Vaccinating the globe: Workers on the Covid-19 vaccine production line at the Serum Institute of India in Pune. The factory is making the Oxford-AstraZeneca vaccination. Photo: Dhiraj Singh/Bloomberg/Getty Images
South Africa’s Covid-19 vaccine roll-out plan practically changed overnight after a small study’s results showed that the AstraZeneca jab, of which the country bought 1.5-million doses, doesn’t provide sufficient protection against mild disease caused by the new SARS-CoV-2 variant, known as 501Y.V2 or B.1.351. This variant is now the dominant form of the virus circulating in the country.
But what exactly does the new plan look like and what science and logistical issues are driving the decisions behind it?
We spoke to experts, looked at the data of studies (the AstraZeneca preprint study has not yet been published, so we worked with the press release and a powerpoint presentation), relistened to Sunday’s health department briefing and incorporated new information from Wednesday morning’s public briefing to break it down. There’s a lot about the new plan that’s still unclear, but we also found some interesting new information along the way.
1. What do the latest AstraZeneca results mean?
The short answer is: the AstraZeneca jab only provides about 10% protection against mild to moderate Covid-19 disease caused by the new variant, 501Y.V2. That’s way too low to make it worthwhile using — the World Health Organisation (WHO) says a Covid shot needs to be at 50% protection to be considered efficacious.
But we don’t yet know if the AstraZeneca vaccine provides sufficient protection against the severe Covid-19 disease that leads to hospitalisation and death as a result of 501Y.V2 — and we need to find out.
The longer answer is more nuanced: the South African arm of the Oxford University study (they are AstraZeneca’s research partner), that was led by Shabir Madhi from the University of the Witwatersrand, looked at whether the AstraZeneca jab could protect young people from mild to moderate Covid-19 disease (where someone displays at least one symptom, like a cough). It included approximately 2 000 participants, who were between the ages of 24-40 (the average age was 31 years) and considered to be at low risk of developing Covid.
The trial began towards the end of August last year before the 501Y.V2 variant was identified and widely circulating in South Africa. According to Madhi’s powerpoint presentation on Sunday, the trial was showing about 75% efficacy after one dose of the vaccine until the end of October.
But, when the new variant became increasingly dominant in November, that efficacy dropped to 10%. This is because, in many cases, people in the group receiving the vaccine had to fight Covid caused by 501Y.V2, as opposed to disease caused by the original form of the variant. Studies have shown that the new variant is not only more infectious than its predecessor, but has changed itself in such a way that it can evade the antibodies that our bodies produce in response to the original version of the virus, which is what the AstraZeneca vaccine was designed around.
In the study group, the overall efficacy (for both the variant and non-variant cases) was found to be almost 22% — with 19 Covid cases appearing in the group that received the jab and 23 in unvaccinated participants. But when researchers calculated the efficacy for the mild to moderate Covid cases caused by 501Y.V2, the efficacy lowered to roughly 10%.
Among the total number of Covid cases (so the variant and non-variant cases), two-thirds were mild Covid disease and the remaining third moderate disease.
Since the study group was young and relatively healthy — a group which is unlikely to develop serious symptoms — scientists couldn’t determine if the vaccine protected the trial participants against severe Covid.
Why would we want to know? Because people with severe Covid disease end up in hospital or die, and that’s what we’d like to prevent. Reducing these cases means our health system doesn’t get overwhelmed, and the effect of the pandemic on our economy when breadwinners, for instance, die, is alleviated. In the case of health workers, who are on the frontline taking care of people with severe Covid and are most vulnerable to infection, we’d want to protect them as much as possible.
Are the vaccines effective against the 501Y V2 variant?
2. Can we trust the recent AstraZeneca study’s results?
The answer is complex — we may need more data for a clear answer. The trial was done in a very small group of about 2 000 people, cautions Salim Abdool Karim, who chairs South Africa’s ministerial advisory committee on Covid-19. Preferably, clinical trials are conducted in larger groups — of between 40 000 to 100 000 — in order to make the findings statistically significant.
In a small sample, the number of symptomatic cases of Covid-19 that will occur in each arm of the trial will be low, says senior pharmacology lecturer at the University of KwaZulu-Natal Andy Gray. As a result, one can be less certain that the results detected in the trial will reflect the real situation in the population.
