key: cord-0785797-pdd5wepz
authors: Antonelli, M.; Penfold, R. S.; Merino, J.; Sudre, C. H.; Molteni, E.; Berry, S.; Canas, L. S.; Graham, M. S.; Klaser, K.; Modat, M.; Murray, B.; Österdahl, M. F.; Cheetham, N. J.; Drew, D. A.; Nguyen, L. A.; Capdeila, J.; Hu, C.; Selvachandran, S.; Polidori, L.; May, A.; Wolf, J.; Chan, A. T.; Hammers, A.; Duncan, E.; Spector, T.; Ourselin, S.; Steves, C. J.
title: Post-vaccination SARS-CoV-2 infection: risk factors and illness profile in a prospective, observational community-based case-control study
date: 2021-05-26
journal: nan
DOI: 10.1101/2021.05.24.21257738
sha: c9ba3843e3f25bd39840cb3754867cc1d67cc105
doc_id: 785797
cord_uid: pdd5wepz

Background: Both BNT162b2 and ChAdOx1 vaccines show good efficacy in clinical trials and real-world data. However, some still contract SARS-CoV-2 post-vaccination. This study identifies risk factors associated with SARS-CoV-2 infection at least 14 days after first vaccination and describes characteristics of post-vaccination illness. Methods: Cases were UK adults reporting post-vaccination SARS-CoV-2 infection between 8th December 2020 and 1st May 2021, reporting on the COVID Symptom Study app. We assessed the associations of age, frailty, comorbidity, area-level deprivation and lifestyle factors with infection (vaccinated cases vs. negative-vaccinated controls); and vaccination with illness profile (vaccinated cases vs positive-unvaccinated controls). Findings: Post-vaccination infection risk was substantially higher in older adults with frailty (OR= 2.78, 95% CI [1.98-3.89], p-value<0.0001) and in individuals living in most deprived areas (OR vs. intermediate group=1.22, 95%CI [1.04-1.43], p-value=0.01). Risk was lower in individuals with a healthier diet (OR=0.73, 95%CI [0.62-0.86], p-value<0.0001) and without obesity (OR=0.6, 95% CI [0.44-0.82], p-value=0.001). Vaccination was associated with reduced odds of hospitalisation (OR=0.36, 95%CI [0.28-0.46], p-value<0.0001), and high acute-symptom burden (OR=0.51, 95%CI [0.42-0.61], p-value<0.0001). In the 60+ age group, risk of >28 days illness was lower following vaccination (OR=0.72 , 95%CI [0.51-1.00], p-value=0.05). Most symptoms were reported less in positive-vaccinated vs. positive-unvaccinated individuals, except sneezing, which was more common post-vaccination (OR=1.24, 95%CI [1.05-1.46], p-value=0.01). Interpretation: Our findings highlight reduced symptom burden and duration in those infected post-vaccination. Whilst reassuring, our data should prompt efforts to boost vaccine effectiveness in at-risk populations; moreover, targeted infection control measures will still be appropriate to minimise SARS-CoV-2 infection.

Evidence before this study:

To identify existing evidence for COVID-19 infection post-vaccination, we searched PubMed for peer-reviewed articles published between December 1, 2020 and May 18, 2021 using keywords ("COVID-19" OR "SARS-CoV-2") AND ("Vaccine" OR "vaccination") AND ("infection") AND ("risk factor*" OR "characteristic*"). We did not restrict our search by language or type of 

This is the first observational study investigating characteristics of and factors associated with SARS-CoV-2 infection after COVID-19 vaccination. We found that vaccinated individuals with frailty had higher rates of infection after vaccination than those without. Adverse determinants of health (e.g. increased social deprivation, obesity or unhealthy lifestyle factors such as less healthy diet) were associated with higher likelihood of infection after vaccination. In comparison with unvaccinated individuals, individuals infected after vaccination had fewer symptoms of COVID- 19 , and more people were entirely asymptomatic. Fewer vaccinated individuals experienced five or more symptoms, required hospitalisation, and, in the older adult group, fewer had prolonged illness duration (symptoms lasting longer than 28 days).

