Meet the Doc Who Put Hydroxychloroquine (and More) on Trial in COVID-19

An Interview With RECOVERY's PI, Martin Landray

; Martin J. Landray, MB ChB, PhD


October 11, 2021

Robert A. Harrington, MD: This is Bob Harrington from Stanford University on Medscape Cardiology and

During the course of COVID, we've tried to talk about some of the science that's going on and some of the other things people might be doing, like writing to relieve some of the stress of COVID. Perhaps one of the most important topics of COVID is what we're going to discuss today: In the midst of a global pandemic, how do you do the high-quality clinical research that's required while under great stress in the healthcare system and under great individual stress of clinicians, investigators, and patients to try to answer the question that we need answered, which is, what do we do?

I can't imagine a better guest to have with me today: Martin Landray from Oxford. Professor Landray is a professor of medicine and epidemiology at Oxford University. I'm pleased to say that he's one of my friends who I can refer to as Sir Landray and have it be a proper title.

Martin, thank you for joining us here today on and thanks for lending your expertise. And maybe most of all, a genuine thank you from the clinical community for the contributions, which have been extraordinary, coming out of your work over the past 18-plus months.

Martin J. Landray, MB ChB, PhD: Thank you so much for having me. I'm just sorry I can't actually be there in person. It seems a long time since I've been to the States, and believe me, I'm missing it.

Harrington: Well, we're all missing you. I remember early in the pandemic, we had you do medical grand rounds here at Stanford when you had the first trial results with dexamethasone. People still talk about that as one of the best grand rounds we've had throughout the pandemic because it made people realize that you can do high-quality clinical investigation in the midst of all the turmoil and stress.

Those are the kind of lessons I want to get to today, but let's start at the beginning. You said you've not been to the States. I've not been to visit my friends in Europe. One of my great disappointments in the past few years is missing one of my favorite meetings of the year, which is the European Society of Cardiology meetings. It's always a great meeting held in great cities.

How are things going in the UK?

Landray: We're in an odd state right now. Vaccination rates in the adult — and particularly in the elderly — community are really very high. I'm not sure what the official statistics are, but probably well over 80% in the at-risk groups are double-vaccinated. That has been very encouraging.

Of course, we recognize that those vaccines are imperfect. A vaccine that has 90% success has 10% failure, so 1 in 10 people will get COVID-19. We are seeing that the impact of COVID-19 is much less severe, so the number of people going into hospital and the number of people dying from COVID is dramatically less than it was previously.

It's what we call autumn here — I guess you call it the fall — and the timing of this year's waxing and waning of the epidemic is such that we came out of lockdown in May, perhaps, and now here we are heading back into school and college. Today is, in fact, the first day that I've been back at the Wellcome Trust building in central London in 18 months.

The last time I caught the number 18 bus was the time that I stood on that bus with Jeremy Farrar, saying, "We better get a trial up and running in a fortnight." That became the RECOVERY trial.

We're heading back indoors. The questions are threefold. One, what will happen as we head indoors with regards to COVID-19, and of course, the worries about future variants, potentially. Two is a huge backlog on the NHS, or the health service, of all the other things that the NHS should be doing but frankly hasn't had the time or capacity to do over the past 18 months. The third is what's going to happen to seasonal influenza this year. I think everybody's bracing themselves for what could be a nasty influenza outbreak on top of those other two factors. We're not out of the woods. We should be mindful of that, but large-scale vaccination has been really important here.

I'd just like to take one reflection, given that we're comparing and contrasting different parts of the world. I think we really do have to recognize that, while the US, the UK, and one or two other places have access to huge quantities of vaccine, many, many places do not.

In the UK and the US, it's a luxury to turn down the vaccine. It's, in my view, not sensible but it's a luxury to even have that option. Many parts of the world simply don't have that option. The real worry is the damage that does to the health and economy in those places, but also, that's where the variants are going to come from. That's where the thing is going to come up from and bite us.

Giving vaccines overseas to the international community is actually a selfish act as well as a generous act. I think we really need to look at that. In the past few days, there have been big announcements from your president, our prime minister, and many others about this. Big announcements are great, but we have to see action. That really is important, as I say, for selfish reasons as well as for humanitarian reasons.

