Coronapod: The divisive hydroxychloroquine study that's triggering mass confusion

Benjamin Thompson

Welcome to Coronapod.

Noah Baker

In this show, we’re going to bring you Nature’s take on the latest COVID-19 developments.

Benjamin Thompson

And we’ll be speaking to experts around the world about research during the pandemic.

Amy Maxmen

I really don’t know how this plays out. We also don’t know a ton about this virus, so there’s so many open questions. I just have a really hard time making predictions because I don’t know how the outbreak is going to change.

Noah Baker

Hello, and welcome to episode 11 of Coronapod. My name is Noah Baker, and I am joined by features editor Richard Van Noorden. Unfortunately, Coronapod regular Amy Maxmen is not able to be with us this week, but Richard and I are going to power on anyway and see if we can bring you some latest news. Richard, how has your week been?

Richard Van Noorden

Yeah, it’s been pretty good, thanks, Noah. I had the day off on Monday and went for a long walk so I’m rejuvenated.

Noah Baker

Rejuvenated, I think as far as one can be rejuvenated at the moment.

Richard Van Noorden

That’s right.

Noah Baker

So, this week, we’re going to talk about something that we have actually talked about before on Coronapod, but we’re going to dig into it a little bit, which is hydroxychloroquine. Now, this is a potential treatment for COVID-19 which has been bandied about by many people. There are lots and lots of trials and studies that have been going and this week, there’s been a whole new development or a whole new series of developments with hydroxychloroquine. So, first off, I think the biggest news has been that there was a paper released from The Lancet of a large observational study. Richard, tell me what’s going on with that.

Richard Van Noorden

Yeah, so this paper was a huge study, but it was observational, as you say, meaning that it’s not a randomised, controlled trial. They’re not dividing their patients randomly into matched groups and giving some hydroxychloroquine and some not without knowing who. That’s the gold standard. That’s some of the trials underway. This, they just looked at many thousands of patients who were in hospital around the world and saw what those patients got, what they were given and what happened to them. And they found that around 11,000 people who didn’t get any treatment, 9% died. Of course, this is people in hospital already. But of 2,000 people on chloroquine, 16% died, and of 3,000 people on hydroxychloroquine, 18% died. These chloroquine and hydroxychloroquine are slight variants of the same old malaria drug. Even more worryingly, perhaps, people who were given chloroquine together with an antibiotic, a macrolide – people thought there would be some kind of synergistic effect between the antibiotic and the chloroquine – in those cases, numbering, again, a few thousand people, 22% died. So, the numbers don’t look great. It looks like the people in hospital who got the chloroquine, or variants thereof, were more likely to die. Now, of course, it’s just an observational study. We don’t really know yet whether it was sicker people who were given the chloroquine, but the people who did this study did try to control for some factors as far as they could. That is that they recorded some things about the patients and they tried to statistically control for those things, like their obesity, whether they were smokers, their age, things like that. So, they tried as best as they could to match the patients up. So, this was quite worrying because it certainly didn’t show that there were any clear benefits and it potentially showed that there was actually harm from giving this drug to people, and that study published in The Lancet has had a huge number of quite confusing knock-on effects this week on all the other randomised clinical trials that are underway. So, as I’m talking to you, the World Health Organization has paused enrolment in its massive trial of the drug. Regulators in the UK and in France and in Australia, they have paused enrolment in most of their trials, although there’s one in the UK that’s still going on, and doctors in France have been told they can’t prescribe this drug outside of clinical trials. So, given that a lot of people were taking this drug, that’s quite a big change. Now, confusingly, in India, this drug is still being recommended to high-risk frontline workers, like healthcare workers and police personnel, as prophylaxis against COVID-19, even though there is no randomised, controlled data for hydroxychloroquine as prophylaxis yet. And India’s medical regulators say that they do have unpublished data that say, according to them, that there are some benefits. So, we’re really in this confusing state where a lot of trials have been stopped, some are still going on and some countries, like India, are still thinking that this drug can show benefits.

Noah Baker

And I think this is one of the things that’s really difficult about studies that are coming out like this really, really quickly at such a pressured time, and there were some things that were immediately picked on by critics saying, ‘Well, you’ve tried to control for this but surely it makes a lot of sense that the people that are the most sick are the ones that are going to get the most sort of treatment and the most treatment options, and so no matter how much you try to control for that, that might skew things.’ And other people in other trials have said, ‘Well, we haven’t seen these kind of safety effects.’ I think the WHO haven’t seen these kinds of problems in their safety trials to date and yet it is still such a significant figure that it’s causing kind of quite a significant impact to other trials around the world.