Some individuals still contract COVID-19 after vaccination and our data suggest that frail older adults and those living in more deprived areas are more at risk. However, in most individuals illness appears less severe, with reduced need for hospitalisation and less risk of prolonged illness populations; moreover, targeted infection control measures will still be appropriate to minimise SARS-CoV-2 infection. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

Vaccination against SARS-CoV-2 is a leading strategy to change the course of the pandemic world-wide. The United Kingdom was the first country internationally to authorise a vaccine against SARS-CoV-2 (1), with three vaccines currently licensed: BNT162b2 ("Pfizer-BioNTech"), mRNA-1273 ("Moderna") and ChAdOx1 nCoV-19 ("Oxford-AstraZeneca"), each with good efficacy in Phase 3 clinical trials (2) (3) (4) (5) . As of 14 May 2021, ~36.3 million (69%) of the UK adult population had received at least one vaccination (6). Thus, UK data presents one of the first windows not only on real-world efficacy but also on the remaining challenges postvaccination.

Previous analysis of community-based individuals using the COVID Symptom Study app showed significant reduction in infection post-vaccination, from 12 days after first vaccine dose (7), findings recapitulated in a UK-based "real world" case-control study (8) . Similarly, national surveillance data from the first four months of Israel's national vaccination campaign showed two doses of BNT162b2 prevented symptomatic and asymptomatic infections, COVID-19 related hospitalisations, severe disease, and death (9) .

None-the-less, some individuals still contract COVID-19 after vaccination, and further virus variants may evolve with increased transmissibility (as with B 1.1.7 and B.1.617.2) (10, 11) .

Further, early data suggest that while COVID-19 is usually milder if contracted post-vaccination, mortality is still high in hospitalised individuals: data from the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC) from April 2021 showed mortality of 28% in individuals hospitalised with COVID-19 >21 days post-vaccination, similar to mortality rates during the first pandemic wave (March-April 2020) (12, 13) .

Identifying and protecting individuals at higher risk of post-vaccination infection will become increasingly salient as a greater proportion of the population is vaccinated. Groups at higher risk of SARS-CoV-2 infection before vaccine availability include frontline healthcare workers and individuals from areas of greater relative deprivation (likely reflecting increased exposure) (14, 15) ; and increasing age, male sex, multi-morbidity and frailty are associated with poorer COVID-19 outcomes (16) (17) (18) . However, to our knowledge there are no studies investigating risk factors for post-vaccination infection.

Individuals with COVID-19 have varying symptoms and differing clinical needs (19) . Elucidating symptom profiles in individuals with COVID-19 post-vaccination has clinical utility: facilitating identification of risk groups for intervention, predicting medical resource requirements and informing appropriate testing guidelines. Finally, some unvaccinated individuals with COVID-19 experience prolonged illness duration ('Long-COVID') (20) . Whether this risk is similar in individuals infected post-vaccination is unknown.

The aims of this study were:

1. To describe individual factors associated with SARS-CoV-2 infection at least 14 days after first vaccination, 2. To assess illness duration, symptom profile and disease severity in individuals with SARS-CoV-2 infection after first vaccination compared to unvaccinated individuals with SARS-CoV-2 infection.

This community-based case-control study used data from the COVID Symptom Study logged through a free smartphone app developed by Zoe Global (London, UK) and King's College London (London, UK) (21) . The app was launched in the UK on 24 March 2020, with nearly 4.5 million unique participants providing data by self-or proxy-report. At registration, each participant reported baseline demographic information (e.g., age, sex, ethnicity, whether a healthcare worker) geographic location, and information on health risk factors including comorbitidies, lifestyle, frailty and visits to hospital. Participants were encouraged to self-report their experience of prespecified symptoms daily, enabling prospective and longitudinal information on incident symptoms. Those experiencing new symptoms were prospectively encouraged to present for a SARS-CoV-2 test through local testing centres. All users were asked to record any COVID-19

testing results (whether prompted by the app or for other reasons), and from December 2020, any administered COVID-19 vaccine(s) and subsequent symptoms (21) .

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

This study used a nested sub-population of individuals who first reported a positive test for SARS-CoV-2 after SARS CoV-2 vaccination (cases) and two control populations matched 1:1 to the cases, all drawn from app-users.