Harrington: You've brought up many directions we could go. Certainly, the notion of the mass-vaccination rates in the UK contrast with the US, where overall we have a high vaccination rate, but it's very heterogeneous. Where I live in Northern California, the two counties that surround Stanford University that make up the primary population that we care for have an exceedingly high vaccination rate.

There are parts of the American South, where I used to live, which are really quite low in terms of their adult vaccination rate. In the US, it's become, as you've undoubtedly read, a political issue as opposed to a medical issue. That's been one of the sad lessons of the pandemic.

Regarding vaccination, we were pleased this week. We know, as part of a global community — and people like you and I spend a lot of time around the globe — none of us are safe if the entire world is not safe. I like how you said it's not just a humanitarian effort but a self-preservation effort, that you have to be able to vaccinate the globe if we're all going to be safe.

We started our mass vaccinations for flu shots this week at our institution, recognizing that it may be potentially a tough winter if we can't get ahead of flu vaccinations.

Let's talk about RECOVERY. I like how you gave the introduction of riding on the bus and saying, "We've got to get a trial going." Was that really how it started? You were just brainstorming about what can we, as clinical investigators, do?

A Memorable London Bus Ride

Landray: I have a part-time role here at the Wellcome Trust. For the past couple of years, I've tried to help sort out what good regulatory guidance should be around clinical trials, so-called good clinical practice, which everybody knows is not good, not clinical, not practical.

That's been one of the themes of my work over the past couple of decades. That's why I was coming to London at the time. It was the 28th of February last year. The outbreak had been and gone, largely, in China. It had been gone mostly in Iran. It was, if you like, other people's problems, to quote a rather unfortunate but probably common sentiment.

Then it hit northern Italy, and I wrote to Jeremy Farrar, who I know well, and said that people are talking about social distancing, travel restrictions, and all these other things. At some point, people are gonna start talking about interventions to treat or to prevent the worst consequences to this disease.

I asked, "Is anybody thinking about randomization? If we're not careful, drugs will get thrown at people in hope and expectation, and at the end of it, we'll be none the wiser about which ones benefited or which ones did not. In fact, now is precisely the time to be setting up randomized trials; it's exactly the time when there's huge public health need, there are bad outcomes, there is no knowledge of meaningful treatments, that actually people are prepared to innovate and be flexible, and that randomized trials could actually deliver solutions."

Pretty much all of that was in that first email right at the end of February last year. He wrote back immediately and said we better start talking to each other. We did the following week. He's the co-lead on the RECOVERY trial.

It was that Monday morning, March 9, 2020. Jeremy Farrar and I were on the bus between the train station and the Wellcome Trust building, comparing notes about that northern Italy outbreak. Jeremy turned to me and said, "You know this is going to come to the UK, in a tsunami in a fortnight. We've got to get the trial up and running in that time." I came into the offices where I am now, started making the phone calls, and 9 days later the first patient was recruited into the trial.

Harrington: That's nothing short of extraordinary in terms of the rapidity with which it happened. I'm going to get into some of the details of what are the key elements of the UK system that allowed that to happen. As you know, our research system in the US is so decentralized. In some ways, that's great. I think it's particularly great for discovery science, but it's not so great when one is faced with a pandemic and you need large-scale trials set up quickly.

Let me ask you a more specific question: How did you pick dexamethasone as the first thing that you wanted to study? What was the aha moment that led to that?

Landray: All of these things are punts or bets. Every drug we studied in this trial — and frankly, in every trial I've ever done — you hope it works. You've got good reasons for thinking it might work. That's why you put it in the trial in the first place. You don't know how it works, and that's why you actually do the trial.

The World Health Organization had done a big filtering process, the government committee in the UK had done a big filtering process, and they got down to a list of 80-odd drugs. It wasn't that much of a filtering process, frankly, but some were bubbling along at the top and kept coming back. Lopinavir was one, remdesivir was another, and dexamethasone had been proposed as a possible treatment for a flu pandemic or flu outbreak.

