Covid-19 Vaccine
In the short time available between the Depp-Heard case ending and the Queen's Jubilee festivities starting I thought I'd have a quick look at some of the scientific papers recently published about the longer term consequences of Covid, with consequential impact on how scared we should be of the disease and what efforts might be effective in protecting ourselves. In this article I will look in-depth at one which considers the impact of vaccination, and in a follow-up article I will look at the other four.

The paper I'm looking at in this article is the one by Al-Aly et al., published in Nature on May 25th, titled "Long Covid after breakthrough SARS-CoV-2 infection". This compared outcomes after Covid infection of almost 35,000 vaccinated U.S. veterans with a set of control groups. To make it clear, 'veterans' aren't, as a colleague of mine once thought, people that look after the health of animals, but instead are members of the U.S. armed forces after they finish active service. As a result, the study looks at individuals aged from approximately 40 to those in their 80's, and the group is biased towards males (though it does include many females).

It is important to note that the study looked at a wide variety of conditions that were present beyond 30 days after the individual's positive test for Covid. Thus the data shown in the paper, and reproduced below, don't include symptoms experienced during the actual acute disease stage of the infection, but they do include sequelae (disease after-effects) that started during the 30 days post-infection, but continued in the weeks and months after this point.

The results? In summary:
  • Long Covid exists at a non-trivial level in those who had symptomatic disease that was serious enough to warrant seeking assistance from their healthcare providers.
  • There are no useful data on the incidence of Long Covid in people that had mild Covid disease.
  • There are no useful data on the protection offered by the vaccines.
  • There is a strong suggestion that there are significant levels of vaccine injury in vaccinated individuals who weren't infected with Covid.
  • Rates of myocarditis appear to be very high in the group that included vaccine side-effects in the six months after vaccination (but where Covid infection occurred at least 30 days before vaccination).
Let's take a closer look at the detail.

The study found a significant increase in sequelae in those having a breakthrough infection (i.e., infection after vaccination) compared with their control groups.
Covid-19 vaccine side effects
That's a complicated graph, so a quick summary: On the left are 'hazard ratios', anything above 1.0 means that the study group (vaccinated infected) had more problems with their health (sequelae) than the control group (that didn't get infected with Covid before or during the study period). For example, around about twice as many people in the vaccinated-infected group had problems with fatigue than in the control group (hazard ratio is about 2.0). On the right are the number of additional cases of each sub-type of sequelae that were seen, per 1,000 people - this is a very important aspect of the results, as it reflects the real-world impact of disease. The number of additional cases of 'negative health conditions' in the infected group was around 10-40 cases per 1,000 individuals for each sequelae category (i.e., an absolute increase in each category by around 1%-4%).

On the face of it that's a rather scary graph - it suggests that around 12% of those infected with Covid after vaccination have at least one 'probably serious health issue' in the months following recovery. Note, however, that the individuals included in the analysis had to have 'interacted' with their healthcare provider at least once in the previous two years to be considered for inclusion in the study, so probably doesn't include the most healthy individuals. I suggest that it is likely that individuals with mild symptoms (the majority of infections) would have lower incidence rates of these sequelae, but this wasn't specifically addressed in the study.

So far their data show that those vaccinated against Covid have a high risk of health issues following an infection with Covid. The question then becomes whether vaccination helps to reduce this risk. Unfortunately, while the study did include data on relative risk of sequelae for the vaccinated-infected versus unvaccinated-infected, these data appear to be highly compromised and I'd suggest shouldn't be used. The problems are:
  • The vaccinated and unvaccinated data come from different times of the year (see supplementary table 1). The data for the vaccinated group are based on infections that occurred between the start of July 2021 and mid-October 2021 (the control group data come from a matched period). The data for the unvaccinated group are based on infections that occurred between mid-January 2021 and the end of July 2021. This means that the vaccinated group were predominantly infected with Delta variant, while the unvaccinated group were were probably infected with original Wuhan variant or the Beta variant. There might also be seasonal effects that could show up in differences between the two groups.
  • As unlikely as it sounds, the 'unvaccinated' group contain individuals that were vaccinated - the only restriction was that they were unvaccinated at the point of infection and 30 days afterwards. The paper doesn't indicate the level of vaccination in the unvaccinated group, but given the very high levels of vaccination achieved in the USA in those aged 40 or over it is likely that the majority of individuals in the 'unvaccinated' group were vaccinated after their infection.
  • The 'vaccinated' group excluded any side-effects that occurred in the weeks after vaccination (within 60 days of vaccination for the mRNA vaccines and 45 days of vaccination for the viral vector vaccines), while the 'unvaccinated' group included side-effects that occurred in the post-vaccination period (for those that were vaccinated after their infection).
  • Indeed, analysis of the characteristics of the vaccinated participants shows that the majority were vaccinated between January and March 2021, many months before they were infected with Covid. The study data suggest that on average there was period of seven months after vaccination before data was collected on their potential side effects (six months before infection and then another 30 days to exclude acute Covid symptoms).
Thus the two groups that they compare aren't 'unvaccinated vs vaccinated', but instead are:
  • Had been vaccinated for some time at point of Covid infection; vaccine side-effects that occurred within around half a year after vaccination aren't included; they were probably infected with Delta variant;
  • Unvaccinated at point of infection and 30 days later; probably vaccinated later; the data include short term (weeks) vaccine side-effects and medium term vaccine side-effects (months); the individuals were probably infected with the original Wuhan or Beta variant in the early part of the year.
I find this approach to the 'unvaccinated' group a bit strange - they had a huge pool of people to gather data from so surely there were a fair few that remained unvaccinated during the study period that they could have worked with? Regardless, the characteristics of their 'unvaccinated' and 'vaccinated' group makes any meaningful comparison impossible. Note that this means the paper certainly doesn't offer robust evidence to support the use of vaccination to reduce the risk of Long Covid.