Inclusion criteria for cases were: 1) age ≥18 years; 2) living in the UK; 3) first dose of a COVID-19 vaccine received between 8 December 2020 and 1 May 2021; 4) at least 14 days of ongoing app usage after vaccination; 5) either a positive RT-PCR or lateral flow antigen test (LFAT) at least fourteen days after receiving first vaccine dose and not after the second dose (if more than one test result was reported, the first positive test was selected), with no previous positive test prevaccination.

To identify risk factors for post-vaccination infection, controls were vaccinated UK adults reporting negative RT-PCR or LFAT before second vaccination, and until 14 May 2021 (date of data extraction) (CG-1), matched 1:1 with cases for date of testing post-vaccination, healthcare worker status, and sex (using a matching algorithm based on minimum Euclidean distance (22) between the vectors of these covariates, with sex as a binary variable multiplied by 100 to ensure balance between covariate strengths). If multiple negative tests were reported, the last test date was used for matching.

To compare symptoms of SARS CoV-2 infection pre-and post-vaccination we selected controls aged ≥18 years, living in the UK, who were unvaccinated and reported a positive SARS-CoV-2 test, regardless of symptoms (CG-2). Controls were matched 1:1 with cases for the day of positive testing, healthcare worker status, sex, BMI, and age, using the method above.

In addition, data from all app users reporting a RT-PCR or LFAT test within the study period (n=1,531,762) were processed to obtain weights for inverse probability of being vaccinated.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) To identify risk factors for testing positive vs. negative for SARS-CoV-2 infection at least 14 days after first vaccination we used case status (self-reported positive RT-PCR test or LFAT for SARS-CoV-2) as the outcome variable.

We assessed age, self-reported comorbidities (cancer, diabetes, asthma, lung disease, heart disease, and kidney disease) analysed individually as binary variables; dependency level (frailty), assessed by the PRISMA7 questionnaire (embedded in the app registration) (23, 24) and classified as a binary variable (PRISMA7 ≥ 3 = frail; PRISMA7 < 3 = not frail) (25) ; local area Index of Multiple Deprivation (IMD), a score, ranging from 1 (most deprived) to 10 (least deprived) providing an estimate of relative locality deprivation derived from postal code on registration, divided into low 

To compare COVID-19 symptoms and severity in vaccinated vs. unvaccinated individuals testing positive for SARS-CoV-2, we assessed: 1) disease severity, assessed as: asymptomatic/symptomatic; >5 symptoms/≤5 symptoms reported in the first week of illness (19) ; and self-reported presentation to hospital/no hospital presentation; 2) illness duration, assessed as duration <28days/ duration≥28days; 3) individual symptom reports. Exposure variable was vaccination status. For cases and CG-2, symptoms were considered within a window from three days before and up to fourteen days after test date for SARS-CoV-2 (see Supplementary Table 2 for complete list).

Data census was 14 May 2021. Data was extracted and preprocessed using ExeTera13, a Python library developed at KCL and openly available on GitHub (28) . Statistical analysis was run using Python 3.7 and the following packages: numpy v1.19.2, pandas v1.1.3, scipy 1.5.2, and statsmodels v0.12.1.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 26, 2021. ; https://doi.org/10.1101/2021.05.24.21257738 doi: medRxiv preprint Differences in proportions and means of covariates between cases and respective controls were assessed using Fisher's exact test for categorical variables and Wilcoxon's test for continuous variables. Univariate logistic regression models adjusted for age, BMI, and sex were used to analyse association between each comorbidity, frailty status, IMD tertile, and environmental and lifestyle factors with post-vaccine infection. As factors associated may differ by age group, all analyses were stratified by sex and age (younger adults: 18-59 years; older adults, ≥60 years).

Multivariable logistic regression adjusted for age, BMI, and sex was used to analyse the combined association of obesity, sedentary lifestyle, healthier diet, IMD tertile and frailty status with the probability of post-vaccine infection.

To examine whether health-conscious behaviour might explain the association of lifestyle factors and infection post-vaccination we adjusted models for reported individual adherence to mask wearing guidance during 2020. Finally, we examined models using inverse probability weighting (IPW) (29) to check for potential index event bias of vaccination using weights derived from probabilities of being vaccinated in the population active on the app in the study period (Supplementary Figure 2) .

Univariate logistic regression models adjusted by age, BMI and sex were used to assess association of individual symptoms, overall illness duration, and disease severity (outcomes), with vaccination status (exposure). Symptoms were examined if reported by >1% of all app users with a positive test. We also provide models additionally adjusting for frailty and comorbidities, given their association with the exposure (vaccination) and outcome (symptomatology) which therefore may confound any observed associations.

