Some new research results on Hydroxychloroquine

This treatment is so controversial. Too bad the president pushed it so hard and oversold it, but I think we have to be careful not to let the natural backlash in response to this keep us from carefully evaluating whether it is effective for some stages of COVID-19 and the documented potential cardiac toxicities are acceptable in that setting.

Science Translational Medicine blogger Derek Lowe just published a blog entry reporting results from two preprints on HCQ.

One paper reports results from a study in Wuhan in which 568 patients "received comparable basic treatments" and 48 of them received oral HCQ (200mg 2x/day for 7-10 days). This cohort is focused on advanced stage patients -- includes all critically ill patients between February 1, 2020 and April 8, 2020 at the study hospitals, who had respiratory failure requiring ventilation, septic shock, or other organ failures requiring ICU treatment.

The paper reports mortality rates of 18.8% (9/48) in the HCQ group and 45.8% (238/520) in the NHCQ group. This study was not randomized, so it is hard to know what to make of these results. The baseline characteristics of HCQ and NHCQ groups were similar, and they did a Cox analysis with covariates and still saw this effect (HR=0.33 for HCQ) -- this is promising but the fact that it is not randomized significantly decreases the value of this evidence. Too bad they didn't do some type of more formal propensity score analysis to try to adjust more precisely for potential selection bias in the HCQ vs. NHCQ arm.

Another interesting element is that they saw a pretty strong decrease in inflammatory cytokine IL-6 after treatment in HCQ arm but not the non-HCQ arm. This type of inflammatory cytokine is elevated during the cytokine storm that seems to induce acute respiratory syndrome that kills many critical COVID patients. This makes biological sense, given that HCQ's use for treating lupus and rheumatoid arthritis is based on related mechanisms. Also, in another related note, it has been previously shown that HCQ has anti-coagulating effect, meaning that part of its mechanism of action may reduce the risk of clots that are also a major killing mechanism of COVID-19.

I don't post this to suggest this article (not peer reviewed yet BTW) proves anything about HCQ, but suggests that if it works, it might be best suited for critical care settings as part of a strategy to prevent ARS. This is the setting in which my undergraduate mentor, Marlin Eby, received it when he was in ICU and his lung performance was going downhill fast, and after treatment immediately recovered (sample of size 1, but anyone who saw how bad he was and how quick he suddenly turned around and recovered would be amazed and wonder whether the treatment played a part). I admit this shapes my viewpoint and makes me want to stay open minded about the potential of HCQ longer than most people. I have no vested interest in the president being right or wrong -- I just want us to best assemble accruing evidence to figure out what treatments might work and in what settings/stage of disease.

People seem to be dismissing HCQ after a couple negative studies (that did not focus on critical patients BTW), which may be premature and unnecessary especially given the enormous number of ongoing studies (189 on covid trial explorer) for HCQ, many of which are randomized and well designed, and which focus on different stages and settings (prevention, mild symptoms, hospitalized, critical), will continue to provide a great deal more information over time than we have now. I propose that if it works, it is in the most critical patients and as an approach to prevent progression to ARS, and it is also in those patients that the cardiac risk of increasing QT interval is likely negligible compared with the imminent risk of death due to COVID-19.

Yes the heart toxicity of HCQ is real, which is why the FDA has limited its use in hospitalization settings, not at-home OTC treatment. But it has been used for years for treatment of other diseases, and at similar doses for lupus and rheumatoid arthritis, with management of the potential cardiac toxicities in those settings. So if it is found to be effective for critical COVID-19, clearly the cardiac toxicity is not a disqualifier.

©2020 by Covid Data Science. Proudly created with