority group, COVID-19-related death rates were three times as high amongst individuals living in care homes compared to those over 80 years living in private residences: 13.37 (95% CI 12.86–13.91) vs 4.64 (95% CI 4.56–4.73), respectively.

There were very few deaths in those under 30 years of age with rates of COVID-19-related death increasing with increased age: 0.02 (95% CI 0.02–0.03) for 16–29 years old vs 5.26 (95% CI 5.17–5.35) for those over 80 years, respectively. Comorbidities with the highest rates of COVID-19-related death included dialysis, transplant and chronic kidney disease: 25.71 (95% CI 21.42–30.86), 18.9 (95% CI 16.47–21.68), and 17.81 (95% CI 15.40–20.61), respectively.

Conclusions

Summary

This descriptive analysis in over 15 million people living in England shows that COVID-19 vaccine breakthrough is occurring; however, events are currently rare and mostly mild in nature. Nevertheless, some individuals are experiencing higher rates of serious illness and death due to COVID-19, such as those in care homes, with chronic kidney disease, on dialysis, with a transplant or who are immunosuppressed. It is important to note that this analysis is a simple descriptive piece of analysis; rates presented here are crude, and are subject to the limitations described below, making interpretation of potential differences in rates of outcomes, such as when stratified by time since vaccination, challenging. While it is possible to adjust rates (as demonstrated for patients with chronic kidney disease in Supplementary file 1) to help inform decisions around rollout of vaccine/booster programme for patients at high risk of adverse outcomes, it is important to stress that this study was not designed to estimate risk factors for COVID-19 vaccine breakthrough and thus cannot be used to answer why certain groups have higher rates of breakthrough cases than others or to estimate vaccine effectiveness.

Strengths and weaknesses

This study used large-scale, routinely-collected primary care records, linked to coronavirus testing surveillance, hospital, and death registry data. This allowed us to describe a substantial proportion of the English population in rich longitudinal detail and to detect variations in COVID-19 vaccine breakthrough cases, as early as possible.

We acknowledge several important limitations of these findings. First, even though the base population consisted of over 15 million fully vaccinated individuals the numbers of COVID-19 vaccine breakthrough cases were relatively small, especially for hospitalisations and deaths. This made comparisons between outcomes, specifically at selected clinical and demographic levels difficult, meaning rates could be imprecisely estimated. Second, due to the targeted rollout of the COVID-19 vaccination programme in England, this cohort represents mostly older and vulnerable populations. In addition, follow-up time is systematically different amongst individuals included in this study and no adjustment for this has been made. Third, asymptomatic testing patterns vary between individuals. Apart from health and care workers, asymptomatic individuals without any underlying health issues or comorbidities are less likely to get tested than those with underlying health issues or comorbidities (i.e., haemodialysis patients) who will undergo asymptomatic testing regularly. Most lateral flow tests conducted at home are not included in our data. The number of reported positive SARS-CoV-2 tests is therefore likely to be an undercount of SARS-CoV-2 amongst fully vaccinated individuals without any underlying health issues or comorbidities, which may have led to underestimation of the corresponding rates. Fourth, while individuals were excluded if they tested positive for SARS-CoV-2 or had been hospitalised or died due to COVID-19 within the two weeks after their second dose, we acknowledge that there is still the potential that some patients will have been included whose outcome was related to a previous infection pre second vaccination. In particular, this may affect some of the highest risk groups such as dialysis patients and those in care homes where there is a high amount of repeated screening in and outcomes stratified by time since vaccination. As a result, there may be a potential increase in the associated rates; as indicated by the high positive SARS-CoV-2 test incidence rate in those 0–4 weeks following full vaccination. Fifth, characteristics linked to COVID-19 vaccine breakthrough in fully vaccinated individuals may be reflective of higher infection rates regardless of vaccination in some groups, not because of vaccination (i.e. due to higher exposure due to behavioural differences) [18]. Finally, this study includes data from December 2020 until November 2021, during which time two different variants of SARS-COV-2 were dominating in England at different times; the Alpha variant (B.1.1.7) between January 2021 and May 2021, and the Delta variant (B.1.617.2) after June 2021 [19]. These have since been supplanted by other variants, in particular the Omicron variant. This is a descriptive study and does not aim to quantify the contribution of vaccine breakthrough due to the change in dominant variant from Alpha to Delta or the extent of waning in the months immediately following vaccination. Further studies are needed to assess vaccine waning and rates of breakthrough COVID-19 between different variants and in newer, more prevalent variants, such as Omicron (B.1.1.529).

Findings in context

There is strong evidence that the UK’s COVID-19 vaccination programme has reduced infection and severe outcomes in vaccinated individuals [4, 5]. However, breakthrough infections after vaccination against SARS-CoV-2 are increasingly reported [7] and there are concerns that the effectiveness of the vaccines may reduce over time, due to the vaccines being less effective against the delta variant, and/or the waning of the immune response [8]. While we did find that breakthrough infections were more frequent in the earlier-vaccinated individuals, potentially due to waning vaccine efficacy or impaired immune response, further studies are needed to gain a clearer picture of the long-term effectiveness of COVID-19 vaccines. In addition, with only a handful of studies investigating risk factors for post-vaccination infection [13, 16, 17], very little is known about how breakthrough COVID-19 varies between key clinical and demographic groups. Our findings on COVID-19 vaccine breakthrough are consistent with patterns observed worldwide; COVID-19 vaccine breakthrough cases are rare and mild. However, there are potentially several groups who are at higher risk of COVID-19 vaccine breakthrough including those in care-homes, with chronic kidney disease, on dialysis, with transplants or who are immunocompromised.

Conclusion

As of 1st November 2021, the majority of COVID-19 vaccine breakthrough cases in England were mild with relatively smaller numbers of fully vaccinated individuals being hospitalised or dying as a result of COVID-19. While these numbers are in line with expectations, and while follow-up time is systematically different amongst fully vaccinated individuals and variation in vaccination coverage between groups and regions will have many complex drivers, some potential differences in rates of breakthrough cases were identified in several clinical groups. The emergence of the Omicron variant of COVID-19 coupled with the number of positive SARS-CoV-2 tests still occurring is concerning and as numbers of fully vaccinated (and boosted) individuals increases and as follow-up time lengthens, so too will the number of COVID-19 breakthrough cases. Additional analyses are therefore, to assess vaccine waning and rates of breakthrough COVID-19 between different variants, and to enable identification of individuals at higher risk, who would require continued strict precautions, and additional vaccination.