The Importance of Consistency and Correct Data Sources when Making a Comparison Between Post-COVID Rates Among Vaccinees and Post-COVID Rates Among Unvaccinated Individuals

by Anette Stahel, MSc

Summary

On October 13, 2023 the meta-analysis The Effectiveness of COVID-19 Vaccine in the Prevention of Post-COVID Conditions: A Systematic Literature Review and Meta-analysis of the Latest Research was published in Antimicrobial Stewardship & Healthcare Epidemiology. Unfortunately, the analysis includes several inconsistencies and errors which make the comparison between post-COVID rates among vaccinees and post-COVID rates among unvaccinated individuals in it incorrect. I'll here explain how come being consistent in your wording and using highly adequate infection figures when conducting such a comparative study is of utmost importance. I'll also bring to light a completely central post-COVID figure regarding the group of vaccinees with a previous infection, a figure which the authors of the analysis have left out.

Introduction

On October 13, 2023 the meta-analysis The Effectiveness of COVID-19 Vaccine in the Prevention of Post-COVID Conditions: A Systematic Literature Review and Meta-analysis of the Latest Research, authored by Marra et al, was published in Antimicrobial Stewardship & Healthcare Epidemiology [1]. It's a large investigation, comprising 32 studies, which evaluates the effect of vaccination on the incidence of the prolonged, painful and debilitating form of COVID-19 called post-COVID. The analysis is very well-cited by other researchers and has been paraphrased in longer articles in reputable magazines such as Scientific American as well as on respected websites such as Medscape [2, 3]. It's also been referred to in advisory texts on post-COVID by the national public health agency of the USA, the CDC [4].

I've now gone through and reviewed this paper and I'm sorry, but this meta-analysis is not correct. Unfortunately, it includes several inconsistencies and errors which make the comparison between post-COVID rates among vaccinees and post-COVID rates among unvaccinated individuals in it inadequate. It also omits to mention and discuss a completely central post-COVID figure regarding the groups of vaccinees with a previous infection in the studies. I'll here explain how come being consistent in your wording as well as using highly adequate infection figures when conducting an analysis like this is of utmost importance. I'll also bring the omitted figure to light and discuss it. 

The importance of consistency regarding the objective and of bringing central figures to light

I'm going to start by quoting the researchers' objective, as this will later be essential to my review. Their aim was to carry out:

"(...) a systematic literature review and meta-analysis on the effectiveness of coronavirus disease 2019 (COVID-19) vaccination against post-COVID conditions (long COVID) among fully vaccinated individuals."

Thus, it's from the beginning clear that it isn't the authors' objective to focus primarily on people who were vaccinated before they contracted the infection, nor that those who were vaccinated after they contracted it would somehow be secondary in the investigation. However, the authors do not stick to their objective throughout the analysis, but digress very soon by withdrawing their focus from the vaccinees with a previous infection and instead focus almost entirely on the vaccinees without one. The expression they use when describing their conclusion regarding the former is ambiguous, "(...) analysis demonstrated no protection against post-COVID conditions among those who received COVID-19 vaccination after COVID-19 infection", and with continued reading it becomes apparent that quite a bit of data has been ignored here. 

When looking closer at this ignored data, I found something highly concerning. That is, all taken together, the figures in the studies showed that vaccination after infection significantly increased the risk of post-COVID conditions, instead of demonstrating a "lack" of protection, as the authors dubbed it [5-9]. On average, the studies showed that if you followed the health authorities' recommendation and got the vaccine despite having been infected (a questionable advice by the way, as you're already equipped with a strong immunity by then), this increased the risk of post-COVID by as much as 37.2% compared to the unvaccinated. In the largest of these studies, in which over 300,000 people who took the vaccine after infection participated, the post-COVID risk increased by a remarkable 105.8% compared to those who declined vaccination [8].

Now worryingly enough, the analysis authors' comment on this post-COVID increase for those who had the infection before vaccination is to ignore it, with the words "(...) it was not possible to estimate VE because it did not prevent post-COVID condition". But that's incorrect, because what happens when you demonstrate an increase in the risk of an infection, is that the vaccine effectiveness (VE) becomes negative - it doesn't just disappear. The researchers here push aside an enormous group of billions of people around the world, who got their vaccine after having had the COVID infection [10]. This part of the paper isn't only inadequate, discriminatory and inconsistent (since it's contradictory to the stated purpose of the analysis), it's also dangerous, because this increase in the risk of post-COVID when getting vaccinated after infection needs attention. Above all, the health authorities immediately need to be informed of this finding so that they, instead of recommending the vaccine as a post-COVID profylaxis, begin warning those who've had the infection but still want vaccination that their choice carries a marked risk increase for post-COVID with it. An explanation from the authors to how come they chose not to shed any light on this significant, negative VE for those vaccinated following infection would be much appreciated.