The evidence see-sawed for dexamethasone. You're suppressing the immune system, you're suppressing inflammation. There are pros and cons from a variety of data coming out over the years. There wasn't a robust answer.

We picked up this flu protocol, and it was unworkable. It was the right question in flu. It turns out it's the right question in COVID-19, but the protocol is impractical. Anybody could design a protocol that won't work. The question is, how do you design a protocol that really will work?

In some ways, the luxury — that's maybe an unfortunate word, but just go with me for a minute — the academic or intellectual luxury in that particular time was that one could actually take the dramatic step change in thought process and say, "What do we really need to know? What do we really need to do? Can we imagine what this is going to be like on the coal face?"

The hospitals are overloaded, the doctors and nurses have no time at all. They're emotionally overloaded and exhausted. The patients are often elderly, they're scared, many of them just can't even breathe for themselves, and they're alone because of social distancing and infection control.

The phrase "war zone" is overused, but it was a very helpful thought process as to what are we going to do if we were to get a tsunami of cases like happened in northern Italy, where frankly they were queuing up to the mortuary, let alone queuing up to the bed.

Harrington: Early on, the stories.

COVID Clinical Trial a Hard or Easy Sell?

Landray: Very difficult decisions about who to put on a ventilator and who to take off a ventilator. You can't think in normal ways. As I say, it's sort of liberating in the sense that you have to think differently if you're going to make this work. Frankly, we should be thinking differently all the time, but it did give us that moment.

Harrington: That's what I want to get at, Martin. You said something that struck me at the beginning. You said, "Well, this wasn't a couple of years' worth of work. This is decades' worth of work." You and I have both worked on large-scale trials for decades. They're hard to do, as you know. They're hard to get going and they're hard to keep going. The 9 days I find nothing short of extraordinary in terms of an accomplishment.

There can sometimes be this tension between clinicians on the front lines who are trying everything they can to treat a patient to keep them alive, and clinician investigators who may be doing some of the same but are also spending some of their time thinking about, as you said, what's really the right way to test what works.

I suspect you had to make many phone calls to people to say, "Hey, we want you to do this and we want you to enroll your patients." Was it a difficult sell, or is the UK ethos different from the American ethos?

Landray: Basically, the story is that we want to make it as easy to randomize people in the trial as it is to prescribe these drugs arbitrarily. Randomize and learn vs distribute the drugs to roughly half the people and learn nothing and continue to make mistakes. It's all very well saying that stuff, but it has to be simple and it has to be doable in these circumstances.

In these circumstances, that meant it had to be almost frictionless. We couldn't remove absolutely everything, but anything that we could we removed. When I was a very junior doctor, and actually, before that when I was a medical student, the trial that made me want to do trials was the ISIS-2 trial of streptokinase, aspirin, both, or neither.

The results are well known and the impact is well known, and there have been other trials in the US and elsewhere, and it led to, frankly, the whole field of big cardiovascular trials. It led to primary PCI. There are whole areas of medicine and therapeutics that came off the back of that.

It was famously a single-page case report form. It was famously a protocol of, I think, 8 pages. The clinic manual was one piece of A5 paper, so half of a piece of paper folded in two, deliberately designed so it would fit into the breast pocket of the doctor's white coat and the name ISIS-2 would stick out the top. Basically, it said, "Randomize me" out of the top. It really did. That was quality by design.

I remember for that trial, I looked at the results and they were beautiful, but I also looked at that trial and thought, Somewhere, there are about a handful of people sitting in a temporary building in the car park of the John Radcliffe Hospital in Oxford, who one day came into work and said, "Not only is this question important, but we're actually going to recruit 18,000 people, and this is how we're going to do it." That's what actually got me really into trials all the way back then.

I had that protocol on my desk. I had the paper from Salim Yusuf, Richard Peto, and Rory Collins on why we need some large, simple trials on my desk, and I went from there. I said, "Actually, what is it we need this time around? From a blank piece of paper, what do we need to write down? Only write down the things that we absolutely must."