The study also compared the risk of sequelae after Covid in the vaccinated with the risk of sequelae after a serious influenza infection. It found significantly greater incidence of sequelae in the vaccinated-infected compared with those infected with influenza (and not vaccinated). The relative contribution of vaccine side-effects and sequelae of the viral infection in this finding remain unexplored in this study.

Thus the data seem to suggest that the sequelae might be associated with the Covid infection itself or might be linked to the vaccination - unfortunately, the paper doesn't directly offer analysis to separate out these two factors. However, the paper's supporting data allow us to do something interesting - compare the rate of the authors' chosen list of health issues in individuals that were vaccinated but who didn't get infected with Covid and the background (historical) rate of these health issues.

Or, to put it more succinctly, the data in this paper allow us to estimate the rate of vaccine side-effects alone, at least medium and longer term ones. Note that the complex method for inclusion of the vaccinated-not-infected group would mean that the data for some individuals would include health conditions that occurred within 30 days of vaccination, but this wouldn't be the case for most individuals in the study - thus any vaccine side-effects identified would likely be biased towards the medium and longer term.

I've computed the excess incidence of their range of negative health outcomes after vaccination compared with the historical control data in the graph below. Note that the data in the study are complex as they applied a weighting to each study group to attempt to remove bias. In the graph below I have presented the estimate of vaccine side-effects based on the raw data in the paper, as well as attempting to apply the same weighting function as the authors used in the paper to remove bias; it is likely that the true rate will fall somewhere between these two values.
Excess incidence of negative health outcomes after vaccination
These data are astounding - they suggest that there are a significant level of serious negative health consequences of vaccination, of the order of 0.5% to 1% (and possibly higher) for each condition, and with around 1.5% to 4% of the vaccinated suffering some condition that resulted in them seeking support from their healthcare provider. This estimate is in line with the findings of the CDC V-Safe survey, which found that 0.9% of recipients of a Pfizer booster sought medical care, those reported by a whistleblower board member of a German insurance company, who said his company's data suggested around 4% of Germans had sought medical care following vaccination, and an Israeli Government survey which found 0.3% Pfizer recipients reported hospitalisation (not just medical care) as a result of vaccine side-effects. It is a shame that the authors missed this important aspect of their data - I'd suggest that it is probably more important than their findings on 'Long Covid', given that the number of vaccinated individuals is much higher than the number of Covid infections, that Omicron appears to be much milder than previous variants (i.e., what were the relative risks of early vaccination versus waiting until less pathogenic variants evolved), and that individuals were deliberately given the Covid vaccines while actual infection with Covid is an act of nature.

There is a silver lining to the very dark cloud indicated in the graph above; vaccination appears to have relieved some mental health risk compared with the historical control. Perhaps we should add Covid vaccination to summer, Buddy Holly and the working folly as reasons to be cheerful?

The data in the graph above raise more questions than they answer - for a start, have there really been deaths in between 0.5% and 1% of those vaccinated, at least in the types of people included in the study? The fact that the study only looked into sequelae in people with sufficiently serious symptomatic disease to seek support from their healthcare provider might help to explain this high death rate.

These data aren't absolutely conclusive evidence for there being significant levels of vaccine side effects, but they are highly suggestive of a problem and are certainly worth reporting and definitely would support further research to fully quantify the effect. However, for some reason there being scary levels of 'Long Covid' appears to be more important than there being 'scary levels' of vaccine-induced injury, and these data weren't reported by the authors.

There's one more rather odd aspect to the data in this paper that I find difficult to explain. Hidden away in two footnotes to two tables in the extended data section of the paper is the information that myocarditis in the vaccinated-infected was double that found in their control group, and in the 'unvaccinated' infected myocarditis was 20-fold higher again than the vaccinated-infected group. However, it is important to remember what 'unvaccinated' means in this paper. Perhaps it would be better to interpret this finding as:
  • Myocarditis risk was double the expected (background) rate in those infected with Covid after vaccination, but ignoring any myocarditis between the point of vaccination and about six months post vaccination.
  • Myocarditis risk was 40-fold the expected rate in those who were unvaccinated at the point where they caught Covid, but who were probably vaccinated after this point and where myocarditis risk includes the weeks and months after vaccination (for the majority that will have been vaccinated).
Is the greatly elevated myocarditis in the latter group a catching-Covid-before-vaccination problem or a risk of vaccination problem? We just can't tell from their data.

What's truly weird is that this is probably the most significant finding in their paper, yet it is relegated to a footnote; there's no mention of myocarditis elsewhere in the paper. The authors clearly knew about it because they did include the data in the footnote, but decided to ignore it anyway. I'm sure that things would be cleared up if the authors could explain why they decided to ignore this important part of their results.

In summary:
  • The authors almost certainly found that the vaccinated that suffer breakthrough infection had higher risk of negative health conditions in the months following vaccination, compared with control groups (those without prior Covid infection).
  • In those not hospitalised, around 8% experienced a negative health condition of some type.
  • Their data weren't robust enough to support the conclusion that the 'unvaccinated' suffered increased levels of sequelae after infection, because it is likely that the majority of their 'unvaccinated' group were vaccinated after infection with Covid.
  • The data strongly suggest a high level of negative health conditions after vaccination in those not infected with Covid, compared with historical norms. The authors decided to not analyse their data in this way.
  • The authors also decided to ignore the significant results they found regarding myocarditis risk.
Amanuensis is an ex-academic and senior Government scientist. He blogs at Bartram's Folly.