Additionally, via the text in Figure 1, we learn that studies were excluded from the final analysis if they didn't report an absolute number of people without a post-COVID condition. And via Table 1, we learn that there were 25 studies which reported absolute numbers of people without post-COVID, as well as absolute numbers with post-COVID, in both the vaccinated and unvaccinated groups. However, the authors of the meta-analysis chose to examine only 24 of these; the one by Arjun et al was left out. Why? It isn't clear how many of the participants of that study were vaccinated before vs after infection, but, another study, the one by Hajjaji et al in which this isn't clear either, was still included in the analysis. It should be noted that this included Hajjaji report showed higher risk figures for the unvaccinated compared to the vaccinated, while the excluded Arjun study on the other hand showed a much higher post-COVID risk in the fully vaccinated compared to the unvaccinated, and if Marra et al had included it in the calculations, it would've reduced the reported VE for the total group of vaccinated people. An explanation from the authors to how come they chose to exclude the Arjun study would be very welcome. 

The importance of correct data sources in relation to the wording of the text

With further reading, it becomes even clearer that this analysis mainly revolves around the group of people who got vaccinated with two doses prior to contracting the COVID infection. And according to the authors, this group had an increased protection against post-COVID compared to the unvaccinated group, with a VE of 36.9%. However, I object to this conclusion, for a number of reasons. 

First of all, the researchers behind the various studies investigated in the analysis had, in violation of scientificity but nevertheless very common among COVID vaccine studies, cut data out for several of the weeks that occurred immediately after vaccination. For instance, in the study by Ayoubkhani et al and the one by Mohr et al, all participants who got their COVID (which later developed into post-COVID) after their first dose were excluded, and those who got it right after their second dose as well; only those who'd been two-dose vaccinated for more than two weeks were included in the investigations. And in the study by Kahlert et al, all participants who contracted their infection one week after an injection were excluded, regardless of the number of the dose. It should be pointed out that the studies upon which Marra et al based a previous meta-analysis mentioned in the text, on the effect of one dose of the vaccine on the incidence of post-COVID, had cut weeks of data out in this manner as well.

So in these reports, they'd excluded data from the vaccine groups by ignoring infections which started the days and weeks right after the injections. But the fact is, that precisely during these first weeks after an injection, if the spread of infection is high, the cases of COVID sharply rise [11-15]. The clearest example of this is a study by the British Government's emergency advisory group, where the VE had plunged to a staggering -431% already the day after the first vaccination [11]. This early spike in COVID cases is mainly observed following the first dose, but to some extent after the other doses as well, and it's been suggested to be caused by the vaccine's leukocyte-weakening effect in combination with a biological response called immune triage (or imbalance effect) [15-18]. The motive for this exclusion is always said to be that the antibody levels haven't properly risen until after this time period, it's claimed that this cutout of data is irrelevant, that the susceptibility to COVID still is the same as in unvaccinated people at that time anyway. But that's incorrect, since as said, many studies have shown that the number of COVID cases spikes during the weeks after vaccination.

But not only that, the authors of these studies had also cut out all data which started to appear around six months after the vaccination, which is a critical period when the COVID VE turns negative again and then continues downward. Many studies have shown this later spike in COVID cases among vaccinated as well [19-23]. An eye-opening example here is a study encompassing all health system employees in all clinics of Cleveland, USA, which during a time period of six months compared the VE of one-dose-, two-dose-, three-dose- and three+-dose-vaccinated to the protection level of the unvaccinated. To the authors' surprise, all vaccine dose levels showed negative VE figures, and paradoxically enough, the figures also revealed that the more doses taken, the lower the VE: -107%, -150%, -210% and -253%, respectively [21].

This later plunge is thought to be due to completely different causes than those of the early one though, of which primarily a combination of so-called original antigenic sin, OAS, and a condition called immune exhaustion, which occurs when the body repeatedly is exposed to the same type of antigens, has been suggested [24, 25]. A third mechanism that's been suggested is ADE, antibody-dependent enhancement [26]. Interestingly enough, a few of the studies included in the analysis had actually at least examined the time around the six-month mark, and in their figures and tables you can clearly see how the post-COVID incidence among the vaccinated started to rise above the level of the unvaccinated at this point, in one of them already around the four-month mark [27-29].

