麻豆传媒

IT IS the ultimate exit strategy from covid-19. A safe and effective vaccine is of 鈥渃ritical importance to world health鈥�, the World Health Organization (WHO) has said.

Vaccine developers are working flat out to make good on that. Last week, the US pharmaceutical giant Pfizer and its German partner BioNTech announced positive-looking results from their ongoing phase III trial, the last stage of testing whether a potential vaccine is safe and effective. The interim results showed a headline success rate of 90 per cent, meaning that nine out of 10 trial participants who caught the new coronavirus had received a placebo rather than the vaccine.

鈥淗ow long will immunity last? The desired answer is 鈥榝orever鈥�, but realistically a year would be positive鈥�

The news got some people very excited indeed. Asked on BBC radio whether these results meant a probable return to normal by early next year, John Bell at the University of Oxford and a member of the UK government鈥檚 coronavirus channelled Meg Ryan in When Harry Met Sally and said: 鈥淵es, yes, yes!鈥� Many listeners no doubt thought: 鈥淚鈥檒l have what he鈥檚 having.鈥�

A few days later, another phase III trial 鈥� this one being run by the in Russia 鈥� reported even better interim results: a success rate of 92 per cent. And earlier this week, US company Moderna announced 95 per cent efficacy from its ongoing phase III trial (Moderna coronavirus vaccine trial produces best results yet).

So things look good. But we are still a long, long way from a vaccine that will get us back to life as normal. That is in no small part due to the huge challenge of manufacturing, distributing and administering one (see What will it take to get a covid-19 vaccine to the world?), plus the reluctance of a significant minority of people to get vaccinated (see Heidi Larson interview: How to stop covid-19 vaccine hesitancy). However, it is also down to trial constraints, which leave a number of questions around safety and effectiveness. If you thought those were the things the trials could give us all the answers to, think again.

Complex question

鈥淚n my line of work, I get asked this nearly every day from my friends and family: will this particular vaccine or that particular vaccine work?鈥� says Susanne Hodgson at , which researches vaccines. 鈥淎nd I鈥檓 always stumped by how to deliver the answer quickly. Because it is a complex question.鈥�

The least complex part of the question is, how long will immunity last? The desired answer is 鈥渇orever鈥�, but realistically a year would be a very positive outcome. In April, the WHO published an of what would constitute a safe and effective covid-19 vaccine. On length of protection, it said its preferred outcome was at least a year, but it would accept a minimum of six months 鈥� though pointing out that this 鈥渕ight not be demonstrated in initial clinical studies鈥�. The US Food and Drug Administration (FDA) has set the same goal, and the UK vaccine task force says it is .

As yet, however, even that six-month bare minimum hasn鈥檛 been attained. The Pfizer and BioNTech phase III study began vaccinating people in late July and has only just finished recruiting volunteers. As a result, it won鈥檛 have an answer until February at the earliest, because the vaccine requires two shots, three weeks apart. We simply don鈥檛 know yet how long protection from any vaccine will last.

鈥�6 months
The minimum protection the WHO requires a vaccine to provide鈥�

鈥�50%
The lowest acceptable level of vaccine protection set by the WHO鈥�

鈥�90%
The protection achieved by Pfizer and BioNTech鈥檚 candidate vaccine, according to early results鈥�

Time isn鈥檛 something that the vaccine developers have control over. But they can control other aspects of trial design, and these raise some major questions, says at the University of Maryland School of Pharmacy. Last month, he wrote an , of which he is an associate editor, entitled 鈥淲ill covid-19 vaccines save lives? Current trials aren鈥檛 designed to tell us.鈥�

How is it possible that these trials aren鈥檛 designed to reveal whether the vaccines on which we are pinning so much hope will actually save lives?

The problem relates to the trials鈥� so-called 鈥渆nd point鈥�, the bar against which success or failure is judged. It is set fairly low. The WHO and FDA have both said they will accept a vaccine that provides at least 50 per cent protection against infection. That means the trials need to show that no more than half as many people who received a vaccine get infected as people who got the placebo.