Richard Van Noorden

Right, I mean it’s possible that these trials will resume quite quickly. As you say, the WHO has said it hasn’t seen signals of problems yet in its trial, and in the UK, the UK’s biggest trial of 10,000 people called the RECOVERY trial, as soon as this Lancet paper came out, they consulted members of their safety board, and the board said, ‘Well, there’s nothing clear in here, according to the board, that says that participants are definitely being harmed by receiving hydroxychloroquine,’ and as a result, this big trial has been allowed to continue enrolling participants. So, it’s not clear what’s going to happen. And the author of the Lancet study, Mandeep Mehra, a cardiologist, he doesn’t support ending the clinical trials. He’s saying, ‘We want clinical trials,’ but he thinks the observational work is helping to bring balance to the public discourse about this drug. He points out that even before this Lancet study came out, there was some data suggesting there wasn’t that much clear benefit from hydroxychloroquine. We need the clinical trials to be sure. But people are really taking it almost with evangelism anyway – Donald Trump perhaps being the most famous person who says that he takes it. So, he views this trial as a useful corrective, but you could argue that it’s just caused even more confusion among the public, and we really don’t know at all what to think about hydroxychloroquine right now and, at the moment, many of the trials that we wanted to happen, to find out, are paused.

Noah Baker

It’s a really interesting one. When we’ve talked about chloroquine and hydroxychloroquine before on the show, we’ve talked about Donald Trump touting it as a wonder drug, as this sort of solution that’s going to come along and fix the crisis, and how that might impact things. And we’ve said, well, the thing is, we need more trials. But often what we’ve been talking about is people like Donald Trump making statements that are not backed up by the evidence yet, and we need to get the evidence here, and now we’ve actually got a kind of a disagreement almost within the people that are generating the evidence. Now, the trials are all a bit confused as well. It’s like the story has muddied yet further around this drug.

Richard Van Noorden

This is medical science. Scientists live in a perpetual state of uncertainty and hoping to narrow that uncertainty to say something definitive and, for the moment, we don’t really have this for this drug. The whole reason that it is attractive as a drug and why, for example, India is still wanting to use it is that it is used to treat people with malaria. It’s also approved to treat people with lupus and rheumatoid arthritis. In fact, when there was so much enthusiasm for the drug last month, people with those conditions were saying, ‘Well, hang on a minute, I can’t get the drug that I need for my condition.’ Now, there’s even been questions about some fundamental things about the Lancet study itself – the numbers in the Lancet study – quite apart from it being an observational study. Some researchers have said they’ve seen some oddities. For example, all the parts of the world that were studied in this analysis, the smoking rates and the obesity rates of all of these people have been reported as nearly equal. That seems very unlikely. What’s going on there? And Mehra says, ‘Oh, we erroneously listed our corrected modelled data for those kinds of things rather than the raw data. We’d already made the comparison and adjusted and we accidentally listed that, not the raw data.’ So, he’s saying things like that, but unfortunately, researchers don’t even have the raw data because the whole work relies on proprietary data that was gathered from medical centres and hospitals by a company called Surgisphere Corporation in Chicago. So, it’s like a perfect storm of raw data not being available, observational study. Important – lots of people in this study but still, really hard to tell what it tells us, really, about hydroxychloroquine. I mean the knock-on effect of this may be that it really dampens down the enthusiasm for people not only wanting to take this drug but even wanting to enrol in trials to get this drug, and this is kind of ironic because a few months ago, we were running a story saying that when people enrolled in clinical trials, they all wanted this drug and they said, ‘Give me this drug. I want hydroxychloroquine, please.’ And they were even worried that they didn’t want to get placebo – of course, they wouldn’t know – and this was causing problems testing other potential treatments in clinical trials. Now, you might think, ‘Well, I don’t want hydroxychloroquine, please. I want the other drug.’ So, that might make it harder to finish the trials that have been starting. Of course, the bigger problem on that line is that in many clinical trial hotspots like the United States and Europe, the pandemic, or the first wave of the pandemic right now, is waning slightly, and that means that the pool of potential clinical trial participants, people coming into hospital, is drying up a bit, so that also is concerning some researchers because it means that we might not have the evidence we need to know which medicines are best when the second wave, if it is to come, follows.