Once again, the purpose of the present meta-analysis was to examine "the effectiveness of coronavirus disease 2019 (COVID-19) vaccination against post-COVID conditions (long COVID) among fully vaccinated individuals". However, when people get vaccinated, they step aboard a roller-coaster of both negative and positive effects against the microbe in question, which in its entirety lasts for many months after the vaccination, sometimes even years, before susceptibility to the disease returns to the level of the unvaccinated. The VE doesn't merely increase, in the sense that you only get a positive, protective effect from the vaccine, but it also drops in different rounds, when the vaccine actually produces a negative effect that entails an increased risk of infection compared to unvaccinated people. And in the case of the COVID vaccine, the effectiveness thus drops sharply for a number of weeks immediately after the first dose (during which the protection becomes negative, i e lower than that of the unvaccinated) [11-15], then it rises for a number of months (it becomes positive) under influence by dose two, and then it drops again (becomes negative again) [19-23] and continues downward for several months, sometimes with short spikes, until it finally reaches a bottom level, after which it turns upward and returns to zero effect, corresponding to the protective effect you've got naturally as unvaccinated.

So, in the present meta-analysis, the effectiveness of the vaccine against post-COVID in fully vaccinated people was to be investigated. But if you put your agenda that way, with a specific wording like that, and if you can suspect that the number of post-COVID cases somewhat follows the number of infection cases - which you can, because it's logical to assume that increased COVID cases leads to increased post-COVID cases, but also by reading the three studies mentioned above [27-29] - then you have to examine this entire roller-coaster period of effectiveness that the fully vaccinated are subjected to. You cannot select out the months during this period when the VE is positive, and only for one of the sub-groups at that, and base your calculations merely on those. Doing so is simply incorrect and unscientific, and this, together with the other errors taken up above, should've resulted in the analysis being either rejected or suggested editing in the peer-review process, followed by either publication denial or a revision suggestion by the scientific journal to which it was submitted. I'm surprised the journal in question here chose not to do that but instead to publish this paper in its current condition.

Anyway, when you consider this VE roller-coaster ride which these people vaccinated prior to infection was subjected to, you understand that the VE figure of 36.9% that the authors of the analysis established for this group in reality was much lower. Of course those who take the vaccine after the infection go for the above VE roller-coaster ride also, which means that the negative VE figure of -37.2% discussed above, calculated for this group on basis of the data in the studies which reported absolute numbers of people with and without post-COVID for vaccine-after-COVID participants, in reality was much lower too - as well as the overall VE figure of 32% stated in the analysis. This ride applies to three- and four-dose-vaccinated as well by the way, and while these doses prolong the positive VE period, the following plunge into negative VE is the steeper [21, 30]. And this all means that the stated figures for total prevalence of post-COVID, 5.3% among the fully vaccinated and 11.8% among the unvaccinated, are incorrect as well. In reality, the former was higher and the latter lower. 

Further, it's true that if you were to return both the aforementioned early and this later risk increase data to the vaccine groups, this part of the analysis would be scientific and correct. Though if you did that, it's likely that its main conclusion, i e that full vaccination reduces the risk of post-COVID if it takes place before the first COVID infection, wouldn't hold up any longer. If data had been collected in a scientific way, that is, if the post-COVID frequency had been measured from the first day of injection and also included the frequency after about six months and onward, then in total, the vaccine groups in these studies of post-injection infection would've likely shown a higher risk of developing post-COVID than the unvaccinated groups. In light of this, and given the very wide spread of this meta-analysis among scientific publications - most importantly the CDC, the very epicenter of vaccine recommendations - it's worrying that the authors chose to include a vaccination advice towards the end of the text, namely this one: 

"(...) the vaccine should be offered to unvaccinated individuals who have not had COVID-19 yet."

The researchers' position here becomes even more contradictory when considering the following: The studies they themselves selected for the analysis show that vaccination after infection increased the risk of post-COVID by 37.2%, while vaccination before infection reduced the risk by a very similar figure, 36.9%. Although I, as stated above, strongly object to the authors' 36.9% figure, it's after all the one that they believe applies. When you think about that, given their advice to unvaccinated who haven't had COVID yet, they ought to give the opposite advice to unvaccinated who have had COVID, but for some reason they omit that. This is one more of the inconsistencies in this analysis and again, an explanation from the authors to how come they chose not to include such an advice to the latter group would be appreciated. 