This study reports on BNT162b2 and ChAdOx1 vaccines only, as there were no positive cases who received the mRNA-1273 vaccine. Post-vaccination infection was similar in these two vaccine types therefore combined analysis was performed.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 

Funders did not have any role in design, analysis or interpretation of the data. Zoe Global, funded by DHSC, made the app available for data collection as a not-for-profit endeavour. Figure 1 shows time in days from the first or second dose and day of positive testing which also reflects changing incidence of COVID-19 infection in the UK population (30) . Table 1 shows demographic information of cases and controls for the risk factor analysis. More participants were female (69.5%). Cases were significantly younger and had higher BMI (p-value< 0.0001 for both variables). Asthma and lung disease were the most commonly reported comorbidities in both cases and controls; there was no significant difference in prevalence of comorbidities between cases and controls, overall or stratified by two age groups ( We also observed a modest but significant positive relationship between BMI and post-vaccination infection especially in younger adults (OR=1.02, 95% CI [1.01-1.03] per unit increase in BMI, p<0.0001).

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Table 4 ). Figure 2 shows the multivariable analysis considering non-obesity, non-sedentary behaviour, nonsmoking, healthy diet, IMD, and frailty, adjusted for age, BMI, and sex (see Supplementary Table   5 Table 5 ). Healthier diet and having non-obese weight were independently associated with reduced risk of post-vaccine infection in younger adults. The above findings were consistent in a sensitivity analysis using inverse probability weighting for factors influencing vaccination (see Supplementary Tables 6 and 7 for univariate and multivariate analysis results, respectively).

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) 

2188 cases met the inclusion criteria and were followed-up for at least 14 days post-vaccination (median duration of follow-up is 77 days, first and third quartile equal to 41 and 101 days, respectively). These were matched 1:1 to 2188 unvaccinated controls using nearest Euclidean distance as per methods, which resulted in a ~20% preponderance of health care workers in the case group ( Figure 4 and Supplementary Table 9 ).

Analyses were unchanged after adjustment for comorbidities and frailty (Supplementary Tables 9   and 10 ).

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ;

age, BMI, sex, frailty and comorbidity status.

As countries around the world vaccinate their populations against COVID-19, there is growing interest in understanding risk factors for and characteristics of post-vaccination infection. Health policies and resource planning in the post-vaccination era need to be guided by evidence of who is most at risk of infection, to reduce morbidity and mortality whilst considering the adverse effects of shielding and isolation on quality of life and health. Here, we present data on 2,394 communitybased adults in the UK who developed test-confirmed SARS-CoV-2infection more than 14 days after first vaccination with either BNT162b2 or ChAdOx1. By 14 days, immunity is starting to develop (31); and infection is unlikely to be due to exposure peri-vaccination (e.g. during travel to the vaccination centre).

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021.

Frail individuals had higher odds of post-vaccine COVID-19 infection than matched controls, highlighting the need for ongoing caution in this vulnerable group. The association was neither diminished by inverse probability weighting for factors influencing vaccination, nor when adjusting for possible confounders such as local area deprivation or lifestyle factors. This finding may, in part, reflect increased exposure: whilst robust older adults may shield (see below), frailer adults may need home care visits or to attend healthcare facilities. Frail adults residing in longterm care are at particular risk of transmission of respiratory illness, and such facilities have been disproportionately affected throughout the pandemic (32) . A second explanation may be altered immune function, or "immunosenescence", a well-established feature of physiological ageing, thought responsible (at least in part) for increased incidence and severity of infection with increasing biological age (33) . Immunosenescence may also be responsible for previously observed ageing-associated decline in immunogenicity following vaccination for other infections (34) . The increased odds of post-vaccine infection in frailer adults may be compounded by more severe outcomes of COVID-19 infection in this group, including delirium (24) and death (17) . NICE recommend systematic frailty assessment in acute settings,

(https://www.nice.org.uk/guidance/NG159), and this should be extended to community settings to facilitate a differential, targeted re-vaccination scheduling, appropriate isolation precautions, case detection, testing and proactive care in this risk group. Research on augmentation of the immune response to vaccination in this group is urgently needed; for example, the impact and timing of booster vaccinations, and appropriate timing for this, could be assessed.