Putting aside the 90 per cent plus results for now, that isn鈥檛 good enough, says Doshi. For one thing, the 50 per cent threshold for the trials could mean that a vaccine that is only actually 30 per cent effective makes it through, as the error bars representing uncertainty in the trial data are quite large.

Another issue is that the success rate obtained in a vaccine clinical trial often exceeds that seen in the real world. As Hodgson puts it, 鈥渧accine efficacy does not always predict vaccine effectiveness鈥�. There are various reasons for this, she says. A major one is that the deployment of a vaccine on the ground, to millions or billions of people, is much more challenging than administering it within a tightly regimented clinical trial.

That is especially true of a two-shot vaccine that relies on people showing up to two appointments, often weeks apart. For this reason, the WHO says it would prefer a one-shot vaccine. However, all but one of the 12 vaccines in phase III trials require a couple of shots. 鈥淚 think it would be prudent to anticipate that we may see some differences between covid vaccine efficacy in clinical trials and real-life settings,鈥� says Hodgson.

According to , at the University of Pennsylvania and a member of the FDA鈥檚 , the FDA is likely to accept six months or even less of efficacy data, even though it usually asks for at least two years and most vaccine trials last even longer. Admittedly, this is an emergency and we have to accept some uncertainty, he says, but we need to be ready to be 鈥渦npleasantly surprised鈥� by a vaccine that delivers weak or short-lived immunity. And the first vaccine to succeed is rarely the best, he warns.

鈥淲e could end up with vaccines that reduce the risk of mild infection but not the risk of death鈥�

Nonetheless, the 90 per cent plus success rates seen so far suggest that these vaccines will easily exceed the 50 per cent threshold, so this issue may just be theoretical. The phase III trials aren鈥檛 complete yet, but it would take a major reversal to erode those high initial figures. Even with a fall to 80 or 70 per cent, a vaccine鈥檚 impact would still be far above the WHO鈥檚 minimum requirement. 鈥淥f course, we all want a vaccine which is as efficacious as possible,鈥� says Hodgson. 鈥淏ut I think given the scale of the pandemic, the rates of transmission and the morbidity and mortality we鈥檙e seeing, even a partially efficacious vaccine could have a really significant impact.鈥�

Mild cases only

Despite this, the trials aren鈥檛 going to tell us what, if any, effect a vaccine has on severe illness, according to Doshi and others. On 22 October, he told a that 鈥渦nless urgent changes are made to the way the trials are designed and evaluated, we could end up with approved vaccines that reduce the risk of a mild infection but do not decrease the risk of hospitalisation, [intensive care unit] use or death.鈥�

This seems outlandish, but again it comes down to the trials鈥� end point. In all the phase III trials, this is defined as the prevention of mild covid-19 symptoms, such as a cough, fever, headache or sore throat. Any participants with these symptoms are tested to confirm whether or not they are infected by the SARS-CoV-2 virus. If there are many more such cases in the placebo group than the control group, we can see that the vaccine is working 鈥� in preventing mild cases, at least. But such a result tells us next to nothing about whether the vaccine is stopping infected people from getting really sick. The issue is compounded by these vaccines being tested in a subset of the population that is predominantly young and healthy, and so at relatively low risk of getting severe covid-19.

鈥淚n a deadly pandemic, we want to see efficacy data demonstrating a reduction in severe disease and long-term consequences,鈥� says Doshi. 鈥淓fficacy against a transient, mild illness in relatively healthy people is far less important than protecting the most vulnerable.鈥�

鈥淭he trials appear designed to answer the easiest questions, not the most important ones鈥�

He accepts that people who are protected against catching the disease cannot, by definition, go on to develop severe covid-19 or die from it. But that isn鈥檛 the point. 鈥淭hat is true, if you are talking about a single person. But a vaccine will not have identical efficacy in all populations,鈥� he says. 鈥淟et鈥檚 say it works really well in healthy adults, but provides very little protection in frail elderly [people], to choose one high-risk group. In this scenario, your trial can demonstrate an effect against mild disease, but you would still have all the serious cases because the vaccine is not protecting the frail elderly.鈥�

More than , but they can be poorly represented in trials.