3. How will the new study results change South Africa’s vaccine roll-out strategy?
South Africa’s initial plan was to roll out vaccines in three stages: first 1.25-million health workers (because they deal with patients they’re the most exposed to the coronavirus) would get inoculated, then a group of 16.6-million people consisting of essential workers, people in congregate settings, those older than 60 years and people over 18 with comorbidities, and lastly everyone else over the age of 18.
The first health workers were supposed to get their first AstraZeneca shots (two doses are required) in about a week, but now we’ve found those shots won’t provide sufficient protection against mild to moderate Covid disease.
So on Sunday, Health Minister Zweli Mkhize announced that South Africa will temporarily halt its vaccine roll-out plan and instead do an implementation study that will assess how three different vaccines — Pfizer, Johnson & Johnson and possibly AstraZeneca — compare when it comes to protecting against severe Covid disease.
“We can still proceed with the roll-out, but we need to do it wisely,” explained Abdool Karim on Sunday. “We can do it wisely by taking a two-step approach.”
How will the vaccines be rolled out?
Abdool Karim told Bhekisisa that each vaccine arm will have the same number of participants. The presentations on Sunday’s briefing mentioned a figure of 100 000 study participants. The reason so many people will need to be included in the implementation study is that hospitalisations make up such a small percentage of total infections, so a bigger sample is needed to pick up whether a vaccine has any effect on how many people are admitted to facilities, Abdool Karim explains.
Johnson & Johnson vaccine
The only vaccine we have data for the level of protection against severe disease is the Johnson & Johnson jab. The trial, which included 43 783 participants across the United States, Latin America and South Africa, showed 85% efficacy against severe disease across all sites — including South Africa where a large proportion (roughly 94%) of cases were due to the new variant.
On Wednesday, Mkhize announced that the implementation study will start with the Johnson & Johnson vaccine (Johnson & Johnson is manufacturing the vaccine and Janssen developed it) and that South Africa could expect the first Johnson & Johnson doses to arrive in the country as early as next week. These doses, Mkhize said, will likely be provided to South Africa for free, as they will be allocated to the country from Johnson & Johnson’s research stock.
Earlier this year, South Africa procured nine million doses from Johnson & Johnson, which, according to a January presentation of President Cyril Ramaphosa, is expected to arrive in the second quarter of this year. Media reports on Tuesday stated that an additional 20-million doses are under negotiation.
Mkhize, however, said on Wednesday, that the Johnson & Johnson doses that are likely to arrive next week, will be a very small proportion of what we would need in total.
Some of the rest of the stock might come from the South African-based company, Aspen Pharmacare, that has been contracted by Johnson & Johnson to pack and fill some of its jabs. “We do anticipate that they [Aspen Pharmacare] will have their first batch manufactured around March. There’s a period of about a month that they’ve got to hold any stock [as a safety precaution],” he said.
Whether South Africa will be allocated some of Aspen’s stock, is, however, up to Johnson & Johnson.
According to the principal investigator of the South African arm of the trial, Glenda Gray, Johnson & Johnson and the national health department will announce a plan later this week explaining how the pharmaceutical company will expedite access to the jab for healthcare workers through the study.
The Johnson & Johnson vaccine has, however, not yet been approved for use outside of a research setting in South Africa. The company has submitted a rolling application with the South African Health Products Authority (Sahpra) but, according to Gray, the approval process is likely to conclude in about three months.
Meanwhile, the company has applied for approval to use its vaccine as part of the implementation study (this type of approval can be given within a few days).
“What we are proposing is to do an expansion as a study because we are waiting for the vaccine to be approved in South Africa. And doing the study rapidly in healthcare workers allows us not only to help healthcare workers, prevent severe disease, death and hospitalisation, but also to help us understand the vaccine as we roll it out in South Africa,” Gray said during the briefing.
This type of research is called an implementation study, Gray says. “In this case, it will provide an alternative route to allow people access to a regulated vaccine that works as soon as possible.”
Gray says that although not part of the country’s original vaccine strategy, this new approach “is a sort of pilot” for the eventual, national roll-out. “This pre-registration use [of the vaccine] has to interact with the national roll-out, it has to be part of it,” she says.
Gray adds: “Rather than duplicating efforts, a small-scale roll-out like this can use systems built for the national roll-out, such as the national vaccine registry, and then use information gathered in this setting to inform South Africa’s overall immunisation strategy.”
For the Pfizer vaccine, we only have lab data for how well it works to protect against 501Y.V2. The data shows that the jab is less effective against the variant than the original form of the virus, but we don’t know how that will play out in real life and if the level of protection for mild to moderate Covid and severe disease will differ.
