1. Introduction
The high mortality from the COVID-19 pandemic led to a desperate search for effective vaccines and of necessity several were given temporary or emergency authorization. However, there is increasing awareness that vaccines exhibit specific and nonspecific effects in both humans and animals [
1,
2,
3,
4,
5,
6]. Thankfully most (but not all) of the nonspecific effects detected in humans and animals have led to wider beneficial effects against all-cause mortality [
1,
2,
3,
4,
5,
6]. Benn et al [
6] have argued that the existence of nonspecific effects has profound implications for the testing, approving, and regulation of vaccines. The specific effects are measured by the efficacy of the vaccine against the targeted pathogen, while the nonspecific effects can be discerned by evaluating the change in all-cause mortality. A fully efficacious vaccine will reduce deaths arising from the targeted pathogen but should have minimal adverse nonspecific effects.
There are two examples of the nonspecific effects of vaccines during COVID-19. During the early stages of the pandemic both influenza and BCG vaccination gave nonspecific protective effects against COVID-19 morbidity and mortality [
7,
8,
9,
10].
The nonspecific effects arise from the ability of pathogen antigens to cause polyclonal immune activation [
11,
12], immunostimulation [
13], antitumor effects [
14], and the ability of pathogen antigens to initiate the mechanisms of pathogen interference, which are mediated by the production of small noncoding RNAs (miRNAs) which comprise the small non-coding RNAs (ncRNAs): miRNA, siRNA, etc. [
4]. The small ncRNAs then regulate gene expression which either enhances or diminishes infection by other pathogens. Vaccines (as a class of antigens) also stimulate the production of miRNAs [
4], and hence create sometimes unexpected, nonspecific outcomes like pathogen interference. Vaccination may also induce antibody-dependent enhancement with negative health consequences [
15,
16,
17,
18].
While it is true that all vaccines in commercial use are effective against the target pathogen, we have recently demonstrated that influenza vaccination has powerful nonspecific effects against all-cause winter mortality [
19]. Indeed, using a data set of nearly 100 countries over a 40-year period no long-term net effect against all-cause winter mortality could be demonstrated [
19,
20]. This was because in some years influenza vaccination was associated with benefit against all-cause mortality, while in others with net disbenefit [
19,
20]. The degree of benefit/disbenefit varied each winter (as does the composition of the vaccine) and between countries. Increasing obesity may be associated with net disbenefit [
20]. Climatic and other variables appear to explain the different levels of international pathogen circulation and diversity over the winter or rainy season near to the equator [
19,
20].
Influenza and SARS-CoV-2 are among the class of RNA pathogens showing high mutation rates [
21,
22,
23,
24]. Each new clade of antigen mutations leads to a unique age profile for each variant which is also associated with the generation of specific miRNAs, further nuances of pathogen interference and epigenetic modifications [
25]. In the UK, the COVID–19 pandemic commenced somewhere in early 2020 with the first laboratory-confirmed death occurring on 2 March 2020 [
26]. However, COVID–19 testing capacity was very low at that time and earlier deaths are possible. Research in the USA suggests that COVID–19 deaths may have started in early January 2020 [
27]. Hence, we have the pre-COVID era which ends in December 2019 through to the ongoing surges as new variants come to the fore [
24,
28,
29,
30].
As for the strains of COVID–19 the original Wuhan strain is predominant during 2020. The Alpha strain (formerly the Kent variant) appears around December 2020 and predominates from January to June 2021, the Delta strain (formerly the Indian variant) commences around May 2021 and predominates from July to December 2021. While Omicron (BA.1) first emerges in November 2021 but begins to spread in December 2021 and dominates from 2022 onward (BA.2 followed by BA.4/5, etc.) [
24,
28,
29,
30]. The Alpha variant caused slightly higher mortality than the original strain and will therefore affect mortality in the winter of 2020/21 [
24,
28,
29,
30]. The Delta variant which mainly affected the winter of 2021/22 had higher transmission and a slightly lower or equal mortality risk [
24,
28,
29,
30].
