In my previous blog in this series I promised to talk about hydroxychloroquine as it's another example of something medical being taken out of context and heavily politicized. As a psychiatrist I never prescribe it, but as a medical doctor I am obviously familiar with this medication. I was surprized to hear that an antimalarial drug can be a treatment for COVID-19 and, given all the publicity around it, I think it's time to review the evidence for its use in this pandemic.
Early clinical impressions
Earlier this year, many of us have heard of this medication as some doctors, like Dr.Zelenko in New York and Dr.Raoult in Paris reported successfully treating multiple patients with hydroxychloroquine. Fortunately or not, but these early successes were heavily politicized as President Trump spoke very highly of this medication and what appears to be a smearing media campaign against this drug emerged, which obviously it didn't help much. So, let us review the actual evidence.
Dr.Didier Raoult ran an open-label non-randomized clinical trial of hydroxychloroquine and azithromycin for treatment on COVID-191. To be honest, having just 20 subjects in a trial like that is somewhat problematic as apparently they had no shortage of people with COVID-19 in France at that point and there was no excuse to have such a small sample size. It's customary to have at least 30 subjects per group, preferably more. They also had just 16 controls, who were much younger and were recruited from other clinical sites. They examined their patients daily for 14 days and used PCR to test to assess viral clearance. In their report they indicated that at day 6, 70% of subjects treated with hydroxychloroquine were "virologically cured" vs. only 12.5% in control group. And even better - 100% of patients treated with hydroxychloroquine and azithromycin were "virologically cured". Sounds like a breakthrough, right? I think it was, but this study was flagged as having several flaws "including questions about its ethical underpinnings, messy confounding variables, missing patients, rushed and conflicted peer review, and confusing data" and it has been heavily criticized. My major questions were related to the sample size, non-randomized design, and absence of the description of clinical outcomes (did they get better?). I don't want to judge the authors. As someone who admittedly is not an expert in infectious diseases, I will say that this study was really important at the moment as it showed some positive signal and a clear need for better studies such as full-scale randomized controlled trials.
Dr.Zev Zelenko from New York, "published" a report of a retrospective case series study, which I find hard to cite as it was published online as a preprint, but not in a scientific journal, at least yet2. Messy sentence, but the report itself appears to be messy too. I have lots of questions to the authors, which I might work on later, in another blog. The gist of their report is that they treated 141 COVID-19 patients with a combination of hydroxychloroquine, azithromycin, and zinc. They also used 377 untreated patients as controls. They reported 1 (0.7%) death and 4 (2.8%) hospitalizations in the treatment group vs 13 (3.5%) deaths and 58 (15.4%) hospitalizations in the non-treatment group. This study definitely showed some promising results, but again it should be corroborated by a properly designed and conducted clinical trial.
Another study, actually the best out of those that I've seen is a retrospective review of fatality in critically ill COVID-19 patients in Wuhan, China3. They have treated 550 patients with severe forms of COVID-19 and 48 of them also received hydroxychloroquine in addition to their standard treatment protocol. They reported 18.8% fatality in the hydroxychloroquine group vs. 47.4% in the control group. Again, another promising set of results, but we need a proper randomized controlled trial to prove that hydroxychloroquine works.
Why observational studies are not good enough?
I must explain why I am saying that there is a need of a proper trial, namely, a randomized placebo-controlled trial (RCT). The data that we get from observational studies are promising, but the outcomes are a function of multiple factors – patients' characteristics such as age, obesity, and severity of disease play a huge role in recovery as well as the parameters of the environment, other treatments they receive and even placebo effect. Each one of these factors may turn out to be a critical confounder for the final outcome, and the only way to isolate the medication is to randomly assign patients to treatment and to use placebo in the control group. Also, the sample size should be large enough to even out the odds of effects of confounding factors. Given the large number of infected people, relatively short course of the disease, and all the incentives and pressure from the public, running such a trial should not be an unsurmountable task. And it wasn't – there were randomized multicenter placebo-controlled trials that aimed to answer the question whether hydroxychloroquine is helpful or not. But before we get to it, I'd like to talk about another (and absolutely irrelevant) study quickly.
