From: Stanford University, via Flickr – https://www.flickr.com/photos/stanfordmedicine/14052409727
On 3 April, researchers from the University of Oxford launched the Randomised Evaluation of COVID-19 Therapy (RECOVERY) trial, the largest randomised clinical trial for Covid-19 treatments in the world.
Almost 1,000 patients were initially enrolled. Over 6,000 patients across 160 or more hospitals have now taken part in trials.
The study is being led by Peter Horby, Professor of Emerging Infectious Diseases and Global Health in the Nuffield Department of Medicine, and Martin Landray, Professor of Medicine and Epidemiology at the Nuffield Department of Population Health and has received around £2 million in funding from the National Institute for Health Research (NIHR).
I spoke to Martin Landray, Deputy Chief Investigator, who gave me some fascinating insights into a real success story of the crisis so far. Excellent institutional frameworks married to keen strategic thinking and knowhow has allowed Britain to make vital early progress in the search for effective treatments.
If these trials do show success, we will have the most comprehensive data on treatment available at present – what works, what is duff. If the disease continues to make progress beyond Europe, easily scalable treatment options are absolutely vital to prevent a human catastrophe.
Infections may well rise once social distancing measures are relaxed and this work, if, if, it yields results, will leave us far better prepared to manage a surge of new cases.
Alastair Benn: How many treatments are being deployed in the trials at present?
Martin Landray: The first thing to say is that this is a new virus. There are no proven treatments. There are many drugs that might work but none that we know do work. Before drugs are used in very large numbers of people then it is critical to know if they actually have benefits. All drugs have some side-effects. If the drugs have no benefit, then clearly the side-effects will outweigh the benefits. We need to know whether these drugs work. We are trialling four different treatments versus standard of care alone. Tomorrow or Friday, it will be five treatments.
There is not going to be a blockbuster drug that suddenly cures Covid. This is not the same as a child with meningitis who can be made better with a dose of penicillin. But in a condition that affects millions, if not billions of people, a modest effect like reducing the risk of dying by one fifth would turn out to be really important. If, for example, for patients being admitted to hospital you could reduce the deaths from 25,000 down to 20,000, that would be 5,000 lives saved. A reduction of one fifth is incredibly valuable.
In due course, we may find that several drugs reduce mortality by one fifth and so we won’t be taking 20,000 out of 25,000 but 15 or 16 thousand. That’s much more realistic than believing that any one treatment cures Covid.
We’ve had similar patterns in other conditions. If you think about HIV, we use multiple drugs in combination to treat it with enormous success. In cardiac disease, people take three or four different drugs to lower blood pressure or cholesterol with some success. In diabetes it is the same. In cancer it is the same.
There are many areas in medicine where we use two, three or four drugs in combination and get big effects but each of those drugs on their own has a pretty modest effect.
At the moment, we don’t know whether giving a drug is good for the patient in front of us or bad for the patient in front of us. So, everybody has their normal conditions dealt with, diabetes, heart conditions, and they get oxygen if they need it or ventilation if they need it and so on.
In addition, patients either might get no extra drug treatment or they might get liponavir/ritonavir, which is an antiviral HIV treatment, or dexamethasone, which is an essentially immune-suppressive drug, a corticosteroid, sometimes used in asthma and other respiratory conditions. Or they might get Hydroxychloroquine, normally an anti-malarial but may have anti-viral properties. Or they might get azithromycin which is normally an antibiotic but may have anti-inflammatory properties.
There are four drugs and a control group. Each patient will get randomly allocated into one of those arms.
For those who have low levels of oxygen in the blood and a lot of inflammation, in addition to one of those options, they may get a drug called Tocilizumab. It blocks part of the inflammatory pathway and is commonly used in rheumatoid arthritis but it is also used to treat what is known as cykotine storm in the treatment of some kinds of blood malignancy.
AB: What is the rationale for those treatment options?
ML: Most patients who get coronavirus have a self-limiting nasty flu-like illness. People stay at home, they isolate, it gets better over a period of a week or ten days. In a sense that’s all there is to it. Of those, a subset becomes sufficiently unwell that they get admitted into hospital. We are interested in starting at that point.
What happens is that a bad viral infection then stimulates the immune system and possibly over-stimulates certain parts of the immune system. That’s the so-called cykotine storm. That accompanies the deterioration from bad viral infection to the lungs not working well and then the need for ventilation and of course in some people to death.
The idea is that blocking part of that immune response may well make it less likely that people are put onto ventilators and die. That’s the theory and all these drugs have a good theory behind them. All of these drugs are used in other settings so we have some information about their safety but we don’t know if any of these drugs actually work in the context of Covid-19. That’s why we’re doing the trial.
AB: Some of the data coming out of trials which have taken place in the rest of the world appears to be pretty partial.
ML: There are quite a large number of not terribly informative studies. There have been a number of studies which don’t actually have a control arm and are not randomised. So, for example, there was a paper on Remdesivir in the New England Journal of Medicine which was published recently and basically said: “We gave sixty patients Remdesivir and this is what happened to them.” That’s all very well but you have no idea what would have happened to them otherwise. It’s uninterpretable.