Under the normal course of events vaccination against something like influenza commences before the influenza season. However, the vaccination schedule for COVID-19 vaccines depended on the dates for approval and the need for widespread vaccination among adults. As a result, individuals were being vaccinated at different times of the year, at points associated with the arrival of new variants, and at occasions of high through to low incidence of COVID-19 infections. If nonspecific effects were to exist, then the unusual circumstances associated with the COVID-19 vaccination campaign offer the greatest opportunity for such effects to be identified and quantified.
In the UK, COVID-19 vaccines were approved in the following order: Pfizer/BioNTech (2 December 2020 - deployed 8 December 2020), AstraZeneca (30 December 2020 - deployed 4 January 2021), Moderna (8 January 2021 - deployed 7 April 2021) [
31,
32,
33]. The proportions of persons vaccinated by age and time from different manufacturers (Pfizer/AstraZeneca/Moderna) does not appear to be publicly available.
Table 1 provides a summary of the timeline for vaccination in England.
COVID–19 vaccination began on 8 Dec 2020 for care home residents, persons aged 80+, and some health care workers, by 18 January 2021 this included age 70+ and persons with very high clinical risk, by 15 February age 65+ and persons with high risk, and by 22 May age 32+ and age 18+ by 18 June 2021 [
31,
32,
33]. Following reports of a rare type of blood clot in late March 2021 for the AstraZeneca vaccine
, persons under 30 years were all given the mRNA vaccine from 7 April 2021 onward, and those aged under 40 from 7 May 2021 onward [
34].
Astra Zeneca was phased out from September 2021. An alternative non-mRNA vaccine Novavax (recombinant protein) was made available from spring 2022 onwards. Some younger children with high clinical risk were vaccinated from January 2021 onward [
33,
34,
35,
36]. Vaccination of persons aged 16 ̶ 17 years was from July 2021 onward, 12 ̶ 15 years from September 2021 onward and 5 ̶ 11 years from February 2022 onwards for those with high risk and for any child aged 5 ̶ 11 from April 2022 onward. The majority aged 12+ were vaccinated during late 2021. All with mRNA as per the age under-40 rule as above. Booster doses began to be delivered from 16 September 2021 and these were all mRNA. Further booster doses were given in February/March 2022, and September 2022 for the winter of 2022/23 respectively. From around spring 2022 onward persons were vaccinated (including booster) with a mix of the mRNA vaccine and the Novavax (a non-MRNA recombinant protein) vaccine.
Healthcare workers in the NHS (who will mostly be under the age of 65) began to be vaccinated from 8 December 2020 (initially with Pfizer/BioNTech) and by March 2021 over 80% of clinical staff had received their first dose and over 39% had received their second dose [
35]. To vaccinate the most people, the timing of the second dose was delayed to approximately 12 weeks [
33]. In practice, vaccination schedules showed local and regional variation. In order not to waste vaccines, toward the end of the day many centers would send social media messages for adults of any age to be vaccinated.
A somewhat neglected 2010 study suggested that optimum vaccination outcomes can only be achieved when the timing of vaccination is adjusted relative to the target and competing pathogens [
36]. The implication is that sub-optimum outcomes are possible.
Table 1 summarizes which vaccines were prevalent in each age band for vaccination during the three variants.
The timing for the approval of COVID-19 vaccines (listed above) meant that the English population (mainly oldest first) only began to be vaccinated during an outbreak of the Alpha variant [
25], and with first dose still being delivered to some people into 2022 and 2023 during the outbreak of the Omicron variant [
37]. Ample opportunities for suboptimum time-based outcomes are therefore present.
While COVID–19 vaccination is clearly effective against COVID–19 mortality per se [
37,
38,
39] there is a paucity of studies using the ‘gold standard’ of a reduction in NCACM. This was achieved using a record-linked whole population study of COVID–19 vaccination in England by the Office for National Statistics (ONS) in 2021 through to May 2023. This study uses age bands, month of death, and vaccination status (first, second, third dose at both up to 12 weeks and greater than 12 weeks post vaccination) [
40].
The all-cause mortality data set used in this study is very large and covers all residents of England who are registered with a GP and were residents in England at the 2011 census [
40]. This allows detailed analysis of 944 000 deaths over a 29-month period by gender, over 7 age bands, and by various stages of vaccination split by less than 21 days, and greater than 21 days post vaccination, and at monthly intervals – which can be grouped by SARS-CoV-2 variant [
25].