Noah Baker

I would never have predicted that we would be in a position now where we’re discussing whether or not clinical trials were going to struggle to get participants when we were on this podcast four or five weeks ago saying clinical trials are struggling to get going because people will only accept chloroquine, and now it’s the polar opposite. Things can happen so quickly. I think it’s also really interesting in a situation like this where scientific trials and preprints have become so tied up with politics as well that political opinions can have such a big impact on science in a way that perhaps they wouldn’t otherwise do. The medical research board in India has continued to approve this hydroxychloroquine for prophylaxis, but there have also been accusations thrown at this decision from scientists who are pointing out that the majority of the world’s chloroquine is produced in India. Is that a conflict of interests there? Is there a reason that there may be a political decision to keep promoting the use of chloroquine because that’s something that would benefit that country? These are questions that don’t usually plague clinical trials of drugs at this stage.

Richard Van Noorden

Yeah, completely. It’s full of politics, which is a real shame because we really need to know whether this drug is going to show benefit. I mean there’s not that many other drugs in the World Health Organization’s trial. They are testing remdesivir which, we should say, is the one drug that has seemed to show promise in a rigorous clinical trial, and remembering again that it doesn’t appear to lessen one’s chance of dying if one is in hospital and one is given remdesivir, but it does seem to speed up the time to recovery. But even that drug has to be administered intravenously over the course of several days, in hospital, so it’s not the same as a relatively cheap and widely available drug like hydroxychloroquine.

Noah Baker

How much of the confusion that we’re currently seeing can be put down to early data that perhaps usually wouldn’t be released on this scale, or certainly reported on this scale, causing confusion? Perhaps, normally, in a trial like this, there would be more time to go through, to do analyses that wouldn’t be jumped on and people wouldn’t react so almost impulsively to pieces of data they find. How much of this is sort of a curse of rushed science?

Richard Van Noorden

Well, it’s true that a huge proportion of the studies that have come out generally about COVID-19 are preprints, meaning they’ve been released before peer review. In fact, it looks like now it’s about a quarter of all the research articles that are coming out are coming out as preprints. This Lancet study did not come out as a preprint first. It was peer reviewed first, so any massive problems ought to have been ironed out, but on the other hand, the peer review itself has been sped up. One study looked at 14 medical journals and found that they published papers on coronavirus, peer-reviewed papers that is, nearly twice as fast as the other papers, largely because they sped up the peer review. Everyone wants to get out, both the kind of dispatches from the frontline as fast as they can and they want to get out the peer-reviewed papers as fast as they can, and I think it’s putting journals under a lot of strain and putting peer reviewers under a lot of strain, and it’s just a really tough balancing act between getting the data out there and having it sufficiently peer reviewed that a message can be sent out to the public and everyone around the world that is not going to be misunderstood or misinterpreted.

Noah Baker

I can’t help but wonder whether or not this paper from The Lancet was rushed through peer review. I’m actually not sure when it was submitted because that information isn’t on the paper anywhere, but there are things in this which are potentially things you’d expect peer reviewers to pick up. Using modelled data instead of raw data is precisely the kind of thing you’d hope a peer reviewer would pick up, unless, perhaps, they were being rushed. But maybe I’m speculating too much here.

Richard Van Noorden

Right, I mean that’s one of the things about peer review is that it very much does not pick up on all these kinds of mistakes. It is simply a set of checks by qualified people who don’t have much time to make those checks and there’s all kinds of suggestions about how peer review can be improved, but we have noticed that peer review has really sped up and we have a sort of series coming out next week on what science might look like after this pandemic, and one of the pieces is on what publishing might look like, and people thought it was impressive, in a way, how fast peer reviewers are working, but there were mistakes and it didn’t seem likely that, really, papers could be peer reviewed this fast all the time. It seemed like a kind of emergency situation that would not hold true in the normal course of science. I don’t know whether in this case The Lancet could have pushed back a bit and said, ‘Well, hang on a minute. You’ve got to make the raw data available here to other researchers to check, even if you can’t publish that raw data because it’s proprietary.’ Maybe there could have been a little bit more asking of these questions like, ‘Was it the sicker people who were given hydroxychloroquine,’ given the enormous impact that this study is going to have. But these are kind of easy questions to see in hindsight and I don’t know if they were obvious at the time the paper was being peer reviewed.