We found an inverse association of age on odds of post-vaccination infection, especially in older adults. Reduced risk of infection in older adults is consistent with previous studies in nonvaccinated individuals showing lower antibody seroprevalence in older adults (35) ; this likely reflects shielding in this age-group in accordance with government classification of over 70s as clinically vulnerable (36). We found some evidence that kidney disease may increase the odds of infection post-vaccination. This is potentially important, as patients with kidney disease were under-represented in the Phase 2 and Phase 3 trials of vaccines (37) . This may reflect increased exposure (e.g. unavoidable attendance at dialysis, despite otherwise shielding), or may reflect impaired immune responses to vaccination in these individuals as already observed for other infections (38, 39) . We also note early evidence suggesting reduced immune response in . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; https://doi.org/10.1101/2021.05.24.21257738 doi: medRxiv preprint individuals taking anti-cancer therapy, with suggestions that augmentation with a third vaccine may be appropriate (40) . In our cohort, neither report of cancer nor other comorbidities were significantly associated with an increased odds of infection. While this is reassuring, given that the comorbidities assessed here conferred higher risk of severe disease, hospitalisation, mechanical ventilation, and mortality from COVID-19 (16, 41) , as with age, ongoing shielding behaviours in these groups may be influencing our results.

Greater area level deprivation was strongly associated with risk of post-vaccination SARS-CoV-2 infection, despite correction for other risk factors. This association persisted following adjustment for compliance with guidance on infection control (specifically, mask-wearing). People living in areas of higher deprivation had increased odds of testing positive for COVID-19 relative to those lower deprivation (42) . Associated factors such as higher population density, and more ethnically diverse populations are all also associated with higher mortality from COVID-19 (43, 44) .

Individuals in more deprived areas also have lower vaccination coverage for COVID-19 (45), and therefore our finding may reflect increased viral transmission in these areas. The current study highlights that disproportionate risk in high deprivation areas may still apply post-vaccination; health policies to mitigate infection will need to be targeted to these areas.

Conversely, this study found that individuals with healthier lifestyles had lower risk of infection post-vaccination. In particular, people reporting a healthier diet were observed to have lower risk of post-vaccination infection. In addition, those without obesity had reduced risk. BMI has been associated with increased risk of hospitalisation and death from COVID-19 (44) . The current study suggests that immune responses post-vaccination may also be influenced by diet quality and obesity. This may in part be due to unadjusted confounding with protective healthy behaviours, including continued adherence to infection control advice post-vaccination; however, the effect was robust when adjusted for compliance with mask-wearing guidance.

Comparing symptoms in vaccinated vs. unvaccinated adults with COVID-19, almost all individual symptoms of COVID-19 disease were less common in the vaccinated population. The exception was sternutation (sneezing), which was reported more commonly in younger adults who contracted infection post-vaccination. We are unaware of previous reports of sneezing being more common post-vaccination for other respiratory illnesses but it is a well-recognised symptom of both . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. We have previously shown that experiencing more than 5 symptoms in the acute period was associated both with severity of disease (19) and duration of symptoms (20) . In the 60+ group, we found lower risk of symptoms lasting for more than 28 days (OR=0.72, 95%CI [0.51-1.00]). Put together, this suggests that severe acute disease is likely to be less common in those who are vaccinated. In the age group most commonly affected by Long-COVID, this consequence may also be less prevalent. However, while absolute numbers of hospitalisations may reduce, early data suggests that mortality in hospitalised patients who contracted COVID-19 after vaccination remains high (12) . Our data would suggest that this may relate to increased risk of post-vaccination infection in older people who are frail.

This study used data from a large population of individuals reporting on a mobile application. This population, while large, was disproportionately female and under-represented individuals of lower socio-economic status as indicated by the skew toward people living in less deprived areas (Table   1) .

Information was self-reported and therefore recording of comorbidities and test results may not be completely accurate. However, previous data from this study have concurred well with population-. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; https://doi.org/10.1101/2021.05.24.21257738 doi: medRxiv preprint based COVID-19 studies (47) , including the influence of socio-demographic factors (42) . A strength of the mobile data collection method is the ability to collect daily information prospectively, on a comprehensive set of symptoms, allowing analysis of both individual symptoms and overall illness duration. We acknowledge also that by virtue of data censoring dates, symptom duration may be underestimated in both cases and controls, as some individuals only had two weeks of logging after their positive test result .