Harrington: I get all that. I, too, like you, around the same age was very excited by those trials coming out of Italy (GISSI), out of the UK (ISIS). I was a fellow at Duke during the GUSTO days. Even GUSTO, which was large and simple, was only a 3-page case report form. Hundreds of manuscripts were then written from those 3 pages. I get all that.

In the US, we are so decentralized with regard to our ability to do things. People would worry about who's going to get credit for this, is my university going to get credit for this, where's the grant going to sit? The politics of clinical research in the US slows things down tremendously.

Why doesn't that happen in the UK? Is it the National Health System and the feeling of responsibility? What's the difference?

Landray: Well, believe me, there's health politics and academic envy and so on in the UK, just as there is in any other part of the planet. That's no different. There are wise people who wrote to the MHRA and said, "You shouldn't be allowing this trial of dexamethasone because it's dangerous to suppress the immune system. You'll kill people." I've got those letters.

Harrington: That's good. If everybody thought it was a good idea, it's probably not a good idea.

Landray: Absolutely. Number one, we did convince the chief medical officer of two things. One was that doing the trial should be seen as part of routine clinical care; not only did we convince the chief medical officer, who has an R&D background, but we also convinced the chief medical officer for the NHS that doing the trial was to be considered part of clinical care.

Secondly, we said that for some of these treatments — hydroxychloroquine is the best example — they are not to be used outside the context of the clinical trial. We're not going to reimburse or allow the use of hydroxychloroquine. We had huge stockpiles of it here in the UK. There must have been tons of this stuff sitting in a warehouse somewhere.

Rather than just say, "Oh, it's there, go use it," we said, "Actually, it's there, and if you want to access it, it's part of the randomization and we'll learn." I think it was 84 days later when we knew the answer. In this context, it's useless. That's an important part of the way of doing it.

We did have to motivate the frontline clinicians, patients, doctors, nurses, pharmacists, and so on. Part of that was building a community spirit around it. We took a very deliberate approach of transparency. Recruitment numbers are all up on the website. All regulatory submissions, IRB submissions, training materials, and patient information are up on the website. Anybody can look at it — you can look at it now. If you're fed up listening to me, you can go look that stuff up.

That created transparency. We took the media with us and said to them, "Look, actually, there is this issue." I recognize it as a doctor myself, previously looking after sick patients in my younger years. The "Help me, Doctor, please help me" from a breathless patient is incredibly powerful.

Helping is not the same as prescribing. Prescribing a drug where you have, frankly, no idea whether it's going to do good or harm or nothing is not doing your patient in front of you any favors or the next one, or the one after that, or the ones in the next ward, or the next hospital, the next county, or the next country.

You have to randomize and learn. Do it fast, do it at scale, and learn the lessons. One of the things you and I have talked about before, perhaps a year ago last fall, was that the US has huge numbers of cases. That's bad. We all wish that wasn't the case. If the ability with everything lined up to answer these sorts of questions in a matter of days or weeks is absolutely there — take convalescent plasma as a great example.

Instead of just giving it to 100,000 people and collecting some simple data — that was a beautiful, large-scale, simple, nonrandomized study. For the sake of a coin toss, we would have had that answer about last July or last August. We'd have known that answer. We just needed the answer. I nearly put the coin in the post and sent it over so that somebody could toss it.

Harrington: I think we both agree that there's a role for observation, but observation shouldn't be the role for understanding therapeutics. Observation should be the role for creating hypotheses that are worth testing. Some convalescent plasma observations are worthwhile, but I agree with you. I think it was an egregious use of resources to have basically given this to tens of thousands of individuals without randomization.

I'm of the mindset, Martin, that clinicians should really only be making two broad categories of decisions when the data are known and clear. The system should help us enforce doing it. Aspirin to the acute MI patient might be an example. When the answer is not known — and that's the great amount of things we do with therapeutics — we should be studying it in a rational way, which is randomization, and not in a way where people, as you've said, willy-nilly pick what they want to do. Ultimately, it gets randomly distributed, but that's not the same as randomization.

Landray: In both cases, you're trying it to see what happens. The difference between randomization and nonrandomization is that in randomization you actually learn what happens. With nonrandomization, you just continue to kid yourself. To be clear, if a treatment has a massive effect, such as benzylpenicillin for bacterial meningitis, you don't need randomization. It's such a great big effect that you're only going to treat three kids like that and you just know it works. Also, insulin for type 1 diabetes and so on.