The unique feature of the ONS data is that mortality is available at monthly intervals – a feature which is very rare in vaccine studies. Such profiles compare the mortality rates by age and gender within a vaccination stage or over time. The shape of the time profile gives an internal consistency check. We thereby avoid arguments regarding the exact value of each data point, since the principal aim of the study is to demonstrate that nonspecific effects do exist and have important consequences.
Our results are illustrative rather than prescriptive for three reasons.
The study uses only seven broad age bands. We have demonstrated that each COVID-19 variant has a unique single-year-of-age profile for mortality [
25] and would argue that age should be a continuous variable. The use of age bands is probably concealing more nuanced behavior.
The study was conducted at a time when vaccines in the UK were based on the original Wuhan stalk antigen. In addition, by early 2023 the Alpha and Delta variants are no longer present, and by the end of the study only Omicron sub-variants were circulating. The results cannot therefore be directly extrapolated into the future should variants other than Omicron arise.
Time since vaccination is split into two groups, namely, up to 21 days and greater than 21 days. The up to 21-day group encompasses the time when immunity is being optimized, however, the greater than 21-day group contains a mix of individuals with differing degrees of vaccine waning.
Hence, we seek to establish the basic principles rather than argue if a certain set of conditions caused a large or very large increase in non-COVID-19 mortality in the vaccinated. Confidence intervals are not shown simply because they only add unnecessary complexity to an already data rich study. They are however available in the ONS data [
40].
The above needs to be understood in terms of system complexity which leads to unexpected outcomes. We have recently proposed that influenza pandemics and epidemics show very high system complexity leading to unexpected all-cause mortality outcomes associated with influenza vaccination in approximately 50% of years [
4,
19,
25]. Such system complexity may well lie behind the reported nonspecific effects of vaccines [
1,
2,
3,
4,
5,
6]. Indeed, it is possible that the age-based schedule of COVID-19 vaccination (as above), along with the specific (and unexplained) single-year-of-age profiles for mortality associated with COVID-19 variants [
25], has inadvertently increased system complexity in unexpected ways. It is hoped that this study will shed light on such issues.
Finally, we will provide an extended discussion of the mechanisms by which such nonspecific effects can occur focusing on the somewhat neglected effects of both the environment, drugs, and vaccines upon the expression of noncoding RNAs which are profoundly powerful regulators of gene expression.
3. Results
3.1. Overview of the net effect of vaccination with time
As a reference point
Figure A1.1 and
Figure A1.2 in the Appendix shows the trend in proportion of total deaths ‘with’ COVID-19. The two major outbreaks peaking at weeks ending 17-Apr (Wuhan) and 8-Jan (Alpha) are evident.
Figure A1 is for all ages but clearly shows the peaks and troughs in COVID-19 deaths. Weekly data has been used for greater definition [
42]. From
Figure A1.2 note three ‘summer’ minima in the proportion of total deaths with COVID-19 in June to Aug-21, Jun-22, and Jul-23. Also note peaks in January 2021, and a series of undulating maxima in Sep-21, Nov-21, Jan-22, Apr-22, Jul-22, Oct-22, Jan-23, and Mar-23. These most likely outbreaks of the various COVID-19 variants and sub-variants. At the other extreme is a trough occurring in April/May of 2021 for age bands below 50 years. The trough represents the non-outbreak tail end of the Alpha variant [
25], followed by the arrival of Delta which specifically targets the younger ages [
25].
Further undulations reflect the relative impact of outbreaks of the three different SARS-CoV-2 variants upon different age groups [
25]. Likewise, Omicron had a disproportionate effect on the groups aged over 80 years [
25]. Hence the overall shapes of the trends are consistent with the independently characterized effects of the variants upon the year-of-age age profiles for mortality [
25].
Figure 1 shows the net effect of COVID-19 vaccination against all-cause mortality (including COVID-19 deaths) for persons aged 18+ receiving one or more doses of the vaccine. Below the blue dashed line is increasing protection, i.e., all-cause mortality is reduced relative to the unvaccinated, while above lies increasing all-cause disbenefit.