In the , 40 per cent of phase III participants are supposed to be 55 years old or over, but the figures released from the trial don鈥檛 include an age breakdown. Neither company responded to 麻豆传媒鈥榮 requests for that information.

The Gamaleya vaccine team told 麻豆传媒 that people aged up to 60 were vaccinated and included in the data, but again provided no actual numbers.

In any case, people aged 55 or even 60 hardly qualify as 鈥渇rail elderly鈥�, who often have weakened immune systems and don鈥檛 respond well to vaccines. Age-related decline in the immune system can kick in as early as 55, but there is huge variation from person to person, says Deborah Dunn-Walters, an immunologist at the University of Surrey, UK.

Some of the vaccines that have yet to report any results are being tested in older groups. In the University of Oxford and AstraZeneca vaccine trial, for instance, .

The Moderna results are more promising because this vaccine was given to people over 65, and some of those who became ill with the disease were in this age bracket.

Another issue is that without regularly testing all participants, a clinical trial could fail to pick up large numbers of asymptomatic infections.

One phase III trial 鈥� of the vaccine being developed by the University of Oxford and AstraZeneca, which Hodgson is working on 鈥� is testing every participant for the virus each week. As a result, its findings may exclude the possibility of missing lots of asymptomatic infections 鈥� at least for its own vaccine, which works in a very different way to Pfizer and BioNTech鈥檚 and Moderna鈥檚.

The decision to omit severe disease as a primary end point is unusual. According to a research paper by an international group of industry, government and academic researchers in the Annals of Internal Medicine, severe disease is an end point 鈥渦sed in virtually all vaccine efficacy trials鈥�. The group urged all vaccine developers to include severe covid-19 as an end point in their trials.

Doshi says the trials appear designed to answer the easiest questions in the least amount of time, not the most clinically important ones.

It is possible to do a covid-19 clinical trial with severe disease as an end point, says Hodgson, but it would be a major undertaking because that outcome is still quite rare. 鈥淭he studies do not have adequate numbers of patients to be able to reliably tell us if they prevent severe disease,鈥� she says. 鈥淲e will need to give these vaccines to much larger populations in order to collect that kind of data and get that output.鈥�

Pfizer has said that it and BioNTech are collecting data on severe disease as a secondary end point 鈥� but the numbers still aren鈥檛 big enough. Hodgson says this may be an issue for all the trials. 鈥淚t鈥檚 unlikely that they鈥檙e going to have sufficiently sized trials to reliably get an indication about whether vaccines prevent severe disease,鈥� she says.

Worse-case scenario

Another key question the current trials are too small to answer is whether a vaccine prevents people from catching and transmitting the virus. This might sound like a crucial feature of a vaccine but it isn鈥檛: a vaccine is designed to prevent people getting ill. It is, however, important because it is necessary (but not sufficient) to achieve herd immunity.

In fact, vaccines could, in theory, make matters worse. If they suppress disease but don鈥檛 stop people from catching and shedding the virus, they effectively convert symptomatic cases into asymptomatic ones. That may lead to large numbers of infected people who aren鈥檛 aware they have the virus going about their daily lives while releasing virus, rather than self-isolating. This 鈥渕ay paradoxically increase transmissions鈥�, the Annals paper says.

鈥淎 worst-case scenario is a vaccine that reduces disease while permitting viral shedding,鈥� wrote Marc Lipsitch at the Harvard T. H. Chan School of Public Health and Natalie Dean at the University of Florida in a recent . 鈥淭his could fail to reduce transmission or conceivably even increase transmission if it suppressed symptoms.鈥�

Yet another thing that the data so far cannot tell us for sure is whether the vaccines are completely safe. Pfizer, Moderna and the Gamaleya National Center all say they haven鈥檛 seen any severe adverse reactions among participants, but are continuing to collect data to be sure that they won鈥檛 occur.