Discussion and conclusion

Now if Marra et al stand firm in their aim to base their meta-analysis on the very limited data in these included studies, there's one way to do it while still being correct and scientific about it. And that's by, firstly, to include a substantial discussion about the vaccine-after-COVID group and its negative VE of 37.2%, and secondly, to use a different wording when describing their findings. Above all, they'd have to clearly underline the fact that the investigation merely concerned the time period between 1-5 (or more) weeks and 4-6 months after vaccination, and highlight the fact that the time periods before and after that were left out. The likelihood of finding very high frequencies of post-COVID cases among these left-out periods if they were to be examined would have to be emphasized also, as well as the likelihood that such an examination would change the total VE substantially, to the vaccine's disadvantage. For instance, a passage like this one...

"Vaccine effectiveness was 36.9% (...) among those who received two doses of COVID-19 vaccine before COVID-19 infection (...) The stratified analysis demonstrated no protection against post-COVID conditions among those who received COVID-19 vaccination after COVID-19 infection." 

...would have to be changed to, for example:

"The stratified analysis showed that during the time period between 1-5 weeks and 4-6 months after vaccination, vaccine effectiveness was -37.2% (...) among those who received two doses of COVID-19 vaccine after COVID-19 infection and +36.9% (...) among those who received two doses of COVID-19 vaccine before COVID-19 infection. That is, the vaccine increased the risk for post-COVID among those who received it after infection and reduced it among those who received it prior to infection. However, it's likely that the time periods before and after this window would show very high frequencies of post-COVID cases if they were to be examined, thus it's likely that such an examination would lower the VE substantially, in both groups." 

A revision and editing of this meta-analysis according to these suggestions is strongly recommended, though it should be said that while admittedly scientific and correct, such a limited analysis would still be potentially misleading and pointless, at least if the intended use was to provide guidance to the health authorities for vaccine recommendations. This is because a window like that fails to give us the big picture, a picture which in this case could very well show a negative VE net figure for post-COVID, even a significant such. And if so, from a post-COVID prophylactic point of view, instead of recommending the vaccine the health authorities would need to begin warning everyone asking for this vaccination that their choice carries a marked risk increase for post-COVID with it.

On the other hand, if Marra et al, the likely implication of a contrary main conclusion notwithstanding, were to realize that they wanted to revise their analysis in the more comprehensive way that I describe above, to attain a full view of the whole period during which the participants' VE in different rounds deviated both negatively and positively from the level of the unvaccinated, then they'd run into problems. And this problem consists of the fact that no studies of the post-COVID frequencies during the first weeks after vaccination, nor of after 4-6 months and onward, seem to be available, at least not at present. 

That no such studies exist is of course very concerning, but perhaps we can look forward to some publications of investigations of this type in the near future, especially considering the increased attention that post-COVID conditions have received in media during the last months? Perhaps critical post-publication peer reviews of post-COVID related vaccine analyses like this one will provide motivation as well, for researchers who strive to conduct vaccine examinations in a fully accurate and meaningful manner? I sincerely hope so, and if such studies do start to appear, my hope is that the authors of the present analysis will revise their investigation and provide us with an updated meta-analysis based on new, more genuine, all-encompassing and valuable data sources.

References

1. Marra, A R, Kobayashi, T, Callado, G Y, Pardo, I, Gutfreund, M C et al The Effectiveness of COVID-19 Vaccine in the Prevention of Post-COVID Conditions: A Systematic Literature Review and Meta-analysis of the Latest Research Antimicrob Steward Healthc Epidemiol 2023; 3(1): e168 https://doi.org/10.1017/ash.2023.447

2. Hall, S Vaccination Dramatically Lowers Long COVID Risk Scientific American; 2024 https://www.scientificamerican.com/article/vaccination-dramatically-lowers-long-covid-risk

3. Waldman, D Some Reasons to Get Off the Fence About COVID Booster Medscape; 2023 https://www.medscape.com/viewarticle/some-reasons-get-fence-about-covid-booster-2023a1000um4?form=fpf

4. Howard, J, Cloeren, M & Vanichkachorn, G Long COVID and Occupational Medicine Practice Centers for Disease Control and Prevention; 2024 https://blogs.cdc.gov/niosh-science-blog/2024/01/22/long-covid-om/ 