As can be seen the COVID-19 vaccines employed in England generally had a net beneficial effect against all-cause mortality (except perhaps in the two youngest age bands, under specific conditions) which diminished with time, seemingly, due to the transition between SARS-CoV-2 variants and sub-variants. This decline in performance reflects the known specific effect of antigenic distance between a vaccine and the prevailing variants and sub-variants [
4]. It will also include any unanticipated nonspecific beneficial/disbenefit effects from vaccination, which are expected to exist [
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14].
Differences between males and females are evident with males seemingly benefiting more from vaccination than females. June and July of 2021 represent a point of minimum COVID-19 mortality as a proportion of all-cause mortality [
42]. COVID-19 mortality does not drop to this low level again until June of 2022 just after the peak of infections due to the arrival of the first Omicron variant, and then again in July 2023 [
25].
Based on the mean/median values for the last six months (Dec-22 to May-23) the best vaccine protection occurs in the interval 50–79 years and deteriorates either side. Disbenefit against all-cause mortality occurs in the youngest and oldest age groups.
Hence, the overall conclusion is that COVID-19 vaccination was generally but not specifically successful at reducing population-wide all-cause mortality (including COVID-19 mortality) but that the effectiveness of the vaccine reduced as new SARAS-CoV-2 variants and sub-variants showed increasing antigenic distance from the vaccines based on the original Wuhan strain. However, age appears to be a key factor as has been previously demonstrated for COVID-19 variants [
25].
Within the above context of general vaccination success (with curious exceptions), the aim of this paper is to avoid investigating the specific effects of vaccination which may include antigenic distance and other factors, hence, to focus primarily on all-cause mortality which excluded COVID-19 deaths (NCACM), to investigate some of the more curious outcomes seen in
Figure 1.
3.2. Unexpected complexity
Figure 2.a and
Figure 2.b show the NCACM rate for the vaccinated during the 21 days and >21 days post vaccination and corresponding unvaccinated NCACM rate. This is shown for the youngest age group, namely, 19–39 years. Note that in both
Figure 2.a and
Figure 2.b the month is the month of death rather than the month of vaccination. For the <21-day post vaccination group, the month of death and vaccination will be mostly identical. As a generalization the male rate is higher than that for females. The only exception is for the group who received their fourth dose earlier than the general population.
The <21-day post vaccination group has fewer data points because it only spans a 21-day period whereas the >21-day group can be for an extended duration, especially for persons who halted their vaccine journey after the first, second or third vaccine dose.
We understand that every data point is subject to statistical uncertainty and the reader is invited to use their knowledge of statistics to interpret the trends. While a 95% confidence interval exists in theory, visual inspection can quickly reveal questionable values. For example, unvaccinated male mortality in Sep-22 and Feb-23 looks to be low. On both occasions the data can be adjusted upwards to somewhere close to the two months on either side. Such an adjustment has a negligible effect on the overall conclusions. The same could be said for female first shot >21-day mortality in Jun-22, and for male first shot > 21-day mortality in Feb-23. Adjustment up to the average of the surrounding points likewise makes a minor effect on interpreting the overall chart, namely, nonspecific effects are highly prevalent.
Should you choose to question the validity of the unvaccinated as an ‘unbiased’ group you can simply shift the lines up or down, with inconsequential effects on interpreting the chart. All high values in the chart remain unquestioned because they are part of a continuous trend. Recall that all values is the real-world actual outcome. As can be seen in
Figure 2.a the vaccination of males with their first dose continued through to March 2022. After this point so few are given their first dose that small numbers preclude meaningful analysis.
Few people aged 19–39 was given a fourth dose delivered mainly in October and November (
Figure 2.b). Note that the corresponding charts for all other age groups are given in the
Appendix Figure A3.1 to 3.6.
The main point is that in the absence of nonspecific effects the lines for the vaccinated and unvaccinated should be one and the same. This is clearly not the case. It would be interesting to review the reasons for early vaccination with the fourth dose from May to September 2023, and their associated risk factors. Whatever the case, they experienced very high mortality. It would be apposite if every NCAM death following the fourth dose in this age band was subject to a retrospective clinical review.
