The risks and opportunities of vaccine hyper-development for Covid-19 Part 4: Learning lessons from history: “everything old is new again”

One of my intended themes in these articles is the analysis of the risk:benefit issues for new vaccines; particularly those leading up to and during mass vaccination campaigns as envisioned for Covid-19. In this article #4, I want to suggest that one of the best ways to understand future risk is to review where things went wrong in the past. I will review two failed vaccine roll-outs: The so-called Cutter Incident and the more recent experience with Dengvaxia? as well as briefly touching on vaccines for RSV, rotavirus and influenza. Before doing so, I want to stress once again that I am a strong pro-vaccine advocate. Vaccines have an excellent safety and efficacy track record and I definitely believe that we need to develop Covid-19 vaccines and that we then need to run global mass vaccination campaigns. My advocacy here is to minimise any risks that we could face in doing that.

In Part 1, I emphasised the huge beneficial role that vaccines have played around the world and through the centuries. They remain the only medical intervention that has eradicated a deadly disease (smallpox) and they continue to reduce significant suffering and deaths from many other major diseases. Vaccines have undoubtably saved millions of lives and reduced untold suffering. I also emphasised the very long history of unfounded resistance to vaccination. The cartoon image above was targeted at Edward Jenner, often referred to as the father of modern vaccines, and was published by the British Anti-Vaccine Society during Jenner’s own lifetime in 1802. Anti-vaccine propaganda has only intensified since then and is now magnified by the echo chambers of the internet. However, community fears of vaccines, although largely misguided, are not entirely without merit. Whilst modern vaccines are exceptionally safe, there have been a number of issues during large scale vaccination programs from which we should all learn, and this is all the more important in a pandemic situation.

Every time there is a misstep in vaccine development or vaccination policy there is the potential to reduce public confidence in vaccines whilst strengthening the hand of the more malicious anti-vaccine pressure groups. We should not expect that the pandemic will offer us any forgiveness in the longer run and for many reasons we should be increasing our vigilance, not decreasing it. As I hope to point out, the majority of risk for new vaccines comes with the speed and scale of the roll-out. The faster that we aim to roll-out a vaccine, and the more people that we intend to vaccinate, the greater the risks that we face.  A rapid, global vaccine roll out for a pandemic is thus at the highest level of risk for any vaccine ever developed.

The first historical example that I want to draw on is the so-called Cutter Incident, which involved Jonas Salk, who is another giant figure in the history of vaccines and vaccination. Salk developed the formalin inactivated vaccine against poliomyelitis or polio in 1955. Before Salk’s eponymous vaccine, polio was considered one of the most serious public health problems in the world and in 1952 an epidemic in the USA killed over 3,000 and left over 21,000 paralysed. Unlike Covid-19, most of those victims of polio were children. Salk’s injected polio vaccine as well as an oral polio vaccine developed by Albert Sabin have helped to reduce the incidence of polio dramatically around the world such that polio is targeted as the second disease (after smallpox) for global eradication. There is no doubt that the Salk and Sabin vaccines are amongst the top causes for improved global health developed in the last century.

Salk’s polio work was funded by the March of Dimes, a charitable organisation founded by US President Franklin D Roosevelt, who himself was left paralysed by polio. The clinical field trial set up to test the Salk vaccine was one of the biggest clinical trials in history, involving 20,000 physicians and public health officers; over 280,000 school personnel and volunteers; and over 1.8 million schoolchildren. On April 12th 1955 the vaccine was approved for use the day after the completion of the clinical trial had been announced. Five different vaccine companies (Eli Lilly, Parke-Davis, Wyeth, Pitman-Moore, and Cutter Laboratories) had been selected to manufacture and distribute the vaccine and in anticipation of approval they had already manufactured and stockpiled the Salk vaccine so that a mass vaccination campaign of Children could begin immediately. Just 13 days after the first children were dosed, reports started to appear of vaccinated children developing polio. It is estimated that some 120,000 children were directly exposed to this risk of whom around 40,000 developed abortive poliomyelitis (polio virus infection but with minimal neurological complications), 56 were severely paralysed and 5 died (https://www.nejm.org/doi/full/10.1056/nejmp048180).