Noah Baker

I think you and I and people at Nature and the researchers that we speak to everyday are very used to robust scientific debate, but then you get in a situation where you are now where every piece of science that’s coming out has the potential to be spread across the front page of a newspaper and it’s almost like the way things are interpreted is very, very different. So, I think science is often seen by members of the general public to be this relatively infallible institution where people say things that are correct and if a scientist said it, it’s right, and if a scientist disagrees then that throws everything into doubt because there isn’t necessarily an understanding that there is a robust scientific debate. That’s part of the way science is done and suddenly it’s splashed out across a newspaper, and I worry that it will impact how much people will even trust the papers that are coming out because what they see as ‘scientists can’t agree’ is actually just normal scientific discourse but being presented differently.

Richard Van Noorden

Yeah, I think that it’s just something that everyone should bear in mind. I would almost recommend to scientists when they’re talking to journalists to say, ‘Yes, we can’t agree but this is what science is. Please write that in the story.’ Because we’re really seeing this just playing out really publicly and often it just isn’t clear. A lot of studies are going up as preprints that are quite well done and scientists debate them and pick holes in them and then it suddenly feels very different when it’s on the front pages of all the newspapers and everyone is shouting at each other. That’s actually happened this week with another study which is from the group in Germany from Christian Drosten, and they were talking about this question of whether children are more susceptible or less susceptible to infection than adults. And Christian Drosten’s group put up a preprint and they said they’d measured the viral load in children and adults and they found it to be, when doing some statistical tests, about the same in children and adults. That didn’t answer all of our questions but it was an interesting observation. And then a statistician, David Speigelhalter, here in the UK, said that the statistical test was the wrong kind of test and that, in fact, he thought the right statistical test would show that children’s viral load was a quarter of that of adults. And this is sort of classic scientist v. scientist. You’d call it good-natured debate. But a German newspaper got hold of this and as blown it up into a sort of attack on Drosten and a big sensational story, and I think Speigelhalter has actually sort of distanced himself from this newspaper story. I’m not sure that the arguments around this new virus that’s hit us are any more vociferous or disagreeable than you would expect at this stage, but we’re sort of seeing this typical progress in science being sort of blown up and put on to the front page and really mattering.

Noah Baker

Okay, well, there’s going to be a lot to keep up with on the various twists and turns in the chloroquine and hydroxychloroquine saga. I’m sure that it will be mentioned again on this podcast. But before the end of this show, let’s turn to our usual ending segment, which is one good thing. I’m going to start this week. So, as regular listeners to Coronapod will know, I’m currently hiding out in the countryside, and being in the countryside has advantages and disadvantages. My internet connection is very slow, but one of the advantages is that I’m surrounded by elder trees, which are very, very common here in the UK, and all of the elder trees are now in flower and I am in a position where I have lots of space and it’s easy to sterilise things, and all these things have come together to make elderflower champagne. So, I’ve been doing some brewing from home. We’ve talked about quarantinis or coronatinis on the show before, and now I am brewing somewhere in the region of 150 litres of elderflower champagne, currently. And I say brewing but really, it’s just cutting a load of elderflowers off some trees and putting them in a bucket with some water. That’s about it. It’s not taking me very much time. But I’m looking forward to having elderflower champagne for the rest of the year. It’s a goodie, a hedgerow brew.

Richard Van Noorden

That is amazing. I really hope I get to taste some.

Noah Baker

I will bring some bottles in. I’m going to have a lot left over, I’m sure.

Richard Van Noorden

I wanted to recommend… it’s not going to be anything as amazing as making my own elderflower champagne. It’s more entertainment on YouTube. So, this is Denis Shiryaev’s YouTube channel, and about at the beginning of this pandemic, he started using open source AI tools to put high resolution and sometimes colour back into century-old film reels, and we’re actually going way back to essentially some of the earliest film reels ever done, way back in the 19^th century and also in the early 20^th century. And it’s incredible what these algorithms that insert this colour, that slow down the video, that sharpen things up slightly artificially. It ends up looking a bit like a modern film but you’re walking through San Francisco before the 1906 earthquake, or you’re looking at some English garden in the 1880s. It’s just brilliant. There’s also a soundtrack on these films as well, which is obviously entirely made up because these are silent film reels, but it is a wonderful YouTube channel and it’s really sort of entertaining to take yourself back a century or more, while some of us are cooped up at home.

Noah Baker

I have to say, when you first sent me a link to this, I was enthralled. I think as a filmmaker, trying to work out how you can take grainy footage, which I deal with quite a lot of because researchers send me things. Who still uses Windows Movie Maker aside from scientific researchers? If I could use some of these algorithms to make them into big, pretty 4K video, I would love it.