Note from
Figure 2.a that outcomes for the second and third dose <21 days after vaccination are mostly beneficial. That for the first dose reaches maximum protection against NCACM in Jun-21, which is around the point of minimum levels of COVID-19 mortality (
Figure 1.b.), implying that very low COVID-19 infection (even as asymptomatic), facilitates this type of nonspecific effect.
Graphs for the other age bands are given in the Appendix as
Figure A2.a to
Figure A2.f. Examples of nonspecific benefit/disbenefit are likewise observed which are age/gender specific. Recall that time represents transitions between COVID-19 variants, transitions between seasons, and fluctuation between high/low COVID-19 deaths (as in
Figure 1.a. and
Figure A1.b.)The significance of such transitions will be covered in the Discussion.
Based on
Figure 2.a and
Figure 2.b plus A2.a to A2.f, nonspecific effects do exist and that under different combinations of age/sex/COVID-19 variant/season/time since vaccination that these can be beneficial or give disbenefit.
Figure 1 is therefore a composite derived from a complex set of highly dynamic interactions.
3.3. The timing of the nonspecific benefit of vaccination during the first 21 days
‘Table 9’ in the ONS data set covering the period January 2021 to May 2022 [
40] provides a useful breakdown of both COVID-19 and non-COVID-19 deaths during the first 11 weeks following vaccination. This material is summarized in
Figure 3 as the ratio of ‘with’ COVID-19 deaths to other non-COVID-19 deaths. This is a composite picture from the first, second and third doses, and is not split by gender. During the first 11 weeks the data encompasses some 2068 deaths for the age 10–39 group, through to 78 925 in the age 80–89 group, and 53 723 in the age 90+ group.
This data is presented as a ratio of ‘with’ COVID-19 deaths to all other deaths. The ratio needs to be interpreted in the light of
Figure 2.b which shows benefit against NCACM for the first 21 days (3 weeks) postvaccination, i.e., the denominator has been reduced.
Also, from
Figure A1.b it should be noted that the ratio of COVID-19 to non-COVID-19 deaths is constantly changing over time and for the prevailing COVID-19 variant [
25]. Depending on the sampling strategy employed by the ONS and the proportion of persons who received each of the three doses up to May-22, the magnitude but not the timing of the peaks seen in
Figure 3 will be affected. However, when compared to
Figure A1.b the 21% ratio of COVID-19 deaths in
Figure 3 is far beyond anything possible from simple COVID-19 infection.
Somewhat surprisingly
Figure 3 appears to reveal nonspecific disbenefit following vaccination relating to ‘with’ COVID-19 deaths. This disbenefit commences in the first week, while the maximum disbenefit occurs during the second week for ages <50 years, and during the third week for ages 80+ and some point between 2 to 3 weeks for ages 50–79. This disbenefit then diminishes and reaches a minimum after around 6 weeks in the younger groups, and up to 10 to 11 weeks in the two oldest age bands. Clearly the extent and timing of the disbenefit is age dependent.
Since this is a composite of different vaccine types, male/female and up to three different vaccine shots, more cannot be discerned, however, it confirms the fact of unanticipated nonspecific effects following COVID-19 vaccination and its age dependence. More detailed analysis using days rather than weeks, males/females, vaccine dose number, and vaccine type/manufacturer is highly recommended.
Antigen production can be excluded since this only occurs around 3 weeks after vaccination. Given the timescale for the increase in COVID-19 mortality we suspect enhancement of asymptomatic COVID-19 infection as the most plausible cause. COVID-19 is a multiorgan disease [
43] and presumably asymptomatic or sub-clinical infections can be triggered to assume greater clinical severity. Further detail will be given in the Discussion.
3.4. All-cause (including COVID-19) mortality during Omicron
Given the results in
Figure 1,
Figure 2,
Figure A2, and
Figure 3 it is illustrative to look at the trend in all-cause (including COVID-19) mortality during Omicron. As shown in
Figure A1.a. and
Figure A1.b. the mortality rate during Omicron is very low, and especially below the age of 70. Depending on your point of view, the rationale for vaccinating ‘healthy’ persons below the age of 65 is a grey area. The Discussion presents the evidence suggesting that use of the experimental vaccines should have been tempered by the possibility of unanticipated disbenefit.