South Africa has a bilateral agreement with Pfizer to buy 20-million doses, but these, according to Ramaphosa’s presentation, will only arrive in the second quarter of this year (between April and June). The country is also getting 117 000 doses via the international procurement mechanism Covax which is anticipated in the first quarter of the year, according to Covax’s interim distribution forecast.
Pfizer has applied for emergency use authorisation with Sahpra, but it’s not clear by when the process would be concluded.
4. What exactly is an implementation study?
At this stage of the pandemic we have a handful of vaccines which have released results and shown that they are able to offer protection against Covid-19. But the emergence of new variants means that there are now gaps in some of the data available.
So far, of the frontrunner candidates, only Johnson & Johnson has shown high efficacy in protecting people from developing severe cases and hospitalisation caused by 501Y.V2.
“Clinical data on severe disease and hospitalisation is essential,” said Abdool Karim during Sunday’s briefing. “We need to do large, simple roll-out implementation studies in South Africa to get that information.”
Usually, testing an intervention, such as a vaccine, means running a clinical trial to test its safety first and then efficacy. But showing how something works in that setting, where variables can be controlled, doesn’t necessarily equate to how effective something will be when it’s rolled out into the real world.
For instance, in a trial you can ensure everyone has the right equipment available, staff know how to vaccinate someone correctly, and the participants are selected to represent a specific group. But these factors may not remain consistent once you start to roll the vaccine out in real-life situations.
In the case of AstraZeneca’s vaccine, we have data that shows the vaccine only provides minimal protection against mild to moderate Covid. But we don’t know how it will perform against severe disease and if it can help reduce hospitalisations.
The solution: A focused implementation study.
Which vaccines will health workers be getting in the upcoming weeks?
According to a 2013 paper published in the BMJ: “Implementation research seeks to understand and work within real world conditions, rather than trying to control for these conditions or to remove their influence as causal effects.”
Instead of trying to find an answer to a broader question, implementation studies are more focused on the application of the research findings among those who most need it.
Doing an implementation study on healthcare workers can help answer key questions about a vaccine in a short time. This is because healthcare workers are at highest risk of getting infected with the coronavirus and would stand to benefit the most from any kind of protection against Covid.
An ongoing assessment of the vaccine in this setting can also help inform how that particular jab should (or shouldn’t) be incorporated into our roll-out.
Abdool Karim explains: “If the vaccine shows to not be effective in reducing hospitalisations, then we would need to offer those individuals another effective vaccine — either a booster of that vaccine with a 501Y.V2 booster, or to give them another vaccine.”
Unlike the clinical trial set-up to measure the efficacy of the vaccines, the implementation study will not have a placebo arm, Abdool Karim explains. In other words, nobody in the study will be given a dummy jab. Instead, researchers will use data from the original clinical trials to determine the point at which the shot provides enough protection against being hospitalised with Covid-19. Then, the jab will get the green light to be included in South Africa’s larger vaccine roll-out.
The study can exclude a placebo arm because it’s only measuring whether the jab has any effect on hospitalisation, he says.
The idea behind the proposed Johnson & Johnson arm of the implementation study is “to confirm the findings of the phase-three study that this vaccine protected against hospitalisation and death”, explains Gray.
“There’s been a proposal, given the uncertainty because of the variant, that it may be a good idea to do this kind of implementation science on all the vaccines so that you can have data about the effectiveness of the vaccine in our country.”
5. What will happen to the one million AstraZeneca vaccines we’ve already bought?
On Sunday, the deputy director general of the national health department Anban Pillay said: “Unfortunately, these vaccines [the one million AstraZeneca vaccines that arrived earlier this month] came through with an expiry date in April, which we only identified upon arrival.”
So will these doses expire before we can use them?
The AstraZeneca candidate consists of two, spaced out doses. The initial trial design allowed for one month between shots, but new information from the United Kingdom shows that efficacy increases if the jabs are given three months apart.
Researchers found that efficacy was 82.4% when participants were given the second dose 12 weeks apart, compared to 54.9% when the second jab was given less than six weeks after the first one.
The preprint study (not yet peer-reviewed) was posted on 1 February — the same day the vaccines arrived in South Africa — and is currently under review by the Lancet.
How far apart doses given in South Africa will be spaced out (if AstraZeneca is indeed one of the vaccines tested in the implementation study) will be up to Sahpra when the medicines regulator registers the jabs. But on Sunday Pillay said that the ministerial advisory committee on Covid vaccines advised that the jabs be given three months apart, in line with the latest findings.