5. El Otmani, H, Nabili, S, Berrada, M, Bellakhdar, S, El Moutawakil, B & Abdoh Rafai, M Prevalence, Characteristics and Risk Factors in a Moroccan Cohort of Long-Covid-19 Neurol Sci Off J Ital Neurol Soc Ital Soc Clin Neurophysiol 2022; 43: 5175-5180 https://doi.org/10.1007/s10072-022-06138-0

6. Jassat, W, Mudara, C, Vika, C, Welch, R, Arendse, T et al A Cohort Study of Post-COVID-19 Condition Across the Beta, Delta, and Omicron Waves in South Africa: 6-month Follow-up of Hospitalized and Nonhospitalized Participants Int J Infect Dis Off Publ Int Soc Infect Dis 2023; 128: 102-111 https://doi.org/10.1016/j.ijid.2022.12.036

7. Kahlert, C R, Strahm, C, Güsewell, S, Cusini, A, Brucher, A et al Post-Acute Sequelae After Severe Acute Respiratory Syndrome Coronavirus 2 Infection by Viral Variant and Vaccination Status: A Multicenter Cross-Sectional Study Clin Infect Dis Off Publ Infect Dis Soc Am 2023; 77: 194-202 https://doi.org/10.1093/cid/ciad143

8. Meza-Torres, B, Delanerolle, G, Okusi, C, Mayor, N, Anand, S et al Differences in Clinical Presentation With Long COVID after Community and Hospital Infection and Associations With All-cause Mortality: English Sentinel Network Database Study JMIR Public Health Surveill 2023; 8(8): e37668 https://doi.org/10.2196/37668

9. Nehme, M, Braillard, O, Chappuis, F, Courvoisier, D S, Kaiser, L et al One-year Persistent Symptoms and Functional Impairment in SARS-CoV-2 Positive and Negative Individuals J Int Med 2022; 292: 103-115 https://doi.org/10.1111/joim.13482

10. Bergeri, I, Whelan, M G, Ware, H, Subissi, L, Nardone, A et al Global SARS-CoV-2 Seroprevalence from January 2020 to April 2022: A Systematic Review and Meta-analysis of Standardized Population-based Studies PLoS Med 2022; 19(11): e1004107 https://doi.org/10.1371/journal.pmed.1004107

11. Egan, C, Knight, S, Baillie, K, Harrison, E, Docherty, A & Semple, C Hospitalised Vaccinated Patients During the Second Wave, Update April ‘21 UK Scientific Advisory Group for Emergencies; 2021 https://assets.publishing.service.gov.uk/media/608ab0d3e90e076ab573a192/S1208_CO-CIN_report_on_impact_of_vaccination_Apr_21.pdf

12. Hunter, P R & Brainard, J Estimating the Effectiveness of the Pfizer COVID-19 BNT162b2 Vaccine after a Single Dose. A Reanalysis of a Study of ‘Real-world’ Vaccination Outcomes from Israel medRxiv 2021.02.01.21250957 https://www.medrxiv.org/content/10.1101/2021.02.01.21250957v1.full.pdf

13. Lopez Bernal, J, Andrews, N, Gower, C, Robertson, C, Stowe, J et al 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 BMJ 2021; 373: n1088 https://doi.org/10.1136/bmj.n1088

14. Rask Moustsen-Helms, I, Emborg, H-D, Nielsen, J, Finderup Nielsen, K, Grove Krause, T et al Vaccine Effectiveness After 1st and 2nd Dose of the BNT162b2 mRNA Covid-19 Vaccine in Long-term Care Facility Residents and Healthcare Workers - a Danish Cohort Study medRxiv 2021.03.08.21252200 https://www.medrxiv.org/content/10.1101/2021.03.08.21252200v1.full.pdf

15. Lataster, R Reply to Fung et al on COVID-19 Vaccine Case-counting Window Biases Overstating Vaccine Effectiveness J Eval Clin Pract 2023; 30: 82-85 https://doi.org/10.1111/jep.13892

16. Craig, C Rapid Response: Thinking Beyond Behavioural Change as an Explanation for Increased COVID Post Vaccination BMJ 2021; 372:n783 https://www.bmj.com/content/372/bmj.n783/rr

17. Jamieson, A M, Crane, M, Xu, Y & Lee, K Immune Triage: Prioritization of Host Immune Responses J Immunol 2016; 196 (1_Supplement): 197.20 https://doi.org/10.4049/jimmunol.196.Supp.197.20