There are, frankly, precious few treatments that are actually like that. When someone comes and shows you one that they think is like that, be very skeptical. The alarm bells should ring every time you see a big effect size. Ask if this is really real. There's a strong touch of used car salesman in that stuff.

Otherwise, you have to randomize. A really big question going forward is, why do we make it so darn hard to do a randomized trial by comparison with either undertreating or overtreating? You can consider it as harming at least some of the patients, either the ones you treat or the ones you don't — patients without any controls at all. That's a mad system. The word "trial" does not mean "risk." Uninformed clinical care is darn risky; randomize-and-learn is a very smart way of making improvements in therapeutics.

Harrington: This is a whole other topic that you and I could come back to, and your remarks touch upon it. Within the bioethics community, there's a big debate as to rethinking the whole notion of informed consent. Allowing doctors to just pick what they think might be best without randomization, with the intent to learn, may actually be unethical. That's a tough argument to have. There are groups in the United States having that argument now.

One of my favorite trials comes out of the UK, which was HEAT-PPCI, which I thought was a fantastic trial of two different types of anticoagulants during primary PCI. They determined that they had equipoise in their practice and they just randomized everybody and gave people information after. I thought this was an efficient way to do research on well-known therapies that people were using, as you say, randomly, but not through randomization.

Landray: We'd like to be talking about cardiovascular medicine, but I think there's much that cardiovascular medicine can learn from COVID-19. A neat example is thinking about all the controversy around the appropriate interval between the first and second dose of the vaccine.

Back at Christmas, the UK government said, "Look, we're going to just do first doses and push the second doses out to 12 weeks." We said, "Well, you could do a trial in which you randomize people as to whether they get their first dose at 12 weeks or 3 weeks." Some people said, "That's great but you need consent." We said, "Hang on a minute; we need consent to bring people back at 3 weeks, which is the licensed dose, but we don't need to consent in order to push it out to 12 weeks on the basis of government decision? That's bonkers."

For that trial, because we put the protocol together, we could have randomized 1 million people for 1 million pounds for 1 year of follow-up to actually understand whether there were pros and cons, and there were good reasons for thinking both, both in the short term and long term of 12 weeks vs 3 weeks. The same is exactly true of who should get a third dose, a booster dose, now or later.

These are examples of very simple questions where there's basically no additional risk above what we know already, but the value of that information would be colossal. That's not your topic.

Harrington: No, I think it's very much on topic.

Landray: Randomization, I think, can make massive differences to really understanding what one's doing. Because of scale, the only differences between COVID-19 and cardiovascular disease — apart from biology; I'll leave that aside — essentially are timing and profile. In COVID-19, all timelines were squashed. We set the trial up in 9 days, but you would see thousands of cases per day.

By comparison, with thousands of cases per week or per month in cardiovascular disease, you would see evolution in science, therapeutics, testing, and all these other things. A period of days and weeks normally would actually be years and decades. Time was one feature, and the second was profile.

COVID-19 was on every page of every newspaper on every website. There was a little ticker tape on the number of cases and number of deaths and all of those other things. It was in every conversation. It was part of everybody's mindset. The profile was high. Cardiovascular disease continues as the sort of hidden juggernaut, and it's not so hidden. It's there, but it's not the thing that you really notice

Harrington: The fact in the US is that mortality is increasing from cardiovascular disease, or at least the decline is decreased. The progress we've made over the decades has slowed for a variety of reasons. My final question for you, Martin, is whether the construct of RECOVERY is applicable to the chronic diseases of hypertension, diabetes, obesity, or lack of physical activity. You've picked a topic that you and I deal with daily clinically. Is the RECOVERY platform appropriate for that?

Landray: Yes, broadly. I don't think every trial needs to look exactly the same. In fact, it definitely shouldn't look exactly the same. Not every trial needs to be a platform, but randomization is part of clinical care. If you don't know what you're doing, randomize so that you learn what you should be doing.