Hence, it is useful to look at the real-world outcomes. Reverting to all-cause mortality removes the issues surrounding subtracting COVID-19 from all-cause mortality to get NCACM. Plus, the number of deaths is slightly larger leading to lower statistical uncertainty. In addition, the net outworking of specific and nonspecific vaccine effects can be observed. The outcome is shown in
Figure 4 which shows the results for age 18–39. Note that deaths from Omicron commence around March 2022 and the transition from the end of Delta is shown for context. A similar chart for age 90+ is shown in the Appendix as
Figure A3.
Regarding statistical uncertainty, male unvaccinated mortality in Sep-22 looks to be a statistical outlier and is probably closer to August and October of 2022, as is observed for females. However, making such an adjustment negligible difference to interpreting the chart. Likewise, female mortality for the fourth dose <21 days in October 2022 looks to be an outlier, but adjustment to somewhere near to the values for September and November 2022 also makes a negligible effect on interpreting the chart.
Hence, all people receiving their first, second or third dose <21 days ago show increasing adverse outcomes beyond Mar-22, May-22, and Aug-22 respectively. The fourth dose <21 days only shows benefit around October/November 2022, but is otherwise not beneficial compared to their unvaccinated colleagues. The fourth dose >21 days is universally not beneficial, etc.
A similar story emerges for those aged 90+ in
Figure A3. Somewhat curiously the disbenefit attached to the fourth dose >21 days appears to diminish with time. This could suggest that vaccine waning is behind this nonspecific effect.
As above, outliers can be quickly spotted, such as low values in August 2022 and January 2023 for males receiving their fourth dose <21 days ago. Adjustment upward turns benefit into borderline disbenefit. The overall observation is that over half of outcomes show disbenefit except for the fourth dose >21 day between April and August 2022 for both males and females. The fourth dose <21 days is generally beneficial up to February 2023, but then leads to disbenefit. The second dose <21 days shows a small benefit between March and May 2022, but otherwise shows disbenefit, etc.
As with NCAM a dynamic interaction between competing biological forces is evident which unfortunately is only revealed retrospectively. As noted in
Figure 1 it is only the age groups 50–79 years which experience a greater proportion of benefit during the period beyond December 2022.
In conclusion, evidence for highly dynamic nonspecific effects from COVID-19 vaccination can be demonstrated. As expected from other vaccination studies [
1,
2,
3,
4,
5,
6,
7,
8,
9,
10] there is a mix of benefit and disbenefit which appears to be highly dependent on the context. The exact balance of the forces can be demonstrated by comparing NCAM and all-cause mortality outcomes.
5. Study summary
This study is not in any way suggesting that COVID-19 vaccination does not provide a measure of protection against COVID-19 disease per se. The interaction between individual health status (age, immunotypes, etc.), COVID–19 infection, SARS-CoV-2 variant, vaccination, and the environment represent a case of exquisite complexity.
It must be emphasized that from around the middle of 2021 the pool of unvaccinated persons stayed relatively constant and the higher proportion of adverse outcomes during Omicron are therefore not due to changes in the unvaccinated baseline.
The second point to emphasize is that the unique age profiles for each COVID-19 variant are not being recognized in most studies. This creates huge problems for age standardization of mortality rates between variants. Our study attempts to avoid this problem by reporting all results by age band.
The necessary urgency to develop and implement vaccines has inadvertently interacted with this complexity in sometimes adverse ways. To deny that such complexity exists is unhelpful and hinders efforts to implement safe vaccination development and implementation. The move to discontinue the use of viral vector vaccines in favor of mRNA vaccines due to increased risk of thromboembolic events is an example of a potentially incorrect decision since the available evidence supports the notion that the original viral vector vaccines had overall lower all-cause mortality. This requires further investigation.
This study confirms that COVID vaccination of the elderly was generally a success – within the limits of certain optimum gender/time/age combinations – and less so during the Omicron period (as in
Figure 1). For unknown reasons the age 90+ group appears to be an exception.