The irrelevant study that everyone cites
There is a meme based on a 2005 article titled "Chloroquine is a potent inhibitor of SARS coronavirus infection and spread"4. The meme states that "Dr.Fauci's NIH knew that hydroxychloroquine is an effective treatment and prevention for coronavirus". I have so much to say about it… It's just unbelievable that someone would be so misguided to propagate this. First, it's not a clinical trial, it's an in vitro study, which used cell cultures, not living organisms. Then, it's not even hydroxychloroquine, it's chloroquine, and, most importantly, it was a study of ANOTHER coronavirus, not the one that causes COVID-19. The study was conducted prior to 2005, when this coronavirus didn't even exist! To those who spread this clear misinformation – please, check your facts first, use common sense!
Randomized controlled trials
Let's talk about the gold standard research initiatives. First of all, one of the major randomized controlled trials conducted so far was the ORCHID study – ORCHID stands for "Outcomes Related to COVID-19 Treated with Hydroxychloroquine among In-patients with symptomatic Disease". The study enrolled 479 participants and randomized them to either hydroxychloroquine or placebo. Exactly how it should be done. It is a registered clinical trial that you can check online (NCT04332991). Every serious research endeavour uses interim analysis to steer their clinical trials and in this case the Data Safety and Monitoring Board (DSMB) decided that based on preliminary data there is no benefit from hydroxychloroquine and therefore there is no point in continuing the trial. It is disappointing, but if after enrolling almost 500 subjects they couldn't show benefit, continuing a trial like that would be unethical and wasteful. At the same time, I would like to see the data from the study (I couldn't find them).
Another randomized controlled trial called RECOVERY (The Randomised Evaluation of COVID-19 therapy) was recently published5 (NCT04381936). It had 1561 patients receiving hydroxychloroquine compared to 3155 patients receiving usual care. The main outcome was 28-day mortality and there was no significant difference between mortality rates in patients receiving hydroxychloroquine (26.8%) and usual care (25.0%). This study was actually large and robust enough to conclude that hydroxychloroquine is ineffective for treatment of COVID-19.
I would expect you to want more evidence, so here is another RCT – a multicenter, open label, randomized controlled trial conducted in Catalonia6. They randomized 293 patients into either hydroxychloroquine group (136 subjects) or control (157 subjects) group. They reported that hydroxychloroquine did not reduce risk of hospitalization (7.1%, control vs. 5.9%, hydroxychloroquine) nor shortened the time to complete resolution of symptoms (12 days, control vs. 10 days, intervention). No relevant treatment-related adverse effects were reported.
Finally, there was a randomized controlled trial aiming at prevention of COVID-19 after exposure to the virus7 conduced in Canada. They enrolled 821 asymptomatic participants who were exposed to a confirmed COVID-19 contact and randomized them into either placebo group (407 subjects) or hydroxychloroquine (414 subjects) group. Again, hydroxychloroquine did not show significant effect in terms of prevention of COVID-19 (11.8% developed COVID-19 in the hydroxychloroquine group vs. 14.3% in the control group), but side effects were more common (40.1% vs. 16.8%). No serious adverse reactions were reported.
So, to summarize all the evidence that I reviewed here – there were some promising results from observational / naturalistic studies and also some indications that hydroxychloroquine would be helpful as it had some effect on other coronaviruses, but after several large scale randomized controlled trials failing to show any benefit, we can conclude that it is not an effective treatment for COVID-19. We need to search in other directions.
As always, I would like to remind you that this blog should not be taken as medical advice and that my opinions do not necessarily represent the official standing of regulatory and public health authorities I belong to. At the same time, I believe that my points are valid, and I would like to encourage you to subscribe to this website, and to my YouTube channel. You are also always welcome to communicate with me via making comments, suggestions, and asking questions, either here, or on social media platforms.
Stay safe, strong, and healthy!