Richard Van Noorden

It is pretty amazing.

Noah Baker

There’s a picture that went around a little while ago that every so often I see pop up again on my social media feeds, and it’s not a video. It’s nothing like as clever as this. It’s a photograph of Charlie Chaplin from 1916 that’s been colourised and cleaned up so it looks like a modern photograph, and every time I see it, I think, ‘My God, you are just a modern-day 26-year-old that lives in Bushwick or Hackney in East London.’ The clothing is the same. The style is the same. It’s really bizarre to sort of see someone from 1916 looking so much like someone I’d expect to see in a trendy bar.

Richard Van Noorden

Wow, I’m just looking at it and, yes, wow. Messy hair and all. That’s incredible.

Noah Baker

Laughs. Messy hair and all.

Richard Van Noorden

That’s what I think of those people.

Noah Baker

The hispters. Okay, well, with that, let’s draw this week to a close. Richard, thank you so much for chatting to me this week on Coronapod, and I will speak to you next week for episode 12, which will be three months into this pandemic. Richard, thank you very much.

Richard Van Noorden

Thanks, Noah.

Noah Baker

As ever in the show, we also have a second half and this week, as I did last week, I’m going to bring you a couple of summaries of the latest COVID-19 research. Remember, you can find all of these summaries over at nature.com. I’ll put a link in the show notes. There’s a blog that’s being updated every day with some of the key papers. First up, we’ve got a preprint uploaded to MedRxiv on 22 May. Now, this paper suggests that a small number of super spreaders are responsible for seeding the virus across Israel. A team from Tel Aviv University sequenced and analysed more than 200 SARS-CoV-2 genomes from people across Israel. Their analysis showed that 80% of infections were transmitted by only 1-10% of people – so-called super spreaders. The analysis also found that travellers from the United States and Europe carried the virus to Israel, but US travellers were responsible for a disproportionate share of the viral spread. One possible explanation for that is that Israel began restricting entry to people from Europe before it banned US arrivals, which gave US visitors more time to spread the virus. The study has not yet been peer reviewed.

Noah Baker

And next up, another study about transmission dynamics, but this time set in New York City. The preprint from the Harvard T. H. Chan School of Public Health drew a link between infection and commuting. In New York City, deaths and hospitalisations from COVID-19 have varied wildly between the city’s neighbourhoods and researchers wanted to find out why. They compiled coronavirus test results from about 1,700 women that came to six city hospitals to give birth. The team analysed the postal codes of the infected women to estimate the disease prevalence in neighbourhoods. Then they compared this information with location data from Facebook about the number of trips every day in and out of each neighbourhood, and there lay the correlation between a neighbourhood’s infection rate and the number of trips taken by its residents. The authors say that many of these commuters are probably essential workers and that they should be protected further to prevent the virus’ spread.

Noah Baker

And finally, we’ve got an update on DNA vaccines. This paper was published in Science on 20 May, and it showed that DNA vaccines protected monkeys from coronavirus. Researchers at Harvard Medical School in Boston explored vaccines composed of DNA. As you heard last week, this type of vaccine prompts the recipient’s cells to make the pathogen or its components and in turn stimulate the immune system, hopefully developing immunity. The researchers developed six DNA vaccines based on the coronavirus spike protein, and tested them in Rhesus macaques. The animals mounted an antibody response similar to that seen in macaques and people who’ve recovered from SARS-CoV-2 infection. The team then gave doses of coronavirus to the vaccinated monkeys, which developed only mild illness. Viral multiplication in the animals was generally lower than in unvaccinated monkeys, probably because the vaccinated animal’s immune systems was keeping the virus in check.

Noah Baker

And that’s it for another episode of Coronapod. You can join us next week for episode 12. Meanwhile, we’re publishing the regular Nature Podcast every Wednesday still, and that’s a coronavirus-free zone if you need a bit of a break. This week, we’ve got stories about a new way to make hydrogen fuel using only light and an answer to a long-standing cosmic mystery. You can find that wherever you found this. If you have any questions or thoughts about Coronapod, or the Nature Podcast for that matter, we’d really like to hear from you. You can get in contact on Twitter – @NaturePodcast – or if you prefer email – podcast@nature.com. In the meantime, I’ll put links to everything we’ve discussed on today’s episode of Coronapod in the show notes so you can read more if you’d like. Until next week, I’m Noah Baker. Stay safe.