Intended for healthcare professionals

Rapid response to:

Feature Medical Research

RECOVERY trial: the UK covid-19 study resetting expectations for clinical trials

BMJ 2020; 369 doi: https://doi.org/10.1136/bmj.m1626 (Published 28 April 2020) Cite this as: BMJ 2020;369:m1626

Read our latest coverage of the coronavirus pandemic

Rapid Response:

Re: RECOVERY trial: the UK covid-19 study resetting expectations for clinical trials

Dear Editor

A short while ago, I listened to a voiced-over doctor on the BBC. She said that some senior doctors had written to the UK’s Chief Medical Officer complaining about restrictions on their ability to prescribe for Covid-19. Apparently, they were being asked not to use hydroxychloroquine (HCQ) outside the relevant experimental arm of the Government RECOVERY project. There was, in their view, little chance that the other two experimental arms would work. They were therefore being asked to randomise some patients to what might well be a preventable death, while withholding a potentially useful treatment from others. As an 82-year-old man, vulnerable to Covid-19 and with experience of non-medical research, I was concerned by this news. This reply reflects my continuing need for reassurance.

The RECOVERY Trial proved to be an impressively large and rapidly commissioned study with an ‘off-the-shelf’ design (one arm per drug tested), in danger of being overtaken by events. One drug tested (Lopinavir-Ritonavir) has now failed in two trials (1, 2). Another (Dexamethasone) is (according to WHO (3)) an ‘adjunctive treatment’ not to be used for Covid-19 outside trials. Dexamethasone reduces inflammation but may embolden viruses, was not shown to work for SARS (4), and used with influenza may kill (5). Azithromycin, also tested, is an antibiotic with anti-inflammatory properties, a logical response to possible cytokine storms and secondary infections, but not to the virus itself. The conflicting and inconclusive evidence on Corticosteroids (6) and Azithromycin (7) relates to their use in combination with other drugs. Why use them on their own rather than in a cross-cutting design (8) where antivirals (relevant to all) are randomly matched with anti-inflammatories (used with discretion) and placebos?

There was better news on HCQ. A small Chinese trial (9) (unfairly dismissed IMHO in a BMJ editorial) suggests that taken at the right time (10) it can stop the disease in its tracks. HCQ provides an arm of the RECOVERY trial for hospital patients (this may be the wrong time for them (11)) but the impressive Principle trial randomises it among selected patients in the community. Sadly, an underlying heart condition means that the recommended dose might kill me (12). It reports in March 2021, too late to influence our lockdown exit. Will this enable me to cuddle my grandchildren? I think not.

Meanwhile there is new evidence from trials of Remdesivir (13,14) and Sovodak (15). Many trials report by July (16). Clinical developments in the delivery of Oxygen (17) may anticipate the results of an 18-month trial (18). Surely, medicine does not advance by trials alone? Practice in my field, and perhaps even in this more scientific one, reflects growing experience and inconclusive but hopefully converging, research. It evolves messily. From my patient’s viewpoint our trials should not hold this back.

Fortunately, the trials are adaptive. They will change with emerging results. Nevertheless, questions remain. Are there strong formal links between these trials and similar international ones (e.g. the French-led DISCOVERY, and WHO SOLIDARITY trials) thus allowing comparable outcome criteria, meta-analyses, and – crucially – quicker recognition that some arms should be stopped? Is there capacity for real-time statistical, and scientific/medical appraisal of both the developing clinical and research evidence? Are there mechanisms for taking decisions on the trials balancing both the internal and external evidence?

In summary, we bet on two trials. Each follows traditional practice in comparing single variables. One has experimental arms unlikely to work; the other an experimental arm, not suitable for everyone and with risks likely to require close monitoring. Practice will vary between these arms because of the treatments involved. Methods for ensuring up-to-date best practice in all arms are not described. Without this, the results are scientifically dubious (if the risks of HCQ require monitoring it may be this that produces a result) and potentially out-of-date. The challenge for the trials is to adapt in such a way that, without loss of rigour, they test whatever is, at the moment, best practice against something potentially better.

1 https://www.nejm.org/doi/full/10.1056/NEJMoa2001282
2 https://www.medrxiv.org/content/10.1101/2020.03.19.20038984v2
3 https://www.who.int/publications-detail/clinical-management-of-severe-ac...(ncov)-infection-is-suspected
4 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1564166/
5 https://ccforum.biomedcentral.com/articles/10.1186/s13054-019-2395-8
6 https://www.medrxiv.org/content/10.1101/2020.03.06.20032342v1
7 https://www.thelancet.com/journals/lanres/article/PIIS2213-2600(20)30172-7/fulltext
8) https://www.thelancet.com/journals/lanres/article/PIIS2213-2600(20)30172-7/fulltext
9) https://www.medrxiv.org/content/10.1101/2020.03.22.20040758v3
10) https://www.who.int/csr/resources/publications/influenza/11_29_01_A.pdf
11) https://www.medrxiv.org/content/10.1101/2020.04.16.20065920v1.full.pdf
12) https://www.acc.org/latest-in-cardiology/articles/2020/04/10/15/06/acc-j...
13) https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31022-9/fulltext
14) https://www.gilead.com/news-and-press/press-room/press-releases/2020/4/g...
15) https://publicaddress.net/hardnews/has-iran-found-an-effective-covid-19-...
16) https://www.biocentury.com/trial-timeline
17) https://www.warringtonguardian.co.uk/news/18402785.warrington-hospital-c...
18) https://www.hra.nhs.uk/covid-19-research/approved-covid-19-research/282338/

Competing interests: No competing interests

07 May 2020
Ian A Sinclair
Professor emeritus
Department of Social Policy and Social Work University of York
ian.sinclair@york.ac.uk