In the search for a treatment for COVID-19, attention has turned to preexisting drugs in the hope to reduce the lead time in finding a cure. Hydroxychloroquine (HCQ), chloroquine (CQ) and macrocodes (MC) are all known to have antiviral effects (first two through endosomal acidification) in other viral diseases and there were promising results in vitro studies with SARS-CoV2. Understandably, there was a rush to start clinical trials in the heat of the worsening pandemic.
The initial letter from China (Gao et al) published as a preprint suggested many clinical trials had shown superiority of HCQ and HQ without giving any data. This was joined by the well-publicized trials from Marseille (Gautret et al, Million et al) with no control arms and hence no way of evaluating the efficacy of these drugs. This has generated huge international interest with the hope that they offer great benefit.
Many countries and institutions have adopted these drugs as part of their treatment regimen for COVID-19 on the basis of the initial promising studies. Since then, there has been a stream of studies published in the scientific literature including 2 studies in the BMJ and 1 in The Lancet last week. There was a stern editorial in the BMJ on the 19 May “Lack of efficacy of hydroxychloroquine in covid-19” and on 25 May, the WHO’s Director General announced that the WHO Solidarity Trial Executive Group has implemented a temporary pause of the HCQ arm while the safety data is reviewed by the Data Safety Monitoring Board following the higher mortality rate in the HCQ/CQ +/- macrolide in The Lancet study.
So, where are we now?
I have reviewed all the clinical studies and meta-analysis available online so far to draw my own conclusions. We have now 16 published studies of which 7 are non-peer-reviewed preprints. There are 1 letter, 3 randomized controlled trials and the remaining 12 retrospective observational studies of varying sizes. I have not included a study submitted as pre-print to the NEJM but then withdrawn by the authors (Barbosa et al – increased ventilator support needed in HCQ group). There are also 3 systematic reviews and meta-analyses.
I have summarized all these studies in 3 colour-coded tables with clickable hyperlinks to the original articles in the following pdf. Make it fullscreen and click away!
What are the clear messages?
- In-vitro findings do not often translate into in-vivo results.
- The era of non-peer-reviewed pre-print publications allow rapid communication to the scientific community at the price of allowing some deeply flawed studies.
- It is impossible to draw any conclusions on studies (Molina et al, Gautret et al, Million et al) with no control arms in a disease where at least 80% of those infected have mild illness and the mortality rate is around 1%.
- Four publications showed benefit of HCQ/CQ: a pre-print letter with no data ((Gao et al), 1 retrospective study with no methodology described (Asraf et al) and 2 small retrospective studies (Yu et al, Kim et al). Interestingly, one of them showed marked improvement when started in patients in severe ARDS requiring ventilation. Is that the mmunomodulatory HCQ effect in action?
- Six studies showed NO difference between treatment with HCQ/CQ and no treatment. These included the 3 RCT (Chen J. et al, Chen Z. et al, Tang et al) and 2 large retrospective studies (Geleris et al, Rosenberg et al) and a small retrospective study (Mahévas et al).
- 3 studies showed HCQ/CQ were harmful: 2 small retrospective studies (Magagnoli et al, Mallat et al) and the largest retrospective study with over 100,000 participants (Mehra et al, The Lancet). The latter showed more a doubling of mortality in the treatment groups. The excess mortality were in patients with heart disease or obesity. These are the patients most at risk from COVID-19 and yet most in need of a wonder drug. Please read footnote.
- It is not clear if there might be a differential effect of HCQ/CQ depending on specific circumstances (prophylaxis, on exposure, on symptom onset, at start of the cytokine storm, ARDS).
- Three systemic reviews (Sarma et al, Kumar et al, Chacko et al) with meta-analysis all showed no benefit of HCQ/CQ. One shoed increased mortality with HCQ/CQ and another significantly higher adverse side effects.
- Results from a large double-blind RCT are still awaited.
- Despite a lack of high quality evidence, an overall picture of no benefit from HCQ/CQ+/-MC has emerged (with serious adverse side effects and increased mortality in one study and one meta-analysis.)
Are we any clearer?
HCQ/CQ arouses passionate discussion. It is possible to pick holes in any of these trials. We are all counting the days and weeks to the results of double-blind randomised control trials. Until then, in the absence of any other treatment for Covid-19, the question is should we continue giving patients HCQ/CQ +/- MC or tell them is no oral medication for the disease? What do you think.
Footnote added on 5 June 2020. Since the above review was posted, there has been increasing criticism of and mounting revelations about the integrity of the data in the large Lancet study by Mehra et al. leading to retraction by 3 of the authors.
This doesn’t change however the conclusion that HCQS and HCQ have no proven benefits in Covid-19 but there is now less evidence of the adverse effects and of excess mortality associated with these drugs.