18. Engwall, A-C & Engwall, B On the Imbalance Effect: Reply to an Article in Dagens Nyheter Published January 11, 2022 Facebook; 2022 https://www.facebook.com/anncathrin.engwall/posts/4647934981922327

19. Xu, Y, Li, H, Kirui, B, Santosa, A, Gisslén, M et al Effectiveness of COVID-19 Vaccines over 13 Months Covering the Period of the Emergence of the Omicron Variant in the Swedish Population Vaccines 2022; 10(12):2074 https://doi.org/10.3390/vaccines10122074

20. Holm Hansen, C, Blicher Schelde, A, Rask Moustsen-Helm, I, Emborg, H-D, Grove Krause, T et al Vaccine Effectiveness Against SARS-CoV-2 Infection with the Omicron or Delta Variants Following a Two-dose or Booster BNT162b2 or mRNA-1273 Vaccination Series: A Danish Cohort Study medRxiv 2021.12.20.21267966 https://www.medrxiv.org/content/10.1101/2021.12.20.21267966v3.full.pdf

21. Shrestha, N K, Burke, P C, Nowacki, A S, Simon, J F, Hagen, A & Gordon, S M Effectiveness of the Coronavirus Disease 2019 Bivalent Vaccine Open Forum Infect Dis 2023; 10(6): ofad209 https://doi.org/10.1093/ofid/ofad209 

22. Nordström, P, Ballin, M & Nordström, A Risk of Infection, Hospitalisation, and Death up to 9 Months after a Second Dose of COVID-19 Vaccine: A Retrospective, Total Population Cohort Study in Sweden Lancet 2022; 399(10327): 814-823 https://doi.org/10.1016/S0140-6736(22)00089-7

23. Lataster, R Rapid Response: We Need Proper Explanations for Apparent COVID-19 Vaccine Negative Effectiveness BMJ 2023; 381: e074404 https://www.bmj.com/content/381/bmj-2022-074404/rr-0

24. Aguilar-Bretones, M, Fouchier, R A M, Koopmans, M P G & van Nierop, G P Impact of  Antigenic Evolution and Original Antigenic Sin on SARS-CoV-2 Immunity J Clin Invest 2023; 133(1): e162192 https://doi.org/10.1172/JCI162192

25. Uversky, V N, Redwan, E M, Makis, W & Rubio-Casillas, A IgG4 Antibodies Induced by Repeated Vaccination May Generate Immune Tolerance to the SARS-CoV-2 Spike Protein Vaccines 2023; 11(5): 991 https://doi.org/10.3390/vaccines11050991

26. Shimizu, J, Sasaki, T, Koketsu, R, Morita, R, Yoshimura, Y et al Reevaluation of Antibody-dependent Enhancement of Infection in Anti-SARS-CoV-2 Therapeutic Antibodies and mRNA-vaccine Antisera Using FcR- and ACE2-positive Cells Sci Rep 2022; 12(1): 15612 https://doi.org/10.1038/s41598-022-19993-w

27. Hernández-Aceituno, A, García-Hernández, A & Larumbe-Zabala, E COVID-19 Long-term Sequelae: Omicron versus Alpha and Delta Variants Infect Dis Now 2023 53(5): 104688 https://doi.org/10.1016/j.idnow.2023.104688 (Table 3)

28. Pinato, D J, Ferrante, D, Aguilar-Company, J, Bower, M, Salazar, R et al Vaccination Against SARS-CoV-2 Protects from Morbidity, Mortality and Sequelae from COVID19 in Patients with Cancer Eur J Cancer 2022; 171: 64-74 https://doi.org/10.1016/j.ejca.2022.04.036 (Supplementary Figure 1) 

29. Taquet, M, Dercon, Q & Harrison, P J Six-month Sequelae of Post-vaccination SARS-CoV-2 Infection: A Retrospective Cohort Study of 10,024 Breakthrough Infections Brain Behav Immun 2022; 103: 154-162 https://doi.org/10.1016/j.bbi.2022.04.013 (Appendix Table A14)

30. Tseng, H F, Ackerson, B K, Bruxvoort, K J, Sy, L S, Tubert, J E et al Effectiveness of mRNA-1273 Vaccination Against SARS-CoV-2 Omicron Subvariants BA.1, BA.2, BA.2.12.1, BA.4, and BA.5 Nat Commun 2023; 14(1): 189 https://doi.org/10.1038/s41467-023-35815-7