Pillay also said that the department might consider “pushing forward” the expiring date. But is that even possible?
The answer is yes, in theory, but only if lab data supports it.
When vaccine manufacturers are making jabs to use as part of a clinical trial, they will already have the kind of data they need about how long the vaccine remains stable, explains Andy Gray.
Jabs used in trials may have a shorter expiration date than the ones used on the market. “There’s not a lot of long-term stability data initially. The company can extend the expiry date as more data becomes available.”
Kerrin Begg, a public health medicine specialist at the University of Cape Town’s faculty of health sciences, says because AstraZeneca trial data only became available in September of 2020, “we’ve only got data available for six months”.
When the April expiry date comes around, and the stock is still unused, the manufacturing company can be asked to determine whether the stock still works, and whether it can be extended, Gray says.
“It’s not unusual, even within clinical trials, for an expiry date to change.”
Are the vaccines still safe?
So will the jabs “go off” on exactly the day of their expiry date?
The answer is no. “The expiry date is not an off switch. It’s a statistical projection that reflects the point where we can no longer be sure that we’ve still got a product which is potent,” Gray says.
To find out for how long a vaccine is usable, companies run a series of tests on the jabs to try and work out how long the shot would last under normal circumstances and what would happen should there be exceptions, such as a change in the temperature at which it is stored.
Although expiry dates can be moved on as more data become available, vaccines (and other medicines) do, however, become unusable eventually.
In the case of a vaccine, that could mean the active ingredient no longer elicits the immune response needed to protect people from infection.
There’s a whole range of other aspects that can go wrong when medicines expire, Gray explains. “In very rare cases, if there is exposure to high temperature a product can be created in the medicine which is toxic. But mostly, it’s that the medication loses its action [doesn’t do what it’s supposed to].”
Gray concludes: “We wouldn’t use this vaccine after April unless we’ve got permission to extend the expiry date from the manufacturer. I would be quite confident to use it if it had an extended expiry date.”
On Wednesday, Mkhize said, South Africa would have used all its AstraZeneca stock by April as part of the country’s original vaccine roll-out plan (we would likely have used all the stock as first doses). With the country’s new, adjusted strategy, we don’t yet know if the AstraZeneca vaccine will be part of the plan. Mkhize said a committee of scientists “will continue with further deliberations on the AstraZeneca vaccine use in South Africa”. If the scientists decide the jab should be included in the implementation study that our new plan will kick off with, “the vaccine will be swapped [for another batch from the Serum Institute] before the expiry date. By exchanging unused vaccines before the expiry date, the department of health will ensure that the acquired AstraZeneca vaccines do not become wasteful and fruitless expenditure”, Mkhize said.
Earlier this year, South Africa also procured close to 3-million AstraZeneca jabs through the international procurement mechanism, Covax. Mkhize said the country will “look at the value [of the AstraZeneca jabs] and what alternatives there are that we can swop them for”. “There is flexibility from Covax,” he said.
Mkhize also said that several governments have contacted South Africa to ask if they could buy our AstraZeneca stock for use in their own countries.
6. Things we don’t know about the new vaccine roll-out plan
- Whether the AstraZeneca vaccine will be included in an implementation study — and at what dosing interval (four weeks or three months) — to determine whether the jab protects against severe disease. A team of scientists is meeting every day this week to make this decision;
- Whether the implementation study will include mixing of vaccines, i.e. a first shot of one jab followed by a second shot of another vaccine;
- How many people will be included in the study overall. It is estimated that there will be 100 000 healthcare workers receiving each vaccine but this could change as the study is underway;
- When the implementation study will start and for how long it will go on. Each vaccine candidate will have to be individually assessed based on its initial clinical trial results to see at what point the jab can move to wider roll out or if it needs to be halted completely;
- Whether other vaccines (on top of Pfizer, Johnson & Johnson and possibly AstraZeneca) will also, eventually, be included in the implementation study. At this stage, it seems scientists will look at the data of each new vaccine option before deciding whether it can just be rolled out or whether it first needs to be part of the implementation study; and
- How many people South Africa aims to have vaccinated by the end of the year with its new roll-out strategy — the original goal was 40-million, which many experts argued is unrealistic.
This story was produced by the Bhekisisa Centre for Health Journalism. Sign up for the newsletter.