Access to data so that you can find out what happens to the people that you're studying. Involvement of the patients and the clinical community, so that you're doing it in partnership — you're doing with people, not to people. National policymaking, which actually is aligned with that overall method of improve and prove.

One final thing is, what's the biggest barrier to actually doing a lot of these things? In my view, a huge barrier is the burden of not only regulation, but the way that regulation is interpreted and applied and inspected, not just by regulators. It's easy to point the finger at the FDA or the MHRA, but actually by all of our institutions. Think of the additional burdens that your IRB places.

Zeke Emanuel did that lovely paper that said that the average word count for the vaccine trial consent forms was 8000 words. It would take you 40 minutes to read it, and you have to be an undergraduate English student in order to actually be able to comprehend it. That's just nonsense.

The burden of regulation is a real, significant barrier. Anything that makes it harder to randomize and learn, we've got to look at that really carefully. I think that's the major thing we have to focus on now.

Harrington: I like the concept, which in some ways is simple and in some ways it's very complex. A central tenet of medical practice needs to be taught from the very beginning, which is, "When you don't know, randomize so you can learn." I think you're spot-on. Somehow, there needs to be a cultural change that that has become the accepted practice.

The other thing that we're struggling with in the US, which you've gotten at, is this notion of alignment of the science with the policymakers. In some ways in the US, we've really separated public health from medical care over the decades. In order to have good public policy, you need them both working in alignment.

Landray: Absolutely. I think the NHS has really woken up to this. Instead of being a consumer of the products of R&D, it is actually a partner in the R&D process itself. I think that can apply to any healthcare system. Instead of just waiting for somebody else to develop the next stent or the next antithrombotic or whatever it might be, why don't we try to work out whether we even need another stent or another antithrombotic? If we do, what do we need to know about it, what's the pricing point, what will that mean for access? What will that mean for disparities? And so on and so forth.

Moving the healthcare system to be a partner in — not a post hoc consumer of — R&D is a really important part of this. On the academic side, we have to think bigger than individuals. It's easy to talk about team science. The thing is to do it.

Overall, 4000 doctors, nurses, and pharmacists — up and down every hospital in the UK — have taken part in COVID-19 research. The headline number — I don't completely believe it, but let's use it for the moment — is that dexamethasone has saved a million lives around the world. Divide a million by 4000, and then work out how many lives each of those individual pharmacists and frontline doctors and nurses have contributed to saving.

Now, think of what they could have done if they had all gone their own way and done their own thing, which is the classic academic model. Teamwork really means collaboration and looking at the big picture and the big impact.

One of the things I have been at pains to do — Peter Horby, my colleague, is also — this is not the Horby and Landray show. This is not a one- or two-man band. This is a whole-system effort, including those frontline doctors and nurses, and including over 43,000 patients and their families. We can't conclude this discussion without mentioning the patients. This is impossible without them, and this is absolutely for them.

When we see in COVID-19, but also in heart disease, data and numbers and statistics and ticker tape of how many hospitalizations and how many deaths and so on, remember that every number is a person who has a family, and sadly for many, is a tragedy. We do this with patients and their families, for patients and their families.

Harrington: Martin, that's a great way to end the conversation because there is this notion of the citizen scientists. That's what those 43,000 individuals are. We always talk about NNT and NNH. Maybe we should talk about the lives saved by investigator. The dexamethasone example is a great one. Those 4000 individuals should take great pride and credit.

I give you a lot of credit because, yes, it is a team effort, but the team needs a leader. Everybody refers to it as RECOVERY, but it doesn't happen without a leader. Congratulations. Martin, I could talk all day to you. This was a fun conversation. You've done extraordinary work and good things for the globe. Thank you. Thanks for joining me here on the and Medscape Cardiology.

Landray: Thank you very much for having me.

Bob Harrington, MD, is chair of medicine at Stanford University and current president of the American Heart Association. (The opinions expressed here are his and not those of the American Heart Association.) He cares deeply about the generation of evidence to guide clinical practice. He's also an over-the-top Boston Red Sox fan.

Follow Bob Harrington on Twitter

Follow | Medscape Cardiology on Twitter

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: