On May 14, the British Medical Journal published an article presenting the results of a multi-center trial of hydroxychloroquine (HCQ) in Chinese patients with COVID-19 in February. It concluded:
“Administration of hydroxychloroquine did not result in a significantly higher probability of negative conversion than standard of care alone in patients admitted to hospital with mainly persistent mild to moderate covid-19. Adverse events were higher in hydroxychloroquine recipients than in non-recipients.”
However, an earlier version of the data told a different story. On April 14, 2020, a preprint manuscript of the study concluded the following:
“The administration of HCQ did not result in a higher negative conversion rate but more alleviation of clinical symptoms than SOC alone in patients hospitalized with COVID-19 without receiving antiviral treatment, possibly through anti-inflammatory effects. Adverse events were significantly increased in HCQ recipients but no apparently increase of serious adverse events.”
The BMJ’s published version on May 14 represents a more narrow and decisively negative conclusion for HCQ.
Why was this positive data removed from the earlier draft of the BMJ paper?
Let’s take a deeper look at the original. From the Results section of the original manuscript on April 14:
A significant efficacy of HCQ on alleviating symptoms was observed when the confounding effects of anti-viral agents were removed in the post-hoc analysis (Hazard ratio, 8.83, 95%CI, 1.09 to 71.3). This was further supported by a significantly greater reduction of CRP (6.986 in SOC plus HCQ versus 2.72 milligram/liter, P=0.045) conferred by the addition of HCQ, which also led to more rapid recovery of lymphopenia, albeit no statistical significance.
Echoed in the Discussion section:
An evident efficacy of HCQ on the alleviation of symptoms was demonstrated (Hazard ratio, 8.83, 95%CI, 1.09 to 71.3) in the subgroup of patients without receiving antiviral treatment in the post-hoc analysis. This was further supported by a significantly greater reduction of CRP (6.986 in SOC plus HCQ versus 2.723 in SOC, milligram/liter, P=0.045) conferred by the addition of HCQ, which also led to more rapid recovery of lymphopenia, albeit no statistical significance.
And in the description of Figure 3:
Shown in panel B are forest plots of subgroup analyses, in which a significantly higher rate of symptoms alleviation was observed in patients with SOC plus HCQ versus SOC in the subgroup of patients without receiving potential anti-SARS-COV-2 drugs. (Hazard ratio, 8.83, 95%CI, 1.09 to 71.3).
Translation: When the analysis controlled for HCQ, the medicine was demonstrably effective in reducing COVID-19 symptoms and helping patients’ immune systems recover more quickly.
Yet in a later version of the manuscript published on May 7 on medRxiv, and subsequently in the final version published by BMJ on May 14, this positive secondary outcome data is gone. Why did this happen? It would appear from the initial manuscript that HCQ had a clinical benefit supported by a mechanism of action.
This mystery comes at a problematic time for COVID-19 patients around the world. HCQ continues to be toxic political football in the American media, while Chinese health data is tainted by the government’s initial efforts to hide the novel coronavirus outbreak from the world.
With more than 180 trials on HCQ for COVID-19 underway, let’s hope we’ll get a clear picture of the data soon.