Peter Marks, who directs the FDA鈥檚 Center for Biologics Evaluation and Research, the body that evaluates applications for vaccine licences and emergency use authorisations (EUAs), has said that he needs to see safety data showing that no volunteer has had a severe adverse reaction within two months of receiving their second shot. The FDA can issue EUAs as a way to fast-track medical products in exceptional cases. Pfizer has said it will have collected this safety data in the coming week, at which point it will apply for one.

Nearly all adverse reactions ought to be picked up within six weeks of a second shot, says Offit, so side effects are probably less of a concern than efficacy, although rare side effects might take longer to spot (see 鈥�Phase IV trials鈥�).

There are a couple of other unknowns too. We don鈥檛 know how people who have had the virus and recovered will respond to any of the vaccines. Pfizer has been vaccinating these so-called seropositive people but excluded their data from the latest analysis. We also don鈥檛 know whether the vaccines will put pressure on the virus to mutate.

All in all, as Hodgson says, the seemingly simple question 鈥渄oes this covid-19 vaccine work?鈥� is surprisingly hard to answer.

In the end, of course, this could all just be speculative bellyaching, and none of these potential problems will actually materialise. Thus far, we have seen interim results from three of the that have reached phase III trials. More will undoubtedly follow.

鈥�200 or more coronavirus vaccines are in development鈥�

鈥�12 of these are in phase III trials鈥�

鈥�3 of which have published early results鈥�

Jeremy Farrar, director of the Wellcome Trust, says we should think of vaccine development as the creation of a portfolio rather than the search for a single magic bullet. Weaknesses in one, such as not working well in older people, may be covered by strengths in others.

Despite her words of caution, Hodgson is optimistic about the future too. 鈥淭here are more than 200 vaccines in development, which is a phenomenal number, and using a variety of vaccine technologies,鈥� she says. 鈥淚t鈥檚 nearly impossible to predict exactly when, but I think the likelihood is we will have a number of candidates that are efficacious.鈥�

So not quite yes, yes, yes! 鈥� at least not yet. But barring some disaster, we will eventually be able to have what the brave volunteers in the trials are having.

Health Check newsletter
Get a weekly round-up of health news in your inbox

Future of Food and Agriculture
A one-day virtual event on 28 November

听

How the Pfizer/BioNTech phase III trial works

More than 43,500 people are recruited to the trial

Around half are given a vaccine, half get a placebo. Neither participants nor researchers know who is in which group

When participants report mild symptoms like a cough or fever they are tested for the coronavirus

Once a certain number of people are confirmed as having had covid-19, called a 鈥渃heckpoint鈥�, the results are 鈥渦nblinded鈥� to reveal whether these positive cases had been given a vaccine or a placebo

So far, of 94 covid-19 cases, 90 per cent were among those in the placebo group

The trial will end when there have been 164 confirmed infections, the final checkpoint

听

Phase IV trials

Once vaccines are approved, they are usually closely monitored to detect any rare but potentially serious side effects that the trials were too small to spot. This evaluation, often called a phase IV trial, usually runs for a year or two because rare adverse reactions may take months or even years to be detected, says Susanne Hodgson at the University of Oxford.

One rare but serious problem is 鈥渧accine enhanced disease鈥�, in which vaccinated people who go on to catch the virus their vaccine targets become more ill than they would have without the vaccine. It occurs when the immune response elicited by a vaccine backfires and actually helps the virus cause disease rather than hinder it.

Hodgson says this was seen in animal experiments on vaccines for SARS and MERS, diseases caused by coronaviruses closely related to SARS-CoV-2, the virus behind covid-19. 鈥淏ut importantly, this hasn鈥檛 been seen in the animal models of covid-19 vaccines to date, and there鈥檚 no signal yet that we鈥檝e seen anything like this in the clinical trials,鈥� she says.

It is also worth noting that Pfizer and BioNTech鈥檚 and Moderna鈥檚 vaccines use an unproven technology, rather than being based on the usual viral proteins or weakened form of the pathogen 鈥� so they could spring new surprises down the road.