So, what went wrong? The Cutter Incident was essentially a failure of vaccine manufacture and control. One of the five vaccine companies, Cutter Laboratories of Berkley California (subsequently acquired by Bayer in the 1970’s), had failed to inactivate the polio virus effectively in the vaccine with formaldehyde. This error only occurred in a single batch of the vaccine but Doctors and nurses using that particular batch were, in effect, injecting live polio virus into children. Since this was a live virus, many of those exposed children then also went on to expose others to the disease, which led to a small epidemic of polio in their families and communities. In total, around 200 were paralyzed and 10 people died.

The Cutter incident should be a lasting memory of the potentially devastating impact of vaccine safety if things do go wrong, but let’s also remember that the Cutter incident was a one-off instance of a manufacture and control failure at a single site and for a single batch that was relatively quickly identified and eliminated. The Salk vaccine itself went on to become one of the most important vaccines that we have today and we can remain hopeful that polio will one day be eradicated from the world. But, what can we learn from the Cutter incident that is relevant for Covid-19 vaccines? Well firstly, I would stress that an incident with exactly the same cause could not occur with most of the vaccines currently being developed for Covid-19. None of the modern recombinant protein, peptide, mRNA or DNA vaccine approaches (either synthetic or virally vectored) have any possibility of producing a Covid-19 infection since they include only a tiny fraction of the causative virus and in the vast majority of cases only a single protein (the spike or s-protein). This is true for those vaccines being developed by Oxford/AstraZeneca, Moderna, BioNTech/Pfizer and most other Western vaccine groups. There is no way for those vaccines to produce Covid-19. However, there is some clinical stage work ongoing with live attenuated vaccines in China and theoretically the risk of failed inactivation could be present for those vaccines. However, even for those vaccines, provided modern vaccine manufacturing processes, systems and controls are applied, a manufacturing control failure is less likely to occur today than in 1955. It is not impossible, but it is much less likely. I will talk more about the importance of vaccine manufacturing in a future article.

What concerns me more about the Cutter incident, which is now mirrored in the current situation, is the intent to manufacture large quantities of vaccine in anticipation of a rapid vaccine approval, to be followed by a mass vaccination campaign. In that situation, if there is a vaccine safety problem, whatever the problem may be and however rare it may be, millions of people may have been vaccinated before that problem is detected. The more rapid the initial roll-out, the greater that potential problem becomes. “Warp Speed” is great if you are heading in the right direction, but less-so if you take even one wrong turn along the way. So, it is key that we acknowledge how important it will be to have truly excellent international pharmacovigilance (drug safety surveillance) in place during each vaccine roll out. That problem will increase exponentially in complexity where vaccine rollouts occur with multiple vaccines in multiple countries in parallel. The Cutter manufacturing issue was actually identified and dealt with surprisingly quickly. Cutter withdrew its vaccine from the market on April 27th 1955, that’s within days of the first reported vaccine-associated cases. Will we be able to do the same in 2020 or 2021 at “warp speed”?

With the Cutter incident, the regulatory authorities in the USA were actually quite lucky, that the effect presented very clearly and quickly and applied only to a single manufacturing batch, so that they were able to stop the problem rapidly and with relatively few cases. What would happen if the vaccine problem were less obvious, or perhaps more delayed? What would happen for example, if the vaccine actually made the disease worse in a small subset of subjects? In that case, the problem would only start to appear when vaccinated subjects were exposed to the disease, which could be weeks or months after the initial vaccination, by which time millions could already be at increased risk. This potential problem of paradoxical vaccine exacerbated disease is not unheard of for new vaccines. One example relates to respiratory syncytial virus (RSV), which causes lung infections mainly in young children but also in older adults. In the 1960s, a formalin-inactivated vaccine adjuvanted with alum was developed for RSV. When tested in a clinical study in children, the vaccine was shown to be immunogenic and produced neutralising antibodies, but it was unfortunately not efficacious (remember Part 2 of these articles, where we distinguished the difference between immunogenicity and efficacy? This was a case in point). Worse still, the vaccine was found to enhance disease significantly when these young children were subsequently exposed to the virus during an RSV outbreak nine months after vaccination. Severe pneumonia was found to be far more prevalent in the vaccinated group than among controls (69% versus 9%) and there were two fatalities attributed to the vaccine. (https://academic.oup.com/aje/article-abstract/89/4/405/198840?redirectedFrom=fulltext).

This RSV vaccine did not progress to approval because the disease exacerbation was so marked. The vaccine failed in clinical efficacy studies and there is still no approved vaccine for RSV to this day. The exact cause of paradoxical disease exacerbation with the RSV vaccine has been the subject of ongoing speculation but it is not an isolated example. A more recent example was seen during the roll-out of the approved vaccine Dengvaxia? in 2017.

Dengvaxia is a vaccine directed against dengue fever developed by Sanofi Pasteur using recombinant DNA technology. Dengue is a mosquito borne viral disease prevalent in South East Asia and South America. It has been estimated that more than half of the world’s population lives at risk of dengue virus infection. Most dengue virus infections are asymptomatic, but it can cause an acute flu-like illness and more than 95 million people a year seek medical treatment following dengue virus infection, although no specific medical treatment exists. The severe form of disease is called dengue fever, which can result in clinically significant bleeding, leading to shock and possibly death. Mortality in dengue fever is between 2 to 5% of those treated but up to 20% if untreated.

Dengvaxia is licensed to prevent dengue virus infection in many countries including the US and the EU. It is a live attenuated tetravalent chimeric vaccine, which means that it is a non-naturally occurring and non-replicating virus consisting of parts of four different dengue viral serotypes “grafted” onto a yellow fever vaccine virus. Before registration, Dengvaxia was tested in over 30 clinical studies, including four phase III non-efficacy studies and two large phase III efficacy studies, one each in South East Asia and South America with over 10,000 and 20,000 healthy children respectively. This was not in any sense a compressed vaccine development schedule. The clinical studies were conducted over a period of roughly a decade and included tens of thousands of subjects and this was preceded and guided by decades of previous clinical and scientific research into dengue fever and dengue virus, including the conduct of viral challenge studies in non-human primates. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6816420/#CIT0052

At the time of registration and before roll-out of Dengvaxia, the summary of all of that prior research and development was:

-         The short-term vaccine safety was similar to that of the yellow fever vaccine platform upon which it was based.

-         The vaccine provided meaningful efficacy against any dengue disease caused by any of the four dengue viral serotypes types as well as efficacy against severe and hospitalized dengue fever.

-         The vaccine efficacy was better in older children compared to younger children.

-         The vaccine efficacy was better in children who already had dengue antibodies compared to those that did not. In other words, if a child had been exposed to dengue before vaccination then they did better than if they had never been exposed to dengue before vaccination. This analysis could only be done on a small subset of children with baseline blood samples.

-         Younger patients, particularly those aged 2 – 5 responded worse to subsequent exposure to dengue virus and both hospitalization and severe dengue disease occurred more commonly in these younger vaccine recipients. However, this effect was only seen clearly in the third year after first vaccination.

As a result of these data, Sanofi Pasteur, national regulators and the WHO initially decided to limit the use of the Dengvaxia vaccine to children of nine years or older. The vaccine was first registered in 2015 and relatively limited localised vaccination programs were then initiated in 2016 in both Brazil and the Philippines.

Meanwhile, Sanofi Pasteur, clinicians and scientists wrestled with the issue of how and why Dengvaxia had caused more severe disease in younger children and whether the same issue could be occurring in older children too. There were two possible answers:

1)     There was an age-specific effect in the under 5’s. In other words, young children somehow reacted differently to the vaccine when compared to older children. If that was true then older children would not be affected.

2)      There was a baseline dengue antibody effect. In other words, children previously not exposed to dengue virus reacted differently to children that had been previously exposed. In this case age would just be a surrogate marker for the likelihood of prior exposure to the virus but older children could possibly face the same problems if they had not been previously infected.

There was a reason to believe that it was the latter, because natural dengue viral infection was known to cause worse disease on the second infection compared to the first. Could the vaccine have been providing a silent but non-natural “primary infection” leading to worse disease later on? At the time, it was hard to tease out those two effects, since the younger children also tended to be those that had not previously been exposed to dengue infections. Making things more difficult, only a subset of children in the phase III studies had baseline blood samples, and so baseline antibody status could not be determined for most children.

Finally, in 2017 using an innovative new assay methodology, Sanofi Pasteur was able to confirm that the issue was not age but baseline antibodies. So, even in children older than 9, if they had no exposure to dengue virus prior to vaccination then they had an increased risk of severe disease and hospitalisation on subsequent infection. Based on this new data, Sanofi, the regulators and the WHO made a change to the vaccine indication in 2017, adding that the vaccine should not be used in anyone, even those aged 9 or older, without evidence of prior dengue virus infection. However, by then the vaccine had been used in hundreds of thousands of children in immunization campaigns in Brazil and the Philippines using the previous indication. There had been records of children hospitalized with dengue fever after vaccination and there were some who subsequently died. However, since these were vaccination campaigns and not clinical studies it is not possible to tell definitively if these severe cases were made worse by the vaccination or not. In the Philippines, where over 800,000 children were vaccinated, around 40 reportedly became seriously ill with subsequent dengue infection and Dengvaxia became a significant social and political problem. The Philippine government subsequently sued both Sanofi Pasteur and certain former government officials over the matter. More worryingly, Philippine public confidence in all vaccines has been considerably eroded by this experience and reduced vaccination rates have been followed by increased disease outbreaks, including over 500% increase in measles cases in 2018. The Philippines also suffered a very severe dengue epidemic in 2019. Not surprisingly the global uptake of Dengvaxia has also been relatively low.

As for the RSV case-study, there have been several possible explanations for the paradoxical effect of Dengvaxia on subsequent disease, but one of the most probable is what is known as Antibody Dependant Enhancement (ADE) where the presence of antibodies (induced either by natural infection or a vaccine) cause enhanced disease in some patients.

Both the RSV vaccine and Dengvaxia are examples of the relatively rare phenomenon of paradoxical disease exacerbation, this is possible for Covid-19 (there is some pre-clinical evidence to suggest that the ADE effect could possibly occur with vaccines against coronavirus) but any new vaccine is far more likely to have non-disease related adverse effects. As for any drug, it is highly unlikely to have a desired biological effect (disease protection) with absolutely no undesirable biological effects (side effects). Vaccines with common and even un-common severe side-effects are usually discontinued during or after large phase III clinical trials but rare vaccine side effects usually only show up during the large-scale vaccination campaigns that usually follow vaccine approval. There are numerous examples of this but perhaps one of the best known is RotaShield?, a vaccine developed by Wyeth Lederle for rotavirus infections in babies.

After 15 years of R&D, Rotashield was launched in 1988 and within 9 months had been administered to over 600,000 babies in the USA. By 1989 pharmacovigilance surveillance showed a possible increased risk of intussusception (a life-threatening intestinal problem requiring surgery) in the first two weeks after vaccination with an estimated occurrence of 1 case of intussusception in every 10,000 children vaccinated. Wyeth immediately withdrew the vaccine from the market. It took another 7 years for second generation rotavirus vaccines to reach the market and after extensive surveillance it appears that the safety signal for Rotashield may well have been spurious.

In the context of pandemic vaccines, two potential pandemic influenza vaccines have also had side effect issues revealed in post-marketing surveillance. In 1976, when a potential pandemic “swine-flu” started circulating in the US shortly before the presidential election, President Gerald Ford raced ahead with a high-profile campaign to vaccinate the nation The swine-flu turned out to be relatively mild but the vaccine may have been associated with Guillain-Barré syndrome (a form of paralysis) in several hundred people. In 2009, another potential swine flu pandemic led to a vaccination campaign with GSK’s Pandemrix to around six million people in the UK before it was found to cause narcolepsy in around one in every 55,000 people vaccinated.

There are numerous potential safety issues for vaccine developers to be aware of, they are sometimes real but often spurious. Unfortunately, even when real, we rarely know the exact mechanisms by which they occur, and they can be highly unpredictable. However, for the sake of this discussion (which to remind you is about lessons learned for the purpose of the upcoming Covid-19 vaccines), it does not matter what the exact cause is, because if something negative does happen for any Covid-19 vaccine, it will probably be unique to that particular vaccine. What is important though are the potential lessons learned from history and how we can apply them to Covid-19 vaccine development and roll-out. Here is my summary of a few key points that I believe we can and should learn from the past:

-         Even multiple large scale pre-clinical trials conducted over many years cannot answer all of the potential safety and efficacy issues for vaccines, because rare things happen rarely. The Dengvaxia program took decades, with over 30 clinical studies and thousands of patients, but still failed to resolve key safety issues before roll-out. Jonas Salk recruited 1.8 million children into his polio trial, but could not avoid the cutter incident. At “warp speed” this becomes far more challenging.

-         Even within clinical trials, safety issues may take time to reveal themselves. It took 9 months to see the worsening of disease for the RSV vaccine and 3 years in the case of Dengvaxia. Longer term follow-up and secondary outcomes measures are important in vaccine clinical trials. Again, the currently projected timelines for Covid-19 vaccine rollout in late 2020 or early 2021 would not allow for that sort of follow-up.

-         Both vaccines and natural infections impact the immune system and how those immunological factors interact with each other within each subject can profoundly affect both efficacy and safety outcomes. This can vary significantly across different age groups and we should not expect Covid-19 vaccines to have the same immunogenicity or risk:benefit in each age group, either in clinical studies or during vaccine roll-out. Each group needs specific and well thought out plans.

-         Roll-out of vaccines should ideally be relatively slow and in well managed programs until safety has been secured in each age group. This is going to be very hard to manage in the face of an ongoing pandemic and needs careful consideration and coordinated management around the globe. We may regret excessive speed.

-         When vaccines fail, for whatever reason, it can take years or decades before a next generation vaccine succeeds. In the case of the Rotashield failure, it took 7 years for next generation rotaviral vaccines to reach the market. In the case of the RSV vaccine failure, we still have no vaccine for RSV. For Covid-19, it is important that we have a diverse range of vaccines in development if we want at least one to succeed. (note: I intend to make diversity of vaccine approaches a theme for a future article).

-         Negative opinions about vaccines are very easy to create but very very hard to erase. What happens in this pandemic will have a profound effect, either positive or negative, on the public perception of vaccines and vaccination for many many years to come. This is not only about the pandemic and we need to play the long game.

Let me finish again by repeating:

-         Vaccines are a tremendous force for good in enhancing global health

-         Vaccines have an excellent safety and efficacy track record over centuries

-         Sometimes things go wrong with vaccine roll-out

-         We need development of Covid-19 vaccines to be rapid and efficient

-         We need mass vaccination campaigns against Covid-19 to be rapid and efficient

-         We need to minimise the risks that we face as we do that

-         This is not only for this pandemic but for the future of vaccination and the health of future generations

Disclaimer: The opinions in this article are entirely my own and are not endorsed by PsiOxus or any other organisation

#covid19 #vaccine #pandemic #pharmaceutical #2020 #hyperdevelopment