PARIS, FRANCE -- Dozens of scientists have raised concerns over a large-scale study of hydroxychloroquine and chloroquine published in the Lancet that led to the World Health Organization suspending clinical trials of the anti-viral drugs as a potential treatment for COVID-19.
Hydroxychloroquine, normally used to treat arthritis, has become one of the most high profile drugs being tested for use against the new coronavirus.
This is partly because of comments by public figures including US President Donald Trump, who announced this month he was taking the drug as a preventative measure.
Mandeep Mehra of the Brigham and Women's Hospital in the US looked at records from 96,000 patients in hundreds of hospitals between December and April and compared those who received treatment with a control group.
The study concluded that treatment with hydroxychloroquine and chloroquine, an anti-malarial, showed no benefit and even increased the likelihood of them dying in hospital.
Both drugs can produce potentially serious side effects, particularly heart arrhythmia.
The research published in the Lancet medical journal followed numerous smaller studies that suggested hydroxychloroquine is ineffective in treating COVID-19 and may even be more dangerous than doing nothing.
Within days the WHO temporarily suspended use of the drugs in its Solidarity Trial, which has seen hundreds of hospitals across several countries enrol patients to test possible treatments for COVID-19.
"This impact has led many researchers around the world to scrutinise in detail the publication in question," said the open letter in response to the study, which was signed by a number of prominent scientists and published Thursday.
It added that this scrutiny raised "both methodological and data integrity concerns".
One of the main concerns was a lack of information about the countries and hospitals that contributed to the data, which was provided by Chicago-based healthcare data analytics firm Surgisphere.
The authors also list "implausible" ratios of use of the drugs in some continents and discrepancies in the data for Australia, where they said there were more deaths recorded in the hospitals covered by the study than official figures for the entire country.
"Surgisphere (the data company) have since stated this was an error of classification of one hospital from Asia. This indicates the need for further error checking throughout the database," the letter said.
'SERIOUS DOUBTS'
Among the signatories are clinicians, epidemiologists and other researchers from around the world, from Harvard to Imperial College London.
Francois Balloux of University College London, said he believed it was his "duty" to add his name to calls for answers to questions about the study and for greater transparency.
"I have serious doubts about the benefit of CQ/HCQ treatment for #COVID19 infection, and cannot wait for the whole drama to be over," he said on Twitter, where #LancetGate was trending among science accounts.
"Though, I believe 'research integrity' cannot be invoked only when a paper doesn't support our preconceptions."
The letter was also signed by French researcher Philippe Parola, a colleague of the Marseille-based professor Didier Raoult, whose work has been at the forefront of promoting hydroxychloroquine and has also been subject to criticisms over methodology.
The Lancet said it had received "several questions" over the study.
"We have referred these questions to the authors and they are working to address the issues that have been raised," it said in a statement.
The coronavirus crisis has put enormous strain on the normally sedate system of scientific publication, with an avalanche of research and the process of peer review considerably accelerated.
A spokesman for the WHO said a comprehensive review of the drugs was expected to reach a conclusion in mid-June.
"Any decision on the hydroxychloroquine arm of the study must be based in a comprehensive analysis of the available evidence, particularly from concluded and ongoing randomised trials," said spokesman Tarik Jasarevic.