There have been other studies that are only randomised in people between a high dose of a drug and a low dose of the same drug. All you can conclude from them is that one dose is better than the other or there is no difference between them. You have no idea if those patients would have been better or worse off if they hadn’t been taking the drug.
Then there’s a group of trials that are very small, perhaps 100 patients vs 100 patients. If you see for example as happened in one that 25 patients died vs 19 patients died, the difference between 19 and 25 is not statistically significant. The conclusion is that there is no significant difference between these two treatments, the treatment and the control. What you can actually conclude is that the study was too small. If that was 19,000 vs 25,000 we would be saying “hallelujah”.
The trials have to be adequately large. They have to be randomised and they have to have complete data on the outcomes that really matter. In this context, that is pretty easy. Does the patient survive? How long is the patient in hospital? Does the patient need ventilation? Those are the key outcomes. The RECOVERY trial is a very large trial – about 6,600 patients have been randomised over the past few weeks. That is probably at least ten times bigger than any other trial that has ever been done on Covid-19.
Various people have claimed, President Trump for example, that hydroxycholoroquine should be used widely but nobody knows if this stuff works. What happens is that very large numbers of people are given these drugs which do have some safety issues and at the end of it you say “well that was fine, but did it do patients any favours? No idea.” That’s why these trials are so important.
Is the commentary on hydroxychloroquine partly driven by how easy it is to produce?
Yes, it’s partly that. It’s been around for forty or fifty years and has been used very effectively to treat malaria. It’s used in some rheumatological conditions as well.
Everybody has a desire to do something. Whether you’re a doctor in front of an individual patient, or whether you’re responsible for a whole health system, you are desperate to do something. But the thing to do is not just to do something when you don’t know the benefits. The something you should do is to work out as rapidly as possible what treatments work and then implement it as quickly as you can thereafter.
There are lots and lots of drugs you could study, including hydroxychloroquine but we’ve started by looking at drugs that have a reasonable theoretical basis and those drugs where we know something about the safety. And if the drug turns out to be a winner we could scale it up to millions and millions of people very rapidly.
Unlike the vaccines, where we may or may not identify a vaccine over the next three of four months, and the timeline to have millions and millions if not billions of doses is still something like 18 months away, the first place we should start with is treatments we can roll out rapidly.
AB: And that would change the picture in a big way on the epidemiological modelling as we go forward.
ML: It’s potentially massively significant if you look at how many deaths there have been globally to date and consider what might happen as infection rolls out to other parts of the world, Africa, South America or other parts of Asia such as Pakistan, and consider that we are all on stage one of this epidemic.
Only one country has really come out of stage one, China, and they are just beginning to relax the social distancing measures. As you relax them, the risks of there being more cases goes back up again. When more cases occur, you really need to have some answers on what treatments work.
What we saw in Ebola, for example, was that many, many different drugs were used during the first outbreak in West Africa but they weren’t used in the context of clinical trials. When the next outbreak occurred in Congo, which is ongoing, at the beginning of that epidemic nobody was any the wiser on how to treat it for all the experience that had been built up during the last outbreak.
It is very typical that the clinical trials get added on at the end of an epidemic. It isn’t seen as a priority. The question is never answered because by the time trials have been set up the outbreak may have blown over. We were very keen in this trial to get it up and running and established in NHS hospitals at the beginning. We went from drafting the protocol, the scientific description of what the trial should look like to recruiting the first patient in nine days. That’s a process that would normally take nine months, possibly even longer in many settings. We then rolled it out across 160 or more hospitals across the whole of the UK.
AB: Do we have certain institutional advantages that have allowed us to move so fast?
ML: On clinical trials, we have a number of long-term structural advantages and short-term tactical advantages. We have one NHS. It has many parts but we do have one NHS. There is a long history of doing clinical trials in the UK. We also have the National Institute for Health Research (NIHR) which is not only a government funding agency for clinical research but also ensures that there are research staff and research nurses in every hospital up and down the country who could be mobilized rapidly to deal with this.
There is an academic sector too where there are units like ours in Oxford where this is what we do for a living. We design and run these kinds of clinical trials.
Those are the structural advantages. Capitalising on those was the ability to move at speed. So the fact that the chief medical officers of all four countries plus the medical director of NHS England wrote to every hospital in the land and said that this trial was a priority and including patients should be seen as part of delivering clinical care was incredibly helpful.
The fact we have one ethics committee that covers the whole country was helpful. We don’t have to go through this several times. The Medicines and Healthcare products Regulator Agency (MHRA), the regulator, was able to turn this around in 24 hours. It’s all very well having a good infrastructure but you need to take advantage of it and mobilize. Those things have happened. The trial is designed to be simple and focussed on the things that matter. It has been straightforward for clinicians to roll out and easy for patients to take part in and get behind.
We’ve been able to bring the world of clinical trials and clinical research right up close to the world of clinical care in the NHS. There are very few places in the world that can do that.
Professor Martin Landray a Professor of Medicine and Epidemiology at the Nuffield Department of Population Health at Oxford University. He is Deputy Director of the Big Data Institute. He has led international trials that have enrolled 65,000 individuals with cardiovascular or kidney disease from 18 countries across 4 continents.