The competition for the “honor” of being the most confusing part of clinical trial results is fierce. I don’t think technical terms…
We’re now into the eighth month since sequencing of the genome of SARS-Cov-2, the virus that causes Covid-19. And this is the third of my month-by-month posts about the vaccine race. This month’s new data confirms that there are likely to be critical differences between vaccines.
Meanwhile, the harsh realities of establishing trials with large numbers of volunteers is starting to slow down some vaccines, while others seem to be creeping ahead. Then there’s the strange case of Sputnik V, the vaccine from Russia that the country’s president shot to worldwide fame, amid lots of claims, that set off a flood of inaccurate reporting. All of which raised another question that’s a theme for me this month: just what do people mean when they say it’s essential to wait till phase 3 trials are “completed”? But before we get to all that, let’s start with the new data.
Vaccines with fewer adverse effects & the possibility of single-dose vaccination?
Those are the 2 themes for me in a batch of 5 major results published or posted since my last post: 2 are clinical trial reports, and the others are preclinical (non-human animal) studies.
I think we need to wait for the phase 3 trials to be able to know much about comparative benefits. Early data on adverse effects of the versions of the vaccine going into phase 3 trials, though, is already raising the possibility that there could be big differences in tolerability between vaccines – if those differences carry through in larger groups of people more representative of the general community.
NVX-Cov2373 by Novavax
First up is a preprint report of a phase 1 trial for the 2-injection vaccine produced by the American company, Novavax. That’s another of the companies in this race that has never gained regulatory approval for a vaccine before. The vaccine is a nanoparticle one – a much smaller size of antigen – and they are using an adjuvant (an additional substance to boost immune response).
The trial was run in Australia, and included 106 people who got the vaccine in one of 2 doses, with or without adjuvant, and 25 who got placebo injections. The report includes data on antibodies and T-cells, and Derek Lowe discusses this in detail. This is the trial that has the most detailed, precise adverse effects results of any so far. Comparisons are risky at this point, but adverse effects were strikingly less common for this one than those reported for vaccines at the higher end of adverse effects so far – if one of these doses is the one that goes ahead.
For some perspective on what the difference is at this stage, here’s one adverse effect: only 1 person had a fever after the Novavax vaccine – a mild fever in 1 of the 25 people on the higher dose with adjuvant (4%). In the (much larger) phase 1/2 trial for the Oxford/AstraZeneca vaccine, 18% of the people who got the vaccine without preventive acetaminophen had fevers; only half of them mild. In the (also small) phase 1 trial for the NIH/Moderna vaccine, 40% of the people on the dose going to phase 3 trial had fevers; mostly mild.
PicoVacc (Coronavac) by Sinovac
This vaccine produced by a company from China is already in phase 3 trial, and it’s the first report of an inactivated (“old school”) Covid-19 vaccine in humans. They have published preclinical results, but not phase 1 trial. Now, though, they posted a preprint of phase 2 trial results. There were 480 people getting 2 injections of the vaccine, and 120 in the placebo group.
This wasn’t as well-reported as some others have been. The lack of phase 1 trial results is probably explained by the authors’ comment that the earlier batch of vaccines produced for that trial had relatively poor immune response compared to the one they have now. The report includes data on antibodies only, not T-cells. After the second injection, 97% of people had seroconverted, although not as often for the people over 50. Again, adverse effects were low, and mostly described as grade 1 (presumably mild). About 3% had fever.
Update shortly after the post was first published (and data added to the table below) – this trial was published with an accompanying editorial: error note – I wasn’t consistent with the way I was counting reports that covered both phase 1 and 2 trials. As of August 14, there were 9 trial reports…
Preclinical results – and that single-dose vaccine
There have been 3 sets of new non-human animal study results that I think are worth considering in particular: INO-4800 from Inovio (a DNA-based vaccine); Ad26.COV2.S from Johnson & Johnson (a single-dose adenovirus-based vaccine); and an unnamed inactivated one from the Chinese Academy of Medical Sciences. That last one is in phase 1/2 trials in China, including one for 471 people 60 years of age or older.
The first 2 are going into phase 3, though: and the Ad26.COV2.S single-dose one is the largest trial planned so far, at 60,000 participants.
A vaccine that needs only a single dose has enormous implications. There is current global capacity to produce at least 2-4 billion doses of Covid-19 vaccine by the end of 2021, according to a survey by the Coalition for Epidemic Preparedness Innovations (CEPI). If global capacity doesn’t increase greatly, even in the best case scenario of successful vaccines aplenty and coordinated and unselfish production, there won’t be anywhere near enough vaccine for the 7.8 billion global population for a long time.
The animal studies aren’t reliably predictive of how things will go in humans. If you’re interested in them, check out Derek Lowe’s post where he has assembled the non-human primate results for 5 vaccines, including those first 2 above.
Overview of phase 3 trials with earlier phase results
There’s good news and bad news this month. It’s distressing that there are so many places in the world with Covid-19 outbreaks, but it does make testing vaccines more feasible. The capacity to mobilize large-scale clinical trials in areas with significant outbreaks is still proving to be a major constraint, though. It looks as though a couple of the early frontrunners might be losing ground to others, but the picture can shift quickly.
CanSino, from China, has had a public commitment for a 5,000-person trial, but that’s only a small proportion out of its planned 40,000-person trial. Non-coronavirus political disputes have hit its plans too: the government their end reportedly hasn’t released the shipment of vaccine for a phase 1 trial in Canada. And the team for the Oxford/AstraZeneca vaccine have been saying since at least July that a 30,000-person phase 3 trial is coming in the US: but that doesn’t seem to have materialized yet. They have also pushed back the completion date for their phase 3 trial in Brazil by a few months.
As you’ll see from the table below, others are pushing forward on larger trials. The table includes vaccines that have published or posted some results, and/or have phase 3 trials registered. That means the vaccine from Russia that’s in the headlines doesn’t qualify – I discuss it separately below.
The phase 3 trials below have either already started, or are expected to start in August or September. Others are planned for October according to their developers, but there is no trial yet registered – for example, one in Canada, backed by the tobacco giant, Phillip Morris, that uses technology based on a plant from the tobacco family. Egypt also looks likely to join in manufacture and trials of one or more of the vaccines from China, but it’s not yet clear which.
Johnson & Johnson
(Brazil, Chile, Colombia, Mexico, Peru, Philippines, South Africa, USA)
? number (Mexico)
(USA – coming: Argentina, Brazil, Germany)
UK (Oxford Uni)
2,000 (apparently extended to 5,000)
(Reportedly seeking approval for trial in India)
|Primate, non-primate||131||n.a.||Unpublished||(Reportedly 30,000 planned)|
|BBIBP-CorV x 2 (1 Wuhan, 1 Beijing)|
|Primate, non-primate (Beijing only?)||n.a.||320 (Wuhan)|
|n.a.||45,000 (Wuhan + Beijing)|
(UAE) (2 records in different trial registers, at 15,000 & 45,000)
6,000 (Reportedly in Bahrain – unclear if additional)
(Reportedly also waiting for approval in Bangladesh)
Vaccines with published or preprint results (and/or Phase 3 trial register entry)
* a combined phase 1/2/3 trial, with 2,000 phase 1/2
** a combined phase 2/3 trial
n.a. = not applicable (where only combined phase 1/2 trial so far reported)
Sources: tagged in my public Zotero collection
Update [August 21]: Note it’s not clear if the US trial below has started injecting people – so it could be more than 3 weeks behind…
Enough eligible volunteers: the next big crunch
The number needed just for the trials in the table above is already approaching a quarter of a million people – and that could double in the near future. Those trials have to fill up, and the sooner the better. That’s not just because the world needs answers fast. The people have to be recruited while SARS-Cov-2 is still circulating in that community at a rate that can show whether the vaccinated groups are actually being protected from the disease and infection. If the infection curve is crushed early in the trial, those numbers won’t be enough to show a difference between the vaccine and placebo even if there is one.
Informing and screening people is a massive task; so large, that it’s one of the most time-consuming parts of a trial. These particular trials have a lot of exclusion criteria, so only a proportion of the people who put their hand up will go the distance – and that proportion might be small.
I could find something about how trials for 4 vaccine groups are going – and it looks likely that at least one of them could be fully recruited, or close to it, by the time I write my next post.
Sinopharm’s BBIBP-CorV vaccines are 2 of the inactivated vaccines from China. The trialists are aiming to recruit 15,000 people in the United Arab Emirates (UAE). They began in early July. A month later, they announced they had vaccinated their 5,000th volunteer. They had recently opened up a major walk-in center at a convention center to speed this up: it’s open 12 hours a day, with the capacity to screen 1,000 people a day.
Update August 16:
One of the other ways that trial, and others, are seeking to speed up recruitment is to expand the number of centers (usually research hospitals) recruiting people across a country. That’s happening fast with another inactivated vaccine from China, Coronavac, in Brazil. As of August 6, there were 12 centers around the country recruiting healthcare workers, and a congressional committee has been told that some of them at least are expected to reach their targets by the end of August. A smaller phase 3 trial for this vaccine in Indonesia started in 6 centers this week.
Next up: the Oxford/AstraZeneca vaccine. Zachary Brennan reported progress data via a spokesperson from Oxford University, and I’ve added the timeframes. The UK phase 2/3 trial that started at the end of May has recruited 7,574 out of 10,260 needed, and the phase 3 trial that started in Brazil at the end of June has recruited 3,032 out of 5,000 needed.
According to contents of an internal trial email reported on by Elizabeth Cohen at CNN, the NIH/Moderna trial started at the end of July and got to 4,536 by August 7, and recruitment is speeding up as more centers come on stream. Moderna apparently said this puts them on track to have their 30,000 by the end of September, but Cohen reports that was met with some skepticism from experts she interviewed. After community pressure, people with HIV will no longer be excluded from participating in this trial.
There’s a story by Phil Galewitz and JoNel Aleccia digging into how recruitment for the NIH/Moderna and BNT/Pfizer trials is being done in the US, but without data on progress.
Update August 23 for the 2 big US trials that started participants’ injections at the end of July:
Interested in putting your hand up for a Covid-19 vaccine trial? There’s a guide in my previous post.
The hair-raising ride of Sputnik V and media reports about it
There’s been a lot of media claiming this is the first approved vaccine for Covid-19, but that seems to be something of an exaggeration, from what I can see. It’s not full approval: it only has a registration certificate for use in limited circumstances in 2020, and can’t be used in the general population till 2021. That sounds more like a form of emergency use authorization, which would put it in the same category as CanSino’s vaccine. That vaccine gained an emergency use authorization for the military in China back on June 29. (See my first Vaccine Race post for more detail on that.)
Back in April, Russia introduced legislation to expedite emergency use of new drugs and vaccines for Covid-19. They weren’t the only ones, although their measures appear to be on the extreme end from the little information internationally that’s easy to find in English. In my previous post, for example, I also reported on the European Union reducing its GMO requirements to allow for genetically-engineered Covid-19 vaccines to navigate the system more quickly.
Just how hair-raising is the Sputnik V ride? It depends on detail we don’t have about what’s really happening with that emergency use, and how much of this is only political theater or PR to drum up vaccine trial and pre-order deals. But it looks like a scary bare-knuckle ride.
Very little information about Sputnik V is publicly available in English. The vaccine’s 2 injections are apparently based on a different adenovirus in each, Ad25 and Ad26. It’s developed by the Gamaleya Research Institute of Epidemiology and Microbiology. There have been 2 phase 1/2 trials, each with 38 people: both were in Moscow, one at a military hospital, and the other at a medical university. The results haven’t been reported. The phase 3 trial is reportedly to be registered “soon”. A website for Sputnik V reportedly says it will be for more than 2,000 people, and was due to start on August 12 in Russia, the UAE, Saudi Arabia, Brazil and Mexico. But it’s not clear that enough vaccine has been manufactured for much use.
This has been yet another case study of the triumph of Covid-19 vaccine bluster and politics over facts. First there was the political theater/PR from Russia that it was the first approved vaccine – when it was apparently neither fully approved, nor the first emergency use authorization. Then there was poor and uncritical reporting blasted across global media, amplifying the PR claim, missing key critical information, and misreporting other parts of the picture.
Take the New York Times report as an example. The vaccine is reported as: being approved without explaining the limitations in the Russian context; having had only a single phase 1/2 trial with 38 people; and being based on a single adenovirus.
Then the reported phase 3 trial is criticized for being in only 2,000 people: “All other Phase 3 trials of coronavirus vaccines currently underway are more than ten times larger than that, with 30,000 volunteers apiece”. As you can see from the table above, that’s not right either. There were some accurate statements, but most of the fundamentals were wrong.
There was also this now-common reporting trope about Covid-19 vaccines: “So far, they have only detected mild or moderate symptoms and no severe side effects”. I wrote about why that’s not only wrong and thus ethically problematic, but also potentially dangerous for public trust in my previous post.
All up, it’s another reminder to exercise maximum care when you read about Covid-19 vaccines!
And what does “when phase 3 trials are complete” even mean, anyway?
It may sound obvious, but I think mileage may vary on this – and dramatically. In the outcry when a prominent US academic argued the US should do what Russia soon did, I found the use of phrases like this frustrating. There’s another outbreak of it now in response to the news about Sputnik V. I think there are at least 4 vague or concrete concepts this could mean:
- When the first data is in after the last person enrolled gets their last injection;
- Some unspecified time after the last person enrolled gets their last injection – a few months, or six months, I suppose;
- Preliminary data at the trial’s primary completion date – that’s the date the last data on a primary outcome measure is collected; and
- When the trial reaches its completion date.
The differences here are huge. For trials of some vaccines, one of the primary effectiveness outcomes is at 2 weeks, so some interim phase 3 trial data could be viewed by data monitoring boards in the next month or so. Whereas actual completion dates for trials are generally when they have collected outcomes a year after the last dose was given. Take the Oxford vaccine trial that started in April. It’s a phase 2/3 trial. Its completion date is listed as May 2021. For the big trials from NIH/Moderna and BNT/Pfizer, for example, it’s currently estimated as late in 2022. (Their primary completion dates are registered as roughly by mid-2021.)
Projected completion can move, if things go more or less quickly than planned. Trials can also be ended early. That could happen if the data monitoring group for the trial decides that the vaccine is causing too much harm. It could also be stopped early if interim data show the vaccine is achieving definitive benefit, and it’s decided that it’s no longer ethical to keep giving people a placebo. (I’ve written about the mess that trials stopping early for benefit can leave behind them, here.)
My position on this? If the pandemic is still raging on, or even escalating as seems possible, then emergency use authorization for particular groups seems both likely and reasonable, after substantial impressive phase 3 data plus equally impressive follow-up safety data from phase 1 and 2 trials – depending on the track record (or lack of one) for the type of vaccine, and the quality assurance for its manufacture.
Even 12-month follow-up from a large phase 3 trial is still provisional to some extent. Our uncertainty is greatly reduced, but it’s not gone. So watch out for exaggerated claims like, “We will know if it’s safe and effective when the phase 3 trial is complete”. These trials don’t have enough people or time to know about very uncommon/rare and longterm adverse effects – or the impact in groups of people who weren’t represented on a large scale (or even at all) in the trial. For that, we need phase 4 trials and excellent post-marketing surveillance studies. On top of that, this is a new disease and we don’t know how long any protection will last.
What worries me most is whether the bar will be set high enough on effectiveness and adverse effects, at whatever the stage the assessment happens.
In other recent news…
- Moderna’s patent complications for its vaccine with the NIH continue, and the company’s recent filings reported that they “cannot be certain that we were the first to make the inventions claimed in our patents or pending patent applications”.
- CanSino debuted on China’s stock market.
- At least one group of people who should know better are McGyvering vaccination.
- Ed Yong wrote an instant classic on the immune system, which is a good backgrounder for reading about vaccines.
- A preprint of a protocol for a living systematic review with meta-analysis for Covid-19 vaccines was posted – and I commented on the adequacy, ironically, of its preprint search strategy.
Update August 14: Added data on recruitment status of the Oxford/AstraZeneca; added Sinopharm’s Wuhan-developed vaccine phase 1/2 trial to the table, and tweet about it embedded.
Update October 16: Corrected table for Sinopharm vaccines – originally had attributed preclinical results to Wuhan vaccine instead of Beijing.
Update on 10 December: Correction to table – BNT/Pfizer results formerly listed as phase 1/2 moved to phase 1. Although titled phase 1/2, no phase 2 results were included. (Grateful to Julia Belluz from Vox, who pointed out this.)
Disclosures: My only interest in Covid-19 trials is as a person worried about the virus, as one of my sons is immunocompromised. I have worked for an institute of the NIH in the past, but not the one working on the vaccine (NIAID). More about me.
Sources for the table are included among the records in my public Zotero collection of Covid-19 vaccines with preclinical or clinical trial results either as publications or preprints, and associated trial registry entries. They are tagged with names and type of record (e.g. Phase 1). Please let me know if I’m missing any! On August 14, it included:
- 25 vaccine groups with published or posted results;
- 30 preclinical reports;
- 2 phase 1 trial reports;
- 4 phase 1/2 reports;
- 2 phase 2 reports;
- 29 trial registry entries associated with these vaccines (some with multiple, some with none).