However, all-cause mortality outcomes for mRNA vaccines during the Omicron variant (the UK had switched to virtually exclusive mRNA vaccination prior to Omicron) were especially poor with most age/gender/vaccine stage/time combinations showing higher all-cause mortality in the vaccinated compared to the unvaccinated.
Generally worse all-cause mortality outcomes after COVID-19 vaccination among persons aged below 40 years are a common theme among this and other studies and this line of investigation should be pursued.
6. Conclusions
This study does not contradict the numerous studies regarding the specific effects of the vaccines against COVID-19 mortality. It merely adds to the growing body of evidence for the nonspecific benefit/disbenefit effects of vaccines [
1,
2,
3,
4,
5,
6,
7,
8,
9,
10,
218].
While the increasing antigenic distance between the Wuhan-based vaccines looks to be the primary driver of declining performance against all-cause mortality which includes COVID-19 (in
Figure 1), it is also evident that other nonspecific mechanisms are involved.
However, our overall conclusion is that the COVID-19 vaccines employed in England during 2021 and 2022 led to unintended selective NCACM harm under specific combinations of age, sex, vaccine history, COVID-19 variant, and time (season, outbreak versus non-outbreak months, etc.). The question is which of the three vaccines employed during 2021 and the those employed in 2022 led to the most unintended NCACM and all-cause mortality harm and under which circumstances.
We again note that the single year of age profiles for mortality in different SARS-CoV-2 variants [
25] presents considerable problems to the issue of age standardization.
Human vaccines must be open to scientific scrutiny, and it is only in such an open framework that vaccines can be continuously improved. It is our observation that researchers questioning aspects of COVID-19 vaccination are not ‘anti-vaxxers’ but are seeking genuine answers to seeming ‘anomalies’ in the data. Indeed, such anomalies can be hidden by the application of certain types of analysis, especially after age-standardization across all age groups.
There is no such thing as ‘perfect’ protection, only a balance between the risks and rewards, which this study demonstrates are far more complex than has been appreciated. We highlight that a ‘good’ vaccine should not disadvantage those (via increased all-cause mortality ) who choose to halt their vaccine journey. The high rate of waning in mRNA vaccines leading to eventual higher mortality in the vaccinated compared to the unvaccinated remains a concern. A very recent study has confirmed that age-based selection occurs between persons receiving the second and third dose of the mRNA vaccine [
205].
It would also appear that the process for reporting/detecting adverse events following vaccination is not achieving its intended purpose since subtle changes in all-cause mortality are going undetected in both influenza [
4,
5,
6] and COVID-19 vaccination. The latter has been elegantly illustrated by the analysis of Seneff et al [
180] and inferred by Pantazatos and Seligmann [
175,
176].
We also recommend that long term surveillance and re-analysis of all-cause disease outcomes following COVID–19 vaccination is actively pursued. Vaccination of the young must be reconsidered based on the comprehensive evaluation of the all-cause mortality vaccination outcomes for the different types and brands of vaccines. In this respect we strongly recommend an international investigation regarding the deaths of any person aged below 40 (adult, teenager, child) who had received COVID-19 vaccination. This should include days since vaccination, potential risk factors, COVID-19 variant(s), and cause(s) of death. We remain unconvinced regarding the clinical justification for the vaccination of any ‘healthy’ person in this age group.
It is also strongly recommended that international studies are conducted to determine the exact contribution of the different types of COVID-19 vaccines against all-cause mortality.
Reporting of outcomes by month as used in the ONS study is the recommended approach as it allows both seasonal, and epidemic and non-epidemic effects to be disentangled along with the effects of different variants. Studies which report outcomes using time (days) after vaccination as a continuous variable are also recommended. We suggest that influenza vaccines be subject to the same all-cause mortality scrutiny [
4].
Indeed, it is entirely possible that the optimum vaccine choice, assuming that the antigenic distance is not too great, is both gender, age, and context specific (timing, latitude, mix of circulating pathogens, ethnicity, and personal risk factors).
We strongly suggest that the evaluation of all-cause and NCACM mortality become a standard for all vaccine trials and that the measurement of noncoding RNA profiles may be of profound benefit.