The design of our study, including matching cases and controls for health-care worker status and time of infection, reduces potential for collider bias, although small differences between the groups remained on matched variables. Risk of reporting a positive SARS-CoV-2 test is higher amongst frontline healthcare workers vs. the general population (14) , reflecting exposure; andappropriately -healthcare workers were prioritised for vaccination in the UK (48). Our data suggests risk of post-vaccination SARS-CoV-2 infection is reduced in older age groups. In order to examine the effect of age on post-vaccination infection, we did not match CG1 by age. However, age was included as a covariate in all analyses other than that looking at effects of age itself, and stratified analyses are presented in two age-groups. While vaccination itself might be considered a potential index event bias, the population of interest in this study is the vaccinated population, and should not be construed as applying to those unvaccinated. Nevertheless, we examined and found no evidence of event bias based on probability of being vaccinated.

Frailty was assessed with the PRISMA-7 questionnaire for app usage. This assessment correlates well with other frailty measures (49) and has the advantage of focusing on functional consequences of frailty, not routinely captured in health records. However, PRISMA-7 has only been validated in older adults; results in younger adults should be interpreted cautiously (23) .

Finally, this study was conducted at the beginning of the post-vaccination period, at a time when incidence of SARS-CoV-2 infection in the UK was rapidly falling. In accordance with recently published ISARIC data, we saw declining incidence of new infections with time after vaccination (12) , which may reflect both increasing immunity and dropping incidence in the population. Our findings may not apply at all time points post-vaccination. Lastly, the small number of individuals who had received a second vaccination precluded study of post-vaccination infection after more than one dose.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; https://doi.org/10.1101/2021.05.24.21257738 doi: medRxiv preprint

We investigated factors associated with increased risk of infection after vaccination and found a substantial increased risk in frail older adults and in individuals living in more deprived areas, and a lower risk of infection in non-obese people and those who reported better diet quality. We found most symptoms post-vaccination were reported less in vaccinated people, except sneezing. Need for hospital assessment was less, and burden of acute symptoms was lower. For older adults, risk of prolonged illness was lower. Our findings may inform policy in the post-vaccination era, in particular with regards to protecting frail older adults, and those individuals living in areas of . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted May 26, 2021. ; . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. [4] [5] [6] [7] . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

Diet was assessed using information obtained from an amended version of the Leeds Short Form Food Frequency Questionnaire that included 27 food items (50) . Participants were asked how often on average they had consumed one portion of each food in a typical week during the month just prior (July 2020) to when they filled out the diet and lifestyle questionnaire. The responses had eight frequency categories ranging from "rarely or never" to "five or more times per day". A healthy diet pattern was ascertained using the Diet Quality Score (DQS), a validated score for adherence to UK dietary guidelines (50) . The DQS was computed from five broad categories including fruits, vegetables, total fat, oily fish, and non-milk extrinsic sugars. Each component was scored from 1 (unhealthiest) to 3 (healthiest) points, with intermediate values scored

proportionally. All component scores were summed to obtain a total score ranging from 5 (lowest diet quality) to 15 (highest) points. We defined a healthier diet pattern as a DQS in the top quartile of the score distribution (score>=12points). To generate the lifestyle scores, the participants received 1 point for each healthy lifestyle factor. The sum of these four scores together gave a healthy lifestyle score ranging from 0 to 4, with higher scores indicating a healthier lifestyle.

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted May 26, 2021. ; https://doi.org/10.1101/2021.05.24.21257738 doi: medRxiv preprint

The UK has approved a COVID vaccine -here's what scientists now want to know

Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine

Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine

Safety and immunogenicity of ChAdOx1 nCoV-19 vaccine administered in a prime-boost regimen in young and old adults (COV002): a single-blind, randomised, controlled, phase 2/3 trial. The Lancet

Safety and efficacy of the ChAdOx1 nCoV-19 vaccine (AZD1222) against SARS-CoV-2: an interim analysis of four randomised controlled trials in Brazil, South Africa, and the UK. The Lancet

Vaccine side-effects and SARS-CoV-2 infection after vaccination in users of the COVID Symptom Study app in the UK: a prospective observational study. The Lancet Infectious Diseases

Effectiveness of the Pfizer-BioNTech and Oxford-AstraZeneca vaccines on covid-19 related symptoms, hospital admissions, and mortality in older adults in England: test negative case-control study

Impact and effectiveness of mRNA BNT162b2 vaccine against SARS-CoV-2 infections and COVID-19 cases, hospitalisations, and deaths following a nationwide vaccination campaign in Israel: an observational study using national surveillance data. The Lancet

Changes in symptomatology, reinfection, and transmissibility associated with the SARS-CoV-2 variant B.1.1.7: an ecological study

SAGE 89 minutes: Coronavirus (COVID-19) response, 13

Calum Semple Hospitalised vaccinated patients during the second wave, update April '21. UK

Changes in hospital mortality in the first wave of COVID-19 in the UK using the ISARIC WHO Clinical Characterisation Protocol: prospective observational cohort study

Risk of COVID-19 among front-line health-care workers and the general community: a prospective cohort study

Risk factors for SARS-CoV-2 among patients in the Oxford Royal College of General Practitioners Research and Surveillance Centre primary care network: a cross-sectional study. The Lancet Infectious Diseases

Comorbidity and its impact on 1590 patients with COVID-19 in China: a nationwide analysis

The effect of frailty on survival in patients with COVID-19 (COPE): a multicentre, European, observational cohort study. The Lancet Public Health

Comparison of two different frailty measurements and risk of hospitalisation or death from COVID-19: findings from UK Biobank

Symptom clusters in COVID-19: A potential clinical prediction tool from the COVID Symptom Study app

Attributes and predictors of long COVID

Real-time tracking of selfreported symptoms to predict potential COVID-19

A Euclidean distance-based matching procedure for nonrandomized comparison studies

PRISMA-7: a case-finding tool to identify older adults with moderate to severe disabilities

Probable delirium is a presenting symptom of COVID-19 in frail, older adults: a cohort study of 322 hospitalised and 535 community-based older adults

Predictive validity of PRISMA-7 as a screening instrument for frailty in a hospital setting

Lifestyle-based prediction model for the prevention of CVD: the Healthy Heart Score

Accessible Data Curation and Analytics for International-Scale Citizen Science

Inverse probability weighting

COVID-19) in the UK: GOV.UK 2021

Spike-antibody responses following first and second doses of ChAdOx1 and BNT162b2 vaccines by age, gender, and clinical factors -a prospective community cohort study (Virus Watch)

Pietro, Economics) PLWaL. COVID-19 mortality and long-term care: a UK comparison

Innate immunesenescence: underlying mechanisms and clinical relevance

Aging and the Immune System: the Impact of Immunosenescence on Viral Infection, Immunity and Vaccine Immunogenicity

):905. 36. The Health Protection (Coronavirus, Restrictions) (England) Regulations 2020, Regulations made by the Secretary of State, laid before Parliament under section 45R of the Public Health

Systematic Review of Safety and Efficacy of COVID-19 Vaccines in Patients With Kidney Disease

Practical Guide to Vaccination in All Stages of CKD, Including Patients Treated by Dialysis or Kidney Transplantation

Immune Dysfunction and Risk of Infection in Chronic Kidney Disease

Immune Responses to COVID-19 mRNA Vaccines in Patients with Solid Tumors on Active, Immunosuppressive Cancer Therapy

Presenting Characteristics, Comorbidities, and Outcomes Among 5700 Patients Hospitalized With COVID-19 in the New York City Area

Geo-social gradients in predicted COVID-19 prevalence in Great Britain: results from 1 960 242 users of the COVID-19 Symptoms Study app

Coronavirus disease 2019 mortality: a multivariate ecological analysis in relation to ethnicity, population density, obesity, deprivation and pollution

Factors associated with COVID-19-related death using OpenSAFELY

Sneeze reflex: facts and fiction

Detecting COVID-19 infection hotspots in England using large-scale self-reported data from a mobile application: a prospective, observational study. The Lancet Public Health

Comparison of Frailty Screening Instruments in the Emergency Department

Can a dietary quality score derived from a short-form FFQ assess dietary quality in UK adult population surveys? Public Health Nutr